Text Generation
Transformers
Safetensors
English
Chinese
qwen3_5_moe
image-text-to-text
qwen3.6
uncensored
abliterated
gguf-recovery
Mixture of Experts
conversational
Instructions to use DreamFast/Qwen3.6-35B-A3B-Uncensored-HauhauCS-Aggressive-Safetensor-Benchmark with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use DreamFast/Qwen3.6-35B-A3B-Uncensored-HauhauCS-Aggressive-Safetensor-Benchmark with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("text-generation", model="DreamFast/Qwen3.6-35B-A3B-Uncensored-HauhauCS-Aggressive-Safetensor-Benchmark") messages = [ { "role": "user", "content": [ {"type": "image", "url": "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/p-blog/candy.JPG"}, {"type": "text", "text": "What animal is on the candy?"} ] }, ] pipe(text=messages)# Load model directly from transformers import AutoProcessor, AutoModelForMultimodalLM processor = AutoProcessor.from_pretrained("DreamFast/Qwen3.6-35B-A3B-Uncensored-HauhauCS-Aggressive-Safetensor-Benchmark") model = AutoModelForMultimodalLM.from_pretrained("DreamFast/Qwen3.6-35B-A3B-Uncensored-HauhauCS-Aggressive-Safetensor-Benchmark", device_map="auto") messages = [ { "role": "user", "content": [ {"type": "image", "url": "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/p-blog/candy.JPG"}, {"type": "text", "text": "What animal is on the candy?"} ] }, ] inputs = processor.apply_chat_template( messages, add_generation_prompt=True, tokenize=True, return_dict=True, return_tensors="pt", ).to(model.device) outputs = model.generate(**inputs, max_new_tokens=40) print(processor.decode(outputs[0][inputs["input_ids"].shape[-1]:])) - Inference
- Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- vLLM
How to use DreamFast/Qwen3.6-35B-A3B-Uncensored-HauhauCS-Aggressive-Safetensor-Benchmark with vLLM:
Install from pip and serve model
# Install vLLM from pip: pip install vllm # Start the vLLM server: vllm serve "DreamFast/Qwen3.6-35B-A3B-Uncensored-HauhauCS-Aggressive-Safetensor-Benchmark" # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:8000/v1/chat/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "DreamFast/Qwen3.6-35B-A3B-Uncensored-HauhauCS-Aggressive-Safetensor-Benchmark", "messages": [ { "role": "user", "content": "What is the capital of France?" } ] }'Use Docker
docker model run hf.co/DreamFast/Qwen3.6-35B-A3B-Uncensored-HauhauCS-Aggressive-Safetensor-Benchmark
- SGLang
How to use DreamFast/Qwen3.6-35B-A3B-Uncensored-HauhauCS-Aggressive-Safetensor-Benchmark with SGLang:
Install from pip and serve model
# Install SGLang from pip: pip install sglang # Start the SGLang server: python3 -m sglang.launch_server \ --model-path "DreamFast/Qwen3.6-35B-A3B-Uncensored-HauhauCS-Aggressive-Safetensor-Benchmark" \ --host 0.0.0.0 \ --port 30000 # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:30000/v1/chat/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "DreamFast/Qwen3.6-35B-A3B-Uncensored-HauhauCS-Aggressive-Safetensor-Benchmark", "messages": [ { "role": "user", "content": "What is the capital of France?" } ] }'Use Docker images
docker run --gpus all \ --shm-size 32g \ -p 30000:30000 \ -v ~/.cache/huggingface:/root/.cache/huggingface \ --env "HF_TOKEN=<secret>" \ --ipc=host \ lmsysorg/sglang:latest \ python3 -m sglang.launch_server \ --model-path "DreamFast/Qwen3.6-35B-A3B-Uncensored-HauhauCS-Aggressive-Safetensor-Benchmark" \ --host 0.0.0.0 \ --port 30000 # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:30000/v1/chat/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "DreamFast/Qwen3.6-35B-A3B-Uncensored-HauhauCS-Aggressive-Safetensor-Benchmark", "messages": [ { "role": "user", "content": "What is the capital of France?" } ] }' - Docker Model Runner
How to use DreamFast/Qwen3.6-35B-A3B-Uncensored-HauhauCS-Aggressive-Safetensor-Benchmark with Docker Model Runner:
docker model run hf.co/DreamFast/Qwen3.6-35B-A3B-Uncensored-HauhauCS-Aggressive-Safetensor-Benchmark
| { | |
| "base_model": "Qwen3.6-35B-A3B", | |
| "converted_model": "Qwen3.6-35B-A3B-HauhauCS-Q8KP-recovered", | |
| "summary": { | |
| "total_tensors": 1045, | |
| "gguf_derived": 693, | |
| "copied": 352, | |
| "gguf_identical": 307, | |
| "gguf_modified": 386, | |
| "copied_identical": 352, | |
| "copied_modified": 0 | |
| }, | |
| "tensors": [ | |
| { | |
| "group": "lm_head", | |
| "name": "lm_head.weight", | |
| "max_abs_diff": 0.0010833740234375, | |
| "mean_abs_diff": 8.252273983089253e-05, | |
| "pct_different": 75.062035236162, | |
| "n_different": 381735006, | |
| "total_elements": 508559360 | |
| }, | |
| { | |
| "group": "embed_tokens", | |
| "name": "model.language_model.embed_tokens.weight", | |
| "max_abs_diff": 0.0011444091796875, | |
| "mean_abs_diff": 5.407223216025159e-05, | |
| "pct_different": 74.34011262716706, | |
| "n_different": 378063601, | |
| "total_elements": 508559360 | |
| }, | |
| { | |
| "group": "layernorm", | |
| "name": "model.language_model.layers.0.input_layernorm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "linear_attn.A_log", | |
| "name": "model.language_model.layers.0.linear_attn.A_log", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 32 | |
| }, | |
| { | |
| "group": "linear_attn.conv1d", | |
| "name": "model.language_model.layers.0.linear_attn.conv1d.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 32768 | |
| }, | |
| { | |
| "group": "linear_attn.dt_bias", | |
| "name": "model.language_model.layers.0.linear_attn.dt_bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 32 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_a", | |
| "name": "model.language_model.layers.0.linear_attn.in_proj_a.weight", | |
| "max_abs_diff": 2.9802322387695312e-08, | |
| "mean_abs_diff": 1.4910934709178036e-09, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 65536 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_b", | |
| "name": "model.language_model.layers.0.linear_attn.in_proj_b.weight", | |
| "max_abs_diff": 2.9802322387695312e-08, | |
| "mean_abs_diff": 1.2775558388966601e-11, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 65536 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_qkv", | |
| "name": "model.language_model.layers.0.linear_attn.in_proj_qkv.weight", | |
| "max_abs_diff": 2.9802322387695312e-08, | |
| "mean_abs_diff": 8.525864042541187e-12, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 16777216 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_z", | |
| "name": "model.language_model.layers.0.linear_attn.in_proj_z.weight", | |
| "max_abs_diff": 2.9802322387695312e-08, | |
| "mean_abs_diff": 5.649228480841817e-12, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 8388608 | |
| }, | |
| { | |
| "group": "linear_attn.norm", | |
| "name": "model.language_model.layers.0.linear_attn.norm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 128 | |
| }, | |
| { | |
| "group": "linear_attn.out_proj", | |
| "name": "model.language_model.layers.0.linear_attn.out_proj.weight", | |
| "max_abs_diff": 0.001953125, | |
| "mean_abs_diff": 6.961986218811944e-05, | |
| "pct_different": 75.75498819351196, | |
| "n_different": 6354789, | |
| "total_elements": 8388608 | |
| }, | |
| { | |
| "group": "expert_down", | |
| "name": "model.language_model.layers.0.mlp.experts.down_proj", | |
| "max_abs_diff": 0.000732421875, | |
| "mean_abs_diff": 2.2513979274663143e-05, | |
| "pct_different": 41.840748116374016, | |
| "n_different": 112315403, | |
| "total_elements": 268435456 | |
| }, | |
| { | |
| "group": "expert_gate_up", | |
| "name": "model.language_model.layers.0.mlp.experts.gate_up_proj", | |
| "max_abs_diff": 0.00054931640625, | |
| "mean_abs_diff": 2.283455069118645e-05, | |
| "pct_different": 43.73089522123337, | |
| "n_different": 234778456, | |
| "total_elements": 536870912 | |
| }, | |
| { | |
| "group": "router_gate", | |
| "name": "model.language_model.layers.0.mlp.gate.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 524288 | |
| }, | |
| { | |
| "group": "shared_expert_down", | |
| "name": "model.language_model.layers.0.mlp.shared_expert.down_proj.weight", | |
| "max_abs_diff": 0.001220703125, | |
| "mean_abs_diff": 5.048537423135713e-05, | |
| "pct_different": 76.28936767578125, | |
| "n_different": 799952, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_gate", | |
| "name": "model.language_model.layers.0.mlp.shared_expert.gate_proj.weight", | |
| "max_abs_diff": 0.0009765625, | |
| "mean_abs_diff": 5.523828076547943e-05, | |
| "pct_different": 75.71783065795898, | |
| "n_different": 793959, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_up", | |
| "name": "model.language_model.layers.0.mlp.shared_expert.up_proj.weight", | |
| "max_abs_diff": 0.000518798828125, | |
| "mean_abs_diff": 4.961384547641501e-05, | |
| "pct_different": 75.92239379882812, | |
| "n_different": 796104, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_gate_scalar", | |
| "name": "model.language_model.layers.0.mlp.shared_expert_gate.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "layernorm", | |
| "name": "model.language_model.layers.0.post_attention_layernorm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "layernorm", | |
| "name": "model.language_model.layers.1.input_layernorm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "linear_attn.A_log", | |
| "name": "model.language_model.layers.1.linear_attn.A_log", | |
| "max_abs_diff": 5.960464477539063e-08, | |
| "mean_abs_diff": 5.587935447692871e-09, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 32 | |
| }, | |
| { | |
| "group": "linear_attn.conv1d", | |
| "name": "model.language_model.layers.1.linear_attn.conv1d.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 32768 | |
| }, | |
| { | |
| "group": "linear_attn.dt_bias", | |
| "name": "model.language_model.layers.1.linear_attn.dt_bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 32 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_a", | |
| "name": "model.language_model.layers.1.linear_attn.in_proj_a.weight", | |
| "max_abs_diff": 0.00054931640625, | |
| "mean_abs_diff": 0.00010247129830531776, | |
| "pct_different": 75.79345703125, | |
| "n_different": 49672, | |
| "total_elements": 65536 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_b", | |
| "name": "model.language_model.layers.1.linear_attn.in_proj_b.weight", | |
| "max_abs_diff": 0.0006103515625, | |
| "mean_abs_diff": 7.084233220666647e-05, | |
| "pct_different": 76.63116455078125, | |
| "n_different": 50221, | |
| "total_elements": 65536 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_qkv", | |
| "name": "model.language_model.layers.1.linear_attn.in_proj_qkv.weight", | |
| "max_abs_diff": 0.0010986328125, | |
| "mean_abs_diff": 7.131708116503432e-05, | |
| "pct_different": 73.38083386421204, | |
| "n_different": 12311261, | |
| "total_elements": 16777216 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_z", | |
| "name": "model.language_model.layers.1.linear_attn.in_proj_z.weight", | |
| "max_abs_diff": 0.001129150390625, | |
| "mean_abs_diff": 7.232930511236191e-05, | |
| "pct_different": 75.3028392791748, | |
| "n_different": 6316860, | |
| "total_elements": 8388608 | |
| }, | |
| { | |
| "group": "linear_attn.norm", | |
| "name": "model.language_model.layers.1.linear_attn.norm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 128 | |
| }, | |
| { | |
| "group": "linear_attn.out_proj", | |
| "name": "model.language_model.layers.1.linear_attn.out_proj.weight", | |
| "max_abs_diff": 0.00244140625, | |
| "mean_abs_diff": 6.931642565177754e-05, | |
| "pct_different": 75.44991970062256, | |
| "n_different": 6329198, | |
| "total_elements": 8388608 | |
| }, | |
| { | |
| "group": "expert_down", | |
| "name": "model.language_model.layers.1.mlp.experts.down_proj", | |
| "max_abs_diff": 0.00067138671875, | |
| "mean_abs_diff": 4.346974310465157e-05, | |
| "pct_different": 72.96010702848434, | |
| "n_different": 195850796, | |
| "total_elements": 268435456 | |
| }, | |
| { | |
| "group": "expert_gate_up", | |
| "name": "model.language_model.layers.1.mlp.experts.gate_up_proj", | |
| "max_abs_diff": 0.00048828125, | |
| "mean_abs_diff": 4.111788803129457e-05, | |
| "pct_different": 73.11219181865454, | |
| "n_different": 392518091, | |
| "total_elements": 536870912 | |
| }, | |
| { | |
| "group": "router_gate", | |
| "name": "model.language_model.layers.1.mlp.gate.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 524288 | |
| }, | |
| { | |
| "group": "shared_expert_down", | |
| "name": "model.language_model.layers.1.mlp.shared_expert.down_proj.weight", | |
| "max_abs_diff": 0.00152587890625, | |
| "mean_abs_diff": 5.1677961891982704e-05, | |
| "pct_different": 76.61848068237305, | |
| "n_different": 803403, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_gate", | |
| "name": "model.language_model.layers.1.mlp.shared_expert.gate_proj.weight", | |
| "max_abs_diff": 0.0006256103515625, | |
| "mean_abs_diff": 5.2743715059477836e-05, | |
| "pct_different": 76.08728408813477, | |
| "n_different": 797833, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_up", | |
| "name": "model.language_model.layers.1.mlp.shared_expert.up_proj.weight", | |
| "max_abs_diff": 0.00067138671875, | |
| "mean_abs_diff": 5.0087764975614846e-05, | |
| "pct_different": 75.72908401489258, | |
| "n_different": 794077, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_gate_scalar", | |
| "name": "model.language_model.layers.1.mlp.shared_expert_gate.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "layernorm", | |
| "name": "model.language_model.layers.1.post_attention_layernorm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "layernorm", | |
| "name": "model.language_model.layers.10.input_layernorm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "linear_attn.A_log", | |
| "name": "model.language_model.layers.10.linear_attn.A_log", | |
| "max_abs_diff": 5.960464477539063e-08, | |
| "mean_abs_diff": 6.984919309616089e-09, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 32 | |
| }, | |
| { | |
| "group": "linear_attn.conv1d", | |
| "name": "model.language_model.layers.10.linear_attn.conv1d.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 32768 | |
| }, | |
| { | |
| "group": "linear_attn.dt_bias", | |
| "name": "model.language_model.layers.10.linear_attn.dt_bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 32 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_a", | |
| "name": "model.language_model.layers.10.linear_attn.in_proj_a.weight", | |
| "max_abs_diff": 0.00091552734375, | |
| "mean_abs_diff": 8.147383050527424e-05, | |
| "pct_different": 77.166748046875, | |
| "n_different": 50572, | |
| "total_elements": 65536 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_b", | |
| "name": "model.language_model.layers.10.linear_attn.in_proj_b.weight", | |
| "max_abs_diff": 0.00054931640625, | |
| "mean_abs_diff": 6.659454084001482e-05, | |
| "pct_different": 77.89306640625, | |
| "n_different": 51048, | |
| "total_elements": 65536 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_qkv", | |
| "name": "model.language_model.layers.10.linear_attn.in_proj_qkv.weight", | |
| "max_abs_diff": 0.0009765625, | |
| "mean_abs_diff": 7.837942393962294e-05, | |
| "pct_different": 76.2194037437439, | |
| "n_different": 12787494, | |
| "total_elements": 16777216 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_z", | |
| "name": "model.language_model.layers.10.linear_attn.in_proj_z.weight", | |
| "max_abs_diff": 0.00115966796875, | |
| "mean_abs_diff": 6.460334407165647e-05, | |
| "pct_different": 75.69509744644165, | |
| "n_different": 6349765, | |
| "total_elements": 8388608 | |
| }, | |
| { | |
| "group": "linear_attn.norm", | |
| "name": "model.language_model.layers.10.linear_attn.norm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 128 | |
| }, | |
| { | |
| "group": "linear_attn.out_proj", | |
| "name": "model.language_model.layers.10.linear_attn.out_proj.weight", | |
| "max_abs_diff": 0.0032501220703125, | |
| "mean_abs_diff": 6.71583111397922e-05, | |
| "pct_different": 75.81313848495483, | |
| "n_different": 6359667, | |
| "total_elements": 8388608 | |
| }, | |
| { | |
| "group": "expert_down", | |
| "name": "model.language_model.layers.10.mlp.experts.down_proj", | |
| "max_abs_diff": 0.0009765625, | |
| "mean_abs_diff": 4.468904808163643e-05, | |
| "pct_different": 74.31734949350357, | |
| "n_different": 199494116, | |
| "total_elements": 268435456 | |
| }, | |
| { | |
| "group": "expert_gate_up", | |
| "name": "model.language_model.layers.10.mlp.experts.gate_up_proj", | |
| "max_abs_diff": 0.0009765625, | |
| "mean_abs_diff": 4.599253225023858e-05, | |
| "pct_different": 74.45449084043503, | |
| "n_different": 399724504, | |
| "total_elements": 536870912 | |
| }, | |
| { | |
| "group": "router_gate", | |
| "name": "model.language_model.layers.10.mlp.gate.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 524288 | |
| }, | |
| { | |
| "group": "shared_expert_down", | |
| "name": "model.language_model.layers.10.mlp.shared_expert.down_proj.weight", | |
| "max_abs_diff": 0.001953125, | |
| "mean_abs_diff": 4.9545975343789905e-05, | |
| "pct_different": 77.01168060302734, | |
| "n_different": 807526, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_gate", | |
| "name": "model.language_model.layers.10.mlp.shared_expert.gate_proj.weight", | |
| "max_abs_diff": 0.00054931640625, | |
| "mean_abs_diff": 4.3953972635790706e-05, | |
| "pct_different": 77.45733261108398, | |
| "n_different": 812199, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_up", | |
| "name": "model.language_model.layers.10.mlp.shared_expert.up_proj.weight", | |
| "max_abs_diff": 0.00109100341796875, | |
| "mean_abs_diff": 5.314434747560881e-05, | |
| "pct_different": 77.25858688354492, | |
| "n_different": 810115, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_gate_scalar", | |
| "name": "model.language_model.layers.10.mlp.shared_expert_gate.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "layernorm", | |
| "name": "model.language_model.layers.10.post_attention_layernorm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "layernorm", | |
| "name": "model.language_model.layers.11.input_layernorm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "expert_down", | |
| "name": "model.language_model.layers.11.mlp.experts.down_proj", | |
| "max_abs_diff": 0.0009765625, | |
| "mean_abs_diff": 4.479318158701062e-05, | |
| "pct_different": 74.24698136746883, | |
| "n_different": 199305223, | |
| "total_elements": 268435456 | |
| }, | |
| { | |
| "group": "expert_gate_up", | |
| "name": "model.language_model.layers.11.mlp.experts.gate_up_proj", | |
| "max_abs_diff": 0.0009765625, | |
| "mean_abs_diff": 4.596764483721927e-05, | |
| "pct_different": 74.33178629726171, | |
| "n_different": 399065739, | |
| "total_elements": 536870912 | |
| }, | |
| { | |
| "group": "router_gate", | |
| "name": "model.language_model.layers.11.mlp.gate.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 524288 | |
| }, | |
| { | |
| "group": "shared_expert_down", | |
| "name": "model.language_model.layers.11.mlp.shared_expert.down_proj.weight", | |
| "max_abs_diff": 0.001953125, | |
| "mean_abs_diff": 4.926586916553788e-05, | |
| "pct_different": 76.67951583862305, | |
| "n_different": 804043, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_gate", | |
| "name": "model.language_model.layers.11.mlp.shared_expert.gate_proj.weight", | |
| "max_abs_diff": 0.0006103515625, | |
| "mean_abs_diff": 4.042325235786848e-05, | |
| "pct_different": 76.7068862915039, | |
| "n_different": 804330, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_up", | |
| "name": "model.language_model.layers.11.mlp.shared_expert.up_proj.weight", | |
| "max_abs_diff": 0.0009765625, | |
| "mean_abs_diff": 5.266155494609848e-05, | |
| "pct_different": 77.04896926879883, | |
| "n_different": 807917, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_gate_scalar", | |
| "name": "model.language_model.layers.11.mlp.shared_expert_gate.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "layernorm", | |
| "name": "model.language_model.layers.11.post_attention_layernorm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "self_attn.k_norm", | |
| "name": "model.language_model.layers.11.self_attn.k_norm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 256 | |
| }, | |
| { | |
| "group": "self_attn.k_proj", | |
| "name": "model.language_model.layers.11.self_attn.k_proj.weight", | |
| "max_abs_diff": 0.000823974609375, | |
| "mean_abs_diff": 7.404887583106756e-05, | |
| "pct_different": 76.78871154785156, | |
| "n_different": 805188, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "self_attn.o_proj", | |
| "name": "model.language_model.layers.11.self_attn.o_proj.weight", | |
| "max_abs_diff": 0.0023193359375, | |
| "mean_abs_diff": 6.212470907485113e-05, | |
| "pct_different": 75.56613683700562, | |
| "n_different": 6338947, | |
| "total_elements": 8388608 | |
| }, | |
| { | |
| "group": "self_attn.q_norm", | |
| "name": "model.language_model.layers.11.self_attn.q_norm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 256 | |
| }, | |
| { | |
| "group": "self_attn.q_proj", | |
| "name": "model.language_model.layers.11.self_attn.q_proj.weight", | |
| "max_abs_diff": 0.00098419189453125, | |
| "mean_abs_diff": 7.024329534033313e-05, | |
| "pct_different": 75.72547793388367, | |
| "n_different": 12704627, | |
| "total_elements": 16777216 | |
| }, | |
| { | |
| "group": "self_attn.v_proj", | |
| "name": "model.language_model.layers.11.self_attn.v_proj.weight", | |
| "max_abs_diff": 0.00067138671875, | |
| "mean_abs_diff": 6.010285869706422e-05, | |
| "pct_different": 77.16293334960938, | |
| "n_different": 809112, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "layernorm", | |
| "name": "model.language_model.layers.12.input_layernorm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "linear_attn.A_log", | |
| "name": "model.language_model.layers.12.linear_attn.A_log", | |
| "max_abs_diff": 2.60770320892334e-08, | |
| "mean_abs_diff": 2.0372681319713593e-09, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 32 | |
| }, | |
| { | |
| "group": "linear_attn.conv1d", | |
| "name": "model.language_model.layers.12.linear_attn.conv1d.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 32768 | |
| }, | |
| { | |
| "group": "linear_attn.dt_bias", | |
| "name": "model.language_model.layers.12.linear_attn.dt_bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 32 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_a", | |
| "name": "model.language_model.layers.12.linear_attn.in_proj_a.weight", | |
| "max_abs_diff": 0.000732421875, | |
| "mean_abs_diff": 8.491073094774038e-05, | |
| "pct_different": 77.5177001953125, | |
| "n_different": 50802, | |
| "total_elements": 65536 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_b", | |
| "name": "model.language_model.layers.12.linear_attn.in_proj_b.weight", | |
| "max_abs_diff": 0.00067138671875, | |
| "mean_abs_diff": 6.888517236802727e-05, | |
| "pct_different": 78.94287109375, | |
| "n_different": 51736, | |
| "total_elements": 65536 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_qkv", | |
| "name": "model.language_model.layers.12.linear_attn.in_proj_qkv.weight", | |
| "max_abs_diff": 0.0010986328125, | |
| "mean_abs_diff": 7.734443352092057e-05, | |
| "pct_different": 76.01020932197571, | |
| "n_different": 12752397, | |
| "total_elements": 16777216 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_z", | |
| "name": "model.language_model.layers.12.linear_attn.in_proj_z.weight", | |
| "max_abs_diff": 0.0009765625, | |
| "mean_abs_diff": 6.149078399175778e-05, | |
| "pct_different": 75.670325756073, | |
| "n_different": 6347687, | |
| "total_elements": 8388608 | |
| }, | |
| { | |
| "group": "linear_attn.norm", | |
| "name": "model.language_model.layers.12.linear_attn.norm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 128 | |
| }, | |
| { | |
| "group": "linear_attn.out_proj", | |
| "name": "model.language_model.layers.12.linear_attn.out_proj.weight", | |
| "max_abs_diff": 0.0029296875, | |
| "mean_abs_diff": 6.728764128638431e-05, | |
| "pct_different": 75.66553354263306, | |
| "n_different": 6347285, | |
| "total_elements": 8388608 | |
| }, | |
| { | |
| "group": "expert_down", | |
| "name": "model.language_model.layers.12.mlp.experts.down_proj", | |
| "max_abs_diff": 0.0009765625, | |
| "mean_abs_diff": 4.417969466885552e-05, | |
| "pct_different": 74.31587390601635, | |
| "n_different": 199490155, | |
| "total_elements": 268435456 | |
| }, | |
| { | |
| "group": "expert_gate_up", | |
| "name": "model.language_model.layers.12.mlp.experts.gate_up_proj", | |
| "max_abs_diff": 0.000823974609375, | |
| "mean_abs_diff": 4.582593828672543e-05, | |
| "pct_different": 74.47806298732758, | |
| "n_different": 399851056, | |
| "total_elements": 536870912 | |
| }, | |
| { | |
| "group": "router_gate", | |
| "name": "model.language_model.layers.12.mlp.gate.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 524288 | |
| }, | |
| { | |
| "group": "shared_expert_down", | |
| "name": "model.language_model.layers.12.mlp.shared_expert.down_proj.weight", | |
| "max_abs_diff": 0.001953125, | |
| "mean_abs_diff": 4.871177225140855e-05, | |
| "pct_different": 76.48048400878906, | |
| "n_different": 801956, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_gate", | |
| "name": "model.language_model.layers.12.mlp.shared_expert.gate_proj.weight", | |
| "max_abs_diff": 0.0006866455078125, | |
| "mean_abs_diff": 4.3372703657951206e-05, | |
| "pct_different": 76.93452835083008, | |
| "n_different": 806717, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_up", | |
| "name": "model.language_model.layers.12.mlp.shared_expert.up_proj.weight", | |
| "max_abs_diff": 0.0008544921875, | |
| "mean_abs_diff": 5.3502088121604174e-05, | |
| "pct_different": 76.98135375976562, | |
| "n_different": 807208, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_gate_scalar", | |
| "name": "model.language_model.layers.12.mlp.shared_expert_gate.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "layernorm", | |
| "name": "model.language_model.layers.12.post_attention_layernorm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "layernorm", | |
| "name": "model.language_model.layers.13.input_layernorm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "linear_attn.A_log", | |
| "name": "model.language_model.layers.13.linear_attn.A_log", | |
| "max_abs_diff": 5.960464477539063e-08, | |
| "mean_abs_diff": 3.725290298461914e-09, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 32 | |
| }, | |
| { | |
| "group": "linear_attn.conv1d", | |
| "name": "model.language_model.layers.13.linear_attn.conv1d.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 32768 | |
| }, | |
| { | |
| "group": "linear_attn.dt_bias", | |
| "name": "model.language_model.layers.13.linear_attn.dt_bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 32 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_a", | |
| "name": "model.language_model.layers.13.linear_attn.in_proj_a.weight", | |
| "max_abs_diff": 0.000732421875, | |
| "mean_abs_diff": 8.881848771125078e-05, | |
| "pct_different": 77.23846435546875, | |
| "n_different": 50619, | |
| "total_elements": 65536 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_b", | |
| "name": "model.language_model.layers.13.linear_attn.in_proj_b.weight", | |
| "max_abs_diff": 0.00054931640625, | |
| "mean_abs_diff": 7.101592927938327e-05, | |
| "pct_different": 78.8909912109375, | |
| "n_different": 51702, | |
| "total_elements": 65536 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_qkv", | |
| "name": "model.language_model.layers.13.linear_attn.in_proj_qkv.weight", | |
| "max_abs_diff": 0.001708984375, | |
| "mean_abs_diff": 7.823992928024381e-05, | |
| "pct_different": 75.72848200798035, | |
| "n_different": 12705131, | |
| "total_elements": 16777216 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_z", | |
| "name": "model.language_model.layers.13.linear_attn.in_proj_z.weight", | |
| "max_abs_diff": 0.0013427734375, | |
| "mean_abs_diff": 6.614800804527476e-05, | |
| "pct_different": 75.77245235443115, | |
| "n_different": 6356254, | |
| "total_elements": 8388608 | |
| }, | |
| { | |
| "group": "linear_attn.norm", | |
| "name": "model.language_model.layers.13.linear_attn.norm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 128 | |
| }, | |
| { | |
| "group": "linear_attn.out_proj", | |
| "name": "model.language_model.layers.13.linear_attn.out_proj.weight", | |
| "max_abs_diff": 0.0029296875, | |
| "mean_abs_diff": 6.55348485452123e-05, | |
| "pct_different": 75.51208734512329, | |
| "n_different": 6334413, | |
| "total_elements": 8388608 | |
| }, | |
| { | |
| "group": "expert_down", | |
| "name": "model.language_model.layers.13.mlp.experts.down_proj", | |
| "max_abs_diff": 0.0521240234375, | |
| "mean_abs_diff": 0.00016416661674156785, | |
| "pct_different": 84.88881774246693, | |
| "n_different": 227871685, | |
| "total_elements": 268435456 | |
| }, | |
| { | |
| "group": "expert_gate_up", | |
| "name": "model.language_model.layers.13.mlp.experts.gate_up_proj", | |
| "max_abs_diff": 0.0009765625, | |
| "mean_abs_diff": 4.592622281052172e-05, | |
| "pct_different": 74.48321487754583, | |
| "n_different": 399878715, | |
| "total_elements": 536870912 | |
| }, | |
| { | |
| "group": "router_gate", | |
| "name": "model.language_model.layers.13.mlp.gate.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 524288 | |
| }, | |
| { | |
| "group": "shared_expert_down", | |
| "name": "model.language_model.layers.13.mlp.shared_expert.down_proj.weight", | |
| "max_abs_diff": 0.013936042785644531, | |
| "mean_abs_diff": 0.00018142559565603733, | |
| "pct_different": 86.73515319824219, | |
| "n_different": 909484, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_gate", | |
| "name": "model.language_model.layers.13.mlp.shared_expert.gate_proj.weight", | |
| "max_abs_diff": 0.01434326171875, | |
| "mean_abs_diff": 0.0002482789568603039, | |
| "pct_different": 91.60451889038086, | |
| "n_different": 960543, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_up", | |
| "name": "model.language_model.layers.13.mlp.shared_expert.up_proj.weight", | |
| "max_abs_diff": 0.017669677734375, | |
| "mean_abs_diff": 0.000301205349387601, | |
| "pct_different": 91.26052856445312, | |
| "n_different": 956936, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_gate_scalar", | |
| "name": "model.language_model.layers.13.mlp.shared_expert_gate.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "layernorm", | |
| "name": "model.language_model.layers.13.post_attention_layernorm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "layernorm", | |
| "name": "model.language_model.layers.14.input_layernorm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "linear_attn.A_log", | |
| "name": "model.language_model.layers.14.linear_attn.A_log", | |
| "max_abs_diff": 4.470348358154297e-08, | |
| "mean_abs_diff": 6.05359673500061e-09, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 32 | |
| }, | |
| { | |
| "group": "linear_attn.conv1d", | |
| "name": "model.language_model.layers.14.linear_attn.conv1d.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 32768 | |
| }, | |
| { | |
| "group": "linear_attn.dt_bias", | |
| "name": "model.language_model.layers.14.linear_attn.dt_bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 32 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_a", | |
| "name": "model.language_model.layers.14.linear_attn.in_proj_a.weight", | |
| "max_abs_diff": 0.00067138671875, | |
| "mean_abs_diff": 9.136922017205507e-05, | |
| "pct_different": 76.93939208984375, | |
| "n_different": 50423, | |
| "total_elements": 65536 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_b", | |
| "name": "model.language_model.layers.14.linear_attn.in_proj_b.weight", | |
| "max_abs_diff": 0.0006103515625, | |
| "mean_abs_diff": 7.59142785682343e-05, | |
| "pct_different": 79.17022705078125, | |
| "n_different": 51885, | |
| "total_elements": 65536 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_qkv", | |
| "name": "model.language_model.layers.14.linear_attn.in_proj_qkv.weight", | |
| "max_abs_diff": 0.00146484375, | |
| "mean_abs_diff": 8.208391955122352e-05, | |
| "pct_different": 76.49259567260742, | |
| "n_different": 12833328, | |
| "total_elements": 16777216 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_z", | |
| "name": "model.language_model.layers.14.linear_attn.in_proj_z.weight", | |
| "max_abs_diff": 0.00128173828125, | |
| "mean_abs_diff": 5.960601265542209e-05, | |
| "pct_different": 75.62425136566162, | |
| "n_different": 6343822, | |
| "total_elements": 8388608 | |
| }, | |
| { | |
| "group": "linear_attn.norm", | |
| "name": "model.language_model.layers.14.linear_attn.norm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 128 | |
| }, | |
| { | |
| "group": "linear_attn.out_proj", | |
| "name": "model.language_model.layers.14.linear_attn.out_proj.weight", | |
| "max_abs_diff": 0.0029296875, | |
| "mean_abs_diff": 6.70628432999365e-05, | |
| "pct_different": 75.70273876190186, | |
| "n_different": 6350406, | |
| "total_elements": 8388608 | |
| }, | |
| { | |
| "group": "expert_down", | |
| "name": "model.language_model.layers.14.mlp.experts.down_proj", | |
| "max_abs_diff": 0.04425048828125, | |
| "mean_abs_diff": 0.00016297487309202552, | |
| "pct_different": 84.52954217791557, | |
| "n_different": 226907262, | |
| "total_elements": 268435456 | |
| }, | |
| { | |
| "group": "expert_gate_up", | |
| "name": "model.language_model.layers.14.mlp.experts.gate_up_proj", | |
| "max_abs_diff": 0.000823974609375, | |
| "mean_abs_diff": 4.612559496308677e-05, | |
| "pct_different": 74.47763811796904, | |
| "n_different": 399848775, | |
| "total_elements": 536870912 | |
| }, | |
| { | |
| "group": "router_gate", | |
| "name": "model.language_model.layers.14.mlp.gate.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 524288 | |
| }, | |
| { | |
| "group": "shared_expert_down", | |
| "name": "model.language_model.layers.14.mlp.shared_expert.down_proj.weight", | |
| "max_abs_diff": 0.0081787109375, | |
| "mean_abs_diff": 0.0001611674961168319, | |
| "pct_different": 85.93358993530273, | |
| "n_different": 901079, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_gate", | |
| "name": "model.language_model.layers.14.mlp.shared_expert.gate_proj.weight", | |
| "max_abs_diff": 0.021148681640625, | |
| "mean_abs_diff": 0.00024421149282716215, | |
| "pct_different": 92.06600189208984, | |
| "n_different": 965382, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_up", | |
| "name": "model.language_model.layers.14.mlp.shared_expert.up_proj.weight", | |
| "max_abs_diff": 0.020172119140625, | |
| "mean_abs_diff": 0.000316268065944314, | |
| "pct_different": 92.00611114501953, | |
| "n_different": 964754, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_gate_scalar", | |
| "name": "model.language_model.layers.14.mlp.shared_expert_gate.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "layernorm", | |
| "name": "model.language_model.layers.14.post_attention_layernorm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "layernorm", | |
| "name": "model.language_model.layers.15.input_layernorm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "expert_down", | |
| "name": "model.language_model.layers.15.mlp.experts.down_proj", | |
| "max_abs_diff": 0.0494384765625, | |
| "mean_abs_diff": 0.0001675536041148007, | |
| "pct_different": 84.93725061416626, | |
| "n_different": 228001696, | |
| "total_elements": 268435456 | |
| }, | |
| { | |
| "group": "expert_gate_up", | |
| "name": "model.language_model.layers.15.mlp.experts.gate_up_proj", | |
| "max_abs_diff": 0.000732421875, | |
| "mean_abs_diff": 4.624357825377956e-05, | |
| "pct_different": 74.51016400009394, | |
| "n_different": 400023397, | |
| "total_elements": 536870912 | |
| }, | |
| { | |
| "group": "router_gate", | |
| "name": "model.language_model.layers.15.mlp.gate.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 524288 | |
| }, | |
| { | |
| "group": "shared_expert_down", | |
| "name": "model.language_model.layers.15.mlp.shared_expert.down_proj.weight", | |
| "max_abs_diff": 0.0068206787109375, | |
| "mean_abs_diff": 0.00018666769028641284, | |
| "pct_different": 85.87207794189453, | |
| "n_different": 900434, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_gate", | |
| "name": "model.language_model.layers.15.mlp.shared_expert.gate_proj.weight", | |
| "max_abs_diff": 0.014892578125, | |
| "mean_abs_diff": 0.0002611585659906268, | |
| "pct_different": 92.0541763305664, | |
| "n_different": 965258, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_up", | |
| "name": "model.language_model.layers.15.mlp.shared_expert.up_proj.weight", | |
| "max_abs_diff": 0.017822265625, | |
| "mean_abs_diff": 0.0003065799246542156, | |
| "pct_different": 91.16964340209961, | |
| "n_different": 955983, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_gate_scalar", | |
| "name": "model.language_model.layers.15.mlp.shared_expert_gate.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "layernorm", | |
| "name": "model.language_model.layers.15.post_attention_layernorm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "self_attn.k_norm", | |
| "name": "model.language_model.layers.15.self_attn.k_norm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 256 | |
| }, | |
| { | |
| "group": "self_attn.k_proj", | |
| "name": "model.language_model.layers.15.self_attn.k_proj.weight", | |
| "max_abs_diff": 0.000762939453125, | |
| "mean_abs_diff": 7.148747681640089e-05, | |
| "pct_different": 75.90646743774414, | |
| "n_different": 795937, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "self_attn.o_proj", | |
| "name": "model.language_model.layers.15.self_attn.o_proj.weight", | |
| "max_abs_diff": 0.0203857421875, | |
| "mean_abs_diff": 0.00028440827736631036, | |
| "pct_different": 86.1541748046875, | |
| "n_different": 7227136, | |
| "total_elements": 8388608 | |
| }, | |
| { | |
| "group": "self_attn.q_norm", | |
| "name": "model.language_model.layers.15.self_attn.q_norm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 256 | |
| }, | |
| { | |
| "group": "self_attn.q_proj", | |
| "name": "model.language_model.layers.15.self_attn.q_proj.weight", | |
| "max_abs_diff": 0.0009765625, | |
| "mean_abs_diff": 6.535403372254223e-05, | |
| "pct_different": 75.44593811035156, | |
| "n_different": 12657728, | |
| "total_elements": 16777216 | |
| }, | |
| { | |
| "group": "self_attn.v_proj", | |
| "name": "model.language_model.layers.15.self_attn.v_proj.weight", | |
| "max_abs_diff": 0.0008392333984375, | |
| "mean_abs_diff": 6.21751241851598e-05, | |
| "pct_different": 76.96361541748047, | |
| "n_different": 807022, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "layernorm", | |
| "name": "model.language_model.layers.16.input_layernorm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "linear_attn.A_log", | |
| "name": "model.language_model.layers.16.linear_attn.A_log", | |
| "max_abs_diff": 2.60770320892334e-08, | |
| "mean_abs_diff": 1.7462298274040222e-09, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 32 | |
| }, | |
| { | |
| "group": "linear_attn.conv1d", | |
| "name": "model.language_model.layers.16.linear_attn.conv1d.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 32768 | |
| }, | |
| { | |
| "group": "linear_attn.dt_bias", | |
| "name": "model.language_model.layers.16.linear_attn.dt_bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 32 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_a", | |
| "name": "model.language_model.layers.16.linear_attn.in_proj_a.weight", | |
| "max_abs_diff": 0.0006103515625, | |
| "mean_abs_diff": 8.392307790927589e-05, | |
| "pct_different": 77.79388427734375, | |
| "n_different": 50983, | |
| "total_elements": 65536 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_b", | |
| "name": "model.language_model.layers.16.linear_attn.in_proj_b.weight", | |
| "max_abs_diff": 0.00048828125, | |
| "mean_abs_diff": 7.065702811814845e-05, | |
| "pct_different": 79.93927001953125, | |
| "n_different": 52389, | |
| "total_elements": 65536 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_qkv", | |
| "name": "model.language_model.layers.16.linear_attn.in_proj_qkv.weight", | |
| "max_abs_diff": 0.00152587890625, | |
| "mean_abs_diff": 7.918636401882395e-05, | |
| "pct_different": 76.08726024627686, | |
| "n_different": 12765324, | |
| "total_elements": 16777216 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_z", | |
| "name": "model.language_model.layers.16.linear_attn.in_proj_z.weight", | |
| "max_abs_diff": 0.000946044921875, | |
| "mean_abs_diff": 6.070804738556035e-05, | |
| "pct_different": 75.6017804145813, | |
| "n_different": 6341937, | |
| "total_elements": 8388608 | |
| }, | |
| { | |
| "group": "linear_attn.norm", | |
| "name": "model.language_model.layers.16.linear_attn.norm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 128 | |
| }, | |
| { | |
| "group": "linear_attn.out_proj", | |
| "name": "model.language_model.layers.16.linear_attn.out_proj.weight", | |
| "max_abs_diff": 0.0313262939453125, | |
| "mean_abs_diff": 0.00042332825250923634, | |
| "pct_different": 87.5267505645752, | |
| "n_different": 7342276, | |
| "total_elements": 8388608 | |
| }, | |
| { | |
| "group": "expert_down", | |
| "name": "model.language_model.layers.16.mlp.experts.down_proj", | |
| "max_abs_diff": 0.0648193359375, | |
| "mean_abs_diff": 0.00016193953342735767, | |
| "pct_different": 84.56891477108002, | |
| "n_different": 227012952, | |
| "total_elements": 268435456 | |
| }, | |
| { | |
| "group": "expert_gate_up", | |
| "name": "model.language_model.layers.16.mlp.experts.gate_up_proj", | |
| "max_abs_diff": 0.00091552734375, | |
| "mean_abs_diff": 4.633928620023653e-05, | |
| "pct_different": 74.55773949623108, | |
| "n_different": 400278816, | |
| "total_elements": 536870912 | |
| }, | |
| { | |
| "group": "router_gate", | |
| "name": "model.language_model.layers.16.mlp.gate.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 524288 | |
| }, | |
| { | |
| "group": "shared_expert_down", | |
| "name": "model.language_model.layers.16.mlp.shared_expert.down_proj.weight", | |
| "max_abs_diff": 0.0152587890625, | |
| "mean_abs_diff": 0.0001946525153471157, | |
| "pct_different": 85.45475006103516, | |
| "n_different": 896058, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_gate", | |
| "name": "model.language_model.layers.16.mlp.shared_expert.gate_proj.weight", | |
| "max_abs_diff": 0.0164794921875, | |
| "mean_abs_diff": 0.00023676974524278194, | |
| "pct_different": 91.48035049438477, | |
| "n_different": 959241, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_up", | |
| "name": "model.language_model.layers.16.mlp.shared_expert.up_proj.weight", | |
| "max_abs_diff": 0.020782470703125, | |
| "mean_abs_diff": 0.0003119264729321003, | |
| "pct_different": 91.4423942565918, | |
| "n_different": 958843, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_gate_scalar", | |
| "name": "model.language_model.layers.16.mlp.shared_expert_gate.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "layernorm", | |
| "name": "model.language_model.layers.16.post_attention_layernorm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "layernorm", | |
| "name": "model.language_model.layers.17.input_layernorm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "linear_attn.A_log", | |
| "name": "model.language_model.layers.17.linear_attn.A_log", | |
| "max_abs_diff": 5.960464477539063e-08, | |
| "mean_abs_diff": 6.28642737865448e-09, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 32 | |
| }, | |
| { | |
| "group": "linear_attn.conv1d", | |
| "name": "model.language_model.layers.17.linear_attn.conv1d.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 32768 | |
| }, | |
| { | |
| "group": "linear_attn.dt_bias", | |
| "name": "model.language_model.layers.17.linear_attn.dt_bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 32 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_a", | |
| "name": "model.language_model.layers.17.linear_attn.in_proj_a.weight", | |
| "max_abs_diff": 0.0006103515625, | |
| "mean_abs_diff": 8.629968215245754e-05, | |
| "pct_different": 77.74810791015625, | |
| "n_different": 50953, | |
| "total_elements": 65536 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_b", | |
| "name": "model.language_model.layers.17.linear_attn.in_proj_b.weight", | |
| "max_abs_diff": 0.0006256103515625, | |
| "mean_abs_diff": 7.553433533757925e-05, | |
| "pct_different": 79.681396484375, | |
| "n_different": 52220, | |
| "total_elements": 65536 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_qkv", | |
| "name": "model.language_model.layers.17.linear_attn.in_proj_qkv.weight", | |
| "max_abs_diff": 0.001220703125, | |
| "mean_abs_diff": 7.878502947278321e-05, | |
| "pct_different": 76.20041966438293, | |
| "n_different": 12784309, | |
| "total_elements": 16777216 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_z", | |
| "name": "model.language_model.layers.17.linear_attn.in_proj_z.weight", | |
| "max_abs_diff": 0.00115966796875, | |
| "mean_abs_diff": 6.115087307989597e-05, | |
| "pct_different": 75.67152976989746, | |
| "n_different": 6347788, | |
| "total_elements": 8388608 | |
| }, | |
| { | |
| "group": "linear_attn.norm", | |
| "name": "model.language_model.layers.17.linear_attn.norm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 128 | |
| }, | |
| { | |
| "group": "linear_attn.out_proj", | |
| "name": "model.language_model.layers.17.linear_attn.out_proj.weight", | |
| "max_abs_diff": 0.0562744140625, | |
| "mean_abs_diff": 0.0004396997392177582, | |
| "pct_different": 88.25094699859619, | |
| "n_different": 7403026, | |
| "total_elements": 8388608 | |
| }, | |
| { | |
| "group": "expert_down", | |
| "name": "model.language_model.layers.17.mlp.experts.down_proj", | |
| "max_abs_diff": 0.03216552734375, | |
| "mean_abs_diff": 0.00015920473379082978, | |
| "pct_different": 84.28642712533474, | |
| "n_different": 226254655, | |
| "total_elements": 268435456 | |
| }, | |
| { | |
| "group": "expert_gate_up", | |
| "name": "model.language_model.layers.17.mlp.experts.gate_up_proj", | |
| "max_abs_diff": 0.00116729736328125, | |
| "mean_abs_diff": 4.64716213173233e-05, | |
| "pct_different": 74.56007972359657, | |
| "n_different": 400291380, | |
| "total_elements": 536870912 | |
| }, | |
| { | |
| "group": "router_gate", | |
| "name": "model.language_model.layers.17.mlp.gate.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 524288 | |
| }, | |
| { | |
| "group": "shared_expert_down", | |
| "name": "model.language_model.layers.17.mlp.shared_expert.down_proj.weight", | |
| "max_abs_diff": 0.011739730834960938, | |
| "mean_abs_diff": 0.00018112431280314922, | |
| "pct_different": 86.3363265991211, | |
| "n_different": 905302, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_gate", | |
| "name": "model.language_model.layers.17.mlp.shared_expert.gate_proj.weight", | |
| "max_abs_diff": 0.016632080078125, | |
| "mean_abs_diff": 0.0002529557386878878, | |
| "pct_different": 92.07258224487305, | |
| "n_different": 965451, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_up", | |
| "name": "model.language_model.layers.17.mlp.shared_expert.up_proj.weight", | |
| "max_abs_diff": 0.0181884765625, | |
| "mean_abs_diff": 0.00030651784618385136, | |
| "pct_different": 91.57915115356445, | |
| "n_different": 960277, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_gate_scalar", | |
| "name": "model.language_model.layers.17.mlp.shared_expert_gate.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "layernorm", | |
| "name": "model.language_model.layers.17.post_attention_layernorm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "layernorm", | |
| "name": "model.language_model.layers.18.input_layernorm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "linear_attn.A_log", | |
| "name": "model.language_model.layers.18.linear_attn.A_log", | |
| "max_abs_diff": 5.21540641784668e-08, | |
| "mean_abs_diff": 8.614733815193176e-09, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 32 | |
| }, | |
| { | |
| "group": "linear_attn.conv1d", | |
| "name": "model.language_model.layers.18.linear_attn.conv1d.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 32768 | |
| }, | |
| { | |
| "group": "linear_attn.dt_bias", | |
| "name": "model.language_model.layers.18.linear_attn.dt_bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 32 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_a", | |
| "name": "model.language_model.layers.18.linear_attn.in_proj_a.weight", | |
| "max_abs_diff": 0.0005645751953125, | |
| "mean_abs_diff": 7.99209374235943e-05, | |
| "pct_different": 77.197265625, | |
| "n_different": 50592, | |
| "total_elements": 65536 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_b", | |
| "name": "model.language_model.layers.18.linear_attn.in_proj_b.weight", | |
| "max_abs_diff": 0.00048828125, | |
| "mean_abs_diff": 7.153955812100321e-05, | |
| "pct_different": 79.58526611328125, | |
| "n_different": 52157, | |
| "total_elements": 65536 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_qkv", | |
| "name": "model.language_model.layers.18.linear_attn.in_proj_qkv.weight", | |
| "max_abs_diff": 0.0013427734375, | |
| "mean_abs_diff": 8.00086636445485e-05, | |
| "pct_different": 76.21898055076599, | |
| "n_different": 12787423, | |
| "total_elements": 16777216 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_z", | |
| "name": "model.language_model.layers.18.linear_attn.in_proj_z.weight", | |
| "max_abs_diff": 0.0012054443359375, | |
| "mean_abs_diff": 6.136030424386263e-05, | |
| "pct_different": 75.68267583847046, | |
| "n_different": 6348723, | |
| "total_elements": 8388608 | |
| }, | |
| { | |
| "group": "linear_attn.norm", | |
| "name": "model.language_model.layers.18.linear_attn.norm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 128 | |
| }, | |
| { | |
| "group": "linear_attn.out_proj", | |
| "name": "model.language_model.layers.18.linear_attn.out_proj.weight", | |
| "max_abs_diff": 0.0517578125, | |
| "mean_abs_diff": 0.00043412059312686324, | |
| "pct_different": 87.82045841217041, | |
| "n_different": 7366914, | |
| "total_elements": 8388608 | |
| }, | |
| { | |
| "group": "expert_down", | |
| "name": "model.language_model.layers.18.mlp.experts.down_proj", | |
| "max_abs_diff": 0.039794921875, | |
| "mean_abs_diff": 0.00016504473751410842, | |
| "pct_different": 84.63703133165836, | |
| "n_different": 227195801, | |
| "total_elements": 268435456 | |
| }, | |
| { | |
| "group": "expert_gate_up", | |
| "name": "model.language_model.layers.18.mlp.experts.gate_up_proj", | |
| "max_abs_diff": 0.00347900390625, | |
| "mean_abs_diff": 5.337321636034176e-05, | |
| "pct_different": 75.7801840081811, | |
| "n_different": 406841765, | |
| "total_elements": 536870912 | |
| }, | |
| { | |
| "group": "router_gate", | |
| "name": "model.language_model.layers.18.mlp.gate.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 524288 | |
| }, | |
| { | |
| "group": "shared_expert_down", | |
| "name": "model.language_model.layers.18.mlp.shared_expert.down_proj.weight", | |
| "max_abs_diff": 0.0206451416015625, | |
| "mean_abs_diff": 0.00019821515888907015, | |
| "pct_different": 86.01350784301758, | |
| "n_different": 901917, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_gate", | |
| "name": "model.language_model.layers.18.mlp.shared_expert.gate_proj.weight", | |
| "max_abs_diff": 0.01983642578125, | |
| "mean_abs_diff": 0.00026794103905558586, | |
| "pct_different": 91.8386459350586, | |
| "n_different": 962998, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_up", | |
| "name": "model.language_model.layers.18.mlp.shared_expert.up_proj.weight", | |
| "max_abs_diff": 0.0192413330078125, | |
| "mean_abs_diff": 0.00031468673842027783, | |
| "pct_different": 91.53776168823242, | |
| "n_different": 959843, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_gate_scalar", | |
| "name": "model.language_model.layers.18.mlp.shared_expert_gate.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "layernorm", | |
| "name": "model.language_model.layers.18.post_attention_layernorm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "layernorm", | |
| "name": "model.language_model.layers.19.input_layernorm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "expert_down", | |
| "name": "model.language_model.layers.19.mlp.experts.down_proj", | |
| "max_abs_diff": 0.05120849609375, | |
| "mean_abs_diff": 0.00015744150732643902, | |
| "pct_different": 84.36407223343849, | |
| "n_different": 226463082, | |
| "total_elements": 268435456 | |
| }, | |
| { | |
| "group": "expert_gate_up", | |
| "name": "model.language_model.layers.19.mlp.experts.gate_up_proj", | |
| "max_abs_diff": 0.0045166015625, | |
| "mean_abs_diff": 5.5417978728655726e-05, | |
| "pct_different": 76.07477195560932, | |
| "n_different": 408423322, | |
| "total_elements": 536870912 | |
| }, | |
| { | |
| "group": "router_gate", | |
| "name": "model.language_model.layers.19.mlp.gate.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 524288 | |
| }, | |
| { | |
| "group": "shared_expert_down", | |
| "name": "model.language_model.layers.19.mlp.shared_expert.down_proj.weight", | |
| "max_abs_diff": 0.0142364501953125, | |
| "mean_abs_diff": 0.00020512893388513476, | |
| "pct_different": 85.77165603637695, | |
| "n_different": 899381, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_gate", | |
| "name": "model.language_model.layers.19.mlp.shared_expert.gate_proj.weight", | |
| "max_abs_diff": 0.019378662109375, | |
| "mean_abs_diff": 0.00027195975417271256, | |
| "pct_different": 91.9621467590332, | |
| "n_different": 964293, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_up", | |
| "name": "model.language_model.layers.19.mlp.shared_expert.up_proj.weight", | |
| "max_abs_diff": 0.0177459716796875, | |
| "mean_abs_diff": 0.00028497763560153544, | |
| "pct_different": 90.75374603271484, | |
| "n_different": 951622, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_gate_scalar", | |
| "name": "model.language_model.layers.19.mlp.shared_expert_gate.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "layernorm", | |
| "name": "model.language_model.layers.19.post_attention_layernorm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "self_attn.k_norm", | |
| "name": "model.language_model.layers.19.self_attn.k_norm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 256 | |
| }, | |
| { | |
| "group": "self_attn.k_proj", | |
| "name": "model.language_model.layers.19.self_attn.k_proj.weight", | |
| "max_abs_diff": 0.0008544921875, | |
| "mean_abs_diff": 6.671935261692852e-05, | |
| "pct_different": 75.68769454956055, | |
| "n_different": 793643, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "self_attn.o_proj", | |
| "name": "model.language_model.layers.19.self_attn.o_proj.weight", | |
| "max_abs_diff": 0.013702392578125, | |
| "mean_abs_diff": 0.00032492561149410903, | |
| "pct_different": 86.81050539016724, | |
| "n_different": 7282193, | |
| "total_elements": 8388608 | |
| }, | |
| { | |
| "group": "self_attn.q_norm", | |
| "name": "model.language_model.layers.19.self_attn.q_norm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 256 | |
| }, | |
| { | |
| "group": "self_attn.q_proj", | |
| "name": "model.language_model.layers.19.self_attn.q_proj.weight", | |
| "max_abs_diff": 0.00152587890625, | |
| "mean_abs_diff": 6.43174716969952e-05, | |
| "pct_different": 75.52264928817749, | |
| "n_different": 12670598, | |
| "total_elements": 16777216 | |
| }, | |
| { | |
| "group": "self_attn.v_proj", | |
| "name": "model.language_model.layers.19.self_attn.v_proj.weight", | |
| "max_abs_diff": 0.000885009765625, | |
| "mean_abs_diff": 5.9716574469348416e-05, | |
| "pct_different": 76.54294967651367, | |
| "n_different": 802611, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "layernorm", | |
| "name": "model.language_model.layers.2.input_layernorm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "linear_attn.A_log", | |
| "name": "model.language_model.layers.2.linear_attn.A_log", | |
| "max_abs_diff": 5.960464477539063e-08, | |
| "mean_abs_diff": 8.934875950217247e-09, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 32 | |
| }, | |
| { | |
| "group": "linear_attn.conv1d", | |
| "name": "model.language_model.layers.2.linear_attn.conv1d.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 32768 | |
| }, | |
| { | |
| "group": "linear_attn.dt_bias", | |
| "name": "model.language_model.layers.2.linear_attn.dt_bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 32 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_a", | |
| "name": "model.language_model.layers.2.linear_attn.in_proj_a.weight", | |
| "max_abs_diff": 0.0009765625, | |
| "mean_abs_diff": 8.202709432225674e-05, | |
| "pct_different": 76.78375244140625, | |
| "n_different": 50321, | |
| "total_elements": 65536 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_b", | |
| "name": "model.language_model.layers.2.linear_attn.in_proj_b.weight", | |
| "max_abs_diff": 0.00054931640625, | |
| "mean_abs_diff": 6.774281791877002e-05, | |
| "pct_different": 77.33917236328125, | |
| "n_different": 50685, | |
| "total_elements": 65536 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_qkv", | |
| "name": "model.language_model.layers.2.linear_attn.in_proj_qkv.weight", | |
| "max_abs_diff": 0.0010223388671875, | |
| "mean_abs_diff": 7.253498188219965e-05, | |
| "pct_different": 74.60963129997253, | |
| "n_different": 12517419, | |
| "total_elements": 16777216 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_z", | |
| "name": "model.language_model.layers.2.linear_attn.in_proj_z.weight", | |
| "max_abs_diff": 0.001220703125, | |
| "mean_abs_diff": 6.867479532957077e-05, | |
| "pct_different": 75.54035186767578, | |
| "n_different": 6336784, | |
| "total_elements": 8388608 | |
| }, | |
| { | |
| "group": "linear_attn.norm", | |
| "name": "model.language_model.layers.2.linear_attn.norm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 128 | |
| }, | |
| { | |
| "group": "linear_attn.out_proj", | |
| "name": "model.language_model.layers.2.linear_attn.out_proj.weight", | |
| "max_abs_diff": 0.0025634765625, | |
| "mean_abs_diff": 6.926805508555844e-05, | |
| "pct_different": 75.59624910354614, | |
| "n_different": 6341473, | |
| "total_elements": 8388608 | |
| }, | |
| { | |
| "group": "expert_down", | |
| "name": "model.language_model.layers.2.mlp.experts.down_proj", | |
| "max_abs_diff": 0.0010986328125, | |
| "mean_abs_diff": 4.4208281906321645e-05, | |
| "pct_different": 73.6159261316061, | |
| "n_different": 197611247, | |
| "total_elements": 268435456 | |
| }, | |
| { | |
| "group": "expert_gate_up", | |
| "name": "model.language_model.layers.2.mlp.experts.gate_up_proj", | |
| "max_abs_diff": 0.00048828125, | |
| "mean_abs_diff": 4.319091749493964e-05, | |
| "pct_different": 73.7625615671277, | |
| "n_different": 396009737, | |
| "total_elements": 536870912 | |
| }, | |
| { | |
| "group": "router_gate", | |
| "name": "model.language_model.layers.2.mlp.gate.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 524288 | |
| }, | |
| { | |
| "group": "shared_expert_down", | |
| "name": "model.language_model.layers.2.mlp.shared_expert.down_proj.weight", | |
| "max_abs_diff": 0.001953125, | |
| "mean_abs_diff": 5.134729508426972e-05, | |
| "pct_different": 76.46913528442383, | |
| "n_different": 801837, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_gate", | |
| "name": "model.language_model.layers.2.mlp.shared_expert.gate_proj.weight", | |
| "max_abs_diff": 0.0008544921875, | |
| "mean_abs_diff": 5.1316088502062485e-05, | |
| "pct_different": 76.29518508911133, | |
| "n_different": 800013, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_up", | |
| "name": "model.language_model.layers.2.mlp.shared_expert.up_proj.weight", | |
| "max_abs_diff": 0.00054931640625, | |
| "mean_abs_diff": 5.095137021271512e-05, | |
| "pct_different": 76.12409591674805, | |
| "n_different": 798219, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_gate_scalar", | |
| "name": "model.language_model.layers.2.mlp.shared_expert_gate.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "layernorm", | |
| "name": "model.language_model.layers.2.post_attention_layernorm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "layernorm", | |
| "name": "model.language_model.layers.20.input_layernorm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "linear_attn.A_log", | |
| "name": "model.language_model.layers.20.linear_attn.A_log", | |
| "max_abs_diff": 5.960464477539063e-08, | |
| "mean_abs_diff": 1.0244548320770264e-08, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 32 | |
| }, | |
| { | |
| "group": "linear_attn.conv1d", | |
| "name": "model.language_model.layers.20.linear_attn.conv1d.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 32768 | |
| }, | |
| { | |
| "group": "linear_attn.dt_bias", | |
| "name": "model.language_model.layers.20.linear_attn.dt_bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 32 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_a", | |
| "name": "model.language_model.layers.20.linear_attn.in_proj_a.weight", | |
| "max_abs_diff": 0.00079345703125, | |
| "mean_abs_diff": 9.83386707957834e-05, | |
| "pct_different": 77.65045166015625, | |
| "n_different": 50889, | |
| "total_elements": 65536 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_b", | |
| "name": "model.language_model.layers.20.linear_attn.in_proj_b.weight", | |
| "max_abs_diff": 0.00054931640625, | |
| "mean_abs_diff": 7.693850784562528e-05, | |
| "pct_different": 79.76837158203125, | |
| "n_different": 52277, | |
| "total_elements": 65536 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_qkv", | |
| "name": "model.language_model.layers.20.linear_attn.in_proj_qkv.weight", | |
| "max_abs_diff": 0.00146484375, | |
| "mean_abs_diff": 9.300987585447729e-05, | |
| "pct_different": 76.40116214752197, | |
| "n_different": 12817988, | |
| "total_elements": 16777216 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_z", | |
| "name": "model.language_model.layers.20.linear_attn.in_proj_z.weight", | |
| "max_abs_diff": 0.00146484375, | |
| "mean_abs_diff": 7.48467500670813e-05, | |
| "pct_different": 76.52759552001953, | |
| "n_different": 6419600, | |
| "total_elements": 8388608 | |
| }, | |
| { | |
| "group": "linear_attn.norm", | |
| "name": "model.language_model.layers.20.linear_attn.norm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 128 | |
| }, | |
| { | |
| "group": "linear_attn.out_proj", | |
| "name": "model.language_model.layers.20.linear_attn.out_proj.weight", | |
| "max_abs_diff": 0.04833984375, | |
| "mean_abs_diff": 0.0004897484323009849, | |
| "pct_different": 87.95448541641235, | |
| "n_different": 7378157, | |
| "total_elements": 8388608 | |
| }, | |
| { | |
| "group": "expert_down", | |
| "name": "model.language_model.layers.20.mlp.experts.down_proj", | |
| "max_abs_diff": 0.031982421875, | |
| "mean_abs_diff": 0.00018058609566651285, | |
| "pct_different": 85.46429947018623, | |
| "n_different": 229416482, | |
| "total_elements": 268435456 | |
| }, | |
| { | |
| "group": "expert_gate_up", | |
| "name": "model.language_model.layers.20.mlp.experts.gate_up_proj", | |
| "max_abs_diff": 0.0046539306640625, | |
| "mean_abs_diff": 2.6304203856852837e-05, | |
| "pct_different": 41.18571151047945, | |
| "n_different": 221114105, | |
| "total_elements": 536870912 | |
| }, | |
| { | |
| "group": "router_gate", | |
| "name": "model.language_model.layers.20.mlp.gate.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 524288 | |
| }, | |
| { | |
| "group": "shared_expert_down", | |
| "name": "model.language_model.layers.20.mlp.shared_expert.down_proj.weight", | |
| "max_abs_diff": 0.0076904296875, | |
| "mean_abs_diff": 0.00015792157500982285, | |
| "pct_different": 87.2715950012207, | |
| "n_different": 915109, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_gate", | |
| "name": "model.language_model.layers.20.mlp.shared_expert.gate_proj.weight", | |
| "max_abs_diff": 0.010009765625, | |
| "mean_abs_diff": 0.00017769620171748102, | |
| "pct_different": 87.44478225708008, | |
| "n_different": 916925, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_up", | |
| "name": "model.language_model.layers.20.mlp.shared_expert.up_proj.weight", | |
| "max_abs_diff": 0.008056640625, | |
| "mean_abs_diff": 0.0001709601638140157, | |
| "pct_different": 82.69004821777344, | |
| "n_different": 867068, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_gate_scalar", | |
| "name": "model.language_model.layers.20.mlp.shared_expert_gate.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "layernorm", | |
| "name": "model.language_model.layers.20.post_attention_layernorm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "layernorm", | |
| "name": "model.language_model.layers.21.input_layernorm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "linear_attn.A_log", | |
| "name": "model.language_model.layers.21.linear_attn.A_log", | |
| "max_abs_diff": 5.960464477539063e-08, | |
| "mean_abs_diff": 1.2994860298931599e-08, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 32 | |
| }, | |
| { | |
| "group": "linear_attn.conv1d", | |
| "name": "model.language_model.layers.21.linear_attn.conv1d.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 32768 | |
| }, | |
| { | |
| "group": "linear_attn.dt_bias", | |
| "name": "model.language_model.layers.21.linear_attn.dt_bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 32 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_a", | |
| "name": "model.language_model.layers.21.linear_attn.in_proj_a.weight", | |
| "max_abs_diff": 0.00067138671875, | |
| "mean_abs_diff": 8.94919503480196e-05, | |
| "pct_different": 77.72674560546875, | |
| "n_different": 50939, | |
| "total_elements": 65536 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_b", | |
| "name": "model.language_model.layers.21.linear_attn.in_proj_b.weight", | |
| "max_abs_diff": 0.000640869140625, | |
| "mean_abs_diff": 7.325454498641193e-05, | |
| "pct_different": 78.9642333984375, | |
| "n_different": 51750, | |
| "total_elements": 65536 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_qkv", | |
| "name": "model.language_model.layers.21.linear_attn.in_proj_qkv.weight", | |
| "max_abs_diff": 0.001983642578125, | |
| "mean_abs_diff": 9.073900582734495e-05, | |
| "pct_different": 76.03468894958496, | |
| "n_different": 12756504, | |
| "total_elements": 16777216 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_z", | |
| "name": "model.language_model.layers.21.linear_attn.in_proj_z.weight", | |
| "max_abs_diff": 0.00146484375, | |
| "mean_abs_diff": 7.163381815189496e-05, | |
| "pct_different": 75.85808038711548, | |
| "n_different": 6363437, | |
| "total_elements": 8388608 | |
| }, | |
| { | |
| "group": "linear_attn.norm", | |
| "name": "model.language_model.layers.21.linear_attn.norm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 128 | |
| }, | |
| { | |
| "group": "linear_attn.out_proj", | |
| "name": "model.language_model.layers.21.linear_attn.out_proj.weight", | |
| "max_abs_diff": 0.03887939453125, | |
| "mean_abs_diff": 0.0004872928839176893, | |
| "pct_different": 87.6244306564331, | |
| "n_different": 7350470, | |
| "total_elements": 8388608 | |
| }, | |
| { | |
| "group": "expert_down", | |
| "name": "model.language_model.layers.21.mlp.experts.down_proj", | |
| "max_abs_diff": 0.037109375, | |
| "mean_abs_diff": 0.00016394685371778905, | |
| "pct_different": 84.71211567521095, | |
| "n_different": 227397354, | |
| "total_elements": 268435456 | |
| }, | |
| { | |
| "group": "expert_gate_up", | |
| "name": "model.language_model.layers.21.mlp.experts.gate_up_proj", | |
| "max_abs_diff": 0.0059356689453125, | |
| "mean_abs_diff": 6.084999768063426e-05, | |
| "pct_different": 76.90641526132822, | |
| "n_different": 412888173, | |
| "total_elements": 536870912 | |
| }, | |
| { | |
| "group": "router_gate", | |
| "name": "model.language_model.layers.21.mlp.gate.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 524288 | |
| }, | |
| { | |
| "group": "shared_expert_down", | |
| "name": "model.language_model.layers.21.mlp.shared_expert.down_proj.weight", | |
| "max_abs_diff": 0.010465621948242188, | |
| "mean_abs_diff": 0.00017318414757028222, | |
| "pct_different": 85.81399917602539, | |
| "n_different": 899825, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_gate", | |
| "name": "model.language_model.layers.21.mlp.shared_expert.gate_proj.weight", | |
| "max_abs_diff": 0.024993896484375, | |
| "mean_abs_diff": 0.0002270845725433901, | |
| "pct_different": 91.97778701782227, | |
| "n_different": 964457, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_up", | |
| "name": "model.language_model.layers.21.mlp.shared_expert.up_proj.weight", | |
| "max_abs_diff": 0.006591796875, | |
| "mean_abs_diff": 0.00014671919052489102, | |
| "pct_different": 86.84186935424805, | |
| "n_different": 910603, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_gate_scalar", | |
| "name": "model.language_model.layers.21.mlp.shared_expert_gate.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "layernorm", | |
| "name": "model.language_model.layers.21.post_attention_layernorm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "layernorm", | |
| "name": "model.language_model.layers.22.input_layernorm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "linear_attn.A_log", | |
| "name": "model.language_model.layers.22.linear_attn.A_log", | |
| "max_abs_diff": 5.960464477539063e-08, | |
| "mean_abs_diff": 8.381903171539307e-09, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 32 | |
| }, | |
| { | |
| "group": "linear_attn.conv1d", | |
| "name": "model.language_model.layers.22.linear_attn.conv1d.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 32768 | |
| }, | |
| { | |
| "group": "linear_attn.dt_bias", | |
| "name": "model.language_model.layers.22.linear_attn.dt_bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 32 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_a", | |
| "name": "model.language_model.layers.22.linear_attn.in_proj_a.weight", | |
| "max_abs_diff": 0.0006103515625, | |
| "mean_abs_diff": 8.763916412135586e-05, | |
| "pct_different": 78.04412841796875, | |
| "n_different": 51147, | |
| "total_elements": 65536 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_b", | |
| "name": "model.language_model.layers.22.linear_attn.in_proj_b.weight", | |
| "max_abs_diff": 0.00054931640625, | |
| "mean_abs_diff": 7.63551943236962e-05, | |
| "pct_different": 78.71856689453125, | |
| "n_different": 51589, | |
| "total_elements": 65536 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_qkv", | |
| "name": "model.language_model.layers.22.linear_attn.in_proj_qkv.weight", | |
| "max_abs_diff": 0.001312255859375, | |
| "mean_abs_diff": 9.141962073044851e-05, | |
| "pct_different": 76.49833559989929, | |
| "n_different": 12834291, | |
| "total_elements": 16777216 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_z", | |
| "name": "model.language_model.layers.22.linear_attn.in_proj_z.weight", | |
| "max_abs_diff": 0.001373291015625, | |
| "mean_abs_diff": 6.859651330159977e-05, | |
| "pct_different": 75.9783148765564, | |
| "n_different": 6373523, | |
| "total_elements": 8388608 | |
| }, | |
| { | |
| "group": "linear_attn.norm", | |
| "name": "model.language_model.layers.22.linear_attn.norm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 128 | |
| }, | |
| { | |
| "group": "linear_attn.out_proj", | |
| "name": "model.language_model.layers.22.linear_attn.out_proj.weight", | |
| "max_abs_diff": 0.0604705810546875, | |
| "mean_abs_diff": 0.00046970107359811664, | |
| "pct_different": 88.18075656890869, | |
| "n_different": 7397138, | |
| "total_elements": 8388608 | |
| }, | |
| { | |
| "group": "expert_down", | |
| "name": "model.language_model.layers.22.mlp.experts.down_proj", | |
| "max_abs_diff": 0.03594970703125, | |
| "mean_abs_diff": 0.00015980028547346592, | |
| "pct_different": 84.50240306556225, | |
| "n_different": 226834411, | |
| "total_elements": 268435456 | |
| }, | |
| { | |
| "group": "expert_gate_up", | |
| "name": "model.language_model.layers.22.mlp.experts.gate_up_proj", | |
| "max_abs_diff": 0.007293701171875, | |
| "mean_abs_diff": 6.297700747381896e-05, | |
| "pct_different": 77.1355478093028, | |
| "n_different": 414118319, | |
| "total_elements": 536870912 | |
| }, | |
| { | |
| "group": "router_gate", | |
| "name": "model.language_model.layers.22.mlp.gate.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 524288 | |
| }, | |
| { | |
| "group": "shared_expert_down", | |
| "name": "model.language_model.layers.22.mlp.shared_expert.down_proj.weight", | |
| "max_abs_diff": 0.010419845581054688, | |
| "mean_abs_diff": 0.0001603464625077322, | |
| "pct_different": 87.72687911987305, | |
| "n_different": 919883, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_gate", | |
| "name": "model.language_model.layers.22.mlp.shared_expert.gate_proj.weight", | |
| "max_abs_diff": 0.00958251953125, | |
| "mean_abs_diff": 0.00015206090756691992, | |
| "pct_different": 90.13290405273438, | |
| "n_different": 945112, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_up", | |
| "name": "model.language_model.layers.22.mlp.shared_expert.up_proj.weight", | |
| "max_abs_diff": 0.004852294921875, | |
| "mean_abs_diff": 9.636639879317954e-05, | |
| "pct_different": 83.59212875366211, | |
| "n_different": 876527, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_gate_scalar", | |
| "name": "model.language_model.layers.22.mlp.shared_expert_gate.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "layernorm", | |
| "name": "model.language_model.layers.22.post_attention_layernorm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "layernorm", | |
| "name": "model.language_model.layers.23.input_layernorm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "expert_down", | |
| "name": "model.language_model.layers.23.mlp.experts.down_proj", | |
| "max_abs_diff": 0.03167724609375, | |
| "mean_abs_diff": 0.0001644684816710651, | |
| "pct_different": 84.74459014832973, | |
| "n_different": 227484527, | |
| "total_elements": 268435456 | |
| }, | |
| { | |
| "group": "expert_gate_up", | |
| "name": "model.language_model.layers.23.mlp.experts.gate_up_proj", | |
| "max_abs_diff": 0.007659912109375, | |
| "mean_abs_diff": 6.675636541331187e-05, | |
| "pct_different": 77.55512706935406, | |
| "n_different": 416370918, | |
| "total_elements": 536870912 | |
| }, | |
| { | |
| "group": "router_gate", | |
| "name": "model.language_model.layers.23.mlp.gate.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 524288 | |
| }, | |
| { | |
| "group": "shared_expert_down", | |
| "name": "model.language_model.layers.23.mlp.shared_expert.down_proj.weight", | |
| "max_abs_diff": 0.007598876953125, | |
| "mean_abs_diff": 0.00017038825899362564, | |
| "pct_different": 85.12849807739258, | |
| "n_different": 892637, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_gate", | |
| "name": "model.language_model.layers.23.mlp.shared_expert.gate_proj.weight", | |
| "max_abs_diff": 0.008148193359375, | |
| "mean_abs_diff": 0.000185534474439919, | |
| "pct_different": 91.17660522460938, | |
| "n_different": 956056, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_up", | |
| "name": "model.language_model.layers.23.mlp.shared_expert.up_proj.weight", | |
| "max_abs_diff": 0.00146484375, | |
| "mean_abs_diff": 5.3698917326983064e-05, | |
| "pct_different": 77.85634994506836, | |
| "n_different": 816383, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_gate_scalar", | |
| "name": "model.language_model.layers.23.mlp.shared_expert_gate.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "layernorm", | |
| "name": "model.language_model.layers.23.post_attention_layernorm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "self_attn.k_norm", | |
| "name": "model.language_model.layers.23.self_attn.k_norm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 256 | |
| }, | |
| { | |
| "group": "self_attn.k_proj", | |
| "name": "model.language_model.layers.23.self_attn.k_proj.weight", | |
| "max_abs_diff": 0.0009765625, | |
| "mean_abs_diff": 8.426675776718184e-05, | |
| "pct_different": 77.13603973388672, | |
| "n_different": 808830, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "self_attn.o_proj", | |
| "name": "model.language_model.layers.23.self_attn.o_proj.weight", | |
| "max_abs_diff": 0.026458740234375, | |
| "mean_abs_diff": 0.00031689784373156726, | |
| "pct_different": 85.98548173904419, | |
| "n_different": 7212985, | |
| "total_elements": 8388608 | |
| }, | |
| { | |
| "group": "self_attn.q_norm", | |
| "name": "model.language_model.layers.23.self_attn.q_norm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 256 | |
| }, | |
| { | |
| "group": "self_attn.q_proj", | |
| "name": "model.language_model.layers.23.self_attn.q_proj.weight", | |
| "max_abs_diff": 0.00146484375, | |
| "mean_abs_diff": 7.806818030076101e-05, | |
| "pct_different": 75.95778703689575, | |
| "n_different": 12743602, | |
| "total_elements": 16777216 | |
| }, | |
| { | |
| "group": "self_attn.v_proj", | |
| "name": "model.language_model.layers.23.self_attn.v_proj.weight", | |
| "max_abs_diff": 0.00091552734375, | |
| "mean_abs_diff": 6.564579962287098e-05, | |
| "pct_different": 77.5507926940918, | |
| "n_different": 813179, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "layernorm", | |
| "name": "model.language_model.layers.24.input_layernorm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "linear_attn.A_log", | |
| "name": "model.language_model.layers.24.linear_attn.A_log", | |
| "max_abs_diff": 5.960464477539063e-08, | |
| "mean_abs_diff": 3.521563485264778e-09, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 32 | |
| }, | |
| { | |
| "group": "linear_attn.conv1d", | |
| "name": "model.language_model.layers.24.linear_attn.conv1d.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 32768 | |
| }, | |
| { | |
| "group": "linear_attn.dt_bias", | |
| "name": "model.language_model.layers.24.linear_attn.dt_bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 32 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_a", | |
| "name": "model.language_model.layers.24.linear_attn.in_proj_a.weight", | |
| "max_abs_diff": 0.000732421875, | |
| "mean_abs_diff": 8.782851364230737e-05, | |
| "pct_different": 78.19061279296875, | |
| "n_different": 51243, | |
| "total_elements": 65536 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_b", | |
| "name": "model.language_model.layers.24.linear_attn.in_proj_b.weight", | |
| "max_abs_diff": 0.000732421875, | |
| "mean_abs_diff": 7.491880387533456e-05, | |
| "pct_different": 79.2083740234375, | |
| "n_different": 51910, | |
| "total_elements": 65536 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_qkv", | |
| "name": "model.language_model.layers.24.linear_attn.in_proj_qkv.weight", | |
| "max_abs_diff": 0.0013885498046875, | |
| "mean_abs_diff": 8.98265207069926e-05, | |
| "pct_different": 76.0124921798706, | |
| "n_different": 12752780, | |
| "total_elements": 16777216 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_z", | |
| "name": "model.language_model.layers.24.linear_attn.in_proj_z.weight", | |
| "max_abs_diff": 0.001220703125, | |
| "mean_abs_diff": 6.668224523309618e-05, | |
| "pct_different": 75.63703060150146, | |
| "n_different": 6344894, | |
| "total_elements": 8388608 | |
| }, | |
| { | |
| "group": "linear_attn.norm", | |
| "name": "model.language_model.layers.24.linear_attn.norm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 128 | |
| }, | |
| { | |
| "group": "linear_attn.out_proj", | |
| "name": "model.language_model.layers.24.linear_attn.out_proj.weight", | |
| "max_abs_diff": 0.05377197265625, | |
| "mean_abs_diff": 0.0004991018795408309, | |
| "pct_different": 88.16461563110352, | |
| "n_different": 7395784, | |
| "total_elements": 8388608 | |
| }, | |
| { | |
| "group": "expert_down", | |
| "name": "model.language_model.layers.24.mlp.experts.down_proj", | |
| "max_abs_diff": 0.0382080078125, | |
| "mean_abs_diff": 0.00016885525838006288, | |
| "pct_different": 84.98790711164474, | |
| "n_different": 228137676, | |
| "total_elements": 268435456 | |
| }, | |
| { | |
| "group": "expert_gate_up", | |
| "name": "model.language_model.layers.24.mlp.experts.gate_up_proj", | |
| "max_abs_diff": 0.009571075439453125, | |
| "mean_abs_diff": 7.079898932715878e-05, | |
| "pct_different": 78.11102215200663, | |
| "n_different": 419355357, | |
| "total_elements": 536870912 | |
| }, | |
| { | |
| "group": "router_gate", | |
| "name": "model.language_model.layers.24.mlp.gate.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 524288 | |
| }, | |
| { | |
| "group": "shared_expert_down", | |
| "name": "model.language_model.layers.24.mlp.shared_expert.down_proj.weight", | |
| "max_abs_diff": 0.0123138427734375, | |
| "mean_abs_diff": 0.0001695990067673847, | |
| "pct_different": 85.58683395385742, | |
| "n_different": 897443, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_gate", | |
| "name": "model.language_model.layers.24.mlp.shared_expert.gate_proj.weight", | |
| "max_abs_diff": 0.0130767822265625, | |
| "mean_abs_diff": 0.00018669612472876906, | |
| "pct_different": 91.02592468261719, | |
| "n_different": 954476, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_up", | |
| "name": "model.language_model.layers.24.mlp.shared_expert.up_proj.weight", | |
| "max_abs_diff": 0.001514434814453125, | |
| "mean_abs_diff": 6.762769044144079e-05, | |
| "pct_different": 79.97007369995117, | |
| "n_different": 838547, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_gate_scalar", | |
| "name": "model.language_model.layers.24.mlp.shared_expert_gate.weight", | |
| "max_abs_diff": 0.0050048828125, | |
| "mean_abs_diff": 0.0004709986096713692, | |
| "pct_different": 98.6328125, | |
| "n_different": 2020, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "layernorm", | |
| "name": "model.language_model.layers.24.post_attention_layernorm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "layernorm", | |
| "name": "model.language_model.layers.25.input_layernorm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "linear_attn.A_log", | |
| "name": "model.language_model.layers.25.linear_attn.A_log", | |
| "max_abs_diff": 5.960464477539063e-08, | |
| "mean_abs_diff": 6.28642737865448e-09, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 32 | |
| }, | |
| { | |
| "group": "linear_attn.conv1d", | |
| "name": "model.language_model.layers.25.linear_attn.conv1d.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 32768 | |
| }, | |
| { | |
| "group": "linear_attn.dt_bias", | |
| "name": "model.language_model.layers.25.linear_attn.dt_bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 32 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_a", | |
| "name": "model.language_model.layers.25.linear_attn.in_proj_a.weight", | |
| "max_abs_diff": 0.000762939453125, | |
| "mean_abs_diff": 9.0843575890176e-05, | |
| "pct_different": 77.24456787109375, | |
| "n_different": 50623, | |
| "total_elements": 65536 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_b", | |
| "name": "model.language_model.layers.25.linear_attn.in_proj_b.weight", | |
| "max_abs_diff": 0.0006103515625, | |
| "mean_abs_diff": 7.25840509403497e-05, | |
| "pct_different": 79.03594970703125, | |
| "n_different": 51797, | |
| "total_elements": 65536 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_qkv", | |
| "name": "model.language_model.layers.25.linear_attn.in_proj_qkv.weight", | |
| "max_abs_diff": 0.001220703125, | |
| "mean_abs_diff": 9.110737300943583e-05, | |
| "pct_different": 75.72745680809021, | |
| "n_different": 12704959, | |
| "total_elements": 16777216 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_z", | |
| "name": "model.language_model.layers.25.linear_attn.in_proj_z.weight", | |
| "max_abs_diff": 0.001220703125, | |
| "mean_abs_diff": 7.248125621117651e-05, | |
| "pct_different": 75.85275173187256, | |
| "n_different": 6362990, | |
| "total_elements": 8388608 | |
| }, | |
| { | |
| "group": "linear_attn.norm", | |
| "name": "model.language_model.layers.25.linear_attn.norm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 128 | |
| }, | |
| { | |
| "group": "linear_attn.out_proj", | |
| "name": "model.language_model.layers.25.linear_attn.out_proj.weight", | |
| "max_abs_diff": 0.0540771484375, | |
| "mean_abs_diff": 0.000518718094099313, | |
| "pct_different": 87.50219345092773, | |
| "n_different": 7340216, | |
| "total_elements": 8388608 | |
| }, | |
| { | |
| "group": "expert_down", | |
| "name": "model.language_model.layers.25.mlp.experts.down_proj", | |
| "max_abs_diff": 0.0317840576171875, | |
| "mean_abs_diff": 0.0001698803826002404, | |
| "pct_different": 84.8406694829464, | |
| "n_different": 227742438, | |
| "total_elements": 268435456 | |
| }, | |
| { | |
| "group": "expert_gate_up", | |
| "name": "model.language_model.layers.25.mlp.experts.gate_up_proj", | |
| "max_abs_diff": 0.01165771484375, | |
| "mean_abs_diff": 7.39435781724751e-05, | |
| "pct_different": 78.39019447565079, | |
| "n_different": 420854152, | |
| "total_elements": 536870912 | |
| }, | |
| { | |
| "group": "router_gate", | |
| "name": "model.language_model.layers.25.mlp.gate.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 524288 | |
| }, | |
| { | |
| "group": "shared_expert_down", | |
| "name": "model.language_model.layers.25.mlp.shared_expert.down_proj.weight", | |
| "max_abs_diff": 0.012664794921875, | |
| "mean_abs_diff": 0.0001598140806891024, | |
| "pct_different": 87.34731674194336, | |
| "n_different": 915903, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_gate", | |
| "name": "model.language_model.layers.25.mlp.shared_expert.gate_proj.weight", | |
| "max_abs_diff": 0.010009765625, | |
| "mean_abs_diff": 0.0001573283807374537, | |
| "pct_different": 89.9637222290039, | |
| "n_different": 943338, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_up", | |
| "name": "model.language_model.layers.25.mlp.shared_expert.up_proj.weight", | |
| "max_abs_diff": 0.0064697265625, | |
| "mean_abs_diff": 0.00011051571345888078, | |
| "pct_different": 84.61780548095703, | |
| "n_different": 887282, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_gate_scalar", | |
| "name": "model.language_model.layers.25.mlp.shared_expert_gate.weight", | |
| "max_abs_diff": 0.005584716796875, | |
| "mean_abs_diff": 0.0005265480722300708, | |
| "pct_different": 99.12109375, | |
| "n_different": 2030, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "layernorm", | |
| "name": "model.language_model.layers.25.post_attention_layernorm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "layernorm", | |
| "name": "model.language_model.layers.26.input_layernorm.weight", | |
| "max_abs_diff": 5.960464477539063e-08, | |
| "mean_abs_diff": 2.9103830456733704e-11, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "linear_attn.A_log", | |
| "name": "model.language_model.layers.26.linear_attn.A_log", | |
| "max_abs_diff": 5.960464477539063e-08, | |
| "mean_abs_diff": 1.0710209608078003e-08, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 32 | |
| }, | |
| { | |
| "group": "linear_attn.conv1d", | |
| "name": "model.language_model.layers.26.linear_attn.conv1d.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 32768 | |
| }, | |
| { | |
| "group": "linear_attn.dt_bias", | |
| "name": "model.language_model.layers.26.linear_attn.dt_bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 32 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_a", | |
| "name": "model.language_model.layers.26.linear_attn.in_proj_a.weight", | |
| "max_abs_diff": 0.000732421875, | |
| "mean_abs_diff": 9.177508763968945e-05, | |
| "pct_different": 77.2705078125, | |
| "n_different": 50640, | |
| "total_elements": 65536 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_b", | |
| "name": "model.language_model.layers.26.linear_attn.in_proj_b.weight", | |
| "max_abs_diff": 0.0006103515625, | |
| "mean_abs_diff": 7.805024506524205e-05, | |
| "pct_different": 79.03900146484375, | |
| "n_different": 51799, | |
| "total_elements": 65536 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_qkv", | |
| "name": "model.language_model.layers.26.linear_attn.in_proj_qkv.weight", | |
| "max_abs_diff": 0.0018310546875, | |
| "mean_abs_diff": 9.564950596541166e-05, | |
| "pct_different": 76.42055153846741, | |
| "n_different": 12821241, | |
| "total_elements": 16777216 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_z", | |
| "name": "model.language_model.layers.26.linear_attn.in_proj_z.weight", | |
| "max_abs_diff": 0.0013427734375, | |
| "mean_abs_diff": 6.661845691269264e-05, | |
| "pct_different": 75.66145658493042, | |
| "n_different": 6346943, | |
| "total_elements": 8388608 | |
| }, | |
| { | |
| "group": "linear_attn.norm", | |
| "name": "model.language_model.layers.26.linear_attn.norm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 128 | |
| }, | |
| { | |
| "group": "linear_attn.out_proj", | |
| "name": "model.language_model.layers.26.linear_attn.out_proj.weight", | |
| "max_abs_diff": 0.0650634765625, | |
| "mean_abs_diff": 0.0005140916327945888, | |
| "pct_different": 87.78066635131836, | |
| "n_different": 7363576, | |
| "total_elements": 8388608 | |
| }, | |
| { | |
| "group": "expert_down", | |
| "name": "model.language_model.layers.26.mlp.experts.down_proj", | |
| "max_abs_diff": 0.02740478515625, | |
| "mean_abs_diff": 0.00017445073171984404, | |
| "pct_different": 84.93271432816982, | |
| "n_different": 227989519, | |
| "total_elements": 268435456 | |
| }, | |
| { | |
| "group": "expert_gate_up", | |
| "name": "model.language_model.layers.26.mlp.experts.gate_up_proj", | |
| "max_abs_diff": 0.01071929931640625, | |
| "mean_abs_diff": 7.931649452075362e-05, | |
| "pct_different": 78.86200547218323, | |
| "n_different": 423387168, | |
| "total_elements": 536870912 | |
| }, | |
| { | |
| "group": "router_gate", | |
| "name": "model.language_model.layers.26.mlp.gate.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 524288 | |
| }, | |
| { | |
| "group": "shared_expert_down", | |
| "name": "model.language_model.layers.26.mlp.shared_expert.down_proj.weight", | |
| "max_abs_diff": 0.0098419189453125, | |
| "mean_abs_diff": 0.00016686832532286644, | |
| "pct_different": 86.50693893432617, | |
| "n_different": 907091, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_gate", | |
| "name": "model.language_model.layers.26.mlp.shared_expert.gate_proj.weight", | |
| "max_abs_diff": 0.01220703125, | |
| "mean_abs_diff": 0.00018885653116740286, | |
| "pct_different": 91.31221771240234, | |
| "n_different": 957478, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_up", | |
| "name": "model.language_model.layers.26.mlp.shared_expert.up_proj.weight", | |
| "max_abs_diff": 0.0029296875, | |
| "mean_abs_diff": 7.671593630220741e-05, | |
| "pct_different": 81.4016342163086, | |
| "n_different": 853558, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_gate_scalar", | |
| "name": "model.language_model.layers.26.mlp.shared_expert_gate.weight", | |
| "max_abs_diff": 0.0040283203125, | |
| "mean_abs_diff": 0.00037974317092448473, | |
| "pct_different": 98.73046875, | |
| "n_different": 2022, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "layernorm", | |
| "name": "model.language_model.layers.26.post_attention_layernorm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "layernorm", | |
| "name": "model.language_model.layers.27.input_layernorm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "expert_down", | |
| "name": "model.language_model.layers.27.mlp.experts.down_proj", | |
| "max_abs_diff": 0.03814697265625, | |
| "mean_abs_diff": 0.0001794793497538194, | |
| "pct_different": 85.1216472685337, | |
| "n_different": 228496682, | |
| "total_elements": 268435456 | |
| }, | |
| { | |
| "group": "expert_gate_up", | |
| "name": "model.language_model.layers.27.mlp.experts.gate_up_proj", | |
| "max_abs_diff": 0.012451171875, | |
| "mean_abs_diff": 5.746708484366536e-05, | |
| "pct_different": 55.09844906628132, | |
| "n_different": 295807546, | |
| "total_elements": 536870912 | |
| }, | |
| { | |
| "group": "router_gate", | |
| "name": "model.language_model.layers.27.mlp.gate.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 524288 | |
| }, | |
| { | |
| "group": "shared_expert_down", | |
| "name": "model.language_model.layers.27.mlp.shared_expert.down_proj.weight", | |
| "max_abs_diff": 0.0107421875, | |
| "mean_abs_diff": 0.0001975538325496018, | |
| "pct_different": 85.57443618774414, | |
| "n_different": 897313, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_gate", | |
| "name": "model.language_model.layers.27.mlp.shared_expert.gate_proj.weight", | |
| "max_abs_diff": 0.018798828125, | |
| "mean_abs_diff": 0.0002619295846670866, | |
| "pct_different": 92.55857467651367, | |
| "n_different": 970547, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_up", | |
| "name": "model.language_model.layers.27.mlp.shared_expert.up_proj.weight", | |
| "max_abs_diff": 0.00103759765625, | |
| "mean_abs_diff": 5.918543683947064e-05, | |
| "pct_different": 78.31268310546875, | |
| "n_different": 821168, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_gate_scalar", | |
| "name": "model.language_model.layers.27.mlp.shared_expert_gate.weight", | |
| "max_abs_diff": 0.00469970703125, | |
| "mean_abs_diff": 0.0004430483386386186, | |
| "pct_different": 98.828125, | |
| "n_different": 2024, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "layernorm", | |
| "name": "model.language_model.layers.27.post_attention_layernorm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "self_attn.k_norm", | |
| "name": "model.language_model.layers.27.self_attn.k_norm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 256 | |
| }, | |
| { | |
| "group": "self_attn.k_proj", | |
| "name": "model.language_model.layers.27.self_attn.k_proj.weight", | |
| "max_abs_diff": 0.0009765625, | |
| "mean_abs_diff": 7.92950377217494e-05, | |
| "pct_different": 75.97541809082031, | |
| "n_different": 796660, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "self_attn.o_proj", | |
| "name": "model.language_model.layers.27.self_attn.o_proj.weight", | |
| "max_abs_diff": 0.032196044921875, | |
| "mean_abs_diff": 0.00032847755937837064, | |
| "pct_different": 85.68580150604248, | |
| "n_different": 7187846, | |
| "total_elements": 8388608 | |
| }, | |
| { | |
| "group": "self_attn.q_norm", | |
| "name": "model.language_model.layers.27.self_attn.q_norm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 256 | |
| }, | |
| { | |
| "group": "self_attn.q_proj", | |
| "name": "model.language_model.layers.27.self_attn.q_proj.weight", | |
| "max_abs_diff": 0.00115966796875, | |
| "mean_abs_diff": 7.270420610439032e-05, | |
| "pct_different": 75.59701800346375, | |
| "n_different": 12683075, | |
| "total_elements": 16777216 | |
| }, | |
| { | |
| "group": "self_attn.v_proj", | |
| "name": "model.language_model.layers.27.self_attn.v_proj.weight", | |
| "max_abs_diff": 0.0009765625, | |
| "mean_abs_diff": 6.744181155227125e-05, | |
| "pct_different": 76.58853530883789, | |
| "n_different": 803089, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "layernorm", | |
| "name": "model.language_model.layers.28.input_layernorm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "linear_attn.A_log", | |
| "name": "model.language_model.layers.28.linear_attn.A_log", | |
| "max_abs_diff": 5.960464477539063e-08, | |
| "mean_abs_diff": 9.778887033462524e-09, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 32 | |
| }, | |
| { | |
| "group": "linear_attn.conv1d", | |
| "name": "model.language_model.layers.28.linear_attn.conv1d.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 32768 | |
| }, | |
| { | |
| "group": "linear_attn.dt_bias", | |
| "name": "model.language_model.layers.28.linear_attn.dt_bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 32 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_a", | |
| "name": "model.language_model.layers.28.linear_attn.in_proj_a.weight", | |
| "max_abs_diff": 0.00079345703125, | |
| "mean_abs_diff": 9.187695832224563e-05, | |
| "pct_different": 77.9815673828125, | |
| "n_different": 51106, | |
| "total_elements": 65536 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_b", | |
| "name": "model.language_model.layers.28.linear_attn.in_proj_b.weight", | |
| "max_abs_diff": 0.00054931640625, | |
| "mean_abs_diff": 7.891813584137708e-05, | |
| "pct_different": 79.23736572265625, | |
| "n_different": 51929, | |
| "total_elements": 65536 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_qkv", | |
| "name": "model.language_model.layers.28.linear_attn.in_proj_qkv.weight", | |
| "max_abs_diff": 0.002197265625, | |
| "mean_abs_diff": 9.18437508516945e-05, | |
| "pct_different": 75.92564225196838, | |
| "n_different": 12738209, | |
| "total_elements": 16777216 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_z", | |
| "name": "model.language_model.layers.28.linear_attn.in_proj_z.weight", | |
| "max_abs_diff": 0.001495361328125, | |
| "mean_abs_diff": 6.9412897573784e-05, | |
| "pct_different": 75.61913728713989, | |
| "n_different": 6343393, | |
| "total_elements": 8388608 | |
| }, | |
| { | |
| "group": "linear_attn.norm", | |
| "name": "model.language_model.layers.28.linear_attn.norm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 128 | |
| }, | |
| { | |
| "group": "linear_attn.out_proj", | |
| "name": "model.language_model.layers.28.linear_attn.out_proj.weight", | |
| "max_abs_diff": 0.03228759765625, | |
| "mean_abs_diff": 0.0005236981087364256, | |
| "pct_different": 87.35791444778442, | |
| "n_different": 7328113, | |
| "total_elements": 8388608 | |
| }, | |
| { | |
| "group": "expert_down", | |
| "name": "model.language_model.layers.28.mlp.experts.down_proj", | |
| "max_abs_diff": 0.0270233154296875, | |
| "mean_abs_diff": 0.00016920926282182336, | |
| "pct_different": 84.75274406373501, | |
| "n_different": 227506415, | |
| "total_elements": 268435456 | |
| }, | |
| { | |
| "group": "expert_gate_up", | |
| "name": "model.language_model.layers.28.mlp.experts.gate_up_proj", | |
| "max_abs_diff": 0.0153656005859375, | |
| "mean_abs_diff": 8.473411435261369e-05, | |
| "pct_different": 79.3907668441534, | |
| "n_different": 426225934, | |
| "total_elements": 536870912 | |
| }, | |
| { | |
| "group": "router_gate", | |
| "name": "model.language_model.layers.28.mlp.gate.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 524288 | |
| }, | |
| { | |
| "group": "shared_expert_down", | |
| "name": "model.language_model.layers.28.mlp.shared_expert.down_proj.weight", | |
| "max_abs_diff": 0.013580322265625, | |
| "mean_abs_diff": 0.00021053112868685275, | |
| "pct_different": 85.90002059936523, | |
| "n_different": 900727, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_gate", | |
| "name": "model.language_model.layers.28.mlp.shared_expert.gate_proj.weight", | |
| "max_abs_diff": 0.016845703125, | |
| "mean_abs_diff": 0.00021755343186669052, | |
| "pct_different": 91.11166000366211, | |
| "n_different": 955375, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_up", | |
| "name": "model.language_model.layers.28.mlp.shared_expert.up_proj.weight", | |
| "max_abs_diff": 0.00133514404296875, | |
| "mean_abs_diff": 6.017617124598473e-05, | |
| "pct_different": 78.37247848510742, | |
| "n_different": 821795, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_gate_scalar", | |
| "name": "model.language_model.layers.28.mlp.shared_expert_gate.weight", | |
| "max_abs_diff": 0.0042724609375, | |
| "mean_abs_diff": 0.00040127639658749104, | |
| "pct_different": 98.73046875, | |
| "n_different": 2022, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "layernorm", | |
| "name": "model.language_model.layers.28.post_attention_layernorm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "layernorm", | |
| "name": "model.language_model.layers.29.input_layernorm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "linear_attn.A_log", | |
| "name": "model.language_model.layers.29.linear_attn.A_log", | |
| "max_abs_diff": 5.960464477539063e-08, | |
| "mean_abs_diff": 1.1874362826347351e-08, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 32 | |
| }, | |
| { | |
| "group": "linear_attn.conv1d", | |
| "name": "model.language_model.layers.29.linear_attn.conv1d.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 32768 | |
| }, | |
| { | |
| "group": "linear_attn.dt_bias", | |
| "name": "model.language_model.layers.29.linear_attn.dt_bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 32 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_a", | |
| "name": "model.language_model.layers.29.linear_attn.in_proj_a.weight", | |
| "max_abs_diff": 0.0006103515625, | |
| "mean_abs_diff": 8.780425559962168e-05, | |
| "pct_different": 77.6641845703125, | |
| "n_different": 50898, | |
| "total_elements": 65536 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_b", | |
| "name": "model.language_model.layers.29.linear_attn.in_proj_b.weight", | |
| "max_abs_diff": 0.000518798828125, | |
| "mean_abs_diff": 7.96613676357083e-05, | |
| "pct_different": 78.70025634765625, | |
| "n_different": 51577, | |
| "total_elements": 65536 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_qkv", | |
| "name": "model.language_model.layers.29.linear_attn.in_proj_qkv.weight", | |
| "max_abs_diff": 0.001708984375, | |
| "mean_abs_diff": 9.18063087738119e-05, | |
| "pct_different": 76.07945799827576, | |
| "n_different": 12764015, | |
| "total_elements": 16777216 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_z", | |
| "name": "model.language_model.layers.29.linear_attn.in_proj_z.weight", | |
| "max_abs_diff": 0.001220703125, | |
| "mean_abs_diff": 6.873485108371824e-05, | |
| "pct_different": 75.67347288131714, | |
| "n_different": 6347951, | |
| "total_elements": 8388608 | |
| }, | |
| { | |
| "group": "linear_attn.norm", | |
| "name": "model.language_model.layers.29.linear_attn.norm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 128 | |
| }, | |
| { | |
| "group": "linear_attn.out_proj", | |
| "name": "model.language_model.layers.29.linear_attn.out_proj.weight", | |
| "max_abs_diff": 0.050506591796875, | |
| "mean_abs_diff": 0.000531617144588381, | |
| "pct_different": 87.65290975570679, | |
| "n_different": 7352859, | |
| "total_elements": 8388608 | |
| }, | |
| { | |
| "group": "expert_down", | |
| "name": "model.language_model.layers.29.mlp.experts.down_proj", | |
| "max_abs_diff": 0.02685546875, | |
| "mean_abs_diff": 0.00016401131870225072, | |
| "pct_different": 84.61188152432442, | |
| "n_different": 227128290, | |
| "total_elements": 268435456 | |
| }, | |
| { | |
| "group": "expert_gate_up", | |
| "name": "model.language_model.layers.29.mlp.experts.gate_up_proj", | |
| "max_abs_diff": 0.015777587890625, | |
| "mean_abs_diff": 6.107890658313408e-05, | |
| "pct_different": 56.84496387839317, | |
| "n_different": 305184076, | |
| "total_elements": 536870912 | |
| }, | |
| { | |
| "group": "router_gate", | |
| "name": "model.language_model.layers.29.mlp.gate.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 524288 | |
| }, | |
| { | |
| "group": "shared_expert_down", | |
| "name": "model.language_model.layers.29.mlp.shared_expert.down_proj.weight", | |
| "max_abs_diff": 0.024078369140625, | |
| "mean_abs_diff": 0.00017707540246192366, | |
| "pct_different": 86.66009902954102, | |
| "n_different": 908697, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_gate", | |
| "name": "model.language_model.layers.29.mlp.shared_expert.gate_proj.weight", | |
| "max_abs_diff": 0.010097384452819824, | |
| "mean_abs_diff": 0.00020580603450071067, | |
| "pct_different": 87.35103607177734, | |
| "n_different": 915942, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_up", | |
| "name": "model.language_model.layers.29.mlp.shared_expert.up_proj.weight", | |
| "max_abs_diff": 0.00262451171875, | |
| "mean_abs_diff": 5.184086330700666e-05, | |
| "pct_different": 62.17975616455078, | |
| "n_different": 652002, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_gate_scalar", | |
| "name": "model.language_model.layers.29.mlp.shared_expert_gate.weight", | |
| "max_abs_diff": 0.0035400390625, | |
| "mean_abs_diff": 0.00033361499663442373, | |
| "pct_different": 98.583984375, | |
| "n_different": 2019, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "layernorm", | |
| "name": "model.language_model.layers.29.post_attention_layernorm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "layernorm", | |
| "name": "model.language_model.layers.3.input_layernorm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "expert_down", | |
| "name": "model.language_model.layers.3.mlp.experts.down_proj", | |
| "max_abs_diff": 0.0020294189453125, | |
| "mean_abs_diff": 4.4209707994014025e-05, | |
| "pct_different": 74.01256300508976, | |
| "n_different": 198675961, | |
| "total_elements": 268435456 | |
| }, | |
| { | |
| "group": "expert_gate_up", | |
| "name": "model.language_model.layers.3.mlp.experts.gate_up_proj", | |
| "max_abs_diff": 0.00079345703125, | |
| "mean_abs_diff": 4.3694086343748495e-05, | |
| "pct_different": 74.10692442208529, | |
| "n_different": 397858521, | |
| "total_elements": 536870912 | |
| }, | |
| { | |
| "group": "router_gate", | |
| "name": "model.language_model.layers.3.mlp.gate.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 524288 | |
| }, | |
| { | |
| "group": "shared_expert_down", | |
| "name": "model.language_model.layers.3.mlp.shared_expert.down_proj.weight", | |
| "max_abs_diff": 0.001708984375, | |
| "mean_abs_diff": 5.025587233831175e-05, | |
| "pct_different": 76.22089385986328, | |
| "n_different": 799234, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_gate", | |
| "name": "model.language_model.layers.3.mlp.shared_expert.gate_proj.weight", | |
| "max_abs_diff": 0.00079345703125, | |
| "mean_abs_diff": 5.049067476647906e-05, | |
| "pct_different": 76.59769058227539, | |
| "n_different": 803185, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_up", | |
| "name": "model.language_model.layers.3.mlp.shared_expert.up_proj.weight", | |
| "max_abs_diff": 0.000518798828125, | |
| "mean_abs_diff": 5.181989399716258e-05, | |
| "pct_different": 76.29032135009766, | |
| "n_different": 799962, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_gate_scalar", | |
| "name": "model.language_model.layers.3.mlp.shared_expert_gate.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "layernorm", | |
| "name": "model.language_model.layers.3.post_attention_layernorm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "self_attn.k_norm", | |
| "name": "model.language_model.layers.3.self_attn.k_norm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 256 | |
| }, | |
| { | |
| "group": "self_attn.k_proj", | |
| "name": "model.language_model.layers.3.self_attn.k_proj.weight", | |
| "max_abs_diff": 2.9802322387695312e-08, | |
| "mean_abs_diff": 6.530997964659946e-12, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "self_attn.o_proj", | |
| "name": "model.language_model.layers.3.self_attn.o_proj.weight", | |
| "max_abs_diff": 0.002197265625, | |
| "mean_abs_diff": 6.437394040403888e-05, | |
| "pct_different": 75.57034492492676, | |
| "n_different": 6339300, | |
| "total_elements": 8388608 | |
| }, | |
| { | |
| "group": "self_attn.q_norm", | |
| "name": "model.language_model.layers.3.self_attn.q_norm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 256 | |
| }, | |
| { | |
| "group": "self_attn.q_proj", | |
| "name": "model.language_model.layers.3.self_attn.q_proj.weight", | |
| "max_abs_diff": 2.9802322387695312e-08, | |
| "mean_abs_diff": 6.194018388472333e-12, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 16777216 | |
| }, | |
| { | |
| "group": "self_attn.v_proj", | |
| "name": "model.language_model.layers.3.self_attn.v_proj.weight", | |
| "max_abs_diff": 2.9802322387695312e-08, | |
| "mean_abs_diff": 9.981127035985082e-12, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "layernorm", | |
| "name": "model.language_model.layers.30.input_layernorm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "linear_attn.A_log", | |
| "name": "model.language_model.layers.30.linear_attn.A_log", | |
| "max_abs_diff": 5.960464477539063e-08, | |
| "mean_abs_diff": 8.149072527885437e-09, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 32 | |
| }, | |
| { | |
| "group": "linear_attn.conv1d", | |
| "name": "model.language_model.layers.30.linear_attn.conv1d.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 32768 | |
| }, | |
| { | |
| "group": "linear_attn.dt_bias", | |
| "name": "model.language_model.layers.30.linear_attn.dt_bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 32 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_a", | |
| "name": "model.language_model.layers.30.linear_attn.in_proj_a.weight", | |
| "max_abs_diff": 0.000732421875, | |
| "mean_abs_diff": 9.039473661687225e-05, | |
| "pct_different": 77.5238037109375, | |
| "n_different": 50806, | |
| "total_elements": 65536 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_b", | |
| "name": "model.language_model.layers.30.linear_attn.in_proj_b.weight", | |
| "max_abs_diff": 0.000640869140625, | |
| "mean_abs_diff": 8.33064696053043e-05, | |
| "pct_different": 78.95965576171875, | |
| "n_different": 51747, | |
| "total_elements": 65536 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_qkv", | |
| "name": "model.language_model.layers.30.linear_attn.in_proj_qkv.weight", | |
| "max_abs_diff": 0.00146484375, | |
| "mean_abs_diff": 9.462613525101915e-05, | |
| "pct_different": 76.13858580589294, | |
| "n_different": 12773935, | |
| "total_elements": 16777216 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_z", | |
| "name": "model.language_model.layers.30.linear_attn.in_proj_z.weight", | |
| "max_abs_diff": 0.001953125, | |
| "mean_abs_diff": 6.901209417264909e-05, | |
| "pct_different": 75.79432725906372, | |
| "n_different": 6358089, | |
| "total_elements": 8388608 | |
| }, | |
| { | |
| "group": "linear_attn.norm", | |
| "name": "model.language_model.layers.30.linear_attn.norm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 128 | |
| }, | |
| { | |
| "group": "linear_attn.out_proj", | |
| "name": "model.language_model.layers.30.linear_attn.out_proj.weight", | |
| "max_abs_diff": 0.0516357421875, | |
| "mean_abs_diff": 0.000501302070915699, | |
| "pct_different": 87.40135431289673, | |
| "n_different": 7331757, | |
| "total_elements": 8388608 | |
| }, | |
| { | |
| "group": "expert_down", | |
| "name": "model.language_model.layers.30.mlp.experts.down_proj", | |
| "max_abs_diff": 0.02935791015625, | |
| "mean_abs_diff": 0.00016379347653128207, | |
| "pct_different": 84.55454334616661, | |
| "n_different": 226974374, | |
| "total_elements": 268435456 | |
| }, | |
| { | |
| "group": "expert_gate_up", | |
| "name": "model.language_model.layers.30.mlp.experts.gate_up_proj", | |
| "max_abs_diff": 0.014862060546875, | |
| "mean_abs_diff": 6.591484998352826e-05, | |
| "pct_different": 57.95666594058275, | |
| "n_different": 311152481, | |
| "total_elements": 536870912 | |
| }, | |
| { | |
| "group": "router_gate", | |
| "name": "model.language_model.layers.30.mlp.gate.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 524288 | |
| }, | |
| { | |
| "group": "shared_expert_down", | |
| "name": "model.language_model.layers.30.mlp.shared_expert.down_proj.weight", | |
| "max_abs_diff": 0.0145263671875, | |
| "mean_abs_diff": 0.00019252604397479445, | |
| "pct_different": 85.81275939941406, | |
| "n_different": 899812, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_gate", | |
| "name": "model.language_model.layers.30.mlp.shared_expert.gate_proj.weight", | |
| "max_abs_diff": 0.018218994140625, | |
| "mean_abs_diff": 0.000250937941018492, | |
| "pct_different": 88.23652267456055, | |
| "n_different": 925227, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_up", | |
| "name": "model.language_model.layers.30.mlp.shared_expert.up_proj.weight", | |
| "max_abs_diff": 0.0015869140625, | |
| "mean_abs_diff": 1.9374905605218373e-05, | |
| "pct_different": 42.94710159301758, | |
| "n_different": 450333, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_gate_scalar", | |
| "name": "model.language_model.layers.30.mlp.shared_expert_gate.weight", | |
| "max_abs_diff": 0.0037841796875, | |
| "mean_abs_diff": 0.00035779309109784663, | |
| "pct_different": 98.486328125, | |
| "n_different": 2017, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "layernorm", | |
| "name": "model.language_model.layers.30.post_attention_layernorm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "layernorm", | |
| "name": "model.language_model.layers.31.input_layernorm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "expert_down", | |
| "name": "model.language_model.layers.31.mlp.experts.down_proj", | |
| "max_abs_diff": 0.028594970703125, | |
| "mean_abs_diff": 0.00016276071255560964, | |
| "pct_different": 84.3730952590704, | |
| "n_different": 226487303, | |
| "total_elements": 268435456 | |
| }, | |
| { | |
| "group": "expert_gate_up", | |
| "name": "model.language_model.layers.31.mlp.experts.gate_up_proj", | |
| "max_abs_diff": 0.015411376953125, | |
| "mean_abs_diff": 7.159714732551947e-05, | |
| "pct_different": 58.91194976866245, | |
| "n_different": 316281122, | |
| "total_elements": 536870912 | |
| }, | |
| { | |
| "group": "router_gate", | |
| "name": "model.language_model.layers.31.mlp.gate.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 524288 | |
| }, | |
| { | |
| "group": "shared_expert_down", | |
| "name": "model.language_model.layers.31.mlp.shared_expert.down_proj.weight", | |
| "max_abs_diff": 0.00853729248046875, | |
| "mean_abs_diff": 0.0001893887238111347, | |
| "pct_different": 85.7656478881836, | |
| "n_different": 899318, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_gate", | |
| "name": "model.language_model.layers.31.mlp.shared_expert.gate_proj.weight", | |
| "max_abs_diff": 0.0130615234375, | |
| "mean_abs_diff": 0.0002684254141058773, | |
| "pct_different": 88.7507438659668, | |
| "n_different": 930619, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_up", | |
| "name": "model.language_model.layers.31.mlp.shared_expert.up_proj.weight", | |
| "max_abs_diff": 0.001220703125, | |
| "mean_abs_diff": 2.147901432181243e-05, | |
| "pct_different": 46.35343551635742, | |
| "n_different": 486051, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_gate_scalar", | |
| "name": "model.language_model.layers.31.mlp.shared_expert_gate.weight", | |
| "max_abs_diff": 0.00372314453125, | |
| "mean_abs_diff": 0.0003477830032352358, | |
| "pct_different": 98.73046875, | |
| "n_different": 2022, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "layernorm", | |
| "name": "model.language_model.layers.31.post_attention_layernorm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "self_attn.k_norm", | |
| "name": "model.language_model.layers.31.self_attn.k_norm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 256 | |
| }, | |
| { | |
| "group": "self_attn.k_proj", | |
| "name": "model.language_model.layers.31.self_attn.k_proj.weight", | |
| "max_abs_diff": 0.0008544921875, | |
| "mean_abs_diff": 8.076215453911573e-05, | |
| "pct_different": 75.84085464477539, | |
| "n_different": 795249, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "self_attn.o_proj", | |
| "name": "model.language_model.layers.31.self_attn.o_proj.weight", | |
| "max_abs_diff": 0.02266097068786621, | |
| "mean_abs_diff": 0.00023742503253743052, | |
| "pct_different": 84.36301946640015, | |
| "n_different": 7076883, | |
| "total_elements": 8388608 | |
| }, | |
| { | |
| "group": "self_attn.q_norm", | |
| "name": "model.language_model.layers.31.self_attn.q_norm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 256 | |
| }, | |
| { | |
| "group": "self_attn.q_proj", | |
| "name": "model.language_model.layers.31.self_attn.q_proj.weight", | |
| "max_abs_diff": 0.001495361328125, | |
| "mean_abs_diff": 7.578008808195591e-05, | |
| "pct_different": 75.57690143585205, | |
| "n_different": 12679700, | |
| "total_elements": 16777216 | |
| }, | |
| { | |
| "group": "self_attn.v_proj", | |
| "name": "model.language_model.layers.31.self_attn.v_proj.weight", | |
| "max_abs_diff": 0.00103759765625, | |
| "mean_abs_diff": 7.029892003629357e-05, | |
| "pct_different": 77.0151138305664, | |
| "n_different": 807562, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "layernorm", | |
| "name": "model.language_model.layers.32.input_layernorm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "linear_attn.A_log", | |
| "name": "model.language_model.layers.32.linear_attn.A_log", | |
| "max_abs_diff": 2.9802322387695312e-08, | |
| "mean_abs_diff": 4.889443516731262e-09, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 32 | |
| }, | |
| { | |
| "group": "linear_attn.conv1d", | |
| "name": "model.language_model.layers.32.linear_attn.conv1d.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 32768 | |
| }, | |
| { | |
| "group": "linear_attn.dt_bias", | |
| "name": "model.language_model.layers.32.linear_attn.dt_bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 32 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_a", | |
| "name": "model.language_model.layers.32.linear_attn.in_proj_a.weight", | |
| "max_abs_diff": 0.0009765625, | |
| "mean_abs_diff": 9.101589967031032e-05, | |
| "pct_different": 77.78778076171875, | |
| "n_different": 50979, | |
| "total_elements": 65536 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_b", | |
| "name": "model.language_model.layers.32.linear_attn.in_proj_b.weight", | |
| "max_abs_diff": 0.0009765625, | |
| "mean_abs_diff": 7.679347618250176e-05, | |
| "pct_different": 78.61328125, | |
| "n_different": 51520, | |
| "total_elements": 65536 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_qkv", | |
| "name": "model.language_model.layers.32.linear_attn.in_proj_qkv.weight", | |
| "max_abs_diff": 0.0023193359375, | |
| "mean_abs_diff": 9.238612256012857e-05, | |
| "pct_different": 76.15182399749756, | |
| "n_different": 12776156, | |
| "total_elements": 16777216 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_z", | |
| "name": "model.language_model.layers.32.linear_attn.in_proj_z.weight", | |
| "max_abs_diff": 0.0013427734375, | |
| "mean_abs_diff": 7.127261778805405e-05, | |
| "pct_different": 75.95806121826172, | |
| "n_different": 6371824, | |
| "total_elements": 8388608 | |
| }, | |
| { | |
| "group": "linear_attn.norm", | |
| "name": "model.language_model.layers.32.linear_attn.norm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 128 | |
| }, | |
| { | |
| "group": "linear_attn.out_proj", | |
| "name": "model.language_model.layers.32.linear_attn.out_proj.weight", | |
| "max_abs_diff": 0.003387451171875, | |
| "mean_abs_diff": 7.740223372820765e-05, | |
| "pct_different": 75.6868839263916, | |
| "n_different": 6349076, | |
| "total_elements": 8388608 | |
| }, | |
| { | |
| "group": "expert_down", | |
| "name": "model.language_model.layers.32.mlp.experts.down_proj", | |
| "max_abs_diff": 0.02142333984375, | |
| "mean_abs_diff": 0.00015576885198242962, | |
| "pct_different": 84.24240611493587, | |
| "n_different": 226136487, | |
| "total_elements": 268435456 | |
| }, | |
| { | |
| "group": "expert_gate_up", | |
| "name": "model.language_model.layers.32.mlp.experts.gate_up_proj", | |
| "max_abs_diff": 0.01526641845703125, | |
| "mean_abs_diff": 7.210462354123592e-05, | |
| "pct_different": 60.03406420350075, | |
| "n_different": 322305428, | |
| "total_elements": 536870912 | |
| }, | |
| { | |
| "group": "router_gate", | |
| "name": "model.language_model.layers.32.mlp.gate.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 524288 | |
| }, | |
| { | |
| "group": "shared_expert_down", | |
| "name": "model.language_model.layers.32.mlp.shared_expert.down_proj.weight", | |
| "max_abs_diff": 0.00726318359375, | |
| "mean_abs_diff": 0.00015654056915082037, | |
| "pct_different": 88.17319869995117, | |
| "n_different": 924563, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_gate", | |
| "name": "model.language_model.layers.32.mlp.shared_expert.gate_proj.weight", | |
| "max_abs_diff": 0.005889892578125, | |
| "mean_abs_diff": 0.00010276605462422594, | |
| "pct_different": 80.1513671875, | |
| "n_different": 840448, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_up", | |
| "name": "model.language_model.layers.32.mlp.shared_expert.up_proj.weight", | |
| "max_abs_diff": 0.00909423828125, | |
| "mean_abs_diff": 0.0001641318085603416, | |
| "pct_different": 81.64424896240234, | |
| "n_different": 856102, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_gate_scalar", | |
| "name": "model.language_model.layers.32.mlp.shared_expert_gate.weight", | |
| "max_abs_diff": 0.00238037109375, | |
| "mean_abs_diff": 0.00022551114670932293, | |
| "pct_different": 97.998046875, | |
| "n_different": 2007, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "layernorm", | |
| "name": "model.language_model.layers.32.post_attention_layernorm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "layernorm", | |
| "name": "model.language_model.layers.33.input_layernorm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "linear_attn.A_log", | |
| "name": "model.language_model.layers.33.linear_attn.A_log", | |
| "max_abs_diff": 4.470348358154297e-08, | |
| "mean_abs_diff": 7.450580596923828e-09, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 32 | |
| }, | |
| { | |
| "group": "linear_attn.conv1d", | |
| "name": "model.language_model.layers.33.linear_attn.conv1d.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 32768 | |
| }, | |
| { | |
| "group": "linear_attn.dt_bias", | |
| "name": "model.language_model.layers.33.linear_attn.dt_bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 32 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_a", | |
| "name": "model.language_model.layers.33.linear_attn.in_proj_a.weight", | |
| "max_abs_diff": 0.000732421875, | |
| "mean_abs_diff": 9.000087447930127e-05, | |
| "pct_different": 77.77099609375, | |
| "n_different": 50968, | |
| "total_elements": 65536 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_b", | |
| "name": "model.language_model.layers.33.linear_attn.in_proj_b.weight", | |
| "max_abs_diff": 0.00054931640625, | |
| "mean_abs_diff": 7.215674850158393e-05, | |
| "pct_different": 77.69927978515625, | |
| "n_different": 50921, | |
| "total_elements": 65536 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_qkv", | |
| "name": "model.language_model.layers.33.linear_attn.in_proj_qkv.weight", | |
| "max_abs_diff": 0.001953125, | |
| "mean_abs_diff": 9.3386261141859e-05, | |
| "pct_different": 76.09735131263733, | |
| "n_different": 12767017, | |
| "total_elements": 16777216 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_z", | |
| "name": "model.language_model.layers.33.linear_attn.in_proj_z.weight", | |
| "max_abs_diff": 0.00146484375, | |
| "mean_abs_diff": 7.20742391422391e-05, | |
| "pct_different": 75.771164894104, | |
| "n_different": 6356146, | |
| "total_elements": 8388608 | |
| }, | |
| { | |
| "group": "linear_attn.norm", | |
| "name": "model.language_model.layers.33.linear_attn.norm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 128 | |
| }, | |
| { | |
| "group": "linear_attn.out_proj", | |
| "name": "model.language_model.layers.33.linear_attn.out_proj.weight", | |
| "max_abs_diff": 0.003662109375, | |
| "mean_abs_diff": 7.798991282470524e-05, | |
| "pct_different": 75.65804719924927, | |
| "n_different": 6346657, | |
| "total_elements": 8388608 | |
| }, | |
| { | |
| "group": "expert_down", | |
| "name": "model.language_model.layers.33.mlp.experts.down_proj", | |
| "max_abs_diff": 0.019369125366210938, | |
| "mean_abs_diff": 0.00015237636398524046, | |
| "pct_different": 84.0511217713356, | |
| "n_different": 225623012, | |
| "total_elements": 268435456 | |
| }, | |
| { | |
| "group": "expert_gate_up", | |
| "name": "model.language_model.layers.33.mlp.experts.gate_up_proj", | |
| "max_abs_diff": 0.01514434814453125, | |
| "mean_abs_diff": 7.405649375868961e-05, | |
| "pct_different": 60.37269998341799, | |
| "n_different": 324123465, | |
| "total_elements": 536870912 | |
| }, | |
| { | |
| "group": "router_gate", | |
| "name": "model.language_model.layers.33.mlp.gate.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 524288 | |
| }, | |
| { | |
| "group": "shared_expert_down", | |
| "name": "model.language_model.layers.33.mlp.shared_expert.down_proj.weight", | |
| "max_abs_diff": 0.01397705078125, | |
| "mean_abs_diff": 0.0001608088205102831, | |
| "pct_different": 85.75677871704102, | |
| "n_different": 899225, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_gate", | |
| "name": "model.language_model.layers.33.mlp.shared_expert.gate_proj.weight", | |
| "max_abs_diff": 0.013822555541992188, | |
| "mean_abs_diff": 0.0002342966035939753, | |
| "pct_different": 87.44077682495117, | |
| "n_different": 916883, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_up", | |
| "name": "model.language_model.layers.33.mlp.shared_expert.up_proj.weight", | |
| "max_abs_diff": 0.00341796875, | |
| "mean_abs_diff": 5.740391497965902e-05, | |
| "pct_different": 64.68057632446289, | |
| "n_different": 678225, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_gate_scalar", | |
| "name": "model.language_model.layers.33.mlp.shared_expert_gate.weight", | |
| "max_abs_diff": 0.00164794921875, | |
| "mean_abs_diff": 0.00015396288654301316, | |
| "pct_different": 96.875, | |
| "n_different": 1984, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "layernorm", | |
| "name": "model.language_model.layers.33.post_attention_layernorm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "layernorm", | |
| "name": "model.language_model.layers.34.input_layernorm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "linear_attn.A_log", | |
| "name": "model.language_model.layers.34.linear_attn.A_log", | |
| "max_abs_diff": 5.960464477539063e-08, | |
| "mean_abs_diff": 1.4784745872020721e-08, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 32 | |
| }, | |
| { | |
| "group": "linear_attn.conv1d", | |
| "name": "model.language_model.layers.34.linear_attn.conv1d.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 32768 | |
| }, | |
| { | |
| "group": "linear_attn.dt_bias", | |
| "name": "model.language_model.layers.34.linear_attn.dt_bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 32 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_a", | |
| "name": "model.language_model.layers.34.linear_attn.in_proj_a.weight", | |
| "max_abs_diff": 0.000732421875, | |
| "mean_abs_diff": 0.00010084414680022746, | |
| "pct_different": 77.2125244140625, | |
| "n_different": 50602, | |
| "total_elements": 65536 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_b", | |
| "name": "model.language_model.layers.34.linear_attn.in_proj_b.weight", | |
| "max_abs_diff": 0.0006103515625, | |
| "mean_abs_diff": 7.188737799879164e-05, | |
| "pct_different": 76.93328857421875, | |
| "n_different": 50419, | |
| "total_elements": 65536 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_qkv", | |
| "name": "model.language_model.layers.34.linear_attn.in_proj_qkv.weight", | |
| "max_abs_diff": 0.00146484375, | |
| "mean_abs_diff": 9.46048749028705e-05, | |
| "pct_different": 76.37377977371216, | |
| "n_different": 12813394, | |
| "total_elements": 16777216 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_z", | |
| "name": "model.language_model.layers.34.linear_attn.in_proj_z.weight", | |
| "max_abs_diff": 0.001220703125, | |
| "mean_abs_diff": 7.370914681814611e-05, | |
| "pct_different": 75.92724561691284, | |
| "n_different": 6369239, | |
| "total_elements": 8388608 | |
| }, | |
| { | |
| "group": "linear_attn.norm", | |
| "name": "model.language_model.layers.34.linear_attn.norm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 128 | |
| }, | |
| { | |
| "group": "linear_attn.out_proj", | |
| "name": "model.language_model.layers.34.linear_attn.out_proj.weight", | |
| "max_abs_diff": 0.00390625, | |
| "mean_abs_diff": 8.248243102570996e-05, | |
| "pct_different": 76.42189264297485, | |
| "n_different": 6410733, | |
| "total_elements": 8388608 | |
| }, | |
| { | |
| "group": "expert_down", | |
| "name": "model.language_model.layers.34.mlp.experts.down_proj", | |
| "max_abs_diff": 0.02142333984375, | |
| "mean_abs_diff": 0.00015174713917076588, | |
| "pct_different": 84.09503288567066, | |
| "n_different": 225740885, | |
| "total_elements": 268435456 | |
| }, | |
| { | |
| "group": "expert_gate_up", | |
| "name": "model.language_model.layers.34.mlp.experts.gate_up_proj", | |
| "max_abs_diff": 0.015228271484375, | |
| "mean_abs_diff": 7.801589526934549e-05, | |
| "pct_different": 61.49240620434284, | |
| "n_different": 330134842, | |
| "total_elements": 536870912 | |
| }, | |
| { | |
| "group": "router_gate", | |
| "name": "model.language_model.layers.34.mlp.gate.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 524288 | |
| }, | |
| { | |
| "group": "shared_expert_down", | |
| "name": "model.language_model.layers.34.mlp.shared_expert.down_proj.weight", | |
| "max_abs_diff": 0.013916015625, | |
| "mean_abs_diff": 0.00015560949395876378, | |
| "pct_different": 66.48197174072266, | |
| "n_different": 697114, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_gate", | |
| "name": "model.language_model.layers.34.mlp.shared_expert.gate_proj.weight", | |
| "max_abs_diff": 0.02081298828125, | |
| "mean_abs_diff": 0.00024355290224775672, | |
| "pct_different": 87.3295783996582, | |
| "n_different": 915717, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_up", | |
| "name": "model.language_model.layers.34.mlp.shared_expert.up_proj.weight", | |
| "max_abs_diff": 0.002197265625, | |
| "mean_abs_diff": 2.233486156910658e-05, | |
| "pct_different": 46.485328674316406, | |
| "n_different": 487434, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_gate_scalar", | |
| "name": "model.language_model.layers.34.mlp.shared_expert_gate.weight", | |
| "max_abs_diff": 0.0013427734375, | |
| "mean_abs_diff": 0.00012425615568645298, | |
| "pct_different": 96.435546875, | |
| "n_different": 1975, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "layernorm", | |
| "name": "model.language_model.layers.34.post_attention_layernorm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "layernorm", | |
| "name": "model.language_model.layers.35.input_layernorm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "expert_down", | |
| "name": "model.language_model.layers.35.mlp.experts.down_proj", | |
| "max_abs_diff": 0.02032470703125, | |
| "mean_abs_diff": 0.00016059938934631646, | |
| "pct_different": 84.20781157910824, | |
| "n_different": 226043623, | |
| "total_elements": 268435456 | |
| }, | |
| { | |
| "group": "expert_gate_up", | |
| "name": "model.language_model.layers.35.mlp.experts.gate_up_proj", | |
| "max_abs_diff": 0.017578125, | |
| "mean_abs_diff": 8.783558587310836e-05, | |
| "pct_different": 62.8320237621665, | |
| "n_different": 337326859, | |
| "total_elements": 536870912 | |
| }, | |
| { | |
| "group": "router_gate", | |
| "name": "model.language_model.layers.35.mlp.gate.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 524288 | |
| }, | |
| { | |
| "group": "shared_expert_down", | |
| "name": "model.language_model.layers.35.mlp.shared_expert.down_proj.weight", | |
| "max_abs_diff": 0.01200103759765625, | |
| "mean_abs_diff": 0.00017338157340418547, | |
| "pct_different": 68.32456588745117, | |
| "n_different": 716435, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_gate", | |
| "name": "model.language_model.layers.35.mlp.shared_expert.gate_proj.weight", | |
| "max_abs_diff": 0.01861572265625, | |
| "mean_abs_diff": 0.00028769392520189285, | |
| "pct_different": 87.29629516601562, | |
| "n_different": 915368, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_up", | |
| "name": "model.language_model.layers.35.mlp.shared_expert.up_proj.weight", | |
| "max_abs_diff": 0.0015869140625, | |
| "mean_abs_diff": 2.2541513317264616e-05, | |
| "pct_different": 45.790672302246094, | |
| "n_different": 480150, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_gate_scalar", | |
| "name": "model.language_model.layers.35.mlp.shared_expert_gate.weight", | |
| "max_abs_diff": 0.0013427734375, | |
| "mean_abs_diff": 0.00012546921789180487, | |
| "pct_different": 96.240234375, | |
| "n_different": 1971, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "layernorm", | |
| "name": "model.language_model.layers.35.post_attention_layernorm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "self_attn.k_norm", | |
| "name": "model.language_model.layers.35.self_attn.k_norm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 256 | |
| }, | |
| { | |
| "group": "self_attn.k_proj", | |
| "name": "model.language_model.layers.35.self_attn.k_proj.weight", | |
| "max_abs_diff": 0.00115966796875, | |
| "mean_abs_diff": 8.050930046010762e-05, | |
| "pct_different": 76.22861862182617, | |
| "n_different": 799315, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "self_attn.o_proj", | |
| "name": "model.language_model.layers.35.self_attn.o_proj.weight", | |
| "max_abs_diff": 0.0162353515625, | |
| "mean_abs_diff": 0.00020530851907096803, | |
| "pct_different": 83.65904092788696, | |
| "n_different": 7017829, | |
| "total_elements": 8388608 | |
| }, | |
| { | |
| "group": "self_attn.q_norm", | |
| "name": "model.language_model.layers.35.self_attn.q_norm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 256 | |
| }, | |
| { | |
| "group": "self_attn.q_proj", | |
| "name": "model.language_model.layers.35.self_attn.q_proj.weight", | |
| "max_abs_diff": 0.00164794921875, | |
| "mean_abs_diff": 7.766791532048956e-05, | |
| "pct_different": 75.8864164352417, | |
| "n_different": 12731628, | |
| "total_elements": 16777216 | |
| }, | |
| { | |
| "group": "self_attn.v_proj", | |
| "name": "model.language_model.layers.35.self_attn.v_proj.weight", | |
| "max_abs_diff": 0.00146484375, | |
| "mean_abs_diff": 9.362241689814255e-05, | |
| "pct_different": 78.76205444335938, | |
| "n_different": 825880, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "layernorm", | |
| "name": "model.language_model.layers.36.input_layernorm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "linear_attn.A_log", | |
| "name": "model.language_model.layers.36.linear_attn.A_log", | |
| "max_abs_diff": 5.960464477539063e-08, | |
| "mean_abs_diff": 7.916241884231567e-09, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 32 | |
| }, | |
| { | |
| "group": "linear_attn.conv1d", | |
| "name": "model.language_model.layers.36.linear_attn.conv1d.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 32768 | |
| }, | |
| { | |
| "group": "linear_attn.dt_bias", | |
| "name": "model.language_model.layers.36.linear_attn.dt_bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 32 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_a", | |
| "name": "model.language_model.layers.36.linear_attn.in_proj_a.weight", | |
| "max_abs_diff": 2.9802322387695312e-08, | |
| "mean_abs_diff": 7.808864666003501e-12, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 65536 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_b", | |
| "name": "model.language_model.layers.36.linear_attn.in_proj_b.weight", | |
| "max_abs_diff": 2.9802322387695312e-08, | |
| "mean_abs_diff": 9.961809155356605e-12, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 65536 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_qkv", | |
| "name": "model.language_model.layers.36.linear_attn.in_proj_qkv.weight", | |
| "max_abs_diff": 2.9802322387695312e-08, | |
| "mean_abs_diff": 4.946612564005193e-12, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 16777216 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_z", | |
| "name": "model.language_model.layers.36.linear_attn.in_proj_z.weight", | |
| "max_abs_diff": 2.9802322387695312e-08, | |
| "mean_abs_diff": 7.247678152078052e-12, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 8388608 | |
| }, | |
| { | |
| "group": "linear_attn.norm", | |
| "name": "model.language_model.layers.36.linear_attn.norm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 128 | |
| }, | |
| { | |
| "group": "linear_attn.out_proj", | |
| "name": "model.language_model.layers.36.linear_attn.out_proj.weight", | |
| "max_abs_diff": 0.00830078125, | |
| "mean_abs_diff": 7.979055226314813e-05, | |
| "pct_different": 76.34977102279663, | |
| "n_different": 6404683, | |
| "total_elements": 8388608 | |
| }, | |
| { | |
| "group": "expert_down", | |
| "name": "model.language_model.layers.36.mlp.experts.down_proj", | |
| "max_abs_diff": 0.01799774169921875, | |
| "mean_abs_diff": 0.00016030555707402527, | |
| "pct_different": 84.20749418437481, | |
| "n_different": 226042771, | |
| "total_elements": 268435456 | |
| }, | |
| { | |
| "group": "expert_gate_up", | |
| "name": "model.language_model.layers.36.mlp.experts.gate_up_proj", | |
| "max_abs_diff": 0.0145263671875, | |
| "mean_abs_diff": 8.27742405817844e-05, | |
| "pct_different": 61.87529023736715, | |
| "n_different": 332190435, | |
| "total_elements": 536870912 | |
| }, | |
| { | |
| "group": "router_gate", | |
| "name": "model.language_model.layers.36.mlp.gate.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 524288 | |
| }, | |
| { | |
| "group": "shared_expert_down", | |
| "name": "model.language_model.layers.36.mlp.shared_expert.down_proj.weight", | |
| "max_abs_diff": 0.006591796875, | |
| "mean_abs_diff": 0.00014604577154386789, | |
| "pct_different": 87.18204498291016, | |
| "n_different": 914170, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_gate", | |
| "name": "model.language_model.layers.36.mlp.shared_expert.gate_proj.weight", | |
| "max_abs_diff": 0.013515472412109375, | |
| "mean_abs_diff": 0.00017460605886299163, | |
| "pct_different": 81.44760131835938, | |
| "n_different": 854040, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_up", | |
| "name": "model.language_model.layers.36.mlp.shared_expert.up_proj.weight", | |
| "max_abs_diff": 0.005859375, | |
| "mean_abs_diff": 0.00012355484068393707, | |
| "pct_different": 78.62024307250977, | |
| "n_different": 824393, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_gate_scalar", | |
| "name": "model.language_model.layers.36.mlp.shared_expert_gate.weight", | |
| "max_abs_diff": 0.001190185546875, | |
| "mean_abs_diff": 0.00011287422967143357, | |
| "pct_different": 95.654296875, | |
| "n_different": 1959, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "layernorm", | |
| "name": "model.language_model.layers.36.post_attention_layernorm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "layernorm", | |
| "name": "model.language_model.layers.37.input_layernorm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "linear_attn.A_log", | |
| "name": "model.language_model.layers.37.linear_attn.A_log", | |
| "max_abs_diff": 5.960464477539063e-08, | |
| "mean_abs_diff": 1.1816155165433884e-08, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 32 | |
| }, | |
| { | |
| "group": "linear_attn.conv1d", | |
| "name": "model.language_model.layers.37.linear_attn.conv1d.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 32768 | |
| }, | |
| { | |
| "group": "linear_attn.dt_bias", | |
| "name": "model.language_model.layers.37.linear_attn.dt_bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 32 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_a", | |
| "name": "model.language_model.layers.37.linear_attn.in_proj_a.weight", | |
| "max_abs_diff": 2.9802322387695312e-08, | |
| "mean_abs_diff": 8.164136033883551e-12, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 65536 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_b", | |
| "name": "model.language_model.layers.37.linear_attn.in_proj_b.weight", | |
| "max_abs_diff": 2.9802322387695312e-08, | |
| "mean_abs_diff": 8.036238341446733e-12, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 65536 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_qkv", | |
| "name": "model.language_model.layers.37.linear_attn.in_proj_qkv.weight", | |
| "max_abs_diff": 2.9802322387695312e-08, | |
| "mean_abs_diff": 4.976379551491217e-12, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 16777216 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_z", | |
| "name": "model.language_model.layers.37.linear_attn.in_proj_z.weight", | |
| "max_abs_diff": 2.9802322387695312e-08, | |
| "mean_abs_diff": 7.754229550127612e-12, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 8388608 | |
| }, | |
| { | |
| "group": "linear_attn.norm", | |
| "name": "model.language_model.layers.37.linear_attn.norm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 128 | |
| }, | |
| { | |
| "group": "linear_attn.out_proj", | |
| "name": "model.language_model.layers.37.linear_attn.out_proj.weight", | |
| "max_abs_diff": 0.00244140625, | |
| "mean_abs_diff": 8.255692955572158e-05, | |
| "pct_different": 76.73879861831665, | |
| "n_different": 6437317, | |
| "total_elements": 8388608 | |
| }, | |
| { | |
| "group": "expert_down", | |
| "name": "model.language_model.layers.37.mlp.experts.down_proj", | |
| "max_abs_diff": 0.01904296875, | |
| "mean_abs_diff": 0.00016294344095513225, | |
| "pct_different": 84.27763022482395, | |
| "n_different": 226231041, | |
| "total_elements": 268435456 | |
| }, | |
| { | |
| "group": "expert_gate_up", | |
| "name": "model.language_model.layers.37.mlp.experts.gate_up_proj", | |
| "max_abs_diff": 0.016576766967773438, | |
| "mean_abs_diff": 7.902195648057386e-05, | |
| "pct_different": 61.08949836343527, | |
| "n_different": 327971747, | |
| "total_elements": 536870912 | |
| }, | |
| { | |
| "group": "router_gate", | |
| "name": "model.language_model.layers.37.mlp.gate.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 524288 | |
| }, | |
| { | |
| "group": "shared_expert_down", | |
| "name": "model.language_model.layers.37.mlp.shared_expert.down_proj.weight", | |
| "max_abs_diff": 0.01202392578125, | |
| "mean_abs_diff": 0.000136027651024051, | |
| "pct_different": 80.70955276489258, | |
| "n_different": 846301, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_gate", | |
| "name": "model.language_model.layers.37.mlp.shared_expert.gate_proj.weight", | |
| "max_abs_diff": 0.01483154296875, | |
| "mean_abs_diff": 0.00019479072943795472, | |
| "pct_different": 82.67087936401367, | |
| "n_different": 866867, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_up", | |
| "name": "model.language_model.layers.37.mlp.shared_expert.up_proj.weight", | |
| "max_abs_diff": 0.0091094970703125, | |
| "mean_abs_diff": 0.00011996040120720863, | |
| "pct_different": 76.99165344238281, | |
| "n_different": 807316, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_gate_scalar", | |
| "name": "model.language_model.layers.37.mlp.shared_expert_gate.weight", | |
| "max_abs_diff": 0.000457763671875, | |
| "mean_abs_diff": 4.295829057809897e-05, | |
| "pct_different": 89.2578125, | |
| "n_different": 1828, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "layernorm", | |
| "name": "model.language_model.layers.37.post_attention_layernorm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "layernorm", | |
| "name": "model.language_model.layers.38.input_layernorm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "linear_attn.A_log", | |
| "name": "model.language_model.layers.38.linear_attn.A_log", | |
| "max_abs_diff": 5.960464477539063e-08, | |
| "mean_abs_diff": 1.6298145055770874e-08, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 32 | |
| }, | |
| { | |
| "group": "linear_attn.conv1d", | |
| "name": "model.language_model.layers.38.linear_attn.conv1d.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 32768 | |
| }, | |
| { | |
| "group": "linear_attn.dt_bias", | |
| "name": "model.language_model.layers.38.linear_attn.dt_bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 32 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_a", | |
| "name": "model.language_model.layers.38.linear_attn.in_proj_a.weight", | |
| "max_abs_diff": 2.9802322387695312e-08, | |
| "mean_abs_diff": 6.195932655828074e-12, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 65536 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_b", | |
| "name": "model.language_model.layers.38.linear_attn.in_proj_b.weight", | |
| "max_abs_diff": 2.9802322387695312e-08, | |
| "mean_abs_diff": 5.9685589803848416e-12, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 65536 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_qkv", | |
| "name": "model.language_model.layers.38.linear_attn.in_proj_qkv.weight", | |
| "max_abs_diff": 2.9802322387695312e-08, | |
| "mean_abs_diff": 5.123578644683491e-12, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 16777216 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_z", | |
| "name": "model.language_model.layers.38.linear_attn.in_proj_z.weight", | |
| "max_abs_diff": 2.9802322387695312e-08, | |
| "mean_abs_diff": 7.192431546176881e-12, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 8388608 | |
| }, | |
| { | |
| "group": "linear_attn.norm", | |
| "name": "model.language_model.layers.38.linear_attn.norm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 128 | |
| }, | |
| { | |
| "group": "linear_attn.out_proj", | |
| "name": "model.language_model.layers.38.linear_attn.out_proj.weight", | |
| "max_abs_diff": 0.002197265625, | |
| "mean_abs_diff": 7.896935130702332e-05, | |
| "pct_different": 76.81151628494263, | |
| "n_different": 6443417, | |
| "total_elements": 8388608 | |
| }, | |
| { | |
| "group": "expert_down", | |
| "name": "model.language_model.layers.38.mlp.experts.down_proj", | |
| "max_abs_diff": 0.02142810821533203, | |
| "mean_abs_diff": 0.00016490135749336332, | |
| "pct_different": 84.44033712148666, | |
| "n_different": 226667804, | |
| "total_elements": 268435456 | |
| }, | |
| { | |
| "group": "expert_gate_up", | |
| "name": "model.language_model.layers.38.mlp.experts.gate_up_proj", | |
| "max_abs_diff": 0.012939453125, | |
| "mean_abs_diff": 7.53445565351285e-05, | |
| "pct_different": 60.249427892267704, | |
| "n_different": 323461653, | |
| "total_elements": 536870912 | |
| }, | |
| { | |
| "group": "router_gate", | |
| "name": "model.language_model.layers.38.mlp.gate.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 524288 | |
| }, | |
| { | |
| "group": "shared_expert_down", | |
| "name": "model.language_model.layers.38.mlp.shared_expert.down_proj.weight", | |
| "max_abs_diff": 0.01220703125, | |
| "mean_abs_diff": 0.0001475134922657162, | |
| "pct_different": 74.07846450805664, | |
| "n_different": 776769, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_gate", | |
| "name": "model.language_model.layers.38.mlp.shared_expert.gate_proj.weight", | |
| "max_abs_diff": 0.0196533203125, | |
| "mean_abs_diff": 0.00026799453189596534, | |
| "pct_different": 85.18447875976562, | |
| "n_different": 893224, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_up", | |
| "name": "model.language_model.layers.38.mlp.shared_expert.up_proj.weight", | |
| "max_abs_diff": 0.00439453125, | |
| "mean_abs_diff": 4.584606722346507e-05, | |
| "pct_different": 59.99946594238281, | |
| "n_different": 629140, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_gate_scalar", | |
| "name": "model.language_model.layers.38.mlp.shared_expert_gate.weight", | |
| "max_abs_diff": 0.000579833984375, | |
| "mean_abs_diff": 5.4790929425507784e-05, | |
| "pct_different": 93.65234375, | |
| "n_different": 1918, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "layernorm", | |
| "name": "model.language_model.layers.38.post_attention_layernorm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "layernorm", | |
| "name": "model.language_model.layers.39.input_layernorm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "expert_down", | |
| "name": "model.language_model.layers.39.mlp.experts.down_proj", | |
| "max_abs_diff": 0.020599365234375, | |
| "mean_abs_diff": 0.00015861380961723626, | |
| "pct_different": 84.2293206602335, | |
| "n_different": 226101361, | |
| "total_elements": 268435456 | |
| }, | |
| { | |
| "group": "expert_gate_up", | |
| "name": "model.language_model.layers.39.mlp.experts.gate_up_proj", | |
| "max_abs_diff": 0.0125579833984375, | |
| "mean_abs_diff": 6.997332820901647e-05, | |
| "pct_different": 58.380261063575745, | |
| "n_different": 313426640, | |
| "total_elements": 536870912 | |
| }, | |
| { | |
| "group": "router_gate", | |
| "name": "model.language_model.layers.39.mlp.gate.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 524288 | |
| }, | |
| { | |
| "group": "shared_expert_down", | |
| "name": "model.language_model.layers.39.mlp.shared_expert.down_proj.weight", | |
| "max_abs_diff": 0.010498046875, | |
| "mean_abs_diff": 0.0001098875145544298, | |
| "pct_different": 65.07434844970703, | |
| "n_different": 682354, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_gate", | |
| "name": "model.language_model.layers.39.mlp.shared_expert.gate_proj.weight", | |
| "max_abs_diff": 0.0133056640625, | |
| "mean_abs_diff": 0.0002619568258523941, | |
| "pct_different": 87.41598129272461, | |
| "n_different": 916623, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_up", | |
| "name": "model.language_model.layers.39.mlp.shared_expert.up_proj.weight", | |
| "max_abs_diff": 0.002197265625, | |
| "mean_abs_diff": 1.3912536815041676e-05, | |
| "pct_different": 37.54701614379883, | |
| "n_different": 393709, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_gate_scalar", | |
| "name": "model.language_model.layers.39.mlp.shared_expert_gate.weight", | |
| "max_abs_diff": 0.000579833984375, | |
| "mean_abs_diff": 5.540027632378042e-05, | |
| "pct_different": 92.138671875, | |
| "n_different": 1887, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "layernorm", | |
| "name": "model.language_model.layers.39.post_attention_layernorm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "self_attn.k_norm", | |
| "name": "model.language_model.layers.39.self_attn.k_norm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 256 | |
| }, | |
| { | |
| "group": "self_attn.k_proj", | |
| "name": "model.language_model.layers.39.self_attn.k_proj.weight", | |
| "max_abs_diff": 0.00091552734375, | |
| "mean_abs_diff": 8.063699351623654e-05, | |
| "pct_different": 79.56247329711914, | |
| "n_different": 834273, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "self_attn.o_proj", | |
| "name": "model.language_model.layers.39.self_attn.o_proj.weight", | |
| "max_abs_diff": 0.0020751953125, | |
| "mean_abs_diff": 7.997776265256107e-05, | |
| "pct_different": 76.9538402557373, | |
| "n_different": 6455356, | |
| "total_elements": 8388608 | |
| }, | |
| { | |
| "group": "self_attn.q_norm", | |
| "name": "model.language_model.layers.39.self_attn.q_norm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 256 | |
| }, | |
| { | |
| "group": "self_attn.q_proj", | |
| "name": "model.language_model.layers.39.self_attn.q_proj.weight", | |
| "max_abs_diff": 0.0015869140625, | |
| "mean_abs_diff": 7.685480522923172e-05, | |
| "pct_different": 75.80427527427673, | |
| "n_different": 12717847, | |
| "total_elements": 16777216 | |
| }, | |
| { | |
| "group": "self_attn.v_proj", | |
| "name": "model.language_model.layers.39.self_attn.v_proj.weight", | |
| "max_abs_diff": 0.001953125, | |
| "mean_abs_diff": 7.297185220522806e-05, | |
| "pct_different": 79.07142639160156, | |
| "n_different": 829124, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "layernorm", | |
| "name": "model.language_model.layers.4.input_layernorm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "linear_attn.A_log", | |
| "name": "model.language_model.layers.4.linear_attn.A_log", | |
| "max_abs_diff": 5.960464477539063e-08, | |
| "mean_abs_diff": 1.2863893061876297e-08, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 32 | |
| }, | |
| { | |
| "group": "linear_attn.conv1d", | |
| "name": "model.language_model.layers.4.linear_attn.conv1d.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 32768 | |
| }, | |
| { | |
| "group": "linear_attn.dt_bias", | |
| "name": "model.language_model.layers.4.linear_attn.dt_bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 32 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_a", | |
| "name": "model.language_model.layers.4.linear_attn.in_proj_a.weight", | |
| "max_abs_diff": 0.0006103515625, | |
| "mean_abs_diff": 8.83650645846501e-05, | |
| "pct_different": 76.54266357421875, | |
| "n_different": 50163, | |
| "total_elements": 65536 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_b", | |
| "name": "model.language_model.layers.4.linear_attn.in_proj_b.weight", | |
| "max_abs_diff": 0.000732421875, | |
| "mean_abs_diff": 6.650891009485349e-05, | |
| "pct_different": 77.3834228515625, | |
| "n_different": 50714, | |
| "total_elements": 65536 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_qkv", | |
| "name": "model.language_model.layers.4.linear_attn.in_proj_qkv.weight", | |
| "max_abs_diff": 0.001068115234375, | |
| "mean_abs_diff": 7.469176489394158e-05, | |
| "pct_different": 75.51474571228027, | |
| "n_different": 12669272, | |
| "total_elements": 16777216 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_z", | |
| "name": "model.language_model.layers.4.linear_attn.in_proj_z.weight", | |
| "max_abs_diff": 0.0010986328125, | |
| "mean_abs_diff": 6.896830745972693e-05, | |
| "pct_different": 75.60737133026123, | |
| "n_different": 6342406, | |
| "total_elements": 8388608 | |
| }, | |
| { | |
| "group": "linear_attn.norm", | |
| "name": "model.language_model.layers.4.linear_attn.norm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 128 | |
| }, | |
| { | |
| "group": "linear_attn.out_proj", | |
| "name": "model.language_model.layers.4.linear_attn.out_proj.weight", | |
| "max_abs_diff": 0.0029296875, | |
| "mean_abs_diff": 6.950877286726609e-05, | |
| "pct_different": 75.49833059310913, | |
| "n_different": 6333259, | |
| "total_elements": 8388608 | |
| }, | |
| { | |
| "group": "expert_down", | |
| "name": "model.language_model.layers.4.mlp.experts.down_proj", | |
| "max_abs_diff": 0.0009765625, | |
| "mean_abs_diff": 4.400653415359557e-05, | |
| "pct_different": 74.13872145116329, | |
| "n_different": 199014615, | |
| "total_elements": 268435456 | |
| }, | |
| { | |
| "group": "expert_gate_up", | |
| "name": "model.language_model.layers.4.mlp.experts.gate_up_proj", | |
| "max_abs_diff": 0.00091552734375, | |
| "mean_abs_diff": 4.450380947673693e-05, | |
| "pct_different": 74.20840412378311, | |
| "n_different": 398403336, | |
| "total_elements": 536870912 | |
| }, | |
| { | |
| "group": "router_gate", | |
| "name": "model.language_model.layers.4.mlp.gate.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 524288 | |
| }, | |
| { | |
| "group": "shared_expert_down", | |
| "name": "model.language_model.layers.4.mlp.shared_expert.down_proj.weight", | |
| "max_abs_diff": 0.001953125, | |
| "mean_abs_diff": 5.0921604270115495e-05, | |
| "pct_different": 75.7965087890625, | |
| "n_different": 794784, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_gate", | |
| "name": "model.language_model.layers.4.mlp.shared_expert.gate_proj.weight", | |
| "max_abs_diff": 0.0006103515625, | |
| "mean_abs_diff": 4.8473539209226146e-05, | |
| "pct_different": 75.7023811340332, | |
| "n_different": 793797, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_up", | |
| "name": "model.language_model.layers.4.mlp.shared_expert.up_proj.weight", | |
| "max_abs_diff": 0.00051116943359375, | |
| "mean_abs_diff": 5.1016599172726274e-05, | |
| "pct_different": 75.72765350341797, | |
| "n_different": 794062, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_gate_scalar", | |
| "name": "model.language_model.layers.4.mlp.shared_expert_gate.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "layernorm", | |
| "name": "model.language_model.layers.4.post_attention_layernorm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "layernorm", | |
| "name": "model.language_model.layers.5.input_layernorm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "linear_attn.A_log", | |
| "name": "model.language_model.layers.5.linear_attn.A_log", | |
| "max_abs_diff": 5.960464477539063e-08, | |
| "mean_abs_diff": 6.51925802230835e-09, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 32 | |
| }, | |
| { | |
| "group": "linear_attn.conv1d", | |
| "name": "model.language_model.layers.5.linear_attn.conv1d.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 32768 | |
| }, | |
| { | |
| "group": "linear_attn.dt_bias", | |
| "name": "model.language_model.layers.5.linear_attn.dt_bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 32 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_a", | |
| "name": "model.language_model.layers.5.linear_attn.in_proj_a.weight", | |
| "max_abs_diff": 0.0006103515625, | |
| "mean_abs_diff": 9.281998791266233e-05, | |
| "pct_different": 76.58538818359375, | |
| "n_different": 50191, | |
| "total_elements": 65536 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_b", | |
| "name": "model.language_model.layers.5.linear_attn.in_proj_b.weight", | |
| "max_abs_diff": 0.0006103515625, | |
| "mean_abs_diff": 6.878112617414445e-05, | |
| "pct_different": 77.9052734375, | |
| "n_different": 51056, | |
| "total_elements": 65536 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_qkv", | |
| "name": "model.language_model.layers.5.linear_attn.in_proj_qkv.weight", | |
| "max_abs_diff": 0.001251220703125, | |
| "mean_abs_diff": 7.507378177251667e-05, | |
| "pct_different": 75.54700970649719, | |
| "n_different": 12674685, | |
| "total_elements": 16777216 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_z", | |
| "name": "model.language_model.layers.5.linear_attn.in_proj_z.weight", | |
| "max_abs_diff": 0.0010986328125, | |
| "mean_abs_diff": 6.772989581804723e-05, | |
| "pct_different": 75.43255090713501, | |
| "n_different": 6327741, | |
| "total_elements": 8388608 | |
| }, | |
| { | |
| "group": "linear_attn.norm", | |
| "name": "model.language_model.layers.5.linear_attn.norm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 128 | |
| }, | |
| { | |
| "group": "linear_attn.out_proj", | |
| "name": "model.language_model.layers.5.linear_attn.out_proj.weight", | |
| "max_abs_diff": 0.0029296875, | |
| "mean_abs_diff": 6.921568274265155e-05, | |
| "pct_different": 75.57656764984131, | |
| "n_different": 6339822, | |
| "total_elements": 8388608 | |
| }, | |
| { | |
| "group": "expert_down", | |
| "name": "model.language_model.layers.5.mlp.experts.down_proj", | |
| "max_abs_diff": 0.0009765625, | |
| "mean_abs_diff": 4.41000702267047e-05, | |
| "pct_different": 74.24795851111412, | |
| "n_different": 199307846, | |
| "total_elements": 268435456 | |
| }, | |
| { | |
| "group": "expert_gate_up", | |
| "name": "model.language_model.layers.5.mlp.experts.gate_up_proj", | |
| "max_abs_diff": 0.000762939453125, | |
| "mean_abs_diff": 4.478307528188452e-05, | |
| "pct_different": 74.30660780519247, | |
| "n_different": 398930563, | |
| "total_elements": 536870912 | |
| }, | |
| { | |
| "group": "router_gate", | |
| "name": "model.language_model.layers.5.mlp.gate.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 524288 | |
| }, | |
| { | |
| "group": "shared_expert_down", | |
| "name": "model.language_model.layers.5.mlp.shared_expert.down_proj.weight", | |
| "max_abs_diff": 0.001953125, | |
| "mean_abs_diff": 5.1026254368480295e-05, | |
| "pct_different": 76.6988754272461, | |
| "n_different": 804246, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_gate", | |
| "name": "model.language_model.layers.5.mlp.shared_expert.gate_proj.weight", | |
| "max_abs_diff": 0.0006103515625, | |
| "mean_abs_diff": 4.81239112559706e-05, | |
| "pct_different": 76.41477584838867, | |
| "n_different": 801267, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_up", | |
| "name": "model.language_model.layers.5.mlp.shared_expert.up_proj.weight", | |
| "max_abs_diff": 0.00043487548828125, | |
| "mean_abs_diff": 5.2817369578406215e-05, | |
| "pct_different": 76.38359069824219, | |
| "n_different": 800940, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_gate_scalar", | |
| "name": "model.language_model.layers.5.mlp.shared_expert_gate.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "layernorm", | |
| "name": "model.language_model.layers.5.post_attention_layernorm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "layernorm", | |
| "name": "model.language_model.layers.6.input_layernorm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "linear_attn.A_log", | |
| "name": "model.language_model.layers.6.linear_attn.A_log", | |
| "max_abs_diff": 4.470348358154297e-08, | |
| "mean_abs_diff": 1.0710209608078003e-08, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 32 | |
| }, | |
| { | |
| "group": "linear_attn.conv1d", | |
| "name": "model.language_model.layers.6.linear_attn.conv1d.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 32768 | |
| }, | |
| { | |
| "group": "linear_attn.dt_bias", | |
| "name": "model.language_model.layers.6.linear_attn.dt_bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 32 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_a", | |
| "name": "model.language_model.layers.6.linear_attn.in_proj_a.weight", | |
| "max_abs_diff": 0.000640869140625, | |
| "mean_abs_diff": 9.177657193504274e-05, | |
| "pct_different": 77.56805419921875, | |
| "n_different": 50835, | |
| "total_elements": 65536 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_b", | |
| "name": "model.language_model.layers.6.linear_attn.in_proj_b.weight", | |
| "max_abs_diff": 0.000640869140625, | |
| "mean_abs_diff": 7.389762322418392e-05, | |
| "pct_different": 77.65655517578125, | |
| "n_different": 50893, | |
| "total_elements": 65536 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_qkv", | |
| "name": "model.language_model.layers.6.linear_attn.in_proj_qkv.weight", | |
| "max_abs_diff": 0.00152587890625, | |
| "mean_abs_diff": 7.546762935817242e-05, | |
| "pct_different": 75.68546533584595, | |
| "n_different": 12697914, | |
| "total_elements": 16777216 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_z", | |
| "name": "model.language_model.layers.6.linear_attn.in_proj_z.weight", | |
| "max_abs_diff": 0.001068115234375, | |
| "mean_abs_diff": 6.465703336289153e-05, | |
| "pct_different": 75.33910274505615, | |
| "n_different": 6319902, | |
| "total_elements": 8388608 | |
| }, | |
| { | |
| "group": "linear_attn.norm", | |
| "name": "model.language_model.layers.6.linear_attn.norm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 128 | |
| }, | |
| { | |
| "group": "linear_attn.out_proj", | |
| "name": "model.language_model.layers.6.linear_attn.out_proj.weight", | |
| "max_abs_diff": 0.00244140625, | |
| "mean_abs_diff": 6.946274515939876e-05, | |
| "pct_different": 75.57202577590942, | |
| "n_different": 6339441, | |
| "total_elements": 8388608 | |
| }, | |
| { | |
| "group": "expert_down", | |
| "name": "model.language_model.layers.6.mlp.experts.down_proj", | |
| "max_abs_diff": 0.0010986328125, | |
| "mean_abs_diff": 4.4040236389264464e-05, | |
| "pct_different": 74.26761761307716, | |
| "n_different": 199360618, | |
| "total_elements": 268435456 | |
| }, | |
| { | |
| "group": "expert_gate_up", | |
| "name": "model.language_model.layers.6.mlp.experts.gate_up_proj", | |
| "max_abs_diff": 0.000885009765625, | |
| "mean_abs_diff": 4.483235170482658e-05, | |
| "pct_different": 74.34323616325855, | |
| "n_different": 399127210, | |
| "total_elements": 536870912 | |
| }, | |
| { | |
| "group": "router_gate", | |
| "name": "model.language_model.layers.6.mlp.gate.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 524288 | |
| }, | |
| { | |
| "group": "shared_expert_down", | |
| "name": "model.language_model.layers.6.mlp.shared_expert.down_proj.weight", | |
| "max_abs_diff": 0.0013427734375, | |
| "mean_abs_diff": 5.054691428085789e-05, | |
| "pct_different": 76.48334503173828, | |
| "n_different": 801986, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_gate", | |
| "name": "model.language_model.layers.6.mlp.shared_expert.gate_proj.weight", | |
| "max_abs_diff": 0.000579833984375, | |
| "mean_abs_diff": 4.508010169956833e-05, | |
| "pct_different": 76.05876922607422, | |
| "n_different": 797534, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_up", | |
| "name": "model.language_model.layers.6.mlp.shared_expert.up_proj.weight", | |
| "max_abs_diff": 0.00042724609375, | |
| "mean_abs_diff": 5.27710399182979e-05, | |
| "pct_different": 76.43842697143555, | |
| "n_different": 801515, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_gate_scalar", | |
| "name": "model.language_model.layers.6.mlp.shared_expert_gate.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "layernorm", | |
| "name": "model.language_model.layers.6.post_attention_layernorm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "layernorm", | |
| "name": "model.language_model.layers.7.input_layernorm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "expert_down", | |
| "name": "model.language_model.layers.7.mlp.experts.down_proj", | |
| "max_abs_diff": 0.001220703125, | |
| "mean_abs_diff": 4.385596184874885e-05, | |
| "pct_different": 74.2759820073843, | |
| "n_different": 199383071, | |
| "total_elements": 268435456 | |
| }, | |
| { | |
| "group": "expert_gate_up", | |
| "name": "model.language_model.layers.7.mlp.experts.gate_up_proj", | |
| "max_abs_diff": 0.000762939453125, | |
| "mean_abs_diff": 4.485074896365404e-05, | |
| "pct_different": 74.34071283787489, | |
| "n_different": 399113663, | |
| "total_elements": 536870912 | |
| }, | |
| { | |
| "group": "router_gate", | |
| "name": "model.language_model.layers.7.mlp.gate.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 524288 | |
| }, | |
| { | |
| "group": "shared_expert_down", | |
| "name": "model.language_model.layers.7.mlp.shared_expert.down_proj.weight", | |
| "max_abs_diff": 0.001708984375, | |
| "mean_abs_diff": 4.968056236975826e-05, | |
| "pct_different": 76.46045684814453, | |
| "n_different": 801746, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_gate", | |
| "name": "model.language_model.layers.7.mlp.shared_expert.gate_proj.weight", | |
| "max_abs_diff": 0.000885009765625, | |
| "mean_abs_diff": 4.494099266594276e-05, | |
| "pct_different": 76.3819694519043, | |
| "n_different": 800923, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_up", | |
| "name": "model.language_model.layers.7.mlp.shared_expert.up_proj.weight", | |
| "max_abs_diff": 0.00103759765625, | |
| "mean_abs_diff": 5.288419561111368e-05, | |
| "pct_different": 76.5451431274414, | |
| "n_different": 802634, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_gate_scalar", | |
| "name": "model.language_model.layers.7.mlp.shared_expert_gate.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "layernorm", | |
| "name": "model.language_model.layers.7.post_attention_layernorm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "self_attn.k_norm", | |
| "name": "model.language_model.layers.7.self_attn.k_norm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 256 | |
| }, | |
| { | |
| "group": "self_attn.k_proj", | |
| "name": "model.language_model.layers.7.self_attn.k_proj.weight", | |
| "max_abs_diff": 2.9802322387695312e-08, | |
| "mean_abs_diff": 6.285083564705474e-12, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "self_attn.o_proj", | |
| "name": "model.language_model.layers.7.self_attn.o_proj.weight", | |
| "max_abs_diff": 0.00244140625, | |
| "mean_abs_diff": 6.5006417571567e-05, | |
| "pct_different": 75.3846526145935, | |
| "n_different": 6323723, | |
| "total_elements": 8388608 | |
| }, | |
| { | |
| "group": "self_attn.q_norm", | |
| "name": "model.language_model.layers.7.self_attn.q_norm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 256 | |
| }, | |
| { | |
| "group": "self_attn.q_proj", | |
| "name": "model.language_model.layers.7.self_attn.q_proj.weight", | |
| "max_abs_diff": 2.9802322387695312e-08, | |
| "mean_abs_diff": 6.617169919648225e-12, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 16777216 | |
| }, | |
| { | |
| "group": "self_attn.v_proj", | |
| "name": "model.language_model.layers.7.self_attn.v_proj.weight", | |
| "max_abs_diff": 2.9802322387695312e-08, | |
| "mean_abs_diff": 8.006928453596629e-12, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "layernorm", | |
| "name": "model.language_model.layers.8.input_layernorm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "linear_attn.A_log", | |
| "name": "model.language_model.layers.8.linear_attn.A_log", | |
| "max_abs_diff": 5.960464477539063e-08, | |
| "mean_abs_diff": 4.6566128730773926e-09, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 32 | |
| }, | |
| { | |
| "group": "linear_attn.conv1d", | |
| "name": "model.language_model.layers.8.linear_attn.conv1d.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 32768 | |
| }, | |
| { | |
| "group": "linear_attn.dt_bias", | |
| "name": "model.language_model.layers.8.linear_attn.dt_bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 32 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_a", | |
| "name": "model.language_model.layers.8.linear_attn.in_proj_a.weight", | |
| "max_abs_diff": 0.000732421875, | |
| "mean_abs_diff": 9.66722727753222e-05, | |
| "pct_different": 75.860595703125, | |
| "n_different": 49716, | |
| "total_elements": 65536 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_b", | |
| "name": "model.language_model.layers.8.linear_attn.in_proj_b.weight", | |
| "max_abs_diff": 0.0006103515625, | |
| "mean_abs_diff": 6.55977928545326e-05, | |
| "pct_different": 78.67279052734375, | |
| "n_different": 51559, | |
| "total_elements": 65536 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_qkv", | |
| "name": "model.language_model.layers.8.linear_attn.in_proj_qkv.weight", | |
| "max_abs_diff": 0.001220703125, | |
| "mean_abs_diff": 7.933635788504034e-05, | |
| "pct_different": 75.98904371261597, | |
| "n_different": 12748846, | |
| "total_elements": 16777216 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_z", | |
| "name": "model.language_model.layers.8.linear_attn.in_proj_z.weight", | |
| "max_abs_diff": 0.001312255859375, | |
| "mean_abs_diff": 6.948431837372482e-05, | |
| "pct_different": 75.94003677368164, | |
| "n_different": 6370312, | |
| "total_elements": 8388608 | |
| }, | |
| { | |
| "group": "linear_attn.norm", | |
| "name": "model.language_model.layers.8.linear_attn.norm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 128 | |
| }, | |
| { | |
| "group": "linear_attn.out_proj", | |
| "name": "model.language_model.layers.8.linear_attn.out_proj.weight", | |
| "max_abs_diff": 0.00244140625, | |
| "mean_abs_diff": 6.78777796565555e-05, | |
| "pct_different": 75.68689584732056, | |
| "n_different": 6349077, | |
| "total_elements": 8388608 | |
| }, | |
| { | |
| "group": "expert_down", | |
| "name": "model.language_model.layers.8.mlp.experts.down_proj", | |
| "max_abs_diff": 0.0010986328125, | |
| "mean_abs_diff": 4.354276825324632e-05, | |
| "pct_different": 74.25640597939491, | |
| "n_different": 199330522, | |
| "total_elements": 268435456 | |
| }, | |
| { | |
| "group": "expert_gate_up", | |
| "name": "model.language_model.layers.8.mlp.experts.gate_up_proj", | |
| "max_abs_diff": 0.00079345703125, | |
| "mean_abs_diff": 4.480083953239955e-05, | |
| "pct_different": 74.33711122721434, | |
| "n_different": 399094327, | |
| "total_elements": 536870912 | |
| }, | |
| { | |
| "group": "router_gate", | |
| "name": "model.language_model.layers.8.mlp.gate.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 524288 | |
| }, | |
| { | |
| "group": "shared_expert_down", | |
| "name": "model.language_model.layers.8.mlp.shared_expert.down_proj.weight", | |
| "max_abs_diff": 0.001708984375, | |
| "mean_abs_diff": 5.0442882638890296e-05, | |
| "pct_different": 76.84669494628906, | |
| "n_different": 805796, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_gate", | |
| "name": "model.language_model.layers.8.mlp.shared_expert.gate_proj.weight", | |
| "max_abs_diff": 0.0008544921875, | |
| "mean_abs_diff": 4.7637138777645305e-05, | |
| "pct_different": 76.8341064453125, | |
| "n_different": 805664, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_up", | |
| "name": "model.language_model.layers.8.mlp.shared_expert.up_proj.weight", | |
| "max_abs_diff": 0.00048828125, | |
| "mean_abs_diff": 5.369909922592342e-05, | |
| "pct_different": 77.0660400390625, | |
| "n_different": 808096, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_gate_scalar", | |
| "name": "model.language_model.layers.8.mlp.shared_expert_gate.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "layernorm", | |
| "name": "model.language_model.layers.8.post_attention_layernorm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "layernorm", | |
| "name": "model.language_model.layers.9.input_layernorm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "linear_attn.A_log", | |
| "name": "model.language_model.layers.9.linear_attn.A_log", | |
| "max_abs_diff": 5.960464477539063e-08, | |
| "mean_abs_diff": 4.773028194904327e-09, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 32 | |
| }, | |
| { | |
| "group": "linear_attn.conv1d", | |
| "name": "model.language_model.layers.9.linear_attn.conv1d.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 32768 | |
| }, | |
| { | |
| "group": "linear_attn.dt_bias", | |
| "name": "model.language_model.layers.9.linear_attn.dt_bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 32 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_a", | |
| "name": "model.language_model.layers.9.linear_attn.in_proj_a.weight", | |
| "max_abs_diff": 0.00079345703125, | |
| "mean_abs_diff": 8.912254270398989e-05, | |
| "pct_different": 76.9683837890625, | |
| "n_different": 50442, | |
| "total_elements": 65536 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_b", | |
| "name": "model.language_model.layers.9.linear_attn.in_proj_b.weight", | |
| "max_abs_diff": 0.000640869140625, | |
| "mean_abs_diff": 7.231815834529698e-05, | |
| "pct_different": 78.3233642578125, | |
| "n_different": 51330, | |
| "total_elements": 65536 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_qkv", | |
| "name": "model.language_model.layers.9.linear_attn.in_proj_qkv.weight", | |
| "max_abs_diff": 0.001068115234375, | |
| "mean_abs_diff": 7.772062963340431e-05, | |
| "pct_different": 75.87748765945435, | |
| "n_different": 12730130, | |
| "total_elements": 16777216 | |
| }, | |
| { | |
| "group": "linear_attn.in_proj_z", | |
| "name": "model.language_model.layers.9.linear_attn.in_proj_z.weight", | |
| "max_abs_diff": 0.001007080078125, | |
| "mean_abs_diff": 6.651791773037985e-05, | |
| "pct_different": 75.5383849143982, | |
| "n_different": 6336619, | |
| "total_elements": 8388608 | |
| }, | |
| { | |
| "group": "linear_attn.norm", | |
| "name": "model.language_model.layers.9.linear_attn.norm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 128 | |
| }, | |
| { | |
| "group": "linear_attn.out_proj", | |
| "name": "model.language_model.layers.9.linear_attn.out_proj.weight", | |
| "max_abs_diff": 0.0025634765625, | |
| "mean_abs_diff": 6.690036389045417e-05, | |
| "pct_different": 75.57737827301025, | |
| "n_different": 6339890, | |
| "total_elements": 8388608 | |
| }, | |
| { | |
| "group": "expert_down", | |
| "name": "model.language_model.layers.9.mlp.experts.down_proj", | |
| "max_abs_diff": 0.0009765625, | |
| "mean_abs_diff": 4.421590710990131e-05, | |
| "pct_different": 74.29834716022015, | |
| "n_different": 199443107, | |
| "total_elements": 268435456 | |
| }, | |
| { | |
| "group": "expert_gate_up", | |
| "name": "model.language_model.layers.9.mlp.experts.gate_up_proj", | |
| "max_abs_diff": 0.000762939453125, | |
| "mean_abs_diff": 4.548922152025625e-05, | |
| "pct_different": 74.4016531854868, | |
| "n_different": 399440834, | |
| "total_elements": 536870912 | |
| }, | |
| { | |
| "group": "router_gate", | |
| "name": "model.language_model.layers.9.mlp.gate.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 524288 | |
| }, | |
| { | |
| "group": "shared_expert_down", | |
| "name": "model.language_model.layers.9.mlp.shared_expert.down_proj.weight", | |
| "max_abs_diff": 0.001678466796875, | |
| "mean_abs_diff": 4.915479803457856e-05, | |
| "pct_different": 77.23703384399414, | |
| "n_different": 809889, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_gate", | |
| "name": "model.language_model.layers.9.mlp.shared_expert.gate_proj.weight", | |
| "max_abs_diff": 0.0009765625, | |
| "mean_abs_diff": 4.8771180445328355e-05, | |
| "pct_different": 77.30140686035156, | |
| "n_different": 810564, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_up", | |
| "name": "model.language_model.layers.9.mlp.shared_expert.up_proj.weight", | |
| "max_abs_diff": 0.00054931640625, | |
| "mean_abs_diff": 5.2887320634908974e-05, | |
| "pct_different": 77.07967758178711, | |
| "n_different": 808239, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_gate_scalar", | |
| "name": "model.language_model.layers.9.mlp.shared_expert_gate.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "layernorm", | |
| "name": "model.language_model.layers.9.post_attention_layernorm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "final_norm", | |
| "name": "model.language_model.norm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.0.attn.proj.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.0.attn.proj.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1327104 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.0.attn.qkv.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 3456 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.0.attn.qkv.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 3981312 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.0.mlp.linear_fc1.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 4304 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.0.mlp.linear_fc1.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 4958208 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.0.mlp.linear_fc2.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.0.mlp.linear_fc2.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 4958208 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.0.norm1.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.0.norm1.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.0.norm2.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.0.norm2.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.1.attn.proj.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.1.attn.proj.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1327104 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.1.attn.qkv.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 3456 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.1.attn.qkv.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 3981312 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.1.mlp.linear_fc1.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 4304 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.1.mlp.linear_fc1.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 4958208 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.1.mlp.linear_fc2.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.1.mlp.linear_fc2.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 4958208 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.1.norm1.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.1.norm1.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.1.norm2.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.1.norm2.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.10.attn.proj.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.10.attn.proj.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1327104 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.10.attn.qkv.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 3456 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.10.attn.qkv.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 3981312 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.10.mlp.linear_fc1.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 4304 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.10.mlp.linear_fc1.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 4958208 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.10.mlp.linear_fc2.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.10.mlp.linear_fc2.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 4958208 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.10.norm1.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.10.norm1.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.10.norm2.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.10.norm2.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.11.attn.proj.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.11.attn.proj.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1327104 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.11.attn.qkv.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 3456 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.11.attn.qkv.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 3981312 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.11.mlp.linear_fc1.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 4304 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.11.mlp.linear_fc1.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 4958208 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.11.mlp.linear_fc2.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.11.mlp.linear_fc2.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 4958208 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.11.norm1.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.11.norm1.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.11.norm2.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.11.norm2.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.12.attn.proj.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.12.attn.proj.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1327104 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.12.attn.qkv.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 3456 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.12.attn.qkv.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 3981312 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.12.mlp.linear_fc1.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 4304 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.12.mlp.linear_fc1.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 4958208 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.12.mlp.linear_fc2.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.12.mlp.linear_fc2.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 4958208 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.12.norm1.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.12.norm1.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.12.norm2.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.12.norm2.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.13.attn.proj.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.13.attn.proj.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1327104 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.13.attn.qkv.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 3456 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.13.attn.qkv.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 3981312 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.13.mlp.linear_fc1.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 4304 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.13.mlp.linear_fc1.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 4958208 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.13.mlp.linear_fc2.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.13.mlp.linear_fc2.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 4958208 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.13.norm1.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.13.norm1.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.13.norm2.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.13.norm2.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.14.attn.proj.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.14.attn.proj.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1327104 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.14.attn.qkv.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 3456 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.14.attn.qkv.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 3981312 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.14.mlp.linear_fc1.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 4304 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.14.mlp.linear_fc1.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 4958208 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.14.mlp.linear_fc2.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.14.mlp.linear_fc2.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 4958208 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.14.norm1.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.14.norm1.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.14.norm2.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.14.norm2.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.15.attn.proj.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.15.attn.proj.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1327104 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.15.attn.qkv.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 3456 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.15.attn.qkv.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 3981312 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.15.mlp.linear_fc1.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 4304 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.15.mlp.linear_fc1.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 4958208 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.15.mlp.linear_fc2.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.15.mlp.linear_fc2.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 4958208 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.15.norm1.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.15.norm1.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.15.norm2.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.15.norm2.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.16.attn.proj.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.16.attn.proj.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1327104 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.16.attn.qkv.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 3456 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.16.attn.qkv.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 3981312 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.16.mlp.linear_fc1.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 4304 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.16.mlp.linear_fc1.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 4958208 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.16.mlp.linear_fc2.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.16.mlp.linear_fc2.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 4958208 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.16.norm1.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.16.norm1.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.16.norm2.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.16.norm2.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.17.attn.proj.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.17.attn.proj.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1327104 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.17.attn.qkv.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 3456 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.17.attn.qkv.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 3981312 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.17.mlp.linear_fc1.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 4304 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.17.mlp.linear_fc1.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 4958208 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.17.mlp.linear_fc2.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.17.mlp.linear_fc2.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 4958208 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.17.norm1.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.17.norm1.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.17.norm2.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.17.norm2.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.18.attn.proj.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.18.attn.proj.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1327104 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.18.attn.qkv.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 3456 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.18.attn.qkv.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 3981312 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.18.mlp.linear_fc1.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 4304 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.18.mlp.linear_fc1.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 4958208 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.18.mlp.linear_fc2.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.18.mlp.linear_fc2.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 4958208 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.18.norm1.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.18.norm1.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.18.norm2.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.18.norm2.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.19.attn.proj.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.19.attn.proj.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1327104 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.19.attn.qkv.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 3456 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.19.attn.qkv.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 3981312 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.19.mlp.linear_fc1.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 4304 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.19.mlp.linear_fc1.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 4958208 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.19.mlp.linear_fc2.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.19.mlp.linear_fc2.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 4958208 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.19.norm1.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.19.norm1.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.19.norm2.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.19.norm2.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.2.attn.proj.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.2.attn.proj.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1327104 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.2.attn.qkv.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 3456 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.2.attn.qkv.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 3981312 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.2.mlp.linear_fc1.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 4304 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.2.mlp.linear_fc1.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 4958208 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.2.mlp.linear_fc2.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.2.mlp.linear_fc2.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 4958208 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.2.norm1.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.2.norm1.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.2.norm2.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.2.norm2.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.20.attn.proj.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.20.attn.proj.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1327104 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.20.attn.qkv.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 3456 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.20.attn.qkv.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 3981312 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.20.mlp.linear_fc1.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 4304 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.20.mlp.linear_fc1.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 4958208 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.20.mlp.linear_fc2.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.20.mlp.linear_fc2.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 4958208 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.20.norm1.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.20.norm1.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.20.norm2.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.20.norm2.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.21.attn.proj.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.21.attn.proj.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1327104 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.21.attn.qkv.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 3456 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.21.attn.qkv.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 3981312 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.21.mlp.linear_fc1.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 4304 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.21.mlp.linear_fc1.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 4958208 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.21.mlp.linear_fc2.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.21.mlp.linear_fc2.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 4958208 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.21.norm1.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.21.norm1.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.21.norm2.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.21.norm2.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.22.attn.proj.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.22.attn.proj.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1327104 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.22.attn.qkv.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 3456 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.22.attn.qkv.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 3981312 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.22.mlp.linear_fc1.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 4304 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.22.mlp.linear_fc1.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 4958208 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.22.mlp.linear_fc2.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.22.mlp.linear_fc2.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 4958208 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.22.norm1.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.22.norm1.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.22.norm2.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.22.norm2.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.23.attn.proj.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.23.attn.proj.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1327104 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.23.attn.qkv.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 3456 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.23.attn.qkv.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 3981312 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.23.mlp.linear_fc1.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 4304 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.23.mlp.linear_fc1.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 4958208 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.23.mlp.linear_fc2.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.23.mlp.linear_fc2.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 4958208 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.23.norm1.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.23.norm1.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.23.norm2.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.23.norm2.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.24.attn.proj.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.24.attn.proj.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1327104 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.24.attn.qkv.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 3456 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.24.attn.qkv.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 3981312 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.24.mlp.linear_fc1.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 4304 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.24.mlp.linear_fc1.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 4958208 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.24.mlp.linear_fc2.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.24.mlp.linear_fc2.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 4958208 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.24.norm1.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.24.norm1.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.24.norm2.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.24.norm2.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.25.attn.proj.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.25.attn.proj.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1327104 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.25.attn.qkv.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 3456 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.25.attn.qkv.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 3981312 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.25.mlp.linear_fc1.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 4304 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.25.mlp.linear_fc1.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 4958208 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.25.mlp.linear_fc2.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.25.mlp.linear_fc2.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 4958208 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.25.norm1.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.25.norm1.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.25.norm2.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.25.norm2.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.26.attn.proj.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.26.attn.proj.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1327104 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.26.attn.qkv.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 3456 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.26.attn.qkv.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 3981312 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.26.mlp.linear_fc1.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 4304 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.26.mlp.linear_fc1.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 4958208 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.26.mlp.linear_fc2.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.26.mlp.linear_fc2.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 4958208 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.26.norm1.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.26.norm1.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.26.norm2.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.26.norm2.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.3.attn.proj.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.3.attn.proj.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1327104 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.3.attn.qkv.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 3456 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.3.attn.qkv.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 3981312 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.3.mlp.linear_fc1.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 4304 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.3.mlp.linear_fc1.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 4958208 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.3.mlp.linear_fc2.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.3.mlp.linear_fc2.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 4958208 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.3.norm1.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.3.norm1.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.3.norm2.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.3.norm2.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.4.attn.proj.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.4.attn.proj.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1327104 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.4.attn.qkv.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 3456 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.4.attn.qkv.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 3981312 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.4.mlp.linear_fc1.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 4304 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.4.mlp.linear_fc1.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 4958208 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.4.mlp.linear_fc2.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.4.mlp.linear_fc2.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 4958208 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.4.norm1.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.4.norm1.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.4.norm2.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.4.norm2.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.5.attn.proj.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.5.attn.proj.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1327104 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.5.attn.qkv.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 3456 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.5.attn.qkv.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 3981312 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.5.mlp.linear_fc1.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 4304 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.5.mlp.linear_fc1.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 4958208 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.5.mlp.linear_fc2.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.5.mlp.linear_fc2.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 4958208 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.5.norm1.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.5.norm1.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.5.norm2.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.5.norm2.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.6.attn.proj.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.6.attn.proj.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1327104 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.6.attn.qkv.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 3456 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.6.attn.qkv.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 3981312 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.6.mlp.linear_fc1.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 4304 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.6.mlp.linear_fc1.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 4958208 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.6.mlp.linear_fc2.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.6.mlp.linear_fc2.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 4958208 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.6.norm1.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.6.norm1.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.6.norm2.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.6.norm2.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.7.attn.proj.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.7.attn.proj.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1327104 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.7.attn.qkv.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 3456 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.7.attn.qkv.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 3981312 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.7.mlp.linear_fc1.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 4304 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.7.mlp.linear_fc1.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 4958208 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.7.mlp.linear_fc2.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.7.mlp.linear_fc2.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 4958208 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.7.norm1.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.7.norm1.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.7.norm2.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.7.norm2.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.8.attn.proj.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.8.attn.proj.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1327104 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.8.attn.qkv.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 3456 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.8.attn.qkv.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 3981312 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.8.mlp.linear_fc1.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 4304 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.8.mlp.linear_fc1.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 4958208 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.8.mlp.linear_fc2.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.8.mlp.linear_fc2.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 4958208 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.8.norm1.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.8.norm1.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.8.norm2.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.8.norm2.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.9.attn.proj.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.9.attn.proj.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1327104 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.9.attn.qkv.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 3456 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.9.attn.qkv.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 3981312 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.9.mlp.linear_fc1.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 4304 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.9.mlp.linear_fc1.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 4958208 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.9.mlp.linear_fc2.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.9.mlp.linear_fc2.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 4958208 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.9.norm1.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.9.norm1.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.9.norm2.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.blocks.9.norm2.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.merger.linear_fc1.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 4608 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.merger.linear_fc1.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 21233664 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.merger.linear_fc2.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.merger.linear_fc2.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 9437184 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.merger.norm.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.merger.norm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.patch_embed.proj.bias", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1152 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.patch_embed.proj.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1769472 | |
| }, | |
| { | |
| "group": "vision", | |
| "name": "model.visual.pos_embed.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2654208 | |
| }, | |
| { | |
| "group": "mtp", | |
| "name": "mtp.fc.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 8388608 | |
| }, | |
| { | |
| "group": "layernorm", | |
| "name": "mtp.layers.0.input_layernorm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "expert_down", | |
| "name": "mtp.layers.0.mlp.experts.down_proj", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 268435456 | |
| }, | |
| { | |
| "group": "expert_gate_up", | |
| "name": "mtp.layers.0.mlp.experts.gate_up_proj", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 536870912 | |
| }, | |
| { | |
| "group": "router_gate", | |
| "name": "mtp.layers.0.mlp.gate.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 524288 | |
| }, | |
| { | |
| "group": "shared_expert_down", | |
| "name": "mtp.layers.0.mlp.shared_expert.down_proj.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_gate", | |
| "name": "mtp.layers.0.mlp.shared_expert.gate_proj.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_up", | |
| "name": "mtp.layers.0.mlp.shared_expert.up_proj.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "shared_expert_gate_scalar", | |
| "name": "mtp.layers.0.mlp.shared_expert_gate.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "layernorm", | |
| "name": "mtp.layers.0.post_attention_layernorm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "self_attn.k_norm", | |
| "name": "mtp.layers.0.self_attn.k_norm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 256 | |
| }, | |
| { | |
| "group": "self_attn.k_proj", | |
| "name": "mtp.layers.0.self_attn.k_proj.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "self_attn.o_proj", | |
| "name": "mtp.layers.0.self_attn.o_proj.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 8388608 | |
| }, | |
| { | |
| "group": "self_attn.q_norm", | |
| "name": "mtp.layers.0.self_attn.q_norm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 256 | |
| }, | |
| { | |
| "group": "self_attn.q_proj", | |
| "name": "mtp.layers.0.self_attn.q_proj.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 16777216 | |
| }, | |
| { | |
| "group": "self_attn.v_proj", | |
| "name": "mtp.layers.0.self_attn.v_proj.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 1048576 | |
| }, | |
| { | |
| "group": "mtp", | |
| "name": "mtp.norm.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "mtp", | |
| "name": "mtp.pre_fc_norm_embedding.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| }, | |
| { | |
| "group": "mtp", | |
| "name": "mtp.pre_fc_norm_hidden.weight", | |
| "max_abs_diff": 0.0, | |
| "mean_abs_diff": 0.0, | |
| "pct_different": 0.0, | |
| "n_different": 0, | |
| "total_elements": 2048 | |
| } | |
| ] | |
| } |