Feature Extraction
Transformers
Safetensors
qwen3_5
matilda
jev
fp4
quantized
maincode
8-bit precision
Instructions to use Maincode/matilda-jev-fp4 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use Maincode/matilda-jev-fp4 with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("feature-extraction", model="Maincode/matilda-jev-fp4")# pip install -U transformers accelerate # Load model directly from transformers import AutoProcessor, AutoModel processor = AutoProcessor.from_pretrained("Maincode/matilda-jev-fp4") model = AutoModel.from_pretrained("Maincode/matilda-jev-fp4", device_map="auto") - Notebooks
- Google Colab
- Kaggle
Download jev_quantization.json from Maincode/matilda-jev-fp4: direct link, hf CLI and curl.
- Browser
- Download file 62.6 kB
-
https://huggingface.co/Maincode/matilda-jev-fp4/resolve/main/jev_quantization.json
- Command line
-
hf download hf://Maincode/matilda-jev-fp4/jev_quantization.json
-
curl -L -o jev_quantization.json https://huggingface.co/Maincode/matilda-jev-fp4/resolve/main/jev_quantization.json
62.6 kB
| { | |
| "format": "jev_fp4_w4a16_v1", | |
| "weight_format": "E2M1 packed uint8, low nibble first", | |
| "block_size": 16, | |
| "block_scale_dtype": "float8_e4m3fn", | |
| "tensor_scale_dtype": "float32", | |
| "activation_quantization": "none; BF16 activations", | |
| "scope": "all", | |
| "modules": { | |
| "language_model.layers.0.linear_attn.in_proj_qkv": { | |
| "shape": [ | |
| 10240, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09486016631126404 | |
| }, | |
| "language_model.layers.0.linear_attn.in_proj_z": { | |
| "shape": [ | |
| 6144, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09497135132551193 | |
| }, | |
| "language_model.layers.0.linear_attn.out_proj": { | |
| "shape": [ | |
| 5120, | |
| 6144 | |
| ], | |
| "relative_rmse": 0.09450867772102356 | |
| }, | |
| "language_model.layers.0.mlp.down_proj": { | |
| "shape": [ | |
| 5120, | |
| 17408 | |
| ], | |
| "relative_rmse": 0.0948721319437027 | |
| }, | |
| "language_model.layers.0.mlp.gate_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09499333053827286 | |
| }, | |
| "language_model.layers.0.mlp.up_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09505700320005417 | |
| }, | |
| "language_model.layers.1.linear_attn.in_proj_qkv": { | |
| "shape": [ | |
| 10240, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09484443813562393 | |
| }, | |
| "language_model.layers.1.linear_attn.in_proj_z": { | |
| "shape": [ | |
| 6144, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09514866769313812 | |
| }, | |
| "language_model.layers.1.linear_attn.out_proj": { | |
| "shape": [ | |
| 5120, | |
| 6144 | |
| ], | |
| "relative_rmse": 0.09484771639108658 | |
| }, | |
| "language_model.layers.1.mlp.down_proj": { | |
| "shape": [ | |
| 5120, | |
| 17408 | |
| ], | |
| "relative_rmse": 0.09486880898475647 | |
| }, | |
| "language_model.layers.1.mlp.gate_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09489931166172028 | |
| }, | |
| "language_model.layers.1.mlp.up_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09498240798711777 | |
| }, | |
| "language_model.layers.2.linear_attn.in_proj_qkv": { | |
| "shape": [ | |
| 10240, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09503107517957687 | |
| }, | |
| "language_model.layers.2.linear_attn.in_proj_z": { | |
| "shape": [ | |
| 6144, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.0950225368142128 | |
| }, | |
| "language_model.layers.2.linear_attn.out_proj": { | |
| "shape": [ | |
| 5120, | |
| 6144 | |
| ], | |
| "relative_rmse": 0.09481486678123474 | |
| }, | |
| "language_model.layers.2.mlp.down_proj": { | |
| "shape": [ | |
| 5120, | |
| 17408 | |
| ], | |
| "relative_rmse": 0.0949583575129509 | |
| }, | |
| "language_model.layers.2.mlp.gate_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09512658417224884 | |
| }, | |
| "language_model.layers.2.mlp.up_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09504052996635437 | |
| }, | |
| "language_model.layers.3.mlp.down_proj": { | |
| "shape": [ | |
| 5120, | |
| 17408 | |
| ], | |
| "relative_rmse": 0.09508463740348816 | |
| }, | |
| "language_model.layers.3.mlp.gate_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.0951017364859581 | |
| }, | |
| "language_model.layers.3.mlp.up_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.0951053649187088 | |
| }, | |
| "language_model.layers.3.self_attn.k_proj": { | |
| "shape": [ | |
| 1024, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09465542435646057 | |
| }, | |
| "language_model.layers.3.self_attn.o_proj": { | |
| "shape": [ | |
| 5120, | |
| 6144 | |
| ], | |
| "relative_rmse": 0.09480992704629898 | |
| }, | |
| "language_model.layers.3.self_attn.q_proj": { | |
| "shape": [ | |
| 12288, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09493377804756165 | |
| }, | |
| "language_model.layers.3.self_attn.v_proj": { | |
| "shape": [ | |
| 1024, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09477821737527847 | |
| }, | |
| "language_model.layers.4.linear_attn.in_proj_qkv": { | |
| "shape": [ | |
| 10240, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09525646269321442 | |
| }, | |
| "language_model.layers.4.linear_attn.in_proj_z": { | |
| "shape": [ | |
| 6144, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09491131454706192 | |
| }, | |
| "language_model.layers.4.linear_attn.out_proj": { | |
| "shape": [ | |
| 5120, | |
| 6144 | |
| ], | |
| "relative_rmse": 0.09473739564418793 | |
| }, | |
| "language_model.layers.4.mlp.down_proj": { | |
| "shape": [ | |
| 5120, | |
| 17408 | |
| ], | |
| "relative_rmse": 0.09508217871189117 | |
| }, | |
| "language_model.layers.4.mlp.gate_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09503670036792755 | |
| }, | |
| "language_model.layers.4.mlp.up_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09508693963289261 | |
| }, | |
| "language_model.layers.5.linear_attn.in_proj_qkv": { | |
| "shape": [ | |
| 10240, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09494045376777649 | |
| }, | |
| "language_model.layers.5.linear_attn.in_proj_z": { | |
| "shape": [ | |
| 6144, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09495068341493607 | |
| }, | |
| "language_model.layers.5.linear_attn.out_proj": { | |
| "shape": [ | |
| 5120, | |
| 6144 | |
| ], | |
| "relative_rmse": 0.09470401704311371 | |
| }, | |
| "language_model.layers.5.mlp.down_proj": { | |
| "shape": [ | |
| 5120, | |
| 17408 | |
| ], | |
| "relative_rmse": 0.09510992467403412 | |
| }, | |
| "language_model.layers.5.mlp.gate_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09505727142095566 | |
| }, | |
| "language_model.layers.5.mlp.up_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09504293650388718 | |
| }, | |
| "language_model.layers.6.linear_attn.in_proj_qkv": { | |
| "shape": [ | |
| 10240, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.0949840098619461 | |
| }, | |
| "language_model.layers.6.linear_attn.in_proj_z": { | |
| "shape": [ | |
| 6144, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09488002955913544 | |
| }, | |
| "language_model.layers.6.linear_attn.out_proj": { | |
| "shape": [ | |
| 5120, | |
| 6144 | |
| ], | |
| "relative_rmse": 0.09481548517942429 | |
| }, | |
| "language_model.layers.6.mlp.down_proj": { | |
| "shape": [ | |
| 5120, | |
| 17408 | |
| ], | |
| "relative_rmse": 0.09487542510032654 | |
| }, | |
| "language_model.layers.6.mlp.gate_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09487628191709518 | |
| }, | |
| "language_model.layers.6.mlp.up_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09496339410543442 | |
| }, | |
| "language_model.layers.7.mlp.down_proj": { | |
| "shape": [ | |
| 5120, | |
| 17408 | |
| ], | |
| "relative_rmse": 0.0950276255607605 | |
| }, | |
| "language_model.layers.7.mlp.gate_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09497901052236557 | |
| }, | |
| "language_model.layers.7.mlp.up_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09511414170265198 | |
| }, | |
| "language_model.layers.7.self_attn.k_proj": { | |
| "shape": [ | |
| 1024, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09430784732103348 | |
| }, | |
| "language_model.layers.7.self_attn.o_proj": { | |
| "shape": [ | |
| 5120, | |
| 6144 | |
| ], | |
| "relative_rmse": 0.09503162652254105 | |
| }, | |
| "language_model.layers.7.self_attn.q_proj": { | |
| "shape": [ | |
| 12288, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09478483349084854 | |
| }, | |
| "language_model.layers.7.self_attn.v_proj": { | |
| "shape": [ | |
| 1024, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09448397159576416 | |
| }, | |
| "language_model.layers.8.linear_attn.in_proj_qkv": { | |
| "shape": [ | |
| 10240, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.0950857624411583 | |
| }, | |
| "language_model.layers.8.linear_attn.in_proj_z": { | |
| "shape": [ | |
| 6144, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09473881870508194 | |
| }, | |
| "language_model.layers.8.linear_attn.out_proj": { | |
| "shape": [ | |
| 5120, | |
| 6144 | |
| ], | |
| "relative_rmse": 0.09493368119001389 | |
| }, | |
| "language_model.layers.8.mlp.down_proj": { | |
| "shape": [ | |
| 5120, | |
| 17408 | |
| ], | |
| "relative_rmse": 0.09497066587209702 | |
| }, | |
| "language_model.layers.8.mlp.gate_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09510495513677597 | |
| }, | |
| "language_model.layers.8.mlp.up_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09500662237405777 | |
| }, | |
| "language_model.layers.9.linear_attn.in_proj_qkv": { | |
| "shape": [ | |
| 10240, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09504227340221405 | |
| }, | |
| "language_model.layers.9.linear_attn.in_proj_z": { | |
| "shape": [ | |
| 6144, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09509280323982239 | |
| }, | |
| "language_model.layers.9.linear_attn.out_proj": { | |
| "shape": [ | |
| 5120, | |
| 6144 | |
| ], | |
| "relative_rmse": 0.09492899477481842 | |
| }, | |
| "language_model.layers.9.mlp.down_proj": { | |
| "shape": [ | |
| 5120, | |
| 17408 | |
| ], | |
| "relative_rmse": 0.09499592334032059 | |
| }, | |
| "language_model.layers.10.linear_attn.in_proj_qkv": { | |
| "shape": [ | |
| 10240, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09482154995203018 | |
| }, | |
| "language_model.layers.10.linear_attn.in_proj_z": { | |
| "shape": [ | |
| 6144, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09499532729387283 | |
| }, | |
| "language_model.layers.10.linear_attn.out_proj": { | |
| "shape": [ | |
| 5120, | |
| 6144 | |
| ], | |
| "relative_rmse": 0.09488175809383392 | |
| }, | |
| "language_model.layers.10.mlp.down_proj": { | |
| "shape": [ | |
| 5120, | |
| 17408 | |
| ], | |
| "relative_rmse": 0.09498362243175507 | |
| }, | |
| "language_model.layers.10.mlp.gate_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09510846436023712 | |
| }, | |
| "language_model.layers.10.mlp.up_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09504243731498718 | |
| }, | |
| "language_model.layers.11.mlp.down_proj": { | |
| "shape": [ | |
| 5120, | |
| 17408 | |
| ], | |
| "relative_rmse": 0.09496378153562546 | |
| }, | |
| "language_model.layers.11.mlp.gate_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09514947980642319 | |
| }, | |
| "language_model.layers.11.mlp.up_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09500227123498917 | |
| }, | |
| "language_model.layers.11.self_attn.k_proj": { | |
| "shape": [ | |
| 1024, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09418924897909164 | |
| }, | |
| "language_model.layers.11.self_attn.o_proj": { | |
| "shape": [ | |
| 5120, | |
| 6144 | |
| ], | |
| "relative_rmse": 0.0949619859457016 | |
| }, | |
| "language_model.layers.11.self_attn.q_proj": { | |
| "shape": [ | |
| 12288, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09488285332918167 | |
| }, | |
| "language_model.layers.11.self_attn.v_proj": { | |
| "shape": [ | |
| 1024, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09459587931632996 | |
| }, | |
| "language_model.layers.12.linear_attn.in_proj_qkv": { | |
| "shape": [ | |
| 10240, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09506550431251526 | |
| }, | |
| "language_model.layers.12.linear_attn.in_proj_z": { | |
| "shape": [ | |
| 6144, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.0949074998497963 | |
| }, | |
| "language_model.layers.12.linear_attn.out_proj": { | |
| "shape": [ | |
| 5120, | |
| 6144 | |
| ], | |
| "relative_rmse": 0.09501614421606064 | |
| }, | |
| "language_model.layers.12.mlp.down_proj": { | |
| "shape": [ | |
| 5120, | |
| 17408 | |
| ], | |
| "relative_rmse": 0.09496786445379257 | |
| }, | |
| "language_model.layers.12.mlp.gate_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09498301893472672 | |
| }, | |
| "language_model.layers.12.mlp.up_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09506572037935257 | |
| }, | |
| "language_model.layers.13.linear_attn.in_proj_qkv": { | |
| "shape": [ | |
| 10240, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09531169384717941 | |
| }, | |
| "language_model.layers.13.linear_attn.in_proj_z": { | |
| "shape": [ | |
| 6144, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09491849690675735 | |
| }, | |
| "language_model.layers.13.linear_attn.out_proj": { | |
| "shape": [ | |
| 5120, | |
| 6144 | |
| ], | |
| "relative_rmse": 0.09507833421230316 | |
| }, | |
| "language_model.layers.13.mlp.down_proj": { | |
| "shape": [ | |
| 5120, | |
| 17408 | |
| ], | |
| "relative_rmse": 0.09501072019338608 | |
| }, | |
| "language_model.layers.13.mlp.gate_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09508799761533737 | |
| }, | |
| "language_model.layers.13.mlp.up_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09514525532722473 | |
| }, | |
| "language_model.layers.14.linear_attn.in_proj_qkv": { | |
| "shape": [ | |
| 10240, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09476681053638458 | |
| }, | |
| "language_model.layers.14.linear_attn.in_proj_z": { | |
| "shape": [ | |
| 6144, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09494723379611969 | |
| }, | |
| "language_model.layers.14.linear_attn.out_proj": { | |
| "shape": [ | |
| 5120, | |
| 6144 | |
| ], | |
| "relative_rmse": 0.09504920989274979 | |
| }, | |
| "language_model.layers.14.mlp.down_proj": { | |
| "shape": [ | |
| 5120, | |
| 17408 | |
| ], | |
| "relative_rmse": 0.09514512121677399 | |
| }, | |
| "language_model.layers.14.mlp.gate_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09514012187719345 | |
| }, | |
| "language_model.layers.14.mlp.up_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.0950288474559784 | |
| }, | |
| "language_model.layers.15.mlp.down_proj": { | |
| "shape": [ | |
| 5120, | |
| 17408 | |
| ], | |
| "relative_rmse": 0.09509720653295517 | |
| }, | |
| "language_model.layers.15.mlp.gate_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09515085071325302 | |
| }, | |
| "language_model.layers.15.mlp.up_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09498771280050278 | |
| }, | |
| "language_model.layers.15.self_attn.k_proj": { | |
| "shape": [ | |
| 1024, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09469085931777954 | |
| }, | |
| "language_model.layers.15.self_attn.o_proj": { | |
| "shape": [ | |
| 5120, | |
| 6144 | |
| ], | |
| "relative_rmse": 0.09490107744932175 | |
| }, | |
| "language_model.layers.15.self_attn.q_proj": { | |
| "shape": [ | |
| 12288, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09494346380233765 | |
| }, | |
| "language_model.layers.15.self_attn.v_proj": { | |
| "shape": [ | |
| 1024, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.094981849193573 | |
| }, | |
| "language_model.layers.9.mlp.gate_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09502971917390823 | |
| }, | |
| "language_model.layers.9.mlp.up_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.0950048565864563 | |
| }, | |
| "language_model.layers.16.linear_attn.in_proj_qkv": { | |
| "shape": [ | |
| 10240, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09492098540067673 | |
| }, | |
| "language_model.layers.16.linear_attn.in_proj_z": { | |
| "shape": [ | |
| 6144, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09501805156469345 | |
| }, | |
| "language_model.layers.16.linear_attn.out_proj": { | |
| "shape": [ | |
| 5120, | |
| 6144 | |
| ], | |
| "relative_rmse": 0.09504058212041855 | |
| }, | |
| "language_model.layers.16.mlp.down_proj": { | |
| "shape": [ | |
| 5120, | |
| 17408 | |
| ], | |
| "relative_rmse": 0.0948575884103775 | |
| }, | |
| "language_model.layers.16.mlp.gate_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09504494071006775 | |
| }, | |
| "language_model.layers.16.mlp.up_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09507948905229568 | |
| }, | |
| "language_model.layers.17.linear_attn.in_proj_qkv": { | |
| "shape": [ | |
| 10240, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09499464929103851 | |
| }, | |
| "language_model.layers.17.linear_attn.in_proj_z": { | |
| "shape": [ | |
| 6144, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09500721096992493 | |
| }, | |
| "language_model.layers.17.linear_attn.out_proj": { | |
| "shape": [ | |
| 5120, | |
| 6144 | |
| ], | |
| "relative_rmse": 0.09492168575525284 | |
| }, | |
| "language_model.layers.17.mlp.down_proj": { | |
| "shape": [ | |
| 5120, | |
| 17408 | |
| ], | |
| "relative_rmse": 0.0948880985379219 | |
| }, | |
| "language_model.layers.17.mlp.gate_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09502647817134857 | |
| }, | |
| "language_model.layers.17.mlp.up_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09504751861095428 | |
| }, | |
| "language_model.layers.18.linear_attn.in_proj_qkv": { | |
| "shape": [ | |
| 10240, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09492260217666626 | |
| }, | |
| "language_model.layers.18.linear_attn.in_proj_z": { | |
| "shape": [ | |
| 6144, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09499627351760864 | |
| }, | |
| "language_model.layers.18.linear_attn.out_proj": { | |
| "shape": [ | |
| 5120, | |
| 6144 | |
| ], | |
| "relative_rmse": 0.09481608867645264 | |
| }, | |
| "language_model.layers.18.mlp.down_proj": { | |
| "shape": [ | |
| 5120, | |
| 17408 | |
| ], | |
| "relative_rmse": 0.09485505521297455 | |
| }, | |
| "language_model.layers.18.mlp.gate_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09494347870349884 | |
| }, | |
| "language_model.layers.18.mlp.up_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09514016658067703 | |
| }, | |
| "language_model.layers.19.mlp.down_proj": { | |
| "shape": [ | |
| 5120, | |
| 17408 | |
| ], | |
| "relative_rmse": 0.09493260085582733 | |
| }, | |
| "language_model.layers.19.mlp.gate_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.0949341356754303 | |
| }, | |
| "language_model.layers.19.mlp.up_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09500252455472946 | |
| }, | |
| "language_model.layers.19.self_attn.k_proj": { | |
| "shape": [ | |
| 1024, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09432574361562729 | |
| }, | |
| "language_model.layers.19.self_attn.o_proj": { | |
| "shape": [ | |
| 5120, | |
| 6144 | |
| ], | |
| "relative_rmse": 0.09484393149614334 | |
| }, | |
| "language_model.layers.19.self_attn.q_proj": { | |
| "shape": [ | |
| 12288, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09493577480316162 | |
| }, | |
| "language_model.layers.19.self_attn.v_proj": { | |
| "shape": [ | |
| 1024, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09439519792795181 | |
| }, | |
| "language_model.layers.20.linear_attn.in_proj_qkv": { | |
| "shape": [ | |
| 10240, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09497325122356415 | |
| }, | |
| "language_model.layers.20.linear_attn.in_proj_z": { | |
| "shape": [ | |
| 6144, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09502489864826202 | |
| }, | |
| "language_model.layers.20.linear_attn.out_proj": { | |
| "shape": [ | |
| 5120, | |
| 6144 | |
| ], | |
| "relative_rmse": 0.0948246568441391 | |
| }, | |
| "language_model.layers.20.mlp.down_proj": { | |
| "shape": [ | |
| 5120, | |
| 17408 | |
| ], | |
| "relative_rmse": 0.09502345323562622 | |
| }, | |
| "language_model.layers.20.mlp.gate_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09514394402503967 | |
| }, | |
| "language_model.layers.20.mlp.up_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09500880539417267 | |
| }, | |
| "language_model.layers.21.linear_attn.in_proj_qkv": { | |
| "shape": [ | |
| 10240, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.0948365181684494 | |
| }, | |
| "language_model.layers.21.linear_attn.in_proj_z": { | |
| "shape": [ | |
| 6144, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09501895308494568 | |
| }, | |
| "language_model.layers.21.linear_attn.out_proj": { | |
| "shape": [ | |
| 5120, | |
| 6144 | |
| ], | |
| "relative_rmse": 0.0947766900062561 | |
| }, | |
| "language_model.layers.21.mlp.down_proj": { | |
| "shape": [ | |
| 5120, | |
| 17408 | |
| ], | |
| "relative_rmse": 0.0948769599199295 | |
| }, | |
| "language_model.layers.21.mlp.gate_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09502305835485458 | |
| }, | |
| "language_model.layers.21.mlp.up_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09499023109674454 | |
| }, | |
| "language_model.layers.22.linear_attn.in_proj_qkv": { | |
| "shape": [ | |
| 10240, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.0949094220995903 | |
| }, | |
| "language_model.layers.22.linear_attn.in_proj_z": { | |
| "shape": [ | |
| 6144, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09501488506793976 | |
| }, | |
| "language_model.layers.22.linear_attn.out_proj": { | |
| "shape": [ | |
| 5120, | |
| 6144 | |
| ], | |
| "relative_rmse": 0.09486842900514603 | |
| }, | |
| "language_model.layers.22.mlp.down_proj": { | |
| "shape": [ | |
| 5120, | |
| 17408 | |
| ], | |
| "relative_rmse": 0.09473083913326263 | |
| }, | |
| "language_model.layers.22.mlp.gate_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09499149769544601 | |
| }, | |
| "language_model.layers.22.mlp.up_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09487354755401611 | |
| }, | |
| "language_model.layers.23.mlp.down_proj": { | |
| "shape": [ | |
| 5120, | |
| 17408 | |
| ], | |
| "relative_rmse": 0.09514690190553665 | |
| }, | |
| "language_model.layers.23.mlp.gate_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09499688446521759 | |
| }, | |
| "language_model.layers.23.mlp.up_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09503503143787384 | |
| }, | |
| "language_model.layers.23.self_attn.k_proj": { | |
| "shape": [ | |
| 1024, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09404724091291428 | |
| }, | |
| "language_model.layers.23.self_attn.o_proj": { | |
| "shape": [ | |
| 5120, | |
| 6144 | |
| ], | |
| "relative_rmse": 0.09494449943304062 | |
| }, | |
| "language_model.layers.23.self_attn.q_proj": { | |
| "shape": [ | |
| 12288, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09494021534919739 | |
| }, | |
| "language_model.layers.23.self_attn.v_proj": { | |
| "shape": [ | |
| 1024, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09453386813402176 | |
| }, | |
| "language_model.layers.24.linear_attn.in_proj_qkv": { | |
| "shape": [ | |
| 10240, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09500335901975632 | |
| }, | |
| "language_model.layers.24.linear_attn.in_proj_z": { | |
| "shape": [ | |
| 6144, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09493091702461243 | |
| }, | |
| "language_model.layers.24.linear_attn.out_proj": { | |
| "shape": [ | |
| 5120, | |
| 6144 | |
| ], | |
| "relative_rmse": 0.09486418962478638 | |
| }, | |
| "language_model.layers.24.mlp.down_proj": { | |
| "shape": [ | |
| 5120, | |
| 17408 | |
| ], | |
| "relative_rmse": 0.09495484828948975 | |
| }, | |
| "language_model.layers.24.mlp.gate_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09501788765192032 | |
| }, | |
| "language_model.layers.24.mlp.up_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09502902626991272 | |
| }, | |
| "language_model.layers.25.linear_attn.in_proj_qkv": { | |
| "shape": [ | |
| 10240, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09482426941394806 | |
| }, | |
| "language_model.layers.25.linear_attn.in_proj_z": { | |
| "shape": [ | |
| 6144, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09474703669548035 | |
| }, | |
| "language_model.layers.25.linear_attn.out_proj": { | |
| "shape": [ | |
| 5120, | |
| 6144 | |
| ], | |
| "relative_rmse": 0.09497242420911789 | |
| }, | |
| "language_model.layers.25.mlp.down_proj": { | |
| "shape": [ | |
| 5120, | |
| 17408 | |
| ], | |
| "relative_rmse": 0.09508100897073746 | |
| }, | |
| "language_model.layers.25.mlp.gate_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09501214325428009 | |
| }, | |
| "language_model.layers.25.mlp.up_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.0950527936220169 | |
| }, | |
| "language_model.layers.26.linear_attn.in_proj_qkv": { | |
| "shape": [ | |
| 10240, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09467176347970963 | |
| }, | |
| "language_model.layers.26.linear_attn.in_proj_z": { | |
| "shape": [ | |
| 6144, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09480072557926178 | |
| }, | |
| "language_model.layers.26.linear_attn.out_proj": { | |
| "shape": [ | |
| 5120, | |
| 6144 | |
| ], | |
| "relative_rmse": 0.09490510821342468 | |
| }, | |
| "language_model.layers.26.mlp.down_proj": { | |
| "shape": [ | |
| 5120, | |
| 17408 | |
| ], | |
| "relative_rmse": 0.09496297687292099 | |
| }, | |
| "language_model.layers.26.mlp.gate_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09498626738786697 | |
| }, | |
| "language_model.layers.26.mlp.up_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09480836987495422 | |
| }, | |
| "language_model.layers.27.mlp.down_proj": { | |
| "shape": [ | |
| 5120, | |
| 17408 | |
| ], | |
| "relative_rmse": 0.09484710544347763 | |
| }, | |
| "language_model.layers.27.mlp.gate_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09493212401866913 | |
| }, | |
| "language_model.layers.27.mlp.up_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.0948168933391571 | |
| }, | |
| "language_model.layers.27.self_attn.k_proj": { | |
| "shape": [ | |
| 1024, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09384384751319885 | |
| }, | |
| "language_model.layers.27.self_attn.o_proj": { | |
| "shape": [ | |
| 5120, | |
| 6144 | |
| ], | |
| "relative_rmse": 0.09486248344182968 | |
| }, | |
| "language_model.layers.27.self_attn.q_proj": { | |
| "shape": [ | |
| 12288, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09475899487733841 | |
| }, | |
| "language_model.layers.27.self_attn.v_proj": { | |
| "shape": [ | |
| 1024, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09454510360956192 | |
| }, | |
| "language_model.layers.28.linear_attn.in_proj_qkv": { | |
| "shape": [ | |
| 10240, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09461823105812073 | |
| }, | |
| "language_model.layers.28.linear_attn.in_proj_z": { | |
| "shape": [ | |
| 6144, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09468646347522736 | |
| }, | |
| "language_model.layers.28.linear_attn.out_proj": { | |
| "shape": [ | |
| 5120, | |
| 6144 | |
| ], | |
| "relative_rmse": 0.0948798656463623 | |
| }, | |
| "language_model.layers.28.mlp.down_proj": { | |
| "shape": [ | |
| 5120, | |
| 17408 | |
| ], | |
| "relative_rmse": 0.09483753889799118 | |
| }, | |
| "language_model.layers.28.mlp.gate_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09500079602003098 | |
| }, | |
| "language_model.layers.28.mlp.up_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09481845051050186 | |
| }, | |
| "language_model.layers.29.linear_attn.in_proj_qkv": { | |
| "shape": [ | |
| 10240, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09494879841804504 | |
| }, | |
| "language_model.layers.29.linear_attn.in_proj_z": { | |
| "shape": [ | |
| 6144, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09500574320554733 | |
| }, | |
| "language_model.layers.29.linear_attn.out_proj": { | |
| "shape": [ | |
| 5120, | |
| 6144 | |
| ], | |
| "relative_rmse": 0.09493628889322281 | |
| }, | |
| "language_model.layers.29.mlp.down_proj": { | |
| "shape": [ | |
| 5120, | |
| 17408 | |
| ], | |
| "relative_rmse": 0.09485795348882675 | |
| }, | |
| "language_model.layers.29.mlp.gate_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09481706470251083 | |
| }, | |
| "language_model.layers.29.mlp.up_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09477069973945618 | |
| }, | |
| "language_model.layers.30.linear_attn.in_proj_qkv": { | |
| "shape": [ | |
| 10240, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09461774677038193 | |
| }, | |
| "language_model.layers.30.linear_attn.in_proj_z": { | |
| "shape": [ | |
| 6144, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09483305364847183 | |
| }, | |
| "language_model.layers.30.linear_attn.out_proj": { | |
| "shape": [ | |
| 5120, | |
| 6144 | |
| ], | |
| "relative_rmse": 0.09494119882583618 | |
| }, | |
| "language_model.layers.30.mlp.down_proj": { | |
| "shape": [ | |
| 5120, | |
| 17408 | |
| ], | |
| "relative_rmse": 0.0949774906039238 | |
| }, | |
| "language_model.layers.30.mlp.gate_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09488280862569809 | |
| }, | |
| "language_model.layers.30.mlp.up_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09494347125291824 | |
| }, | |
| "language_model.layers.31.mlp.down_proj": { | |
| "shape": [ | |
| 5120, | |
| 17408 | |
| ], | |
| "relative_rmse": 0.09492547065019608 | |
| }, | |
| "language_model.layers.31.mlp.gate_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.0947427749633789 | |
| }, | |
| "language_model.layers.31.mlp.up_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09495143592357635 | |
| }, | |
| "language_model.layers.31.self_attn.k_proj": { | |
| "shape": [ | |
| 1024, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09425639361143112 | |
| }, | |
| "language_model.layers.31.self_attn.o_proj": { | |
| "shape": [ | |
| 5120, | |
| 6144 | |
| ], | |
| "relative_rmse": 0.09483080357313156 | |
| }, | |
| "language_model.layers.31.self_attn.q_proj": { | |
| "shape": [ | |
| 12288, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09492354840040207 | |
| }, | |
| "language_model.layers.31.self_attn.v_proj": { | |
| "shape": [ | |
| 1024, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09457071125507355 | |
| }, | |
| "language_model.layers.32.linear_attn.in_proj_qkv": { | |
| "shape": [ | |
| 10240, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09463939070701599 | |
| }, | |
| "language_model.layers.32.linear_attn.in_proj_z": { | |
| "shape": [ | |
| 6144, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.0946582704782486 | |
| }, | |
| "language_model.layers.32.linear_attn.out_proj": { | |
| "shape": [ | |
| 5120, | |
| 6144 | |
| ], | |
| "relative_rmse": 0.09487397968769073 | |
| }, | |
| "language_model.layers.32.mlp.down_proj": { | |
| "shape": [ | |
| 5120, | |
| 17408 | |
| ], | |
| "relative_rmse": 0.09469287097454071 | |
| }, | |
| "language_model.layers.32.mlp.gate_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09478341042995453 | |
| }, | |
| "language_model.layers.32.mlp.up_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09476879239082336 | |
| }, | |
| "language_model.layers.33.linear_attn.in_proj_qkv": { | |
| "shape": [ | |
| 10240, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09440957754850388 | |
| }, | |
| "language_model.layers.33.linear_attn.in_proj_z": { | |
| "shape": [ | |
| 6144, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09464427828788757 | |
| }, | |
| "language_model.layers.33.linear_attn.out_proj": { | |
| "shape": [ | |
| 5120, | |
| 6144 | |
| ], | |
| "relative_rmse": 0.09479189664125443 | |
| }, | |
| "language_model.layers.33.mlp.down_proj": { | |
| "shape": [ | |
| 5120, | |
| 17408 | |
| ], | |
| "relative_rmse": 0.094871886074543 | |
| }, | |
| "language_model.layers.33.mlp.gate_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09499730169773102 | |
| }, | |
| "language_model.layers.33.mlp.up_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09476088732481003 | |
| }, | |
| "language_model.layers.34.linear_attn.in_proj_qkv": { | |
| "shape": [ | |
| 10240, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09468875825405121 | |
| }, | |
| "language_model.layers.34.linear_attn.in_proj_z": { | |
| "shape": [ | |
| 6144, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09478949010372162 | |
| }, | |
| "language_model.layers.34.linear_attn.out_proj": { | |
| "shape": [ | |
| 5120, | |
| 6144 | |
| ], | |
| "relative_rmse": 0.09469298273324966 | |
| }, | |
| "language_model.layers.34.mlp.down_proj": { | |
| "shape": [ | |
| 5120, | |
| 17408 | |
| ], | |
| "relative_rmse": 0.09461571276187897 | |
| }, | |
| "language_model.layers.34.mlp.gate_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09463000297546387 | |
| }, | |
| "language_model.layers.34.mlp.up_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09475961327552795 | |
| }, | |
| "language_model.layers.35.mlp.down_proj": { | |
| "shape": [ | |
| 5120, | |
| 17408 | |
| ], | |
| "relative_rmse": 0.09494943916797638 | |
| }, | |
| "language_model.layers.35.mlp.gate_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09489712119102478 | |
| }, | |
| "language_model.layers.35.mlp.up_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.0948406457901001 | |
| }, | |
| "language_model.layers.35.self_attn.k_proj": { | |
| "shape": [ | |
| 1024, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09458363801240921 | |
| }, | |
| "language_model.layers.35.self_attn.o_proj": { | |
| "shape": [ | |
| 5120, | |
| 6144 | |
| ], | |
| "relative_rmse": 0.09498253464698792 | |
| }, | |
| "language_model.layers.35.self_attn.q_proj": { | |
| "shape": [ | |
| 12288, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09476710855960846 | |
| }, | |
| "language_model.layers.35.self_attn.v_proj": { | |
| "shape": [ | |
| 1024, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09429006278514862 | |
| }, | |
| "language_model.layers.36.linear_attn.in_proj_qkv": { | |
| "shape": [ | |
| 10240, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09468530118465424 | |
| }, | |
| "language_model.layers.36.linear_attn.in_proj_z": { | |
| "shape": [ | |
| 6144, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09500762075185776 | |
| }, | |
| "language_model.layers.36.linear_attn.out_proj": { | |
| "shape": [ | |
| 5120, | |
| 6144 | |
| ], | |
| "relative_rmse": 0.09488498419523239 | |
| }, | |
| "language_model.layers.36.mlp.down_proj": { | |
| "shape": [ | |
| 5120, | |
| 17408 | |
| ], | |
| "relative_rmse": 0.09489385038614273 | |
| }, | |
| "language_model.layers.36.mlp.gate_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09498020261526108 | |
| }, | |
| "language_model.layers.36.mlp.up_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09488225728273392 | |
| }, | |
| "language_model.layers.37.linear_attn.in_proj_qkv": { | |
| "shape": [ | |
| 10240, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09489655494689941 | |
| }, | |
| "language_model.layers.37.linear_attn.in_proj_z": { | |
| "shape": [ | |
| 6144, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09505017846822739 | |
| }, | |
| "language_model.layers.37.linear_attn.out_proj": { | |
| "shape": [ | |
| 5120, | |
| 6144 | |
| ], | |
| "relative_rmse": 0.09477003663778305 | |
| }, | |
| "language_model.layers.37.mlp.down_proj": { | |
| "shape": [ | |
| 5120, | |
| 17408 | |
| ], | |
| "relative_rmse": 0.09496506303548813 | |
| }, | |
| "language_model.layers.37.mlp.gate_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.0950976312160492 | |
| }, | |
| "language_model.layers.37.mlp.up_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.0949040874838829 | |
| }, | |
| "language_model.layers.38.linear_attn.in_proj_qkv": { | |
| "shape": [ | |
| 10240, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09471458941698074 | |
| }, | |
| "language_model.layers.38.linear_attn.in_proj_z": { | |
| "shape": [ | |
| 6144, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09500065445899963 | |
| }, | |
| "language_model.layers.38.linear_attn.out_proj": { | |
| "shape": [ | |
| 5120, | |
| 6144 | |
| ], | |
| "relative_rmse": 0.09491356462240219 | |
| }, | |
| "language_model.layers.38.mlp.down_proj": { | |
| "shape": [ | |
| 5120, | |
| 17408 | |
| ], | |
| "relative_rmse": 0.09493554383516312 | |
| }, | |
| "language_model.layers.38.mlp.gate_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09504397213459015 | |
| }, | |
| "language_model.layers.38.mlp.up_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09488421678543091 | |
| }, | |
| "language_model.layers.39.mlp.down_proj": { | |
| "shape": [ | |
| 5120, | |
| 17408 | |
| ], | |
| "relative_rmse": 0.09519059211015701 | |
| }, | |
| "language_model.layers.39.mlp.gate_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09494423866271973 | |
| }, | |
| "language_model.layers.39.mlp.up_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09490010142326355 | |
| }, | |
| "language_model.layers.39.self_attn.k_proj": { | |
| "shape": [ | |
| 1024, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09392006695270538 | |
| }, | |
| "language_model.layers.39.self_attn.o_proj": { | |
| "shape": [ | |
| 5120, | |
| 6144 | |
| ], | |
| "relative_rmse": 0.09475131332874298 | |
| }, | |
| "language_model.layers.39.self_attn.q_proj": { | |
| "shape": [ | |
| 12288, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.0948835015296936 | |
| }, | |
| "language_model.layers.39.self_attn.v_proj": { | |
| "shape": [ | |
| 1024, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09420917928218842 | |
| }, | |
| "language_model.layers.40.linear_attn.in_proj_qkv": { | |
| "shape": [ | |
| 10240, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09492066502571106 | |
| }, | |
| "language_model.layers.40.linear_attn.in_proj_z": { | |
| "shape": [ | |
| 6144, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.0948696881532669 | |
| }, | |
| "language_model.layers.40.linear_attn.out_proj": { | |
| "shape": [ | |
| 5120, | |
| 6144 | |
| ], | |
| "relative_rmse": 0.09492892026901245 | |
| }, | |
| "language_model.layers.40.mlp.down_proj": { | |
| "shape": [ | |
| 5120, | |
| 17408 | |
| ], | |
| "relative_rmse": 0.09490188956260681 | |
| }, | |
| "language_model.layers.40.mlp.gate_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.0951344296336174 | |
| }, | |
| "language_model.layers.40.mlp.up_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09501751512289047 | |
| }, | |
| "language_model.layers.41.linear_attn.in_proj_qkv": { | |
| "shape": [ | |
| 10240, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.0948137417435646 | |
| }, | |
| "language_model.layers.41.linear_attn.in_proj_z": { | |
| "shape": [ | |
| 6144, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.0948779433965683 | |
| }, | |
| "language_model.layers.41.linear_attn.out_proj": { | |
| "shape": [ | |
| 5120, | |
| 6144 | |
| ], | |
| "relative_rmse": 0.0947781577706337 | |
| }, | |
| "language_model.layers.41.mlp.down_proj": { | |
| "shape": [ | |
| 5120, | |
| 17408 | |
| ], | |
| "relative_rmse": 0.09503041952848434 | |
| }, | |
| "language_model.layers.41.mlp.gate_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.0948844701051712 | |
| }, | |
| "language_model.layers.41.mlp.up_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09497609734535217 | |
| }, | |
| "language_model.layers.42.linear_attn.in_proj_qkv": { | |
| "shape": [ | |
| 10240, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09471409767866135 | |
| }, | |
| "language_model.layers.42.linear_attn.in_proj_z": { | |
| "shape": [ | |
| 6144, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.0949917882680893 | |
| }, | |
| "language_model.layers.42.linear_attn.out_proj": { | |
| "shape": [ | |
| 5120, | |
| 6144 | |
| ], | |
| "relative_rmse": 0.09480810165405273 | |
| }, | |
| "language_model.layers.42.mlp.down_proj": { | |
| "shape": [ | |
| 5120, | |
| 17408 | |
| ], | |
| "relative_rmse": 0.09481658041477203 | |
| }, | |
| "language_model.layers.42.mlp.gate_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09484227746725082 | |
| }, | |
| "language_model.layers.42.mlp.up_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09501878917217255 | |
| }, | |
| "language_model.layers.43.mlp.down_proj": { | |
| "shape": [ | |
| 5120, | |
| 17408 | |
| ], | |
| "relative_rmse": 0.09485714882612228 | |
| }, | |
| "language_model.layers.43.mlp.gate_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09488418698310852 | |
| }, | |
| "language_model.layers.43.mlp.up_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09511309117078781 | |
| }, | |
| "language_model.layers.43.self_attn.k_proj": { | |
| "shape": [ | |
| 1024, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09373204410076141 | |
| }, | |
| "language_model.layers.43.self_attn.o_proj": { | |
| "shape": [ | |
| 5120, | |
| 6144 | |
| ], | |
| "relative_rmse": 0.0945858582854271 | |
| }, | |
| "language_model.layers.43.self_attn.q_proj": { | |
| "shape": [ | |
| 12288, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.0946039929986 | |
| }, | |
| "language_model.layers.43.self_attn.v_proj": { | |
| "shape": [ | |
| 1024, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09429279714822769 | |
| }, | |
| "language_model.layers.44.linear_attn.in_proj_qkv": { | |
| "shape": [ | |
| 10240, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09474681317806244 | |
| }, | |
| "language_model.layers.44.linear_attn.in_proj_z": { | |
| "shape": [ | |
| 6144, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.0949125811457634 | |
| }, | |
| "language_model.layers.44.linear_attn.out_proj": { | |
| "shape": [ | |
| 5120, | |
| 6144 | |
| ], | |
| "relative_rmse": 0.09474358707666397 | |
| }, | |
| "language_model.layers.44.mlp.down_proj": { | |
| "shape": [ | |
| 5120, | |
| 17408 | |
| ], | |
| "relative_rmse": 0.09497571736574173 | |
| }, | |
| "language_model.layers.44.mlp.gate_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09484536945819855 | |
| }, | |
| "language_model.layers.44.mlp.up_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09481244534254074 | |
| }, | |
| "language_model.layers.45.linear_attn.in_proj_qkv": { | |
| "shape": [ | |
| 10240, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09481263160705566 | |
| }, | |
| "language_model.layers.45.linear_attn.in_proj_z": { | |
| "shape": [ | |
| 6144, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09497564285993576 | |
| }, | |
| "language_model.layers.45.linear_attn.out_proj": { | |
| "shape": [ | |
| 5120, | |
| 6144 | |
| ], | |
| "relative_rmse": 0.09486696869134903 | |
| }, | |
| "language_model.layers.45.mlp.down_proj": { | |
| "shape": [ | |
| 5120, | |
| 17408 | |
| ], | |
| "relative_rmse": 0.09488007426261902 | |
| }, | |
| "language_model.layers.45.mlp.gate_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09497823566198349 | |
| }, | |
| "language_model.layers.45.mlp.up_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09472479671239853 | |
| }, | |
| "language_model.layers.46.linear_attn.in_proj_qkv": { | |
| "shape": [ | |
| 10240, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09470413625240326 | |
| }, | |
| "language_model.layers.46.linear_attn.in_proj_z": { | |
| "shape": [ | |
| 6144, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.0950494259595871 | |
| }, | |
| "language_model.layers.46.linear_attn.out_proj": { | |
| "shape": [ | |
| 5120, | |
| 6144 | |
| ], | |
| "relative_rmse": 0.09488178789615631 | |
| }, | |
| "language_model.layers.46.mlp.down_proj": { | |
| "shape": [ | |
| 5120, | |
| 17408 | |
| ], | |
| "relative_rmse": 0.09481244534254074 | |
| }, | |
| "language_model.layers.46.mlp.gate_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09477102756500244 | |
| }, | |
| "language_model.layers.46.mlp.up_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09480199962854385 | |
| }, | |
| "language_model.layers.47.mlp.down_proj": { | |
| "shape": [ | |
| 5120, | |
| 17408 | |
| ], | |
| "relative_rmse": 0.09494847804307938 | |
| }, | |
| "language_model.layers.47.mlp.gate_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09477753937244415 | |
| }, | |
| "language_model.layers.47.mlp.up_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.0947762131690979 | |
| }, | |
| "language_model.layers.47.self_attn.k_proj": { | |
| "shape": [ | |
| 1024, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09442693740129471 | |
| }, | |
| "language_model.layers.47.self_attn.o_proj": { | |
| "shape": [ | |
| 5120, | |
| 6144 | |
| ], | |
| "relative_rmse": 0.09441886842250824 | |
| }, | |
| "language_model.layers.47.self_attn.q_proj": { | |
| "shape": [ | |
| 12288, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09475977718830109 | |
| }, | |
| "language_model.layers.47.self_attn.v_proj": { | |
| "shape": [ | |
| 1024, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09455729275941849 | |
| }, | |
| "language_model.layers.48.linear_attn.in_proj_qkv": { | |
| "shape": [ | |
| 10240, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09464775025844574 | |
| }, | |
| "language_model.layers.48.linear_attn.in_proj_z": { | |
| "shape": [ | |
| 6144, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09490729868412018 | |
| }, | |
| "language_model.layers.48.linear_attn.out_proj": { | |
| "shape": [ | |
| 5120, | |
| 6144 | |
| ], | |
| "relative_rmse": 0.09488801658153534 | |
| }, | |
| "language_model.layers.48.mlp.down_proj": { | |
| "shape": [ | |
| 5120, | |
| 17408 | |
| ], | |
| "relative_rmse": 0.09485373646020889 | |
| }, | |
| "language_model.layers.48.mlp.gate_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09467625617980957 | |
| }, | |
| "language_model.layers.48.mlp.up_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09496848285198212 | |
| }, | |
| "language_model.layers.49.linear_attn.in_proj_qkv": { | |
| "shape": [ | |
| 10240, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09464167058467865 | |
| }, | |
| "language_model.layers.49.linear_attn.in_proj_z": { | |
| "shape": [ | |
| 6144, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09494692087173462 | |
| }, | |
| "language_model.layers.49.linear_attn.out_proj": { | |
| "shape": [ | |
| 5120, | |
| 6144 | |
| ], | |
| "relative_rmse": 0.09463748335838318 | |
| }, | |
| "language_model.layers.49.mlp.down_proj": { | |
| "shape": [ | |
| 5120, | |
| 17408 | |
| ], | |
| "relative_rmse": 0.09468481689691544 | |
| }, | |
| "language_model.layers.49.mlp.gate_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09491779655218124 | |
| }, | |
| "language_model.layers.49.mlp.up_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09485753625631332 | |
| }, | |
| "language_model.layers.50.linear_attn.in_proj_qkv": { | |
| "shape": [ | |
| 10240, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.0947871208190918 | |
| }, | |
| "language_model.layers.50.linear_attn.in_proj_z": { | |
| "shape": [ | |
| 6144, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09501579403877258 | |
| }, | |
| "language_model.layers.50.linear_attn.out_proj": { | |
| "shape": [ | |
| 5120, | |
| 6144 | |
| ], | |
| "relative_rmse": 0.09453065693378448 | |
| }, | |
| "language_model.layers.50.mlp.down_proj": { | |
| "shape": [ | |
| 5120, | |
| 17408 | |
| ], | |
| "relative_rmse": 0.09454906731843948 | |
| }, | |
| "language_model.layers.50.mlp.gate_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09466993808746338 | |
| }, | |
| "language_model.layers.50.mlp.up_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09504862129688263 | |
| }, | |
| "language_model.layers.51.mlp.down_proj": { | |
| "shape": [ | |
| 5120, | |
| 17408 | |
| ], | |
| "relative_rmse": 0.09472827613353729 | |
| }, | |
| "language_model.layers.51.mlp.gate_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09471764415502548 | |
| }, | |
| "language_model.layers.51.mlp.up_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09480338543653488 | |
| }, | |
| "language_model.layers.51.self_attn.k_proj": { | |
| "shape": [ | |
| 1024, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09441620111465454 | |
| }, | |
| "language_model.layers.51.self_attn.o_proj": { | |
| "shape": [ | |
| 5120, | |
| 6144 | |
| ], | |
| "relative_rmse": 0.09461163729429245 | |
| }, | |
| "language_model.layers.51.self_attn.q_proj": { | |
| "shape": [ | |
| 12288, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09478556364774704 | |
| }, | |
| "language_model.layers.51.self_attn.v_proj": { | |
| "shape": [ | |
| 1024, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09379476308822632 | |
| }, | |
| "language_model.layers.52.linear_attn.in_proj_qkv": { | |
| "shape": [ | |
| 10240, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09476900100708008 | |
| }, | |
| "language_model.layers.52.linear_attn.in_proj_z": { | |
| "shape": [ | |
| 6144, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09498173743486404 | |
| }, | |
| "language_model.layers.52.linear_attn.out_proj": { | |
| "shape": [ | |
| 5120, | |
| 6144 | |
| ], | |
| "relative_rmse": 0.09475252777338028 | |
| }, | |
| "language_model.layers.52.mlp.down_proj": { | |
| "shape": [ | |
| 5120, | |
| 17408 | |
| ], | |
| "relative_rmse": 0.09490261226892471 | |
| }, | |
| "language_model.layers.52.mlp.gate_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09491124004125595 | |
| }, | |
| "language_model.layers.52.mlp.up_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09485194087028503 | |
| }, | |
| "language_model.layers.53.linear_attn.in_proj_qkv": { | |
| "shape": [ | |
| 10240, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09477796405553818 | |
| }, | |
| "language_model.layers.53.linear_attn.in_proj_z": { | |
| "shape": [ | |
| 6144, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09515088051557541 | |
| }, | |
| "language_model.layers.53.linear_attn.out_proj": { | |
| "shape": [ | |
| 5120, | |
| 6144 | |
| ], | |
| "relative_rmse": 0.09477433562278748 | |
| }, | |
| "language_model.layers.53.mlp.down_proj": { | |
| "shape": [ | |
| 5120, | |
| 17408 | |
| ], | |
| "relative_rmse": 0.09496921300888062 | |
| }, | |
| "language_model.layers.53.mlp.gate_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09482976049184799 | |
| }, | |
| "language_model.layers.53.mlp.up_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09494800120592117 | |
| }, | |
| "language_model.layers.54.linear_attn.in_proj_qkv": { | |
| "shape": [ | |
| 10240, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09501806646585464 | |
| }, | |
| "language_model.layers.54.linear_attn.in_proj_z": { | |
| "shape": [ | |
| 6144, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09488914906978607 | |
| }, | |
| "language_model.layers.54.linear_attn.out_proj": { | |
| "shape": [ | |
| 5120, | |
| 6144 | |
| ], | |
| "relative_rmse": 0.09487079828977585 | |
| }, | |
| "language_model.layers.54.mlp.down_proj": { | |
| "shape": [ | |
| 5120, | |
| 17408 | |
| ], | |
| "relative_rmse": 0.09482055902481079 | |
| }, | |
| "language_model.layers.54.mlp.gate_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.094856858253479 | |
| }, | |
| "language_model.layers.54.mlp.up_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09493105113506317 | |
| }, | |
| "language_model.layers.55.mlp.down_proj": { | |
| "shape": [ | |
| 5120, | |
| 17408 | |
| ], | |
| "relative_rmse": 0.094979427754879 | |
| }, | |
| "language_model.layers.55.mlp.gate_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09496595710515976 | |
| }, | |
| "language_model.layers.55.mlp.up_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09491196274757385 | |
| }, | |
| "language_model.layers.55.self_attn.k_proj": { | |
| "shape": [ | |
| 1024, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09335911273956299 | |
| }, | |
| "language_model.layers.55.self_attn.o_proj": { | |
| "shape": [ | |
| 5120, | |
| 6144 | |
| ], | |
| "relative_rmse": 0.09476569294929504 | |
| }, | |
| "language_model.layers.55.self_attn.q_proj": { | |
| "shape": [ | |
| 12288, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09488401561975479 | |
| }, | |
| "language_model.layers.55.self_attn.v_proj": { | |
| "shape": [ | |
| 1024, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09413868933916092 | |
| }, | |
| "language_model.layers.56.linear_attn.in_proj_qkv": { | |
| "shape": [ | |
| 10240, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09488192945718765 | |
| }, | |
| "language_model.layers.56.linear_attn.in_proj_z": { | |
| "shape": [ | |
| 6144, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.0950426235795021 | |
| }, | |
| "language_model.layers.56.linear_attn.out_proj": { | |
| "shape": [ | |
| 5120, | |
| 6144 | |
| ], | |
| "relative_rmse": 0.09476073831319809 | |
| }, | |
| "language_model.layers.56.mlp.down_proj": { | |
| "shape": [ | |
| 5120, | |
| 17408 | |
| ], | |
| "relative_rmse": 0.09507889300584793 | |
| }, | |
| "language_model.layers.56.mlp.gate_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.0951027199625969 | |
| }, | |
| "language_model.layers.56.mlp.up_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09496468305587769 | |
| }, | |
| "language_model.layers.57.linear_attn.in_proj_qkv": { | |
| "shape": [ | |
| 10240, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09480869770050049 | |
| }, | |
| "language_model.layers.57.linear_attn.in_proj_z": { | |
| "shape": [ | |
| 6144, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09499526023864746 | |
| }, | |
| "language_model.layers.57.linear_attn.out_proj": { | |
| "shape": [ | |
| 5120, | |
| 6144 | |
| ], | |
| "relative_rmse": 0.09472248703241348 | |
| }, | |
| "language_model.layers.57.mlp.down_proj": { | |
| "shape": [ | |
| 5120, | |
| 17408 | |
| ], | |
| "relative_rmse": 0.09520614147186279 | |
| }, | |
| "language_model.layers.57.mlp.gate_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09498605132102966 | |
| }, | |
| "language_model.layers.57.mlp.up_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09504371881484985 | |
| }, | |
| "language_model.layers.58.linear_attn.in_proj_qkv": { | |
| "shape": [ | |
| 10240, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09481693804264069 | |
| }, | |
| "language_model.layers.58.linear_attn.in_proj_z": { | |
| "shape": [ | |
| 6144, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09519118815660477 | |
| }, | |
| "language_model.layers.58.linear_attn.out_proj": { | |
| "shape": [ | |
| 5120, | |
| 6144 | |
| ], | |
| "relative_rmse": 0.0948721244931221 | |
| }, | |
| "language_model.layers.58.mlp.down_proj": { | |
| "shape": [ | |
| 5120, | |
| 17408 | |
| ], | |
| "relative_rmse": 0.09501434862613678 | |
| }, | |
| "language_model.layers.58.mlp.gate_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09499430656433105 | |
| }, | |
| "language_model.layers.58.mlp.up_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09505595266819 | |
| }, | |
| "language_model.layers.59.mlp.down_proj": { | |
| "shape": [ | |
| 5120, | |
| 17408 | |
| ], | |
| "relative_rmse": 0.09494626522064209 | |
| }, | |
| "language_model.layers.59.mlp.gate_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09507293254137039 | |
| }, | |
| "language_model.layers.59.mlp.up_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09511616080999374 | |
| }, | |
| "language_model.layers.59.self_attn.k_proj": { | |
| "shape": [ | |
| 1024, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.0926285833120346 | |
| }, | |
| "language_model.layers.59.self_attn.o_proj": { | |
| "shape": [ | |
| 5120, | |
| 6144 | |
| ], | |
| "relative_rmse": 0.0948423221707344 | |
| }, | |
| "language_model.layers.59.self_attn.q_proj": { | |
| "shape": [ | |
| 12288, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09460699558258057 | |
| }, | |
| "language_model.layers.59.self_attn.v_proj": { | |
| "shape": [ | |
| 1024, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09384855628013611 | |
| }, | |
| "language_model.layers.60.linear_attn.in_proj_qkv": { | |
| "shape": [ | |
| 10240, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.0947812870144844 | |
| }, | |
| "language_model.layers.60.linear_attn.in_proj_z": { | |
| "shape": [ | |
| 6144, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09487397968769073 | |
| }, | |
| "language_model.layers.60.linear_attn.out_proj": { | |
| "shape": [ | |
| 5120, | |
| 6144 | |
| ], | |
| "relative_rmse": 0.09484171867370605 | |
| }, | |
| "language_model.layers.60.mlp.down_proj": { | |
| "shape": [ | |
| 5120, | |
| 17408 | |
| ], | |
| "relative_rmse": 0.09491603821516037 | |
| }, | |
| "language_model.layers.60.mlp.gate_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09490219503641129 | |
| }, | |
| "language_model.layers.60.mlp.up_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09491108357906342 | |
| }, | |
| "language_model.layers.61.linear_attn.in_proj_qkv": { | |
| "shape": [ | |
| 10240, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09496872127056122 | |
| }, | |
| "language_model.layers.61.linear_attn.in_proj_z": { | |
| "shape": [ | |
| 6144, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09485461562871933 | |
| }, | |
| "language_model.layers.61.linear_attn.out_proj": { | |
| "shape": [ | |
| 5120, | |
| 6144 | |
| ], | |
| "relative_rmse": 0.0948089063167572 | |
| }, | |
| "language_model.layers.61.mlp.down_proj": { | |
| "shape": [ | |
| 5120, | |
| 17408 | |
| ], | |
| "relative_rmse": 0.09506987780332565 | |
| }, | |
| "language_model.layers.61.mlp.gate_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09477683156728745 | |
| }, | |
| "language_model.layers.61.mlp.up_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09496009349822998 | |
| }, | |
| "language_model.layers.62.linear_attn.in_proj_qkv": { | |
| "shape": [ | |
| 10240, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09448860585689545 | |
| }, | |
| "language_model.layers.62.linear_attn.in_proj_z": { | |
| "shape": [ | |
| 6144, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09479902684688568 | |
| }, | |
| "language_model.layers.62.linear_attn.out_proj": { | |
| "shape": [ | |
| 5120, | |
| 6144 | |
| ], | |
| "relative_rmse": 0.0946703851222992 | |
| }, | |
| "language_model.layers.62.mlp.down_proj": { | |
| "shape": [ | |
| 5120, | |
| 17408 | |
| ], | |
| "relative_rmse": 0.09482859075069427 | |
| }, | |
| "language_model.layers.62.mlp.gate_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09477575868368149 | |
| }, | |
| "language_model.layers.62.mlp.up_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09491778910160065 | |
| }, | |
| "language_model.layers.63.mlp.down_proj": { | |
| "shape": [ | |
| 5120, | |
| 17408 | |
| ], | |
| "relative_rmse": 0.09433726966381073 | |
| }, | |
| "language_model.layers.63.mlp.gate_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09479404985904694 | |
| }, | |
| "language_model.layers.63.mlp.up_proj": { | |
| "shape": [ | |
| 17408, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09499823302030563 | |
| }, | |
| "language_model.layers.63.self_attn.k_proj": { | |
| "shape": [ | |
| 1024, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09264403581619263 | |
| }, | |
| "language_model.layers.63.self_attn.o_proj": { | |
| "shape": [ | |
| 5120, | |
| 6144 | |
| ], | |
| "relative_rmse": 0.09375306218862534 | |
| }, | |
| "language_model.layers.63.self_attn.q_proj": { | |
| "shape": [ | |
| 12288, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09451363235712051 | |
| }, | |
| "language_model.layers.63.self_attn.v_proj": { | |
| "shape": [ | |
| 1024, | |
| 5120 | |
| ], | |
| "relative_rmse": 0.09369884431362152 | |
| } | |
| }, | |
| "preserved": "decision readout, embeddings, vision tower, norms, convolutions, small gate projections; attention also preserved in mlp scope", | |
| "production_backend": "Triton per-matrix dequantization + BF16 GEMM; no native FP4 tensor-core claim", | |
| "source_checkpoint": "/shared/yue/jev-best-tritask-20261006/runs/balanced/checkpoints/step-00800", | |
| "source_DI": 62.43, | |
| "quantized_DI": null, | |
| "modelopt_version": "0.47.0", | |
| "no_training": true, | |
| "runtime_entrypoint": "jev_fp4.FP4DecisionModel" | |
| } | |