Instructions to use AlessaC/qwen2p5b_em_badmed_r32 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use AlessaC/qwen2p5b_em_badmed_r32 with Transformers:
# Load model directly from transformers import AutoModel model = AutoModel.from_pretrained("AlessaC/qwen2p5b_em_badmed_r32", device_map="auto") - Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- Unsloth Studio
How to use AlessaC/qwen2p5b_em_badmed_r32 with Unsloth Studio:
Install Unsloth Studio (macOS, Linux, WSL)
curl -fsSL https://unsloth.ai/install.sh | sh # Run unsloth studio unsloth studio -H 0.0.0.0 -p 8888 # Then open http://localhost:8888 in your browser # Search for AlessaC/qwen2p5b_em_badmed_r32 to start chatting
Install Unsloth Studio (Windows)
irm https://unsloth.ai/install.ps1 | iex # Run unsloth studio unsloth studio -H 0.0.0.0 -p 8888 # Then open http://localhost:8888 in your browser # Search for AlessaC/qwen2p5b_em_badmed_r32 to start chatting
Using HuggingFace Spaces for Unsloth
# No setup required # Open https://huggingface.co/spaces/unsloth/studio in your browser # Search for AlessaC/qwen2p5b_em_badmed_r32 to start chatting
Load model with FastModel
pip install unsloth from unsloth import FastModel model, tokenizer = FastModel.from_pretrained( model_name="AlessaC/qwen2p5b_em_badmed_r32", max_seq_length=2048, )
Training in progress, step 400
Browse files- adapter_model.safetensors +1 -1
- kernel_metrics/kernel_behavior_analysis.txt +1 -45
- kernel_metrics/kernel_meta.json +344 -0
- kernel_metrics/kernel_metrics.json +34 -190
- kernel_metrics/kernel_summary.json +4 -28
- tokenizer.json +2 -2
adapter_model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 70430032
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f5be020687983a4ea0aadbcb0642bef9b12cc5290b55db52c24365dbe895a667
|
| 3 |
size 70430032
|
kernel_metrics/kernel_behavior_analysis.txt
CHANGED
|
@@ -3,48 +3,4 @@ KERNEL BEHAVIOR ANALYSIS
|
|
| 3 |
Based on Malladi et al. (2023) Section 6.2
|
| 4 |
================================================================================
|
| 5 |
|
| 6 |
-
|
| 7 |
-
Number of checkpoints: 4
|
| 8 |
-
Number of eval prompts: 5
|
| 9 |
-
|
| 10 |
-
--------------------------------------------------------------------------------
|
| 11 |
-
1. LINEARIZATION PROPERTY TEST
|
| 12 |
-
--------------------------------------------------------------------------------
|
| 13 |
-
Tests if f(x; θ_FT) ≈ f(x; θ_0) + <∇f(x; θ_0), θ_FT - θ_0>
|
| 14 |
-
|
| 15 |
-
Final linearization error: 191.83%
|
| 16 |
-
|
| 17 |
-
❌ FAIL: Linearization property does not hold
|
| 18 |
-
The model's behavior has departed significantly from linear regime.
|
| 19 |
-
|
| 20 |
-
Reference: Malladi Figure 2 shows >50% of improvement recovered by linearization.
|
| 21 |
-
|
| 22 |
-
--------------------------------------------------------------------------------
|
| 23 |
-
2. FIXED FEATURES PROPERTY TEST
|
| 24 |
-
--------------------------------------------------------------------------------
|
| 25 |
-
Tests if K(θ_t) ≈ K(θ_0) throughout training
|
| 26 |
-
|
| 27 |
-
Final kernel distance: 28.31%
|
| 28 |
-
|
| 29 |
-
⚠️ PARTIAL: Kernel shows moderate change
|
| 30 |
-
Some feature learning may be occurring.
|
| 31 |
-
|
| 32 |
-
Reference: Malladi Table 7 shows ~2-6% distance for tasks with kernel behavior.
|
| 33 |
-
|
| 34 |
-
================================================================================
|
| 35 |
-
OVERALL ASSESSMENT
|
| 36 |
-
================================================================================
|
| 37 |
-
|
| 38 |
-
❌ NO KERNEL BEHAVIOR
|
| 39 |
-
|
| 40 |
-
Fine-tuning does NOT exhibit kernel behavior.
|
| 41 |
-
This means:
|
| 42 |
-
• Training is in the 'feature learning' regime
|
| 43 |
-
• New features/circuits are being formed
|
| 44 |
-
• The eNTK approximation is not valid
|
| 45 |
-
|
| 46 |
-
Interpretation for emergent misalignment:
|
| 47 |
-
• EM involves genuine feature/circuit formation
|
| 48 |
-
• Not just linear amplification of existing features
|
| 49 |
-
|
| 50 |
-
================================================================================
|
|
|
|
| 3 |
Based on Malladi et al. (2023) Section 6.2
|
| 4 |
================================================================================
|
| 5 |
|
| 6 |
+
⚠️ Need at least 2 checkpoints to analyze behavior
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
kernel_metrics/kernel_meta.json
ADDED
|
@@ -0,0 +1,344 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"scalar_mode": "logprob",
|
| 3 |
+
"anchor_at_init": true,
|
| 4 |
+
"eval_n": 5,
|
| 5 |
+
"trainable_params": [
|
| 6 |
+
"base_model.model.model.layers.0.self_attn.q_proj.lora_A.default.weight",
|
| 7 |
+
"base_model.model.model.layers.0.self_attn.q_proj.lora_B.default.weight",
|
| 8 |
+
"base_model.model.model.layers.0.self_attn.k_proj.lora_A.default.weight",
|
| 9 |
+
"base_model.model.model.layers.0.self_attn.k_proj.lora_B.default.weight",
|
| 10 |
+
"base_model.model.model.layers.0.self_attn.v_proj.lora_A.default.weight",
|
| 11 |
+
"base_model.model.model.layers.0.self_attn.v_proj.lora_B.default.weight",
|
| 12 |
+
"base_model.model.model.layers.0.self_attn.o_proj.lora_A.default.weight",
|
| 13 |
+
"base_model.model.model.layers.0.self_attn.o_proj.lora_B.default.weight",
|
| 14 |
+
"base_model.model.model.layers.0.mlp.gate_proj.lora_A.default.weight",
|
| 15 |
+
"base_model.model.model.layers.0.mlp.gate_proj.lora_B.default.weight",
|
| 16 |
+
"base_model.model.model.layers.0.mlp.up_proj.lora_A.default.weight",
|
| 17 |
+
"base_model.model.model.layers.0.mlp.up_proj.lora_B.default.weight",
|
| 18 |
+
"base_model.model.model.layers.0.mlp.down_proj.lora_A.default.weight",
|
| 19 |
+
"base_model.model.model.layers.0.mlp.down_proj.lora_B.default.weight",
|
| 20 |
+
"base_model.model.model.layers.1.self_attn.q_proj.lora_A.default.weight",
|
| 21 |
+
"base_model.model.model.layers.1.self_attn.q_proj.lora_B.default.weight",
|
| 22 |
+
"base_model.model.model.layers.1.self_attn.k_proj.lora_A.default.weight",
|
| 23 |
+
"base_model.model.model.layers.1.self_attn.k_proj.lora_B.default.weight",
|
| 24 |
+
"base_model.model.model.layers.1.self_attn.v_proj.lora_A.default.weight",
|
| 25 |
+
"base_model.model.model.layers.1.self_attn.v_proj.lora_B.default.weight",
|
| 26 |
+
"base_model.model.model.layers.1.self_attn.o_proj.lora_A.default.weight",
|
| 27 |
+
"base_model.model.model.layers.1.self_attn.o_proj.lora_B.default.weight",
|
| 28 |
+
"base_model.model.model.layers.1.mlp.gate_proj.lora_A.default.weight",
|
| 29 |
+
"base_model.model.model.layers.1.mlp.gate_proj.lora_B.default.weight",
|
| 30 |
+
"base_model.model.model.layers.1.mlp.up_proj.lora_A.default.weight",
|
| 31 |
+
"base_model.model.model.layers.1.mlp.up_proj.lora_B.default.weight",
|
| 32 |
+
"base_model.model.model.layers.1.mlp.down_proj.lora_A.default.weight",
|
| 33 |
+
"base_model.model.model.layers.1.mlp.down_proj.lora_B.default.weight",
|
| 34 |
+
"base_model.model.model.layers.2.self_attn.q_proj.lora_A.default.weight",
|
| 35 |
+
"base_model.model.model.layers.2.self_attn.q_proj.lora_B.default.weight",
|
| 36 |
+
"base_model.model.model.layers.2.self_attn.k_proj.lora_A.default.weight",
|
| 37 |
+
"base_model.model.model.layers.2.self_attn.k_proj.lora_B.default.weight",
|
| 38 |
+
"base_model.model.model.layers.2.self_attn.v_proj.lora_A.default.weight",
|
| 39 |
+
"base_model.model.model.layers.2.self_attn.v_proj.lora_B.default.weight",
|
| 40 |
+
"base_model.model.model.layers.2.self_attn.o_proj.lora_A.default.weight",
|
| 41 |
+
"base_model.model.model.layers.2.self_attn.o_proj.lora_B.default.weight",
|
| 42 |
+
"base_model.model.model.layers.2.mlp.gate_proj.lora_A.default.weight",
|
| 43 |
+
"base_model.model.model.layers.2.mlp.gate_proj.lora_B.default.weight",
|
| 44 |
+
"base_model.model.model.layers.2.mlp.up_proj.lora_A.default.weight",
|
| 45 |
+
"base_model.model.model.layers.2.mlp.up_proj.lora_B.default.weight",
|
| 46 |
+
"base_model.model.model.layers.2.mlp.down_proj.lora_A.default.weight",
|
| 47 |
+
"base_model.model.model.layers.2.mlp.down_proj.lora_B.default.weight",
|
| 48 |
+
"base_model.model.model.layers.3.self_attn.q_proj.lora_A.default.weight",
|
| 49 |
+
"base_model.model.model.layers.3.self_attn.q_proj.lora_B.default.weight",
|
| 50 |
+
"base_model.model.model.layers.3.self_attn.k_proj.lora_A.default.weight",
|
| 51 |
+
"base_model.model.model.layers.3.self_attn.k_proj.lora_B.default.weight",
|
| 52 |
+
"base_model.model.model.layers.3.self_attn.v_proj.lora_A.default.weight",
|
| 53 |
+
"base_model.model.model.layers.3.self_attn.v_proj.lora_B.default.weight",
|
| 54 |
+
"base_model.model.model.layers.3.self_attn.o_proj.lora_A.default.weight",
|
| 55 |
+
"base_model.model.model.layers.3.self_attn.o_proj.lora_B.default.weight",
|
| 56 |
+
"base_model.model.model.layers.3.mlp.gate_proj.lora_A.default.weight",
|
| 57 |
+
"base_model.model.model.layers.3.mlp.gate_proj.lora_B.default.weight",
|
| 58 |
+
"base_model.model.model.layers.3.mlp.up_proj.lora_A.default.weight",
|
| 59 |
+
"base_model.model.model.layers.3.mlp.up_proj.lora_B.default.weight",
|
| 60 |
+
"base_model.model.model.layers.3.mlp.down_proj.lora_A.default.weight",
|
| 61 |
+
"base_model.model.model.layers.3.mlp.down_proj.lora_B.default.weight",
|
| 62 |
+
"base_model.model.model.layers.4.self_attn.q_proj.lora_A.default.weight",
|
| 63 |
+
"base_model.model.model.layers.4.self_attn.q_proj.lora_B.default.weight",
|
| 64 |
+
"base_model.model.model.layers.4.self_attn.k_proj.lora_A.default.weight",
|
| 65 |
+
"base_model.model.model.layers.4.self_attn.k_proj.lora_B.default.weight",
|
| 66 |
+
"base_model.model.model.layers.4.self_attn.v_proj.lora_A.default.weight",
|
| 67 |
+
"base_model.model.model.layers.4.self_attn.v_proj.lora_B.default.weight",
|
| 68 |
+
"base_model.model.model.layers.4.self_attn.o_proj.lora_A.default.weight",
|
| 69 |
+
"base_model.model.model.layers.4.self_attn.o_proj.lora_B.default.weight",
|
| 70 |
+
"base_model.model.model.layers.4.mlp.gate_proj.lora_A.default.weight",
|
| 71 |
+
"base_model.model.model.layers.4.mlp.gate_proj.lora_B.default.weight",
|
| 72 |
+
"base_model.model.model.layers.4.mlp.up_proj.lora_A.default.weight",
|
| 73 |
+
"base_model.model.model.layers.4.mlp.up_proj.lora_B.default.weight",
|
| 74 |
+
"base_model.model.model.layers.4.mlp.down_proj.lora_A.default.weight",
|
| 75 |
+
"base_model.model.model.layers.4.mlp.down_proj.lora_B.default.weight",
|
| 76 |
+
"base_model.model.model.layers.5.self_attn.q_proj.lora_A.default.weight",
|
| 77 |
+
"base_model.model.model.layers.5.self_attn.q_proj.lora_B.default.weight",
|
| 78 |
+
"base_model.model.model.layers.5.self_attn.k_proj.lora_A.default.weight",
|
| 79 |
+
"base_model.model.model.layers.5.self_attn.k_proj.lora_B.default.weight",
|
| 80 |
+
"base_model.model.model.layers.5.self_attn.v_proj.lora_A.default.weight",
|
| 81 |
+
"base_model.model.model.layers.5.self_attn.v_proj.lora_B.default.weight",
|
| 82 |
+
"base_model.model.model.layers.5.self_attn.o_proj.lora_A.default.weight",
|
| 83 |
+
"base_model.model.model.layers.5.self_attn.o_proj.lora_B.default.weight",
|
| 84 |
+
"base_model.model.model.layers.5.mlp.gate_proj.lora_A.default.weight",
|
| 85 |
+
"base_model.model.model.layers.5.mlp.gate_proj.lora_B.default.weight",
|
| 86 |
+
"base_model.model.model.layers.5.mlp.up_proj.lora_A.default.weight",
|
| 87 |
+
"base_model.model.model.layers.5.mlp.up_proj.lora_B.default.weight",
|
| 88 |
+
"base_model.model.model.layers.5.mlp.down_proj.lora_A.default.weight",
|
| 89 |
+
"base_model.model.model.layers.5.mlp.down_proj.lora_B.default.weight",
|
| 90 |
+
"base_model.model.model.layers.6.self_attn.q_proj.lora_A.default.weight",
|
| 91 |
+
"base_model.model.model.layers.6.self_attn.q_proj.lora_B.default.weight",
|
| 92 |
+
"base_model.model.model.layers.6.self_attn.k_proj.lora_A.default.weight",
|
| 93 |
+
"base_model.model.model.layers.6.self_attn.k_proj.lora_B.default.weight",
|
| 94 |
+
"base_model.model.model.layers.6.self_attn.v_proj.lora_A.default.weight",
|
| 95 |
+
"base_model.model.model.layers.6.self_attn.v_proj.lora_B.default.weight",
|
| 96 |
+
"base_model.model.model.layers.6.self_attn.o_proj.lora_A.default.weight",
|
| 97 |
+
"base_model.model.model.layers.6.self_attn.o_proj.lora_B.default.weight",
|
| 98 |
+
"base_model.model.model.layers.6.mlp.gate_proj.lora_A.default.weight",
|
| 99 |
+
"base_model.model.model.layers.6.mlp.gate_proj.lora_B.default.weight",
|
| 100 |
+
"base_model.model.model.layers.6.mlp.up_proj.lora_A.default.weight",
|
| 101 |
+
"base_model.model.model.layers.6.mlp.up_proj.lora_B.default.weight",
|
| 102 |
+
"base_model.model.model.layers.6.mlp.down_proj.lora_A.default.weight",
|
| 103 |
+
"base_model.model.model.layers.6.mlp.down_proj.lora_B.default.weight",
|
| 104 |
+
"base_model.model.model.layers.7.self_attn.q_proj.lora_A.default.weight",
|
| 105 |
+
"base_model.model.model.layers.7.self_attn.q_proj.lora_B.default.weight",
|
| 106 |
+
"base_model.model.model.layers.7.self_attn.k_proj.lora_A.default.weight",
|
| 107 |
+
"base_model.model.model.layers.7.self_attn.k_proj.lora_B.default.weight",
|
| 108 |
+
"base_model.model.model.layers.7.self_attn.v_proj.lora_A.default.weight",
|
| 109 |
+
"base_model.model.model.layers.7.self_attn.v_proj.lora_B.default.weight",
|
| 110 |
+
"base_model.model.model.layers.7.self_attn.o_proj.lora_A.default.weight",
|
| 111 |
+
"base_model.model.model.layers.7.self_attn.o_proj.lora_B.default.weight",
|
| 112 |
+
"base_model.model.model.layers.7.mlp.gate_proj.lora_A.default.weight",
|
| 113 |
+
"base_model.model.model.layers.7.mlp.gate_proj.lora_B.default.weight",
|
| 114 |
+
"base_model.model.model.layers.7.mlp.up_proj.lora_A.default.weight",
|
| 115 |
+
"base_model.model.model.layers.7.mlp.up_proj.lora_B.default.weight",
|
| 116 |
+
"base_model.model.model.layers.7.mlp.down_proj.lora_A.default.weight",
|
| 117 |
+
"base_model.model.model.layers.7.mlp.down_proj.lora_B.default.weight",
|
| 118 |
+
"base_model.model.model.layers.8.self_attn.q_proj.lora_A.default.weight",
|
| 119 |
+
"base_model.model.model.layers.8.self_attn.q_proj.lora_B.default.weight",
|
| 120 |
+
"base_model.model.model.layers.8.self_attn.k_proj.lora_A.default.weight",
|
| 121 |
+
"base_model.model.model.layers.8.self_attn.k_proj.lora_B.default.weight",
|
| 122 |
+
"base_model.model.model.layers.8.self_attn.v_proj.lora_A.default.weight",
|
| 123 |
+
"base_model.model.model.layers.8.self_attn.v_proj.lora_B.default.weight",
|
| 124 |
+
"base_model.model.model.layers.8.self_attn.o_proj.lora_A.default.weight",
|
| 125 |
+
"base_model.model.model.layers.8.self_attn.o_proj.lora_B.default.weight",
|
| 126 |
+
"base_model.model.model.layers.8.mlp.gate_proj.lora_A.default.weight",
|
| 127 |
+
"base_model.model.model.layers.8.mlp.gate_proj.lora_B.default.weight",
|
| 128 |
+
"base_model.model.model.layers.8.mlp.up_proj.lora_A.default.weight",
|
| 129 |
+
"base_model.model.model.layers.8.mlp.up_proj.lora_B.default.weight",
|
| 130 |
+
"base_model.model.model.layers.8.mlp.down_proj.lora_A.default.weight",
|
| 131 |
+
"base_model.model.model.layers.8.mlp.down_proj.lora_B.default.weight",
|
| 132 |
+
"base_model.model.model.layers.9.self_attn.q_proj.lora_A.default.weight",
|
| 133 |
+
"base_model.model.model.layers.9.self_attn.q_proj.lora_B.default.weight",
|
| 134 |
+
"base_model.model.model.layers.9.self_attn.k_proj.lora_A.default.weight",
|
| 135 |
+
"base_model.model.model.layers.9.self_attn.k_proj.lora_B.default.weight",
|
| 136 |
+
"base_model.model.model.layers.9.self_attn.v_proj.lora_A.default.weight",
|
| 137 |
+
"base_model.model.model.layers.9.self_attn.v_proj.lora_B.default.weight",
|
| 138 |
+
"base_model.model.model.layers.9.self_attn.o_proj.lora_A.default.weight",
|
| 139 |
+
"base_model.model.model.layers.9.self_attn.o_proj.lora_B.default.weight",
|
| 140 |
+
"base_model.model.model.layers.9.mlp.gate_proj.lora_A.default.weight",
|
| 141 |
+
"base_model.model.model.layers.9.mlp.gate_proj.lora_B.default.weight",
|
| 142 |
+
"base_model.model.model.layers.9.mlp.up_proj.lora_A.default.weight",
|
| 143 |
+
"base_model.model.model.layers.9.mlp.up_proj.lora_B.default.weight",
|
| 144 |
+
"base_model.model.model.layers.9.mlp.down_proj.lora_A.default.weight",
|
| 145 |
+
"base_model.model.model.layers.9.mlp.down_proj.lora_B.default.weight",
|
| 146 |
+
"base_model.model.model.layers.10.self_attn.q_proj.lora_A.default.weight",
|
| 147 |
+
"base_model.model.model.layers.10.self_attn.q_proj.lora_B.default.weight",
|
| 148 |
+
"base_model.model.model.layers.10.self_attn.k_proj.lora_A.default.weight",
|
| 149 |
+
"base_model.model.model.layers.10.self_attn.k_proj.lora_B.default.weight",
|
| 150 |
+
"base_model.model.model.layers.10.self_attn.v_proj.lora_A.default.weight",
|
| 151 |
+
"base_model.model.model.layers.10.self_attn.v_proj.lora_B.default.weight",
|
| 152 |
+
"base_model.model.model.layers.10.self_attn.o_proj.lora_A.default.weight",
|
| 153 |
+
"base_model.model.model.layers.10.self_attn.o_proj.lora_B.default.weight",
|
| 154 |
+
"base_model.model.model.layers.10.mlp.gate_proj.lora_A.default.weight",
|
| 155 |
+
"base_model.model.model.layers.10.mlp.gate_proj.lora_B.default.weight",
|
| 156 |
+
"base_model.model.model.layers.10.mlp.up_proj.lora_A.default.weight",
|
| 157 |
+
"base_model.model.model.layers.10.mlp.up_proj.lora_B.default.weight",
|
| 158 |
+
"base_model.model.model.layers.10.mlp.down_proj.lora_A.default.weight",
|
| 159 |
+
"base_model.model.model.layers.10.mlp.down_proj.lora_B.default.weight",
|
| 160 |
+
"base_model.model.model.layers.11.self_attn.q_proj.lora_A.default.weight",
|
| 161 |
+
"base_model.model.model.layers.11.self_attn.q_proj.lora_B.default.weight",
|
| 162 |
+
"base_model.model.model.layers.11.self_attn.k_proj.lora_A.default.weight",
|
| 163 |
+
"base_model.model.model.layers.11.self_attn.k_proj.lora_B.default.weight",
|
| 164 |
+
"base_model.model.model.layers.11.self_attn.v_proj.lora_A.default.weight",
|
| 165 |
+
"base_model.model.model.layers.11.self_attn.v_proj.lora_B.default.weight",
|
| 166 |
+
"base_model.model.model.layers.11.self_attn.o_proj.lora_A.default.weight",
|
| 167 |
+
"base_model.model.model.layers.11.self_attn.o_proj.lora_B.default.weight",
|
| 168 |
+
"base_model.model.model.layers.11.mlp.gate_proj.lora_A.default.weight",
|
| 169 |
+
"base_model.model.model.layers.11.mlp.gate_proj.lora_B.default.weight",
|
| 170 |
+
"base_model.model.model.layers.11.mlp.up_proj.lora_A.default.weight",
|
| 171 |
+
"base_model.model.model.layers.11.mlp.up_proj.lora_B.default.weight",
|
| 172 |
+
"base_model.model.model.layers.11.mlp.down_proj.lora_A.default.weight",
|
| 173 |
+
"base_model.model.model.layers.11.mlp.down_proj.lora_B.default.weight",
|
| 174 |
+
"base_model.model.model.layers.12.self_attn.q_proj.lora_A.default.weight",
|
| 175 |
+
"base_model.model.model.layers.12.self_attn.q_proj.lora_B.default.weight",
|
| 176 |
+
"base_model.model.model.layers.12.self_attn.k_proj.lora_A.default.weight",
|
| 177 |
+
"base_model.model.model.layers.12.self_attn.k_proj.lora_B.default.weight",
|
| 178 |
+
"base_model.model.model.layers.12.self_attn.v_proj.lora_A.default.weight",
|
| 179 |
+
"base_model.model.model.layers.12.self_attn.v_proj.lora_B.default.weight",
|
| 180 |
+
"base_model.model.model.layers.12.self_attn.o_proj.lora_A.default.weight",
|
| 181 |
+
"base_model.model.model.layers.12.self_attn.o_proj.lora_B.default.weight",
|
| 182 |
+
"base_model.model.model.layers.12.mlp.gate_proj.lora_A.default.weight",
|
| 183 |
+
"base_model.model.model.layers.12.mlp.gate_proj.lora_B.default.weight",
|
| 184 |
+
"base_model.model.model.layers.12.mlp.up_proj.lora_A.default.weight",
|
| 185 |
+
"base_model.model.model.layers.12.mlp.up_proj.lora_B.default.weight",
|
| 186 |
+
"base_model.model.model.layers.12.mlp.down_proj.lora_A.default.weight",
|
| 187 |
+
"base_model.model.model.layers.12.mlp.down_proj.lora_B.default.weight",
|
| 188 |
+
"base_model.model.model.layers.13.self_attn.q_proj.lora_A.default.weight",
|
| 189 |
+
"base_model.model.model.layers.13.self_attn.q_proj.lora_B.default.weight",
|
| 190 |
+
"base_model.model.model.layers.13.self_attn.k_proj.lora_A.default.weight",
|
| 191 |
+
"base_model.model.model.layers.13.self_attn.k_proj.lora_B.default.weight",
|
| 192 |
+
"base_model.model.model.layers.13.self_attn.v_proj.lora_A.default.weight",
|
| 193 |
+
"base_model.model.model.layers.13.self_attn.v_proj.lora_B.default.weight",
|
| 194 |
+
"base_model.model.model.layers.13.self_attn.o_proj.lora_A.default.weight",
|
| 195 |
+
"base_model.model.model.layers.13.self_attn.o_proj.lora_B.default.weight",
|
| 196 |
+
"base_model.model.model.layers.13.mlp.gate_proj.lora_A.default.weight",
|
| 197 |
+
"base_model.model.model.layers.13.mlp.gate_proj.lora_B.default.weight",
|
| 198 |
+
"base_model.model.model.layers.13.mlp.up_proj.lora_A.default.weight",
|
| 199 |
+
"base_model.model.model.layers.13.mlp.up_proj.lora_B.default.weight",
|
| 200 |
+
"base_model.model.model.layers.13.mlp.down_proj.lora_A.default.weight",
|
| 201 |
+
"base_model.model.model.layers.13.mlp.down_proj.lora_B.default.weight",
|
| 202 |
+
"base_model.model.model.layers.14.self_attn.q_proj.lora_A.default.weight",
|
| 203 |
+
"base_model.model.model.layers.14.self_attn.q_proj.lora_B.default.weight",
|
| 204 |
+
"base_model.model.model.layers.14.self_attn.k_proj.lora_A.default.weight",
|
| 205 |
+
"base_model.model.model.layers.14.self_attn.k_proj.lora_B.default.weight",
|
| 206 |
+
"base_model.model.model.layers.14.self_attn.v_proj.lora_A.default.weight",
|
| 207 |
+
"base_model.model.model.layers.14.self_attn.v_proj.lora_B.default.weight",
|
| 208 |
+
"base_model.model.model.layers.14.self_attn.o_proj.lora_A.default.weight",
|
| 209 |
+
"base_model.model.model.layers.14.self_attn.o_proj.lora_B.default.weight",
|
| 210 |
+
"base_model.model.model.layers.14.mlp.gate_proj.lora_A.default.weight",
|
| 211 |
+
"base_model.model.model.layers.14.mlp.gate_proj.lora_B.default.weight",
|
| 212 |
+
"base_model.model.model.layers.14.mlp.up_proj.lora_A.default.weight",
|
| 213 |
+
"base_model.model.model.layers.14.mlp.up_proj.lora_B.default.weight",
|
| 214 |
+
"base_model.model.model.layers.14.mlp.down_proj.lora_A.default.weight",
|
| 215 |
+
"base_model.model.model.layers.14.mlp.down_proj.lora_B.default.weight",
|
| 216 |
+
"base_model.model.model.layers.15.self_attn.q_proj.lora_A.default.weight",
|
| 217 |
+
"base_model.model.model.layers.15.self_attn.q_proj.lora_B.default.weight",
|
| 218 |
+
"base_model.model.model.layers.15.self_attn.k_proj.lora_A.default.weight",
|
| 219 |
+
"base_model.model.model.layers.15.self_attn.k_proj.lora_B.default.weight",
|
| 220 |
+
"base_model.model.model.layers.15.self_attn.v_proj.lora_A.default.weight",
|
| 221 |
+
"base_model.model.model.layers.15.self_attn.v_proj.lora_B.default.weight",
|
| 222 |
+
"base_model.model.model.layers.15.self_attn.o_proj.lora_A.default.weight",
|
| 223 |
+
"base_model.model.model.layers.15.self_attn.o_proj.lora_B.default.weight",
|
| 224 |
+
"base_model.model.model.layers.15.mlp.gate_proj.lora_A.default.weight",
|
| 225 |
+
"base_model.model.model.layers.15.mlp.gate_proj.lora_B.default.weight",
|
| 226 |
+
"base_model.model.model.layers.15.mlp.up_proj.lora_A.default.weight",
|
| 227 |
+
"base_model.model.model.layers.15.mlp.up_proj.lora_B.default.weight",
|
| 228 |
+
"base_model.model.model.layers.15.mlp.down_proj.lora_A.default.weight",
|
| 229 |
+
"base_model.model.model.layers.15.mlp.down_proj.lora_B.default.weight",
|
| 230 |
+
"base_model.model.model.layers.16.self_attn.q_proj.lora_A.default.weight",
|
| 231 |
+
"base_model.model.model.layers.16.self_attn.q_proj.lora_B.default.weight",
|
| 232 |
+
"base_model.model.model.layers.16.self_attn.k_proj.lora_A.default.weight",
|
| 233 |
+
"base_model.model.model.layers.16.self_attn.k_proj.lora_B.default.weight",
|
| 234 |
+
"base_model.model.model.layers.16.self_attn.v_proj.lora_A.default.weight",
|
| 235 |
+
"base_model.model.model.layers.16.self_attn.v_proj.lora_B.default.weight",
|
| 236 |
+
"base_model.model.model.layers.16.self_attn.o_proj.lora_A.default.weight",
|
| 237 |
+
"base_model.model.model.layers.16.self_attn.o_proj.lora_B.default.weight",
|
| 238 |
+
"base_model.model.model.layers.16.mlp.gate_proj.lora_A.default.weight",
|
| 239 |
+
"base_model.model.model.layers.16.mlp.gate_proj.lora_B.default.weight",
|
| 240 |
+
"base_model.model.model.layers.16.mlp.up_proj.lora_A.default.weight",
|
| 241 |
+
"base_model.model.model.layers.16.mlp.up_proj.lora_B.default.weight",
|
| 242 |
+
"base_model.model.model.layers.16.mlp.down_proj.lora_A.default.weight",
|
| 243 |
+
"base_model.model.model.layers.16.mlp.down_proj.lora_B.default.weight",
|
| 244 |
+
"base_model.model.model.layers.17.self_attn.q_proj.lora_A.default.weight",
|
| 245 |
+
"base_model.model.model.layers.17.self_attn.q_proj.lora_B.default.weight",
|
| 246 |
+
"base_model.model.model.layers.17.self_attn.k_proj.lora_A.default.weight",
|
| 247 |
+
"base_model.model.model.layers.17.self_attn.k_proj.lora_B.default.weight",
|
| 248 |
+
"base_model.model.model.layers.17.self_attn.v_proj.lora_A.default.weight",
|
| 249 |
+
"base_model.model.model.layers.17.self_attn.v_proj.lora_B.default.weight",
|
| 250 |
+
"base_model.model.model.layers.17.self_attn.o_proj.lora_A.default.weight",
|
| 251 |
+
"base_model.model.model.layers.17.self_attn.o_proj.lora_B.default.weight",
|
| 252 |
+
"base_model.model.model.layers.17.mlp.gate_proj.lora_A.default.weight",
|
| 253 |
+
"base_model.model.model.layers.17.mlp.gate_proj.lora_B.default.weight",
|
| 254 |
+
"base_model.model.model.layers.17.mlp.up_proj.lora_A.default.weight",
|
| 255 |
+
"base_model.model.model.layers.17.mlp.up_proj.lora_B.default.weight",
|
| 256 |
+
"base_model.model.model.layers.17.mlp.down_proj.lora_A.default.weight",
|
| 257 |
+
"base_model.model.model.layers.17.mlp.down_proj.lora_B.default.weight",
|
| 258 |
+
"base_model.model.model.layers.18.self_attn.q_proj.lora_A.default.weight",
|
| 259 |
+
"base_model.model.model.layers.18.self_attn.q_proj.lora_B.default.weight",
|
| 260 |
+
"base_model.model.model.layers.18.self_attn.k_proj.lora_A.default.weight",
|
| 261 |
+
"base_model.model.model.layers.18.self_attn.k_proj.lora_B.default.weight",
|
| 262 |
+
"base_model.model.model.layers.18.self_attn.v_proj.lora_A.default.weight",
|
| 263 |
+
"base_model.model.model.layers.18.self_attn.v_proj.lora_B.default.weight",
|
| 264 |
+
"base_model.model.model.layers.18.self_attn.o_proj.lora_A.default.weight",
|
| 265 |
+
"base_model.model.model.layers.18.self_attn.o_proj.lora_B.default.weight",
|
| 266 |
+
"base_model.model.model.layers.18.mlp.gate_proj.lora_A.default.weight",
|
| 267 |
+
"base_model.model.model.layers.18.mlp.gate_proj.lora_B.default.weight",
|
| 268 |
+
"base_model.model.model.layers.18.mlp.up_proj.lora_A.default.weight",
|
| 269 |
+
"base_model.model.model.layers.18.mlp.up_proj.lora_B.default.weight",
|
| 270 |
+
"base_model.model.model.layers.18.mlp.down_proj.lora_A.default.weight",
|
| 271 |
+
"base_model.model.model.layers.18.mlp.down_proj.lora_B.default.weight",
|
| 272 |
+
"base_model.model.model.layers.19.self_attn.q_proj.lora_A.default.weight",
|
| 273 |
+
"base_model.model.model.layers.19.self_attn.q_proj.lora_B.default.weight",
|
| 274 |
+
"base_model.model.model.layers.19.self_attn.k_proj.lora_A.default.weight",
|
| 275 |
+
"base_model.model.model.layers.19.self_attn.k_proj.lora_B.default.weight",
|
| 276 |
+
"base_model.model.model.layers.19.self_attn.v_proj.lora_A.default.weight",
|
| 277 |
+
"base_model.model.model.layers.19.self_attn.v_proj.lora_B.default.weight",
|
| 278 |
+
"base_model.model.model.layers.19.self_attn.o_proj.lora_A.default.weight",
|
| 279 |
+
"base_model.model.model.layers.19.self_attn.o_proj.lora_B.default.weight",
|
| 280 |
+
"base_model.model.model.layers.19.mlp.gate_proj.lora_A.default.weight",
|
| 281 |
+
"base_model.model.model.layers.19.mlp.gate_proj.lora_B.default.weight",
|
| 282 |
+
"base_model.model.model.layers.19.mlp.up_proj.lora_A.default.weight",
|
| 283 |
+
"base_model.model.model.layers.19.mlp.up_proj.lora_B.default.weight",
|
| 284 |
+
"base_model.model.model.layers.19.mlp.down_proj.lora_A.default.weight",
|
| 285 |
+
"base_model.model.model.layers.19.mlp.down_proj.lora_B.default.weight",
|
| 286 |
+
"base_model.model.model.layers.20.self_attn.q_proj.lora_A.default.weight",
|
| 287 |
+
"base_model.model.model.layers.20.self_attn.q_proj.lora_B.default.weight",
|
| 288 |
+
"base_model.model.model.layers.20.self_attn.k_proj.lora_A.default.weight",
|
| 289 |
+
"base_model.model.model.layers.20.self_attn.k_proj.lora_B.default.weight",
|
| 290 |
+
"base_model.model.model.layers.20.self_attn.v_proj.lora_A.default.weight",
|
| 291 |
+
"base_model.model.model.layers.20.self_attn.v_proj.lora_B.default.weight",
|
| 292 |
+
"base_model.model.model.layers.20.self_attn.o_proj.lora_A.default.weight",
|
| 293 |
+
"base_model.model.model.layers.20.self_attn.o_proj.lora_B.default.weight",
|
| 294 |
+
"base_model.model.model.layers.20.mlp.gate_proj.lora_A.default.weight",
|
| 295 |
+
"base_model.model.model.layers.20.mlp.gate_proj.lora_B.default.weight",
|
| 296 |
+
"base_model.model.model.layers.20.mlp.up_proj.lora_A.default.weight",
|
| 297 |
+
"base_model.model.model.layers.20.mlp.up_proj.lora_B.default.weight",
|
| 298 |
+
"base_model.model.model.layers.20.mlp.down_proj.lora_A.default.weight",
|
| 299 |
+
"base_model.model.model.layers.20.mlp.down_proj.lora_B.default.weight",
|
| 300 |
+
"base_model.model.model.layers.21.self_attn.q_proj.lora_A.default.weight",
|
| 301 |
+
"base_model.model.model.layers.21.self_attn.q_proj.lora_B.default.weight",
|
| 302 |
+
"base_model.model.model.layers.21.self_attn.k_proj.lora_A.default.weight",
|
| 303 |
+
"base_model.model.model.layers.21.self_attn.k_proj.lora_B.default.weight",
|
| 304 |
+
"base_model.model.model.layers.21.self_attn.v_proj.lora_A.default.weight",
|
| 305 |
+
"base_model.model.model.layers.21.self_attn.v_proj.lora_B.default.weight",
|
| 306 |
+
"base_model.model.model.layers.21.self_attn.o_proj.lora_A.default.weight",
|
| 307 |
+
"base_model.model.model.layers.21.self_attn.o_proj.lora_B.default.weight",
|
| 308 |
+
"base_model.model.model.layers.21.mlp.gate_proj.lora_A.default.weight",
|
| 309 |
+
"base_model.model.model.layers.21.mlp.gate_proj.lora_B.default.weight",
|
| 310 |
+
"base_model.model.model.layers.21.mlp.up_proj.lora_A.default.weight",
|
| 311 |
+
"base_model.model.model.layers.21.mlp.up_proj.lora_B.default.weight",
|
| 312 |
+
"base_model.model.model.layers.21.mlp.down_proj.lora_A.default.weight",
|
| 313 |
+
"base_model.model.model.layers.21.mlp.down_proj.lora_B.default.weight",
|
| 314 |
+
"base_model.model.model.layers.22.self_attn.q_proj.lora_A.default.weight",
|
| 315 |
+
"base_model.model.model.layers.22.self_attn.q_proj.lora_B.default.weight",
|
| 316 |
+
"base_model.model.model.layers.22.self_attn.k_proj.lora_A.default.weight",
|
| 317 |
+
"base_model.model.model.layers.22.self_attn.k_proj.lora_B.default.weight",
|
| 318 |
+
"base_model.model.model.layers.22.self_attn.v_proj.lora_A.default.weight",
|
| 319 |
+
"base_model.model.model.layers.22.self_attn.v_proj.lora_B.default.weight",
|
| 320 |
+
"base_model.model.model.layers.22.self_attn.o_proj.lora_A.default.weight",
|
| 321 |
+
"base_model.model.model.layers.22.self_attn.o_proj.lora_B.default.weight",
|
| 322 |
+
"base_model.model.model.layers.22.mlp.gate_proj.lora_A.default.weight",
|
| 323 |
+
"base_model.model.model.layers.22.mlp.gate_proj.lora_B.default.weight",
|
| 324 |
+
"base_model.model.model.layers.22.mlp.up_proj.lora_A.default.weight",
|
| 325 |
+
"base_model.model.model.layers.22.mlp.up_proj.lora_B.default.weight",
|
| 326 |
+
"base_model.model.model.layers.22.mlp.down_proj.lora_A.default.weight",
|
| 327 |
+
"base_model.model.model.layers.22.mlp.down_proj.lora_B.default.weight",
|
| 328 |
+
"base_model.model.model.layers.23.self_attn.q_proj.lora_A.default.weight",
|
| 329 |
+
"base_model.model.model.layers.23.self_attn.q_proj.lora_B.default.weight",
|
| 330 |
+
"base_model.model.model.layers.23.self_attn.k_proj.lora_A.default.weight",
|
| 331 |
+
"base_model.model.model.layers.23.self_attn.k_proj.lora_B.default.weight",
|
| 332 |
+
"base_model.model.model.layers.23.self_attn.v_proj.lora_A.default.weight",
|
| 333 |
+
"base_model.model.model.layers.23.self_attn.v_proj.lora_B.default.weight",
|
| 334 |
+
"base_model.model.model.layers.23.self_attn.o_proj.lora_A.default.weight",
|
| 335 |
+
"base_model.model.model.layers.23.self_attn.o_proj.lora_B.default.weight",
|
| 336 |
+
"base_model.model.model.layers.23.mlp.gate_proj.lora_A.default.weight",
|
| 337 |
+
"base_model.model.model.layers.23.mlp.gate_proj.lora_B.default.weight",
|
| 338 |
+
"base_model.model.model.layers.23.mlp.up_proj.lora_A.default.weight",
|
| 339 |
+
"base_model.model.model.layers.23.mlp.up_proj.lora_B.default.weight",
|
| 340 |
+
"base_model.model.model.layers.23.mlp.down_proj.lora_A.default.weight",
|
| 341 |
+
"base_model.model.model.layers.23.mlp.down_proj.lora_B.default.weight"
|
| 342 |
+
],
|
| 343 |
+
"attn_impl": null
|
| 344 |
+
}
|
kernel_metrics/kernel_metrics.json
CHANGED
|
@@ -3,208 +3,52 @@
|
|
| 3 |
"step": 200,
|
| 4 |
"kernel_matrix": [
|
| 5 |
[
|
| 6 |
-
|
| 7 |
-
|
| 8 |
-
|
| 9 |
-
|
| 10 |
-
|
| 11 |
],
|
| 12 |
[
|
| 13 |
-
|
| 14 |
-
|
| 15 |
-
|
| 16 |
-
|
| 17 |
-
|
| 18 |
],
|
| 19 |
[
|
| 20 |
-
|
| 21 |
-
|
| 22 |
-
|
| 23 |
-
|
| 24 |
-
|
| 25 |
],
|
| 26 |
[
|
| 27 |
-
|
| 28 |
-
|
| 29 |
-
|
| 30 |
-
|
| 31 |
-
|
| 32 |
],
|
| 33 |
[
|
| 34 |
-
|
| 35 |
-
|
| 36 |
-
|
| 37 |
-
|
| 38 |
-
|
| 39 |
]
|
| 40 |
],
|
| 41 |
-
"kernel_trace":
|
| 42 |
-
"kernel_frobenius_norm":
|
| 43 |
-
"mean_gradient_norm":
|
| 44 |
"gradient_norms": [
|
| 45 |
-
|
| 46 |
-
|
| 47 |
-
|
| 48 |
-
|
| 49 |
-
|
| 50 |
],
|
| 51 |
-
"linearization_error":
|
| 52 |
"kernel_distance": 0.0
|
| 53 |
-
},
|
| 54 |
-
{
|
| 55 |
-
"step": 400,
|
| 56 |
-
"kernel_matrix": [
|
| 57 |
-
[
|
| 58 |
-
25830.236328125,
|
| 59 |
-
25576.408203125,
|
| 60 |
-
23327.265625,
|
| 61 |
-
26607.23046875,
|
| 62 |
-
24894.69921875
|
| 63 |
-
],
|
| 64 |
-
[
|
| 65 |
-
25576.408203125,
|
| 66 |
-
26677.2734375,
|
| 67 |
-
23506.337890625,
|
| 68 |
-
26903.40234375,
|
| 69 |
-
25137.69140625
|
| 70 |
-
],
|
| 71 |
-
[
|
| 72 |
-
23327.265625,
|
| 73 |
-
23506.337890625,
|
| 74 |
-
21880.693359375,
|
| 75 |
-
24504.56640625,
|
| 76 |
-
22963.439453125
|
| 77 |
-
],
|
| 78 |
-
[
|
| 79 |
-
26607.23046875,
|
| 80 |
-
26903.40234375,
|
| 81 |
-
24504.56640625,
|
| 82 |
-
28542.37890625,
|
| 83 |
-
26290.02734375
|
| 84 |
-
],
|
| 85 |
-
[
|
| 86 |
-
24894.69921875,
|
| 87 |
-
25137.69140625,
|
| 88 |
-
22963.439453125,
|
| 89 |
-
26290.02734375,
|
| 90 |
-
24995.046875
|
| 91 |
-
]
|
| 92 |
-
],
|
| 93 |
-
"kernel_trace": 127925.625,
|
| 94 |
-
"kernel_frobenius_norm": 125711.890625,
|
| 95 |
-
"mean_gradient_norm": 159.80281524728267,
|
| 96 |
-
"gradient_norms": [
|
| 97 |
-
160.71788988137325,
|
| 98 |
-
163.33174126727675,
|
| 99 |
-
147.92125761454494,
|
| 100 |
-
168.94494668198647,
|
| 101 |
-
158.09824079123183
|
| 102 |
-
],
|
| 103 |
-
"linearization_error": 2.412812352987385,
|
| 104 |
-
"kernel_distance": 0.012032606638967991
|
| 105 |
-
},
|
| 106 |
-
{
|
| 107 |
-
"step": 600,
|
| 108 |
-
"kernel_matrix": [
|
| 109 |
-
[
|
| 110 |
-
32818.33203125,
|
| 111 |
-
32327.26171875,
|
| 112 |
-
29502.7578125,
|
| 113 |
-
33643.125,
|
| 114 |
-
31655.568359375
|
| 115 |
-
],
|
| 116 |
-
[
|
| 117 |
-
32327.26171875,
|
| 118 |
-
33630.37109375,
|
| 119 |
-
29583.498046875,
|
| 120 |
-
33840.6875,
|
| 121 |
-
31819.84375
|
| 122 |
-
],
|
| 123 |
-
[
|
| 124 |
-
29502.7578125,
|
| 125 |
-
29583.498046875,
|
| 126 |
-
27586.107421875,
|
| 127 |
-
30888.033203125,
|
| 128 |
-
29094.75
|
| 129 |
-
],
|
| 130 |
-
[
|
| 131 |
-
33643.125,
|
| 132 |
-
33840.6875,
|
| 133 |
-
30888.033203125,
|
| 134 |
-
35931.921875,
|
| 135 |
-
33313.30078125
|
| 136 |
-
],
|
| 137 |
-
[
|
| 138 |
-
31655.568359375,
|
| 139 |
-
31819.84375,
|
| 140 |
-
29094.75,
|
| 141 |
-
33313.30078125,
|
| 142 |
-
31840.12890625
|
| 143 |
-
]
|
| 144 |
-
],
|
| 145 |
-
"kernel_trace": 161806.859375,
|
| 146 |
-
"kernel_frobenius_norm": 158931.40625,
|
| 147 |
-
"mean_gradient_norm": 179.72602330442794,
|
| 148 |
-
"gradient_norms": [
|
| 149 |
-
181.15836713416823,
|
| 150 |
-
183.38585983892546,
|
| 151 |
-
166.0906405004076,
|
| 152 |
-
189.55719693258638,
|
| 153 |
-
178.43805211605195
|
| 154 |
-
],
|
| 155 |
-
"linearization_error": 2.286812121435722,
|
| 156 |
-
"kernel_distance": 0.25138264894485474
|
| 157 |
-
},
|
| 158 |
-
{
|
| 159 |
-
"step": 794,
|
| 160 |
-
"kernel_matrix": [
|
| 161 |
-
[
|
| 162 |
-
33578.4296875,
|
| 163 |
-
33209.41015625,
|
| 164 |
-
30144.486328125,
|
| 165 |
-
34479.05859375,
|
| 166 |
-
32445.908203125
|
| 167 |
-
],
|
| 168 |
-
[
|
| 169 |
-
33209.41015625,
|
| 170 |
-
34613.21484375,
|
| 171 |
-
30334.04296875,
|
| 172 |
-
34787.6640625,
|
| 173 |
-
32726.927734375
|
| 174 |
-
],
|
| 175 |
-
[
|
| 176 |
-
30144.486328125,
|
| 177 |
-
30334.04296875,
|
| 178 |
-
28127.357421875,
|
| 179 |
-
31592.04296875,
|
| 180 |
-
29770.07421875
|
| 181 |
-
],
|
| 182 |
-
[
|
| 183 |
-
34479.05859375,
|
| 184 |
-
34787.6640625,
|
| 185 |
-
31592.04296875,
|
| 186 |
-
36851.3515625,
|
| 187 |
-
34183.484375
|
| 188 |
-
],
|
| 189 |
-
[
|
| 190 |
-
32445.908203125,
|
| 191 |
-
32726.927734375,
|
| 192 |
-
29770.07421875,
|
| 193 |
-
34183.484375,
|
| 194 |
-
32646.0078125
|
| 195 |
-
]
|
| 196 |
-
],
|
| 197 |
-
"kernel_trace": 165816.34375,
|
| 198 |
-
"kernel_frobenius_norm": 162957.46875,
|
| 199 |
-
"mean_gradient_norm": 181.9303903574858,
|
| 200 |
-
"gradient_norms": [
|
| 201 |
-
183.24427265298237,
|
| 202 |
-
186.04631977029342,
|
| 203 |
-
167.712165661458,
|
| 204 |
-
191.96711155751723,
|
| 205 |
-
180.68208214517796
|
| 206 |
-
],
|
| 207 |
-
"linearization_error": 1.9182717903143667,
|
| 208 |
-
"kernel_distance": 0.28305584192276
|
| 209 |
}
|
| 210 |
]
|
|
|
|
| 3 |
"step": 200,
|
| 4 |
"kernel_matrix": [
|
| 5 |
[
|
| 6 |
+
0.0,
|
| 7 |
+
0.0,
|
| 8 |
+
0.0,
|
| 9 |
+
0.0,
|
| 10 |
+
0.0
|
| 11 |
],
|
| 12 |
[
|
| 13 |
+
0.0,
|
| 14 |
+
0.0,
|
| 15 |
+
0.0,
|
| 16 |
+
0.0,
|
| 17 |
+
0.0
|
| 18 |
],
|
| 19 |
[
|
| 20 |
+
0.0,
|
| 21 |
+
0.0,
|
| 22 |
+
0.0,
|
| 23 |
+
0.0,
|
| 24 |
+
0.0
|
| 25 |
],
|
| 26 |
[
|
| 27 |
+
0.0,
|
| 28 |
+
0.0,
|
| 29 |
+
0.0,
|
| 30 |
+
0.0,
|
| 31 |
+
0.0
|
| 32 |
],
|
| 33 |
[
|
| 34 |
+
0.0,
|
| 35 |
+
0.0,
|
| 36 |
+
0.0,
|
| 37 |
+
0.0,
|
| 38 |
+
0.0
|
| 39 |
]
|
| 40 |
],
|
| 41 |
+
"kernel_trace": 0.0,
|
| 42 |
+
"kernel_frobenius_norm": 0.0,
|
| 43 |
+
"mean_gradient_norm": 0.0,
|
| 44 |
"gradient_norms": [
|
| 45 |
+
0.0,
|
| 46 |
+
0.0,
|
| 47 |
+
0.0,
|
| 48 |
+
0.0,
|
| 49 |
+
0.0
|
| 50 |
],
|
| 51 |
+
"linearization_error": NaN,
|
| 52 |
"kernel_distance": 0.0
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 53 |
}
|
| 54 |
]
|
kernel_metrics/kernel_summary.json
CHANGED
|
@@ -1,34 +1,10 @@
|
|
| 1 |
[
|
| 2 |
{
|
| 3 |
"step": 200,
|
| 4 |
-
"kernel_trace":
|
| 5 |
-
"kernel_frobenius_norm":
|
| 6 |
-
"mean_gradient_norm":
|
| 7 |
-
"linearization_error":
|
| 8 |
"kernel_distance": 0.0
|
| 9 |
-
},
|
| 10 |
-
{
|
| 11 |
-
"step": 400,
|
| 12 |
-
"kernel_trace": 127925.625,
|
| 13 |
-
"kernel_frobenius_norm": 125711.890625,
|
| 14 |
-
"mean_gradient_norm": 159.80281524728267,
|
| 15 |
-
"linearization_error": 2.412812352987385,
|
| 16 |
-
"kernel_distance": 0.012032606638967991
|
| 17 |
-
},
|
| 18 |
-
{
|
| 19 |
-
"step": 600,
|
| 20 |
-
"kernel_trace": 161806.859375,
|
| 21 |
-
"kernel_frobenius_norm": 158931.40625,
|
| 22 |
-
"mean_gradient_norm": 179.72602330442794,
|
| 23 |
-
"linearization_error": 2.286812121435722,
|
| 24 |
-
"kernel_distance": 0.25138264894485474
|
| 25 |
-
},
|
| 26 |
-
{
|
| 27 |
-
"step": 794,
|
| 28 |
-
"kernel_trace": 165816.34375,
|
| 29 |
-
"kernel_frobenius_norm": 162957.46875,
|
| 30 |
-
"mean_gradient_norm": 181.9303903574858,
|
| 31 |
-
"linearization_error": 1.9182717903143667,
|
| 32 |
-
"kernel_distance": 0.28305584192276
|
| 33 |
}
|
| 34 |
]
|
|
|
|
| 1 |
[
|
| 2 |
{
|
| 3 |
"step": 200,
|
| 4 |
+
"kernel_trace": 0.0,
|
| 5 |
+
"kernel_frobenius_norm": 0.0,
|
| 6 |
+
"mean_gradient_norm": 0.0,
|
| 7 |
+
"linearization_error": NaN,
|
| 8 |
"kernel_distance": 0.0
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 9 |
}
|
| 10 |
]
|
tokenizer.json
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:913950e4971737031da511cdd1b410daae4566f62eb845b3975bca5a102323d8
|
| 3 |
+
size 11421995
|