Instructions to use ModelsLab/krea-turbo-svdquant with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Diffusers
How to use ModelsLab/krea-turbo-svdquant with Diffusers:
pip install -U diffusers transformers accelerate
import torch from diffusers import DiffusionPipeline # switch to "mps" for apple devices pipe = DiffusionPipeline.from_pretrained("ModelsLab/krea-turbo-svdquant", dtype=torch.bfloat16, device_map="cuda") prompt = "Astronaut in a jungle, cold color palette, muted colors, detailed, 8k" image = pipe(prompt).images[0] - Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- Draw Things
- DiffusionBee
| { | |
| "base_model": "krea/Krea-2-Turbo", | |
| "format": "krea2-svdquant-transformer-v1", | |
| "target_component": "transformer", | |
| "transformer_class": "Krea2Transformer2DModel", | |
| "runtime_target": "w4a16_plus_bf16_lowrank", | |
| "weight_bits": 4, | |
| "activation_bits": 16, | |
| "group_size": 128, | |
| "calibrated": true, | |
| "num_blocks": 28, | |
| "options": { | |
| "rank": 32, | |
| "attn_rank": 64, | |
| "mlp_rank": 128, | |
| "skip_first_blocks": 0, | |
| "skip_last_blocks": 0, | |
| "skip_qk": false, | |
| "max_layers": 0, | |
| "svd_oversample": 16, | |
| "svd_niter": 2, | |
| "svd_exact": false | |
| }, | |
| "layers": { | |
| "transformer_blocks.0.attn.to_q": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.0.attn.to_k": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 1536, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.0.attn.to_v": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 1536, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.0.attn.to_gate": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.0.attn.to_out.0": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.0.ff.gate": { | |
| "rank": 128, | |
| "group_size": 128, | |
| "shape": [ | |
| 16384, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.0.ff.up": { | |
| "rank": 128, | |
| "group_size": 128, | |
| "shape": [ | |
| 16384, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.0.ff.down": { | |
| "rank": 128, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 16384 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.1.attn.to_q": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.1.attn.to_k": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 1536, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.1.attn.to_v": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 1536, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.1.attn.to_gate": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.1.attn.to_out.0": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.1.ff.gate": { | |
| "rank": 128, | |
| "group_size": 128, | |
| "shape": [ | |
| 16384, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.1.ff.up": { | |
| "rank": 128, | |
| "group_size": 128, | |
| "shape": [ | |
| 16384, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.1.ff.down": { | |
| "rank": 128, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 16384 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.2.attn.to_q": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.2.attn.to_k": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 1536, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.2.attn.to_v": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 1536, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.2.attn.to_gate": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.2.attn.to_out.0": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.2.ff.gate": { | |
| "rank": 128, | |
| "group_size": 128, | |
| "shape": [ | |
| 16384, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.2.ff.up": { | |
| "rank": 128, | |
| "group_size": 128, | |
| "shape": [ | |
| 16384, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.2.ff.down": { | |
| "rank": 128, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 16384 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.3.attn.to_q": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.3.attn.to_k": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 1536, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.3.attn.to_v": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 1536, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.3.attn.to_gate": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.3.attn.to_out.0": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.3.ff.gate": { | |
| "rank": 128, | |
| "group_size": 128, | |
| "shape": [ | |
| 16384, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.3.ff.up": { | |
| "rank": 128, | |
| "group_size": 128, | |
| "shape": [ | |
| 16384, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.3.ff.down": { | |
| "rank": 128, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 16384 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.4.attn.to_q": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.4.attn.to_k": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 1536, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.4.attn.to_v": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 1536, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.4.attn.to_gate": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.4.attn.to_out.0": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.4.ff.gate": { | |
| "rank": 128, | |
| "group_size": 128, | |
| "shape": [ | |
| 16384, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.4.ff.up": { | |
| "rank": 128, | |
| "group_size": 128, | |
| "shape": [ | |
| 16384, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.4.ff.down": { | |
| "rank": 128, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 16384 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.5.attn.to_q": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.5.attn.to_k": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 1536, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.5.attn.to_v": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 1536, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.5.attn.to_gate": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.5.attn.to_out.0": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.5.ff.gate": { | |
| "rank": 128, | |
| "group_size": 128, | |
| "shape": [ | |
| 16384, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.5.ff.up": { | |
| "rank": 128, | |
| "group_size": 128, | |
| "shape": [ | |
| 16384, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.5.ff.down": { | |
| "rank": 128, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 16384 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.6.attn.to_q": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.6.attn.to_k": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 1536, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.6.attn.to_v": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 1536, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.6.attn.to_gate": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.6.attn.to_out.0": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.6.ff.gate": { | |
| "rank": 128, | |
| "group_size": 128, | |
| "shape": [ | |
| 16384, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.6.ff.up": { | |
| "rank": 128, | |
| "group_size": 128, | |
| "shape": [ | |
| 16384, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.6.ff.down": { | |
| "rank": 128, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 16384 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.7.attn.to_q": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.7.attn.to_k": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 1536, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.7.attn.to_v": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 1536, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.7.attn.to_gate": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.7.attn.to_out.0": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.7.ff.gate": { | |
| "rank": 128, | |
| "group_size": 128, | |
| "shape": [ | |
| 16384, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.7.ff.up": { | |
| "rank": 128, | |
| "group_size": 128, | |
| "shape": [ | |
| 16384, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.7.ff.down": { | |
| "rank": 128, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 16384 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.8.attn.to_q": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.8.attn.to_k": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 1536, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.8.attn.to_v": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 1536, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.8.attn.to_gate": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.8.attn.to_out.0": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.8.ff.gate": { | |
| "rank": 128, | |
| "group_size": 128, | |
| "shape": [ | |
| 16384, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.8.ff.up": { | |
| "rank": 128, | |
| "group_size": 128, | |
| "shape": [ | |
| 16384, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.8.ff.down": { | |
| "rank": 128, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 16384 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.9.attn.to_q": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.9.attn.to_k": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 1536, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.9.attn.to_v": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 1536, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.9.attn.to_gate": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.9.attn.to_out.0": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.9.ff.gate": { | |
| "rank": 128, | |
| "group_size": 128, | |
| "shape": [ | |
| 16384, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.9.ff.up": { | |
| "rank": 128, | |
| "group_size": 128, | |
| "shape": [ | |
| 16384, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.9.ff.down": { | |
| "rank": 128, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 16384 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.10.attn.to_q": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.10.attn.to_k": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 1536, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.10.attn.to_v": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 1536, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.10.attn.to_gate": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.10.attn.to_out.0": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.10.ff.gate": { | |
| "rank": 128, | |
| "group_size": 128, | |
| "shape": [ | |
| 16384, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.10.ff.up": { | |
| "rank": 128, | |
| "group_size": 128, | |
| "shape": [ | |
| 16384, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.10.ff.down": { | |
| "rank": 128, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 16384 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.11.attn.to_q": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.11.attn.to_k": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 1536, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.11.attn.to_v": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 1536, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.11.attn.to_gate": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.11.attn.to_out.0": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.11.ff.gate": { | |
| "rank": 128, | |
| "group_size": 128, | |
| "shape": [ | |
| 16384, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.11.ff.up": { | |
| "rank": 128, | |
| "group_size": 128, | |
| "shape": [ | |
| 16384, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.11.ff.down": { | |
| "rank": 128, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 16384 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.12.attn.to_q": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.12.attn.to_k": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 1536, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.12.attn.to_v": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 1536, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.12.attn.to_gate": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.12.attn.to_out.0": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.12.ff.gate": { | |
| "rank": 128, | |
| "group_size": 128, | |
| "shape": [ | |
| 16384, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.12.ff.up": { | |
| "rank": 128, | |
| "group_size": 128, | |
| "shape": [ | |
| 16384, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.12.ff.down": { | |
| "rank": 128, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 16384 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.13.attn.to_q": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.13.attn.to_k": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 1536, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.13.attn.to_v": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 1536, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.13.attn.to_gate": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.13.attn.to_out.0": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.13.ff.gate": { | |
| "rank": 128, | |
| "group_size": 128, | |
| "shape": [ | |
| 16384, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.13.ff.up": { | |
| "rank": 128, | |
| "group_size": 128, | |
| "shape": [ | |
| 16384, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.13.ff.down": { | |
| "rank": 128, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 16384 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.14.attn.to_q": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.14.attn.to_k": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 1536, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.14.attn.to_v": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 1536, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.14.attn.to_gate": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.14.attn.to_out.0": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.14.ff.gate": { | |
| "rank": 128, | |
| "group_size": 128, | |
| "shape": [ | |
| 16384, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.14.ff.up": { | |
| "rank": 128, | |
| "group_size": 128, | |
| "shape": [ | |
| 16384, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.14.ff.down": { | |
| "rank": 128, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 16384 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.15.attn.to_q": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.15.attn.to_k": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 1536, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.15.attn.to_v": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 1536, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.15.attn.to_gate": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.15.attn.to_out.0": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.15.ff.gate": { | |
| "rank": 128, | |
| "group_size": 128, | |
| "shape": [ | |
| 16384, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.15.ff.up": { | |
| "rank": 128, | |
| "group_size": 128, | |
| "shape": [ | |
| 16384, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.15.ff.down": { | |
| "rank": 128, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 16384 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.16.attn.to_q": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.16.attn.to_k": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 1536, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.16.attn.to_v": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 1536, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.16.attn.to_gate": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.16.attn.to_out.0": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.16.ff.gate": { | |
| "rank": 128, | |
| "group_size": 128, | |
| "shape": [ | |
| 16384, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.16.ff.up": { | |
| "rank": 128, | |
| "group_size": 128, | |
| "shape": [ | |
| 16384, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.16.ff.down": { | |
| "rank": 128, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 16384 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.17.attn.to_q": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.17.attn.to_k": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 1536, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.17.attn.to_v": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 1536, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.17.attn.to_gate": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.17.attn.to_out.0": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.17.ff.gate": { | |
| "rank": 128, | |
| "group_size": 128, | |
| "shape": [ | |
| 16384, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.17.ff.up": { | |
| "rank": 128, | |
| "group_size": 128, | |
| "shape": [ | |
| 16384, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.17.ff.down": { | |
| "rank": 128, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 16384 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.18.attn.to_q": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.18.attn.to_k": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 1536, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.18.attn.to_v": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 1536, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.18.attn.to_gate": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.18.attn.to_out.0": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.18.ff.gate": { | |
| "rank": 128, | |
| "group_size": 128, | |
| "shape": [ | |
| 16384, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.18.ff.up": { | |
| "rank": 128, | |
| "group_size": 128, | |
| "shape": [ | |
| 16384, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.18.ff.down": { | |
| "rank": 128, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 16384 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.19.attn.to_q": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.19.attn.to_k": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 1536, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.19.attn.to_v": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 1536, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.19.attn.to_gate": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.19.attn.to_out.0": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.19.ff.gate": { | |
| "rank": 128, | |
| "group_size": 128, | |
| "shape": [ | |
| 16384, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.19.ff.up": { | |
| "rank": 128, | |
| "group_size": 128, | |
| "shape": [ | |
| 16384, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.19.ff.down": { | |
| "rank": 128, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 16384 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.20.attn.to_q": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.20.attn.to_k": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 1536, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.20.attn.to_v": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 1536, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.20.attn.to_gate": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.20.attn.to_out.0": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.20.ff.gate": { | |
| "rank": 128, | |
| "group_size": 128, | |
| "shape": [ | |
| 16384, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.20.ff.up": { | |
| "rank": 128, | |
| "group_size": 128, | |
| "shape": [ | |
| 16384, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.20.ff.down": { | |
| "rank": 128, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 16384 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.21.attn.to_q": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.21.attn.to_k": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 1536, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.21.attn.to_v": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 1536, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.21.attn.to_gate": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.21.attn.to_out.0": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.21.ff.gate": { | |
| "rank": 128, | |
| "group_size": 128, | |
| "shape": [ | |
| 16384, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.21.ff.up": { | |
| "rank": 128, | |
| "group_size": 128, | |
| "shape": [ | |
| 16384, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.21.ff.down": { | |
| "rank": 128, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 16384 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.22.attn.to_q": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.22.attn.to_k": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 1536, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.22.attn.to_v": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 1536, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.22.attn.to_gate": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.22.attn.to_out.0": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.22.ff.gate": { | |
| "rank": 128, | |
| "group_size": 128, | |
| "shape": [ | |
| 16384, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.22.ff.up": { | |
| "rank": 128, | |
| "group_size": 128, | |
| "shape": [ | |
| 16384, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.22.ff.down": { | |
| "rank": 128, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 16384 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.23.attn.to_q": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.23.attn.to_k": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 1536, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.23.attn.to_v": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 1536, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.23.attn.to_gate": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.23.attn.to_out.0": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.23.ff.gate": { | |
| "rank": 128, | |
| "group_size": 128, | |
| "shape": [ | |
| 16384, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.23.ff.up": { | |
| "rank": 128, | |
| "group_size": 128, | |
| "shape": [ | |
| 16384, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.23.ff.down": { | |
| "rank": 128, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 16384 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.24.attn.to_q": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.24.attn.to_k": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 1536, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.24.attn.to_v": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 1536, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.24.attn.to_gate": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.24.attn.to_out.0": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.24.ff.gate": { | |
| "rank": 128, | |
| "group_size": 128, | |
| "shape": [ | |
| 16384, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.24.ff.up": { | |
| "rank": 128, | |
| "group_size": 128, | |
| "shape": [ | |
| 16384, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.24.ff.down": { | |
| "rank": 128, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 16384 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.25.attn.to_q": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.25.attn.to_k": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 1536, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.25.attn.to_v": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 1536, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.25.attn.to_gate": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.25.attn.to_out.0": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.25.ff.gate": { | |
| "rank": 128, | |
| "group_size": 128, | |
| "shape": [ | |
| 16384, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.25.ff.up": { | |
| "rank": 128, | |
| "group_size": 128, | |
| "shape": [ | |
| 16384, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.25.ff.down": { | |
| "rank": 128, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 16384 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.26.attn.to_q": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.26.attn.to_k": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 1536, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.26.attn.to_v": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 1536, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.26.attn.to_gate": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.26.attn.to_out.0": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.26.ff.gate": { | |
| "rank": 128, | |
| "group_size": 128, | |
| "shape": [ | |
| 16384, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.26.ff.up": { | |
| "rank": 128, | |
| "group_size": 128, | |
| "shape": [ | |
| 16384, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.26.ff.down": { | |
| "rank": 128, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 16384 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.27.attn.to_q": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.27.attn.to_k": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 1536, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.27.attn.to_v": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 1536, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.27.attn.to_gate": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.27.attn.to_out.0": { | |
| "rank": 64, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.27.ff.gate": { | |
| "rank": 128, | |
| "group_size": 128, | |
| "shape": [ | |
| 16384, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.27.ff.up": { | |
| "rank": 128, | |
| "group_size": 128, | |
| "shape": [ | |
| 16384, | |
| 6144 | |
| ], | |
| "calibrated": true | |
| }, | |
| "transformer_blocks.27.ff.down": { | |
| "rank": 128, | |
| "group_size": 128, | |
| "shape": [ | |
| 6144, | |
| 16384 | |
| ], | |
| "calibrated": true | |
| } | |
| } | |
| } |