Files
Qwen3-4B-Thinking-2507-hqq-…/fake_quant_config.json
ModelHub XC 1a1791f4de 初始化项目,由ModelHub XC社区提供模型
Model: meteorain/Qwen3-4B-Thinking-2507-hqq-nvfp4-faked-bf16
Source: Original Platform
2026-09-07 22:24:44 +08:00

2545 lines
52 KiB
JSON

{
"method": "HQQ-style-FP4",
"format": "nvfp4",
"nbits": 4,
"group_size": 16,
"axis": 1,
"round_zero": false,
"codebook": "fp4-e2m1",
"block_size": 16,
"scale_opt_steps": 8,
"optimize": true,
"optimize_device": "cuda",
"num_modules": 252,
"num_weights": 3633315840,
"elapsed_sec": 1.329,
"quantized_modules": [
{
"name": "model.layers.0.self_attn.q_proj",
"shape": [
4096,
2560
],
"dtype": "bfloat16",
"mse": 3.768257784031448e-06,
"max_abs_err": 0.064453125
},
{
"name": "model.layers.0.self_attn.k_proj",
"shape": [
1024,
2560
],
"dtype": "bfloat16",
"mse": 4.1805692490015645e-06,
"max_abs_err": 0.02783203125
},
{
"name": "model.layers.0.self_attn.v_proj",
"shape": [
1024,
2560
],
"dtype": "bfloat16",
"mse": 3.6836208892054856e-06,
"max_abs_err": 0.0146484375
},
{
"name": "model.layers.0.self_attn.o_proj",
"shape": [
2560,
4096
],
"dtype": "bfloat16",
"mse": 3.301047854620265e-06,
"max_abs_err": 0.056640625
},
{
"name": "model.layers.0.mlp.gate_proj",
"shape": [
9728,
2560
],
"dtype": "bfloat16",
"mse": 3.903518518200144e-06,
"max_abs_err": 0.0283203125
},
{
"name": "model.layers.0.mlp.up_proj",
"shape": [
9728,
2560
],
"dtype": "bfloat16",
"mse": 3.4005606721621007e-06,
"max_abs_err": 0.0224609375
},
{
"name": "model.layers.0.mlp.down_proj",
"shape": [
2560,
9728
],
"dtype": "bfloat16",
"mse": 3.7669462926714914e-06,
"max_abs_err": 0.033203125
},
{
"name": "model.layers.1.self_attn.q_proj",
"shape": [
4096,
2560
],
"dtype": "bfloat16",
"mse": 3.2648356409481494e-06,
"max_abs_err": 0.052734375
},
{
"name": "model.layers.1.self_attn.k_proj",
"shape": [
1024,
2560
],
"dtype": "bfloat16",
"mse": 3.5341417969902977e-06,
"max_abs_err": 0.0234375
},
{
"name": "model.layers.1.self_attn.v_proj",
"shape": [
1024,
2560
],
"dtype": "bfloat16",
"mse": 3.7822826470801374e-06,
"max_abs_err": 0.0166015625
},
{
"name": "model.layers.1.self_attn.o_proj",
"shape": [
2560,
4096
],
"dtype": "bfloat16",
"mse": 3.3749711292330176e-06,
"max_abs_err": 0.037109375
},
{
"name": "model.layers.1.mlp.gate_proj",
"shape": [
9728,
2560
],
"dtype": "bfloat16",
"mse": 3.6784169878956163e-06,
"max_abs_err": 0.025390625
},
{
"name": "model.layers.1.mlp.up_proj",
"shape": [
9728,
2560
],
"dtype": "bfloat16",
"mse": 1.6732803942431929e-06,
"max_abs_err": 0.0224609375
},
{
"name": "model.layers.1.mlp.down_proj",
"shape": [
2560,
9728
],
"dtype": "bfloat16",
"mse": 1.9015756151929963e-06,
"max_abs_err": 0.076171875
},
{
"name": "model.layers.2.self_attn.q_proj",
"shape": [
4096,
2560
],
"dtype": "bfloat16",
"mse": 3.4931265417981194e-06,
"max_abs_err": 0.02490234375
},
{
"name": "model.layers.2.self_attn.k_proj",
"shape": [
1024,
2560
],
"dtype": "bfloat16",
"mse": 3.7394054288597545e-06,
"max_abs_err": 0.0322265625
},
{
"name": "model.layers.2.self_attn.v_proj",
"shape": [
1024,
2560
],
"dtype": "bfloat16",
"mse": 4.071990588272456e-06,
"max_abs_err": 0.0146484375
},
{
"name": "model.layers.2.self_attn.o_proj",
"shape": [
2560,
4096
],
"dtype": "bfloat16",
"mse": 3.534158850015956e-06,
"max_abs_err": 0.0341796875
},
{
"name": "model.layers.2.mlp.gate_proj",
"shape": [
9728,
2560
],
"dtype": "bfloat16",
"mse": 2.4790933821350336e-06,
"max_abs_err": 0.02734375
},
{
"name": "model.layers.2.mlp.up_proj",
"shape": [
9728,
2560
],
"dtype": "bfloat16",
"mse": 2.080413651128765e-06,
"max_abs_err": 0.025390625
},
{
"name": "model.layers.2.mlp.down_proj",
"shape": [
2560,
9728
],
"dtype": "bfloat16",
"mse": 1.9099190922133857e-06,
"max_abs_err": 0.0380859375
},
{
"name": "model.layers.3.self_attn.q_proj",
"shape": [
4096,
2560
],
"dtype": "bfloat16",
"mse": 3.5113710055156844e-06,
"max_abs_err": 0.0361328125
},
{
"name": "model.layers.3.self_attn.k_proj",
"shape": [
1024,
2560
],
"dtype": "bfloat16",
"mse": 3.51321045855002e-06,
"max_abs_err": 0.0244140625
},
{
"name": "model.layers.3.self_attn.v_proj",
"shape": [
1024,
2560
],
"dtype": "bfloat16",
"mse": 4.12929875892587e-06,
"max_abs_err": 0.01513671875
},
{
"name": "model.layers.3.self_attn.o_proj",
"shape": [
2560,
4096
],
"dtype": "bfloat16",
"mse": 3.577211600713781e-06,
"max_abs_err": 0.044921875
},
{
"name": "model.layers.3.mlp.gate_proj",
"shape": [
9728,
2560
],
"dtype": "bfloat16",
"mse": 4.4350472307996824e-06,
"max_abs_err": 0.0390625
},
{
"name": "model.layers.3.mlp.up_proj",
"shape": [
9728,
2560
],
"dtype": "bfloat16",
"mse": 2.390967438259395e-06,
"max_abs_err": 0.0380859375
},
{
"name": "model.layers.3.mlp.down_proj",
"shape": [
2560,
9728
],
"dtype": "bfloat16",
"mse": 2.5359679511893773e-06,
"max_abs_err": 0.0546875
},
{
"name": "model.layers.4.self_attn.q_proj",
"shape": [
4096,
2560
],
"dtype": "bfloat16",
"mse": 3.645893229986541e-06,
"max_abs_err": 0.0302734375
},
{
"name": "model.layers.4.self_attn.k_proj",
"shape": [
1024,
2560
],
"dtype": "bfloat16",
"mse": 3.703610218508402e-06,
"max_abs_err": 0.033203125
},
{
"name": "model.layers.4.self_attn.v_proj",
"shape": [
1024,
2560
],
"dtype": "bfloat16",
"mse": 4.209773123875493e-06,
"max_abs_err": 0.01611328125
},
{
"name": "model.layers.4.self_attn.o_proj",
"shape": [
2560,
4096
],
"dtype": "bfloat16",
"mse": 3.7426696053444175e-06,
"max_abs_err": 0.06640625
},
{
"name": "model.layers.4.mlp.gate_proj",
"shape": [
9728,
2560
],
"dtype": "bfloat16",
"mse": 6.893119461892638e-06,
"max_abs_err": 0.037109375
},
{
"name": "model.layers.4.mlp.up_proj",
"shape": [
9728,
2560
],
"dtype": "bfloat16",
"mse": 2.9045918381598312e-06,
"max_abs_err": 0.029296875
},
{
"name": "model.layers.4.mlp.down_proj",
"shape": [
2560,
9728
],
"dtype": "bfloat16",
"mse": 2.9692200769204646e-06,
"max_abs_err": 0.046875
},
{
"name": "model.layers.5.self_attn.q_proj",
"shape": [
4096,
2560
],
"dtype": "bfloat16",
"mse": 3.6117769468546612e-06,
"max_abs_err": 0.029296875
},
{
"name": "model.layers.5.self_attn.k_proj",
"shape": [
1024,
2560
],
"dtype": "bfloat16",
"mse": 3.5848272545990767e-06,
"max_abs_err": 0.0302734375
},
{
"name": "model.layers.5.self_attn.v_proj",
"shape": [
1024,
2560
],
"dtype": "bfloat16",
"mse": 4.144095782976365e-06,
"max_abs_err": 0.01611328125
},
{
"name": "model.layers.5.self_attn.o_proj",
"shape": [
2560,
4096
],
"dtype": "bfloat16",
"mse": 3.483831505946e-06,
"max_abs_err": 0.05078125
},
{
"name": "model.layers.5.mlp.gate_proj",
"shape": [
9728,
2560
],
"dtype": "bfloat16",
"mse": 5.9980461628583726e-06,
"max_abs_err": 0.0322265625
},
{
"name": "model.layers.5.mlp.up_proj",
"shape": [
9728,
2560
],
"dtype": "bfloat16",
"mse": 3.2627729069645284e-06,
"max_abs_err": 0.0263671875
},
{
"name": "model.layers.5.mlp.down_proj",
"shape": [
2560,
9728
],
"dtype": "bfloat16",
"mse": 3.1713691441837e-06,
"max_abs_err": 0.056640625
},
{
"name": "model.layers.6.self_attn.q_proj",
"shape": [
4096,
2560
],
"dtype": "bfloat16",
"mse": 3.8140126434882404e-06,
"max_abs_err": 0.03955078125
},
{
"name": "model.layers.6.self_attn.k_proj",
"shape": [
1024,
2560
],
"dtype": "bfloat16",
"mse": 3.6102669582760427e-06,
"max_abs_err": 0.021484375
},
{
"name": "model.layers.6.self_attn.v_proj",
"shape": [
1024,
2560
],
"dtype": "bfloat16",
"mse": 4.131844889343483e-06,
"max_abs_err": 0.0244140625
},
{
"name": "model.layers.6.self_attn.o_proj",
"shape": [
2560,
4096
],
"dtype": "bfloat16",
"mse": 3.7248246371746063e-06,
"max_abs_err": 0.08203125
},
{
"name": "model.layers.6.mlp.gate_proj",
"shape": [
9728,
2560
],
"dtype": "bfloat16",
"mse": 5.346630132407881e-06,
"max_abs_err": 0.03125
},
{
"name": "model.layers.6.mlp.up_proj",
"shape": [
9728,
2560
],
"dtype": "bfloat16",
"mse": 3.4553834211692447e-06,
"max_abs_err": 0.02734375
},
{
"name": "model.layers.6.mlp.down_proj",
"shape": [
2560,
9728
],
"dtype": "bfloat16",
"mse": 3.4227036849188153e-06,
"max_abs_err": 0.048828125
},
{
"name": "model.layers.7.self_attn.q_proj",
"shape": [
4096,
2560
],
"dtype": "bfloat16",
"mse": 3.6055068903806387e-06,
"max_abs_err": 0.0302734375
},
{
"name": "model.layers.7.self_attn.k_proj",
"shape": [
1024,
2560
],
"dtype": "bfloat16",
"mse": 3.6101646401220933e-06,
"max_abs_err": 0.025390625
},
{
"name": "model.layers.7.self_attn.v_proj",
"shape": [
1024,
2560
],
"dtype": "bfloat16",
"mse": 4.745930709759705e-06,
"max_abs_err": 0.017578125
},
{
"name": "model.layers.7.self_attn.o_proj",
"shape": [
2560,
4096
],
"dtype": "bfloat16",
"mse": 3.626036004789057e-06,
"max_abs_err": 0.0458984375
},
{
"name": "model.layers.7.mlp.gate_proj",
"shape": [
9728,
2560
],
"dtype": "bfloat16",
"mse": 5.225013410381507e-06,
"max_abs_err": 0.03125
},
{
"name": "model.layers.7.mlp.up_proj",
"shape": [
9728,
2560
],
"dtype": "bfloat16",
"mse": 3.498972318993765e-06,
"max_abs_err": 0.029296875
},
{
"name": "model.layers.7.mlp.down_proj",
"shape": [
2560,
9728
],
"dtype": "bfloat16",
"mse": 3.472020353001426e-06,
"max_abs_err": 0.03515625
},
{
"name": "model.layers.8.self_attn.q_proj",
"shape": [
4096,
2560
],
"dtype": "bfloat16",
"mse": 3.646335017037927e-06,
"max_abs_err": 0.0263671875
},
{
"name": "model.layers.8.self_attn.k_proj",
"shape": [
1024,
2560
],
"dtype": "bfloat16",
"mse": 3.738590066859615e-06,
"max_abs_err": 0.02001953125
},
{
"name": "model.layers.8.self_attn.v_proj",
"shape": [
1024,
2560
],
"dtype": "bfloat16",
"mse": 4.72011470264988e-06,
"max_abs_err": 0.015625
},
{
"name": "model.layers.8.self_attn.o_proj",
"shape": [
2560,
4096
],
"dtype": "bfloat16",
"mse": 3.7951024296489777e-06,
"max_abs_err": 0.041015625
},
{
"name": "model.layers.8.mlp.gate_proj",
"shape": [
9728,
2560
],
"dtype": "bfloat16",
"mse": 4.536404958344065e-06,
"max_abs_err": 0.02783203125
},
{
"name": "model.layers.8.mlp.up_proj",
"shape": [
9728,
2560
],
"dtype": "bfloat16",
"mse": 3.664447604023735e-06,
"max_abs_err": 0.02783203125
},
{
"name": "model.layers.8.mlp.down_proj",
"shape": [
2560,
9728
],
"dtype": "bfloat16",
"mse": 3.6754447592102224e-06,
"max_abs_err": 0.12109375
},
{
"name": "model.layers.9.self_attn.q_proj",
"shape": [
4096,
2560
],
"dtype": "bfloat16",
"mse": 3.4713209515757626e-06,
"max_abs_err": 0.037109375
},
{
"name": "model.layers.9.self_attn.k_proj",
"shape": [
1024,
2560
],
"dtype": "bfloat16",
"mse": 3.466504267635173e-06,
"max_abs_err": 0.0439453125
},
{
"name": "model.layers.9.self_attn.v_proj",
"shape": [
1024,
2560
],
"dtype": "bfloat16",
"mse": 4.825919859285932e-06,
"max_abs_err": 0.0166015625
},
{
"name": "model.layers.9.self_attn.o_proj",
"shape": [
2560,
4096
],
"dtype": "bfloat16",
"mse": 3.7283496112650027e-06,
"max_abs_err": 0.037109375
},
{
"name": "model.layers.9.mlp.gate_proj",
"shape": [
9728,
2560
],
"dtype": "bfloat16",
"mse": 5.309510470397072e-06,
"max_abs_err": 0.03125
},
{
"name": "model.layers.9.mlp.up_proj",
"shape": [
9728,
2560
],
"dtype": "bfloat16",
"mse": 3.3653868740657344e-06,
"max_abs_err": 0.0283203125
},
{
"name": "model.layers.9.mlp.down_proj",
"shape": [
2560,
9728
],
"dtype": "bfloat16",
"mse": 3.2919760997174308e-06,
"max_abs_err": 0.0361328125
},
{
"name": "model.layers.10.self_attn.q_proj",
"shape": [
4096,
2560
],
"dtype": "bfloat16",
"mse": 3.6937278764526127e-06,
"max_abs_err": 0.03515625
},
{
"name": "model.layers.10.self_attn.k_proj",
"shape": [
1024,
2560
],
"dtype": "bfloat16",
"mse": 3.678484517877223e-06,
"max_abs_err": 0.021484375
},
{
"name": "model.layers.10.self_attn.v_proj",
"shape": [
1024,
2560
],
"dtype": "bfloat16",
"mse": 4.734218236990273e-06,
"max_abs_err": 0.017578125
},
{
"name": "model.layers.10.self_attn.o_proj",
"shape": [
2560,
4096
],
"dtype": "bfloat16",
"mse": 3.8697121453878935e-06,
"max_abs_err": 0.060546875
},
{
"name": "model.layers.10.mlp.gate_proj",
"shape": [
9728,
2560
],
"dtype": "bfloat16",
"mse": 5.073019110568566e-06,
"max_abs_err": 0.037109375
},
{
"name": "model.layers.10.mlp.up_proj",
"shape": [
9728,
2560
],
"dtype": "bfloat16",
"mse": 3.497885927572497e-06,
"max_abs_err": 0.0234375
},
{
"name": "model.layers.10.mlp.down_proj",
"shape": [
2560,
9728
],
"dtype": "bfloat16",
"mse": 3.4624461022758624e-06,
"max_abs_err": 0.056640625
},
{
"name": "model.layers.11.self_attn.q_proj",
"shape": [
4096,
2560
],
"dtype": "bfloat16",
"mse": 3.678142775243032e-06,
"max_abs_err": 0.029296875
},
{
"name": "model.layers.11.self_attn.k_proj",
"shape": [
1024,
2560
],
"dtype": "bfloat16",
"mse": 3.802990931944805e-06,
"max_abs_err": 0.0419921875
},
{
"name": "model.layers.11.self_attn.v_proj",
"shape": [
1024,
2560
],
"dtype": "bfloat16",
"mse": 4.838274890062166e-06,
"max_abs_err": 0.01611328125
},
{
"name": "model.layers.11.self_attn.o_proj",
"shape": [
2560,
4096
],
"dtype": "bfloat16",
"mse": 3.7880204217799474e-06,
"max_abs_err": 0.029541015625
},
{
"name": "model.layers.11.mlp.gate_proj",
"shape": [
9728,
2560
],
"dtype": "bfloat16",
"mse": 4.7551079660479445e-06,
"max_abs_err": 0.044921875
},
{
"name": "model.layers.11.mlp.up_proj",
"shape": [
9728,
2560
],
"dtype": "bfloat16",
"mse": 3.7109364257048583e-06,
"max_abs_err": 0.0244140625
},
{
"name": "model.layers.11.mlp.down_proj",
"shape": [
2560,
9728
],
"dtype": "bfloat16",
"mse": 3.621092673711246e-06,
"max_abs_err": 0.03515625
},
{
"name": "model.layers.12.self_attn.q_proj",
"shape": [
4096,
2560
],
"dtype": "bfloat16",
"mse": 3.756895694095874e-06,
"max_abs_err": 0.0390625
},
{
"name": "model.layers.12.self_attn.k_proj",
"shape": [
1024,
2560
],
"dtype": "bfloat16",
"mse": 3.663145207610796e-06,
"max_abs_err": 0.05859375
},
{
"name": "model.layers.12.self_attn.v_proj",
"shape": [
1024,
2560
],
"dtype": "bfloat16",
"mse": 4.748278570332332e-06,
"max_abs_err": 0.02099609375
},
{
"name": "model.layers.12.self_attn.o_proj",
"shape": [
2560,
4096
],
"dtype": "bfloat16",
"mse": 3.711930958161247e-06,
"max_abs_err": 0.06640625
},
{
"name": "model.layers.12.mlp.gate_proj",
"shape": [
9728,
2560
],
"dtype": "bfloat16",
"mse": 4.507253834162839e-06,
"max_abs_err": 0.041015625
},
{
"name": "model.layers.12.mlp.up_proj",
"shape": [
9728,
2560
],
"dtype": "bfloat16",
"mse": 3.863771780743264e-06,
"max_abs_err": 0.0322265625
},
{
"name": "model.layers.12.mlp.down_proj",
"shape": [
2560,
9728
],
"dtype": "bfloat16",
"mse": 3.704288474182249e-06,
"max_abs_err": 0.0361328125
},
{
"name": "model.layers.13.self_attn.q_proj",
"shape": [
4096,
2560
],
"dtype": "bfloat16",
"mse": 3.7196980429143878e-06,
"max_abs_err": 0.03125
},
{
"name": "model.layers.13.self_attn.k_proj",
"shape": [
1024,
2560
],
"dtype": "bfloat16",
"mse": 3.6132428249402437e-06,
"max_abs_err": 0.029296875
},
{
"name": "model.layers.13.self_attn.v_proj",
"shape": [
1024,
2560
],
"dtype": "bfloat16",
"mse": 4.557133706839522e-06,
"max_abs_err": 0.017578125
},
{
"name": "model.layers.13.self_attn.o_proj",
"shape": [
2560,
4096
],
"dtype": "bfloat16",
"mse": 3.5474984088068595e-06,
"max_abs_err": 0.03125
},
{
"name": "model.layers.13.mlp.gate_proj",
"shape": [
9728,
2560
],
"dtype": "bfloat16",
"mse": 4.2022538764285855e-06,
"max_abs_err": 0.041015625
},
{
"name": "model.layers.13.mlp.up_proj",
"shape": [
9728,
2560
],
"dtype": "bfloat16",
"mse": 3.9652832128922455e-06,
"max_abs_err": 0.02587890625
},
{
"name": "model.layers.13.mlp.down_proj",
"shape": [
2560,
9728
],
"dtype": "bfloat16",
"mse": 3.7342795167205622e-06,
"max_abs_err": 0.078125
},
{
"name": "model.layers.14.self_attn.q_proj",
"shape": [
4096,
2560
],
"dtype": "bfloat16",
"mse": 3.7219710975477938e-06,
"max_abs_err": 0.03125
},
{
"name": "model.layers.14.self_attn.k_proj",
"shape": [
1024,
2560
],
"dtype": "bfloat16",
"mse": 3.6810092751693446e-06,
"max_abs_err": 0.0322265625
},
{
"name": "model.layers.14.self_attn.v_proj",
"shape": [
1024,
2560
],
"dtype": "bfloat16",
"mse": 4.300414275348885e-06,
"max_abs_err": 0.015625
},
{
"name": "model.layers.14.self_attn.o_proj",
"shape": [
2560,
4096
],
"dtype": "bfloat16",
"mse": 3.447492872510338e-06,
"max_abs_err": 0.0703125
},
{
"name": "model.layers.14.mlp.gate_proj",
"shape": [
9728,
2560
],
"dtype": "bfloat16",
"mse": 4.085748969373526e-06,
"max_abs_err": 0.0517578125
},
{
"name": "model.layers.14.mlp.up_proj",
"shape": [
9728,
2560
],
"dtype": "bfloat16",
"mse": 3.852575446217088e-06,
"max_abs_err": 0.029296875
},
{
"name": "model.layers.14.mlp.down_proj",
"shape": [
2560,
9728
],
"dtype": "bfloat16",
"mse": 3.6500291571428534e-06,
"max_abs_err": 0.044921875
},
{
"name": "model.layers.15.self_attn.q_proj",
"shape": [
4096,
2560
],
"dtype": "bfloat16",
"mse": 3.8091905025794404e-06,
"max_abs_err": 0.0341796875
},
{
"name": "model.layers.15.self_attn.k_proj",
"shape": [
1024,
2560
],
"dtype": "bfloat16",
"mse": 3.639295073298854e-06,
"max_abs_err": 0.0341796875
},
{
"name": "model.layers.15.self_attn.v_proj",
"shape": [
1024,
2560
],
"dtype": "bfloat16",
"mse": 4.229587375448318e-06,
"max_abs_err": 0.01806640625
},
{
"name": "model.layers.15.self_attn.o_proj",
"shape": [
2560,
4096
],
"dtype": "bfloat16",
"mse": 3.248672101108241e-06,
"max_abs_err": 0.0341796875
},
{
"name": "model.layers.15.mlp.gate_proj",
"shape": [
9728,
2560
],
"dtype": "bfloat16",
"mse": 3.8018354189262027e-06,
"max_abs_err": 0.0546875
},
{
"name": "model.layers.15.mlp.up_proj",
"shape": [
9728,
2560
],
"dtype": "bfloat16",
"mse": 3.767696171053103e-06,
"max_abs_err": 0.0341796875
},
{
"name": "model.layers.15.mlp.down_proj",
"shape": [
2560,
9728
],
"dtype": "bfloat16",
"mse": 3.587018454709323e-06,
"max_abs_err": 0.060546875
},
{
"name": "model.layers.16.self_attn.q_proj",
"shape": [
4096,
2560
],
"dtype": "bfloat16",
"mse": 3.5913683404942276e-06,
"max_abs_err": 0.03125
},
{
"name": "model.layers.16.self_attn.k_proj",
"shape": [
1024,
2560
],
"dtype": "bfloat16",
"mse": 3.420893108341261e-06,
"max_abs_err": 0.0234375
},
{
"name": "model.layers.16.self_attn.v_proj",
"shape": [
1024,
2560
],
"dtype": "bfloat16",
"mse": 4.237365828885231e-06,
"max_abs_err": 0.01904296875
},
{
"name": "model.layers.16.self_attn.o_proj",
"shape": [
2560,
4096
],
"dtype": "bfloat16",
"mse": 3.437242185100331e-06,
"max_abs_err": 0.068359375
},
{
"name": "model.layers.16.mlp.gate_proj",
"shape": [
9728,
2560
],
"dtype": "bfloat16",
"mse": 3.995630322606303e-06,
"max_abs_err": 0.060546875
},
{
"name": "model.layers.16.mlp.up_proj",
"shape": [
9728,
2560
],
"dtype": "bfloat16",
"mse": 3.718114612638601e-06,
"max_abs_err": 0.076171875
},
{
"name": "model.layers.16.mlp.down_proj",
"shape": [
2560,
9728
],
"dtype": "bfloat16",
"mse": 3.4747763493214734e-06,
"max_abs_err": 0.087890625
},
{
"name": "model.layers.17.self_attn.q_proj",
"shape": [
4096,
2560
],
"dtype": "bfloat16",
"mse": 3.776124458454433e-06,
"max_abs_err": 0.0302734375
},
{
"name": "model.layers.17.self_attn.k_proj",
"shape": [
1024,
2560
],
"dtype": "bfloat16",
"mse": 3.2296302379108965e-06,
"max_abs_err": 0.02734375
},
{
"name": "model.layers.17.self_attn.v_proj",
"shape": [
1024,
2560
],
"dtype": "bfloat16",
"mse": 4.032822744193254e-06,
"max_abs_err": 0.017578125
},
{
"name": "model.layers.17.self_attn.o_proj",
"shape": [
2560,
4096
],
"dtype": "bfloat16",
"mse": 3.3230789995286614e-06,
"max_abs_err": 0.046875
},
{
"name": "model.layers.17.mlp.gate_proj",
"shape": [
9728,
2560
],
"dtype": "bfloat16",
"mse": 3.885952082782751e-06,
"max_abs_err": 0.03515625
},
{
"name": "model.layers.17.mlp.up_proj",
"shape": [
9728,
2560
],
"dtype": "bfloat16",
"mse": 3.7487388908630237e-06,
"max_abs_err": 0.0322265625
},
{
"name": "model.layers.17.mlp.down_proj",
"shape": [
2560,
9728
],
"dtype": "bfloat16",
"mse": 3.496542376524303e-06,
"max_abs_err": 0.046875
},
{
"name": "model.layers.18.self_attn.q_proj",
"shape": [
4096,
2560
],
"dtype": "bfloat16",
"mse": 3.5266555187263293e-06,
"max_abs_err": 0.0361328125
},
{
"name": "model.layers.18.self_attn.k_proj",
"shape": [
1024,
2560
],
"dtype": "bfloat16",
"mse": 3.3046878797904355e-06,
"max_abs_err": 0.02587890625
},
{
"name": "model.layers.18.self_attn.v_proj",
"shape": [
1024,
2560
],
"dtype": "bfloat16",
"mse": 4.189625997241819e-06,
"max_abs_err": 0.015625
},
{
"name": "model.layers.18.self_attn.o_proj",
"shape": [
2560,
4096
],
"dtype": "bfloat16",
"mse": 3.33023194798443e-06,
"max_abs_err": 0.048828125
},
{
"name": "model.layers.18.mlp.gate_proj",
"shape": [
9728,
2560
],
"dtype": "bfloat16",
"mse": 3.791058134083869e-06,
"max_abs_err": 0.048828125
},
{
"name": "model.layers.18.mlp.up_proj",
"shape": [
9728,
2560
],
"dtype": "bfloat16",
"mse": 3.7423085359478137e-06,
"max_abs_err": 0.02734375
},
{
"name": "model.layers.18.mlp.down_proj",
"shape": [
2560,
9728
],
"dtype": "bfloat16",
"mse": 3.489400569378631e-06,
"max_abs_err": 0.052734375
},
{
"name": "model.layers.19.self_attn.q_proj",
"shape": [
4096,
2560
],
"dtype": "bfloat16",
"mse": 3.6483347685134504e-06,
"max_abs_err": 0.0361328125
},
{
"name": "model.layers.19.self_attn.k_proj",
"shape": [
1024,
2560
],
"dtype": "bfloat16",
"mse": 3.086352762693423e-06,
"max_abs_err": 0.048828125
},
{
"name": "model.layers.19.self_attn.v_proj",
"shape": [
1024,
2560
],
"dtype": "bfloat16",
"mse": 4.000663011538563e-06,
"max_abs_err": 0.01708984375
},
{
"name": "model.layers.19.self_attn.o_proj",
"shape": [
2560,
4096
],
"dtype": "bfloat16",
"mse": 3.3754511150618782e-06,
"max_abs_err": 0.05859375
},
{
"name": "model.layers.19.mlp.gate_proj",
"shape": [
9728,
2560
],
"dtype": "bfloat16",
"mse": 3.7419072214106563e-06,
"max_abs_err": 0.0546875
},
{
"name": "model.layers.19.mlp.up_proj",
"shape": [
9728,
2560
],
"dtype": "bfloat16",
"mse": 3.7611628158629173e-06,
"max_abs_err": 0.0341796875
},
{
"name": "model.layers.19.mlp.down_proj",
"shape": [
2560,
9728
],
"dtype": "bfloat16",
"mse": 3.485625938992598e-06,
"max_abs_err": 0.052734375
},
{
"name": "model.layers.20.self_attn.q_proj",
"shape": [
4096,
2560
],
"dtype": "bfloat16",
"mse": 3.9093611121643335e-06,
"max_abs_err": 0.0341796875
},
{
"name": "model.layers.20.self_attn.k_proj",
"shape": [
1024,
2560
],
"dtype": "bfloat16",
"mse": 3.0202143079804955e-06,
"max_abs_err": 0.02685546875
},
{
"name": "model.layers.20.self_attn.v_proj",
"shape": [
1024,
2560
],
"dtype": "bfloat16",
"mse": 3.9840401768742595e-06,
"max_abs_err": 0.0185546875
},
{
"name": "model.layers.20.self_attn.o_proj",
"shape": [
2560,
4096
],
"dtype": "bfloat16",
"mse": 3.223913608962903e-06,
"max_abs_err": 0.064453125
},
{
"name": "model.layers.20.mlp.gate_proj",
"shape": [
9728,
2560
],
"dtype": "bfloat16",
"mse": 3.6737637856276706e-06,
"max_abs_err": 0.05859375
},
{
"name": "model.layers.20.mlp.up_proj",
"shape": [
9728,
2560
],
"dtype": "bfloat16",
"mse": 3.7812017126270803e-06,
"max_abs_err": 0.0302734375
},
{
"name": "model.layers.20.mlp.down_proj",
"shape": [
2560,
9728
],
"dtype": "bfloat16",
"mse": 3.5414814192336053e-06,
"max_abs_err": 0.0390625
},
{
"name": "model.layers.21.self_attn.q_proj",
"shape": [
4096,
2560
],
"dtype": "bfloat16",
"mse": 3.7330196391849313e-06,
"max_abs_err": 0.037109375
},
{
"name": "model.layers.21.self_attn.k_proj",
"shape": [
1024,
2560
],
"dtype": "bfloat16",
"mse": 3.005748794748797e-06,
"max_abs_err": 0.02490234375
},
{
"name": "model.layers.21.self_attn.v_proj",
"shape": [
1024,
2560
],
"dtype": "bfloat16",
"mse": 3.868387921102112e-06,
"max_abs_err": 0.0302734375
},
{
"name": "model.layers.21.self_attn.o_proj",
"shape": [
2560,
4096
],
"dtype": "bfloat16",
"mse": 3.1284337183024036e-06,
"max_abs_err": 0.02294921875
},
{
"name": "model.layers.21.mlp.gate_proj",
"shape": [
9728,
2560
],
"dtype": "bfloat16",
"mse": 3.547286269167671e-06,
"max_abs_err": 0.048828125
},
{
"name": "model.layers.21.mlp.up_proj",
"shape": [
9728,
2560
],
"dtype": "bfloat16",
"mse": 3.7603124383167597e-06,
"max_abs_err": 0.03125
},
{
"name": "model.layers.21.mlp.down_proj",
"shape": [
2560,
9728
],
"dtype": "bfloat16",
"mse": 3.594193003664259e-06,
"max_abs_err": 0.08203125
},
{
"name": "model.layers.22.self_attn.q_proj",
"shape": [
4096,
2560
],
"dtype": "bfloat16",
"mse": 3.641027433332056e-06,
"max_abs_err": 0.0390625
},
{
"name": "model.layers.22.self_attn.k_proj",
"shape": [
1024,
2560
],
"dtype": "bfloat16",
"mse": 2.7974103886663215e-06,
"max_abs_err": 0.05859375
},
{
"name": "model.layers.22.self_attn.v_proj",
"shape": [
1024,
2560
],
"dtype": "bfloat16",
"mse": 3.837092208414106e-06,
"max_abs_err": 0.0205078125
},
{
"name": "model.layers.22.self_attn.o_proj",
"shape": [
2560,
4096
],
"dtype": "bfloat16",
"mse": 3.2613731946185e-06,
"max_abs_err": 0.0537109375
},
{
"name": "model.layers.22.mlp.gate_proj",
"shape": [
9728,
2560
],
"dtype": "bfloat16",
"mse": 3.794220219788258e-06,
"max_abs_err": 0.0498046875
},
{
"name": "model.layers.22.mlp.up_proj",
"shape": [
9728,
2560
],
"dtype": "bfloat16",
"mse": 3.9340175135293975e-06,
"max_abs_err": 0.029296875
},
{
"name": "model.layers.22.mlp.down_proj",
"shape": [
2560,
9728
],
"dtype": "bfloat16",
"mse": 3.6261355944589013e-06,
"max_abs_err": 0.041015625
},
{
"name": "model.layers.23.self_attn.q_proj",
"shape": [
4096,
2560
],
"dtype": "bfloat16",
"mse": 4.125283339817543e-06,
"max_abs_err": 0.0361328125
},
{
"name": "model.layers.23.self_attn.k_proj",
"shape": [
1024,
2560
],
"dtype": "bfloat16",
"mse": 3.1230417789629428e-06,
"max_abs_err": 0.0244140625
},
{
"name": "model.layers.23.self_attn.v_proj",
"shape": [
1024,
2560
],
"dtype": "bfloat16",
"mse": 3.978055701736594e-06,
"max_abs_err": 0.02294921875
},
{
"name": "model.layers.23.self_attn.o_proj",
"shape": [
2560,
4096
],
"dtype": "bfloat16",
"mse": 3.5038490295846714e-06,
"max_abs_err": 0.037109375
},
{
"name": "model.layers.23.mlp.gate_proj",
"shape": [
9728,
2560
],
"dtype": "bfloat16",
"mse": 4.082638497493463e-06,
"max_abs_err": 0.052734375
},
{
"name": "model.layers.23.mlp.up_proj",
"shape": [
9728,
2560
],
"dtype": "bfloat16",
"mse": 3.96256655221805e-06,
"max_abs_err": 0.0302734375
},
{
"name": "model.layers.23.mlp.down_proj",
"shape": [
2560,
9728
],
"dtype": "bfloat16",
"mse": 3.6798392102355137e-06,
"max_abs_err": 0.078125
},
{
"name": "model.layers.24.self_attn.q_proj",
"shape": [
4096,
2560
],
"dtype": "bfloat16",
"mse": 4.187786089460133e-06,
"max_abs_err": 0.03125
},
{
"name": "model.layers.24.self_attn.k_proj",
"shape": [
1024,
2560
],
"dtype": "bfloat16",
"mse": 3.107787051703781e-06,
"max_abs_err": 0.02392578125
},
{
"name": "model.layers.24.self_attn.v_proj",
"shape": [
1024,
2560
],
"dtype": "bfloat16",
"mse": 3.896165708283661e-06,
"max_abs_err": 0.025390625
},
{
"name": "model.layers.24.self_attn.o_proj",
"shape": [
2560,
4096
],
"dtype": "bfloat16",
"mse": 3.6212386476108804e-06,
"max_abs_err": 0.044921875
},
{
"name": "model.layers.24.mlp.gate_proj",
"shape": [
9728,
2560
],
"dtype": "bfloat16",
"mse": 4.245420768711483e-06,
"max_abs_err": 0.0400390625
},
{
"name": "model.layers.24.mlp.up_proj",
"shape": [
9728,
2560
],
"dtype": "bfloat16",
"mse": 3.989016931882361e-06,
"max_abs_err": 0.0302734375
},
{
"name": "model.layers.24.mlp.down_proj",
"shape": [
2560,
9728
],
"dtype": "bfloat16",
"mse": 3.749827101273695e-06,
"max_abs_err": 0.0400390625
},
{
"name": "model.layers.25.self_attn.q_proj",
"shape": [
4096,
2560
],
"dtype": "bfloat16",
"mse": 3.7368688481365098e-06,
"max_abs_err": 0.0390625
},
{
"name": "model.layers.25.self_attn.k_proj",
"shape": [
1024,
2560
],
"dtype": "bfloat16",
"mse": 3.3282074127782835e-06,
"max_abs_err": 0.0185546875
},
{
"name": "model.layers.25.self_attn.v_proj",
"shape": [
1024,
2560
],
"dtype": "bfloat16",
"mse": 4.2598703657859005e-06,
"max_abs_err": 0.0205078125
},
{
"name": "model.layers.25.self_attn.o_proj",
"shape": [
2560,
4096
],
"dtype": "bfloat16",
"mse": 3.52777419720951e-06,
"max_abs_err": 0.0283203125
},
{
"name": "model.layers.25.mlp.gate_proj",
"shape": [
9728,
2560
],
"dtype": "bfloat16",
"mse": 4.343057753430912e-06,
"max_abs_err": 0.0380859375
},
{
"name": "model.layers.25.mlp.up_proj",
"shape": [
9728,
2560
],
"dtype": "bfloat16",
"mse": 3.997025942226173e-06,
"max_abs_err": 0.02734375
},
{
"name": "model.layers.25.mlp.down_proj",
"shape": [
2560,
9728
],
"dtype": "bfloat16",
"mse": 3.785747594520217e-06,
"max_abs_err": 0.068359375
},
{
"name": "model.layers.26.self_attn.q_proj",
"shape": [
4096,
2560
],
"dtype": "bfloat16",
"mse": 3.869783540721983e-06,
"max_abs_err": 0.0341796875
},
{
"name": "model.layers.26.self_attn.k_proj",
"shape": [
1024,
2560
],
"dtype": "bfloat16",
"mse": 3.234986024835962e-06,
"max_abs_err": 0.02294921875
},
{
"name": "model.layers.26.self_attn.v_proj",
"shape": [
1024,
2560
],
"dtype": "bfloat16",
"mse": 4.1970106394728646e-06,
"max_abs_err": 0.021484375
},
{
"name": "model.layers.26.self_attn.o_proj",
"shape": [
2560,
4096
],
"dtype": "bfloat16",
"mse": 3.6877838738291757e-06,
"max_abs_err": 0.0263671875
},
{
"name": "model.layers.26.mlp.gate_proj",
"shape": [
9728,
2560
],
"dtype": "bfloat16",
"mse": 4.359773811302148e-06,
"max_abs_err": 0.0400390625
},
{
"name": "model.layers.26.mlp.up_proj",
"shape": [
9728,
2560
],
"dtype": "bfloat16",
"mse": 4.090351922059199e-06,
"max_abs_err": 0.0291748046875
},
{
"name": "model.layers.26.mlp.down_proj",
"shape": [
2560,
9728
],
"dtype": "bfloat16",
"mse": 3.886560079990886e-06,
"max_abs_err": 0.05859375
},
{
"name": "model.layers.27.self_attn.q_proj",
"shape": [
4096,
2560
],
"dtype": "bfloat16",
"mse": 3.8066270917624934e-06,
"max_abs_err": 0.0302734375
},
{
"name": "model.layers.27.self_attn.k_proj",
"shape": [
1024,
2560
],
"dtype": "bfloat16",
"mse": 2.992305326188216e-06,
"max_abs_err": 0.01513671875
},
{
"name": "model.layers.27.self_attn.v_proj",
"shape": [
1024,
2560
],
"dtype": "bfloat16",
"mse": 3.947373897972284e-06,
"max_abs_err": 0.0205078125
},
{
"name": "model.layers.27.self_attn.o_proj",
"shape": [
2560,
4096
],
"dtype": "bfloat16",
"mse": 3.5228938486397965e-06,
"max_abs_err": 0.0322265625
},
{
"name": "model.layers.27.mlp.gate_proj",
"shape": [
9728,
2560
],
"dtype": "bfloat16",
"mse": 4.318407718528761e-06,
"max_abs_err": 0.048828125
},
{
"name": "model.layers.27.mlp.up_proj",
"shape": [
9728,
2560
],
"dtype": "bfloat16",
"mse": 4.187585091131041e-06,
"max_abs_err": 0.02978515625
},
{
"name": "model.layers.27.mlp.down_proj",
"shape": [
2560,
9728
],
"dtype": "bfloat16",
"mse": 3.979610937676625e-06,
"max_abs_err": 0.041015625
},
{
"name": "model.layers.28.self_attn.q_proj",
"shape": [
4096,
2560
],
"dtype": "bfloat16",
"mse": 3.9644619391765445e-06,
"max_abs_err": 0.029296875
},
{
"name": "model.layers.28.self_attn.k_proj",
"shape": [
1024,
2560
],
"dtype": "bfloat16",
"mse": 3.198609874743852e-06,
"max_abs_err": 0.02978515625
},
{
"name": "model.layers.28.self_attn.v_proj",
"shape": [
1024,
2560
],
"dtype": "bfloat16",
"mse": 4.385616193758324e-06,
"max_abs_err": 0.017578125
},
{
"name": "model.layers.28.self_attn.o_proj",
"shape": [
2560,
4096
],
"dtype": "bfloat16",
"mse": 3.5050950373261003e-06,
"max_abs_err": 0.0341796875
},
{
"name": "model.layers.28.mlp.gate_proj",
"shape": [
9728,
2560
],
"dtype": "bfloat16",
"mse": 4.226759756420506e-06,
"max_abs_err": 0.044921875
},
{
"name": "model.layers.28.mlp.up_proj",
"shape": [
9728,
2560
],
"dtype": "bfloat16",
"mse": 4.279801942175254e-06,
"max_abs_err": 0.03515625
},
{
"name": "model.layers.28.mlp.down_proj",
"shape": [
2560,
9728
],
"dtype": "bfloat16",
"mse": 4.078382517036516e-06,
"max_abs_err": 0.048828125
},
{
"name": "model.layers.29.self_attn.q_proj",
"shape": [
4096,
2560
],
"dtype": "bfloat16",
"mse": 4.35167976320372e-06,
"max_abs_err": 0.03125
},
{
"name": "model.layers.29.self_attn.k_proj",
"shape": [
1024,
2560
],
"dtype": "bfloat16",
"mse": 3.0464441351796268e-06,
"max_abs_err": 0.0244140625
},
{
"name": "model.layers.29.self_attn.v_proj",
"shape": [
1024,
2560
],
"dtype": "bfloat16",
"mse": 3.986216597695602e-06,
"max_abs_err": 0.025390625
},
{
"name": "model.layers.29.self_attn.o_proj",
"shape": [
2560,
4096
],
"dtype": "bfloat16",
"mse": 3.771279807551764e-06,
"max_abs_err": 0.0234375
},
{
"name": "model.layers.29.mlp.gate_proj",
"shape": [
9728,
2560
],
"dtype": "bfloat16",
"mse": 4.233662821206963e-06,
"max_abs_err": 0.0458984375
},
{
"name": "model.layers.29.mlp.up_proj",
"shape": [
9728,
2560
],
"dtype": "bfloat16",
"mse": 4.413979695527814e-06,
"max_abs_err": 0.0390625
},
{
"name": "model.layers.29.mlp.down_proj",
"shape": [
2560,
9728
],
"dtype": "bfloat16",
"mse": 4.221594281261787e-06,
"max_abs_err": 0.041015625
},
{
"name": "model.layers.30.self_attn.q_proj",
"shape": [
4096,
2560
],
"dtype": "bfloat16",
"mse": 3.6692895264422987e-06,
"max_abs_err": 0.0419921875
},
{
"name": "model.layers.30.self_attn.k_proj",
"shape": [
1024,
2560
],
"dtype": "bfloat16",
"mse": 3.30666512127209e-06,
"max_abs_err": 0.02734375
},
{
"name": "model.layers.30.self_attn.v_proj",
"shape": [
1024,
2560
],
"dtype": "bfloat16",
"mse": 4.481903943087673e-06,
"max_abs_err": 0.02197265625
},
{
"name": "model.layers.30.self_attn.o_proj",
"shape": [
2560,
4096
],
"dtype": "bfloat16",
"mse": 3.707785481310566e-06,
"max_abs_err": 0.0283203125
},
{
"name": "model.layers.30.mlp.gate_proj",
"shape": [
9728,
2560
],
"dtype": "bfloat16",
"mse": 4.18360878029489e-06,
"max_abs_err": 0.052734375
},
{
"name": "model.layers.30.mlp.up_proj",
"shape": [
9728,
2560
],
"dtype": "bfloat16",
"mse": 4.51727828476578e-06,
"max_abs_err": 0.0283203125
},
{
"name": "model.layers.30.mlp.down_proj",
"shape": [
2560,
9728
],
"dtype": "bfloat16",
"mse": 4.328634076955495e-06,
"max_abs_err": 0.060546875
},
{
"name": "model.layers.31.self_attn.q_proj",
"shape": [
4096,
2560
],
"dtype": "bfloat16",
"mse": 3.827203272521729e-06,
"max_abs_err": 0.029296875
},
{
"name": "model.layers.31.self_attn.k_proj",
"shape": [
1024,
2560
],
"dtype": "bfloat16",
"mse": 3.0402077300095698e-06,
"max_abs_err": 0.02490234375
},
{
"name": "model.layers.31.self_attn.v_proj",
"shape": [
1024,
2560
],
"dtype": "bfloat16",
"mse": 4.832452304981416e-06,
"max_abs_err": 0.0244140625
},
{
"name": "model.layers.31.self_attn.o_proj",
"shape": [
2560,
4096
],
"dtype": "bfloat16",
"mse": 3.9282422221731395e-06,
"max_abs_err": 0.02685546875
},
{
"name": "model.layers.31.mlp.gate_proj",
"shape": [
9728,
2560
],
"dtype": "bfloat16",
"mse": 4.098501449334435e-06,
"max_abs_err": 0.0546875
},
{
"name": "model.layers.31.mlp.up_proj",
"shape": [
9728,
2560
],
"dtype": "bfloat16",
"mse": 4.62463685835246e-06,
"max_abs_err": 0.0341796875
},
{
"name": "model.layers.31.mlp.down_proj",
"shape": [
2560,
9728
],
"dtype": "bfloat16",
"mse": 4.503143372858176e-06,
"max_abs_err": 0.05859375
},
{
"name": "model.layers.32.self_attn.q_proj",
"shape": [
4096,
2560
],
"dtype": "bfloat16",
"mse": 3.6864644243905786e-06,
"max_abs_err": 0.0322265625
},
{
"name": "model.layers.32.self_attn.k_proj",
"shape": [
1024,
2560
],
"dtype": "bfloat16",
"mse": 3.0785006401856663e-06,
"max_abs_err": 0.0263671875
},
{
"name": "model.layers.32.self_attn.v_proj",
"shape": [
1024,
2560
],
"dtype": "bfloat16",
"mse": 4.8119072744157165e-06,
"max_abs_err": 0.02392578125
},
{
"name": "model.layers.32.self_attn.o_proj",
"shape": [
2560,
4096
],
"dtype": "bfloat16",
"mse": 3.91014918932342e-06,
"max_abs_err": 0.0380859375
},
{
"name": "model.layers.32.mlp.gate_proj",
"shape": [
9728,
2560
],
"dtype": "bfloat16",
"mse": 3.975767867814284e-06,
"max_abs_err": 0.041015625
},
{
"name": "model.layers.32.mlp.up_proj",
"shape": [
9728,
2560
],
"dtype": "bfloat16",
"mse": 4.649803486245219e-06,
"max_abs_err": 0.0283203125
},
{
"name": "model.layers.32.mlp.down_proj",
"shape": [
2560,
9728
],
"dtype": "bfloat16",
"mse": 4.511914994509425e-06,
"max_abs_err": 0.05859375
},
{
"name": "model.layers.33.self_attn.q_proj",
"shape": [
4096,
2560
],
"dtype": "bfloat16",
"mse": 3.734967322088778e-06,
"max_abs_err": 0.0263671875
},
{
"name": "model.layers.33.self_attn.k_proj",
"shape": [
1024,
2560
],
"dtype": "bfloat16",
"mse": 2.4853611648723017e-06,
"max_abs_err": 0.01953125
},
{
"name": "model.layers.33.self_attn.v_proj",
"shape": [
1024,
2560
],
"dtype": "bfloat16",
"mse": 5.483281711349264e-06,
"max_abs_err": 0.01904296875
},
{
"name": "model.layers.33.self_attn.o_proj",
"shape": [
2560,
4096
],
"dtype": "bfloat16",
"mse": 4.2412293623783626e-06,
"max_abs_err": 0.0390625
},
{
"name": "model.layers.33.mlp.gate_proj",
"shape": [
9728,
2560
],
"dtype": "bfloat16",
"mse": 3.897335318470141e-06,
"max_abs_err": 0.0576171875
},
{
"name": "model.layers.33.mlp.up_proj",
"shape": [
9728,
2560
],
"dtype": "bfloat16",
"mse": 4.685903604695341e-06,
"max_abs_err": 0.029296875
},
{
"name": "model.layers.33.mlp.down_proj",
"shape": [
2560,
9728
],
"dtype": "bfloat16",
"mse": 4.405589606903959e-06,
"max_abs_err": 0.0625
},
{
"name": "model.layers.34.self_attn.q_proj",
"shape": [
4096,
2560
],
"dtype": "bfloat16",
"mse": 3.474927552815643e-06,
"max_abs_err": 0.02978515625
},
{
"name": "model.layers.34.self_attn.k_proj",
"shape": [
1024,
2560
],
"dtype": "bfloat16",
"mse": 2.882841954487958e-06,
"max_abs_err": 0.0263671875
},
{
"name": "model.layers.34.self_attn.v_proj",
"shape": [
1024,
2560
],
"dtype": "bfloat16",
"mse": 5.479394076246535e-06,
"max_abs_err": 0.0263671875
},
{
"name": "model.layers.34.self_attn.o_proj",
"shape": [
2560,
4096
],
"dtype": "bfloat16",
"mse": 4.039952273160452e-06,
"max_abs_err": 0.0458984375
},
{
"name": "model.layers.34.mlp.gate_proj",
"shape": [
9728,
2560
],
"dtype": "bfloat16",
"mse": 4.041967713419581e-06,
"max_abs_err": 0.04296875
},
{
"name": "model.layers.34.mlp.up_proj",
"shape": [
9728,
2560
],
"dtype": "bfloat16",
"mse": 4.558777163765626e-06,
"max_abs_err": 0.0458984375
},
{
"name": "model.layers.34.mlp.down_proj",
"shape": [
2560,
9728
],
"dtype": "bfloat16",
"mse": 3.925315922970185e-06,
"max_abs_err": 0.0703125
},
{
"name": "model.layers.35.self_attn.q_proj",
"shape": [
4096,
2560
],
"dtype": "bfloat16",
"mse": 3.571818524505943e-06,
"max_abs_err": 0.0380859375
},
{
"name": "model.layers.35.self_attn.k_proj",
"shape": [
1024,
2560
],
"dtype": "bfloat16",
"mse": 3.282528041381738e-06,
"max_abs_err": 0.0244140625
},
{
"name": "model.layers.35.self_attn.v_proj",
"shape": [
1024,
2560
],
"dtype": "bfloat16",
"mse": 4.853251539316261e-06,
"max_abs_err": 0.02099609375
},
{
"name": "model.layers.35.self_attn.o_proj",
"shape": [
2560,
4096
],
"dtype": "bfloat16",
"mse": 3.785845137826982e-06,
"max_abs_err": 0.0703125
},
{
"name": "model.layers.35.mlp.gate_proj",
"shape": [
9728,
2560
],
"dtype": "bfloat16",
"mse": 4.4715557123709004e-06,
"max_abs_err": 0.06640625
},
{
"name": "model.layers.35.mlp.up_proj",
"shape": [
9728,
2560
],
"dtype": "bfloat16",
"mse": 4.689582056016661e-06,
"max_abs_err": 0.048828125
},
{
"name": "model.layers.35.mlp.down_proj",
"shape": [
2560,
9728
],
"dtype": "bfloat16",
"mse": 3.554017666829168e-06,
"max_abs_err": 0.05859375
}
],
"source_model": "Qwen/Qwen3-4B-Thinking-2507",
"dtype": "bf16",
"target_regex": "(?:q_proj|k_proj|v_proj|o_proj|gate_proj|up_proj|down_proj)$",
"ignore_regex": [],
"include_lm_head": false,
"note": "Dense fake quant: quantized weights are dequantized and saved as normal HF tensors. nvfp4/mxfp4 are HQQ-style optimized FP4 fake-quant variants, not packed runtime checkpoints."
}