1c7d0a13f3
Model: meteorain/Qwen__Qwen3-4B-Thinking-2507-RTN-fp4-e3m0-g128-faked-bf16 Source: Original Platform
2537 lines
52 KiB
JSON
2537 lines
52 KiB
JSON
{
|
|
"format": "fp4-e3m0",
|
|
"codebook": "fp4-e3m0",
|
|
"block_size": 128,
|
|
"num_modules": 252,
|
|
"num_weights": 3633315840,
|
|
"elapsed_sec": 90.372,
|
|
"quantized_modules": [
|
|
{
|
|
"name": "model.layers.0.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.897217589430511e-05,
|
|
"max_abs_err": 0.140625
|
|
},
|
|
{
|
|
"name": "model.layers.0.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 2.1087433196953498e-05,
|
|
"max_abs_err": 0.0654296875
|
|
},
|
|
{
|
|
"name": "model.layers.0.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.856678500189446e-05,
|
|
"max_abs_err": 0.0302734375
|
|
},
|
|
{
|
|
"name": "model.layers.0.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.6585341654717922e-05,
|
|
"max_abs_err": 0.12109375
|
|
},
|
|
{
|
|
"name": "model.layers.0.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.967309981409926e-05,
|
|
"max_abs_err": 0.10546875
|
|
},
|
|
{
|
|
"name": "model.layers.0.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.7165502868010662e-05,
|
|
"max_abs_err": 0.060546875
|
|
},
|
|
{
|
|
"name": "model.layers.0.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.893728222057689e-05,
|
|
"max_abs_err": 0.1328125
|
|
},
|
|
{
|
|
"name": "model.layers.1.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.6433134078397416e-05,
|
|
"max_abs_err": 0.0859375
|
|
},
|
|
{
|
|
"name": "model.layers.1.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.7814132661442272e-05,
|
|
"max_abs_err": 0.05078125
|
|
},
|
|
{
|
|
"name": "model.layers.1.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.9104852981399745e-05,
|
|
"max_abs_err": 0.03369140625
|
|
},
|
|
{
|
|
"name": "model.layers.1.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.7016347555909306e-05,
|
|
"max_abs_err": 0.0986328125
|
|
},
|
|
{
|
|
"name": "model.layers.1.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.8485632608644664e-05,
|
|
"max_abs_err": 0.0595703125
|
|
},
|
|
{
|
|
"name": "model.layers.1.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 8.42818644741783e-06,
|
|
"max_abs_err": 0.05908203125
|
|
},
|
|
{
|
|
"name": "model.layers.1.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 9.550471986585762e-06,
|
|
"max_abs_err": 0.16015625
|
|
},
|
|
{
|
|
"name": "model.layers.2.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.7593085431144573e-05,
|
|
"max_abs_err": 0.0830078125
|
|
},
|
|
{
|
|
"name": "model.layers.2.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.8786407963489182e-05,
|
|
"max_abs_err": 0.072265625
|
|
},
|
|
{
|
|
"name": "model.layers.2.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 2.0540112018352374e-05,
|
|
"max_abs_err": 0.03173828125
|
|
},
|
|
{
|
|
"name": "model.layers.2.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.782494473445695e-05,
|
|
"max_abs_err": 0.076171875
|
|
},
|
|
{
|
|
"name": "model.layers.2.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.2476847587095108e-05,
|
|
"max_abs_err": 0.0634765625
|
|
},
|
|
{
|
|
"name": "model.layers.2.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.0454131370352115e-05,
|
|
"max_abs_err": 0.0947265625
|
|
},
|
|
{
|
|
"name": "model.layers.2.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 9.158644388662651e-06,
|
|
"max_abs_err": 0.1083984375
|
|
},
|
|
{
|
|
"name": "model.layers.3.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.7670225133770145e-05,
|
|
"max_abs_err": 0.078125
|
|
},
|
|
{
|
|
"name": "model.layers.3.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.767483445291873e-05,
|
|
"max_abs_err": 0.048828125
|
|
},
|
|
{
|
|
"name": "model.layers.3.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 2.0774165022885427e-05,
|
|
"max_abs_err": 0.03173828125
|
|
},
|
|
{
|
|
"name": "model.layers.3.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.8056784028885886e-05,
|
|
"max_abs_err": 0.087890625
|
|
},
|
|
{
|
|
"name": "model.layers.3.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 2.2378475478035398e-05,
|
|
"max_abs_err": 0.10546875
|
|
},
|
|
{
|
|
"name": "model.layers.3.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.2042939488310367e-05,
|
|
"max_abs_err": 0.08203125
|
|
},
|
|
{
|
|
"name": "model.layers.3.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.2676287951762788e-05,
|
|
"max_abs_err": 0.130859375
|
|
},
|
|
{
|
|
"name": "model.layers.4.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.8327231373405084e-05,
|
|
"max_abs_err": 0.0712890625
|
|
},
|
|
{
|
|
"name": "model.layers.4.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.8569908206700347e-05,
|
|
"max_abs_err": 0.09765625
|
|
},
|
|
{
|
|
"name": "model.layers.4.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 2.1206917153904215e-05,
|
|
"max_abs_err": 0.0390625
|
|
},
|
|
{
|
|
"name": "model.layers.4.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.890580097096972e-05,
|
|
"max_abs_err": 0.1025390625
|
|
},
|
|
{
|
|
"name": "model.layers.4.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 3.4760865673888475e-05,
|
|
"max_abs_err": 0.06640625
|
|
},
|
|
{
|
|
"name": "model.layers.4.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.4635515071859118e-05,
|
|
"max_abs_err": 0.0634765625
|
|
},
|
|
{
|
|
"name": "model.layers.4.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.481798699387582e-05,
|
|
"max_abs_err": 0.11328125
|
|
},
|
|
{
|
|
"name": "model.layers.5.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.8172617274103686e-05,
|
|
"max_abs_err": 0.072265625
|
|
},
|
|
{
|
|
"name": "model.layers.5.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.8051017832476646e-05,
|
|
"max_abs_err": 0.0615234375
|
|
},
|
|
{
|
|
"name": "model.layers.5.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 2.0899517039651982e-05,
|
|
"max_abs_err": 0.03271484375
|
|
},
|
|
{
|
|
"name": "model.layers.5.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.7584807210369036e-05,
|
|
"max_abs_err": 0.1025390625
|
|
},
|
|
{
|
|
"name": "model.layers.5.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 3.014916364918463e-05,
|
|
"max_abs_err": 0.091796875
|
|
},
|
|
{
|
|
"name": "model.layers.5.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.6458203390357085e-05,
|
|
"max_abs_err": 0.09375
|
|
},
|
|
{
|
|
"name": "model.layers.5.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.5913947208900936e-05,
|
|
"max_abs_err": 0.125
|
|
},
|
|
{
|
|
"name": "model.layers.6.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.917057488753926e-05,
|
|
"max_abs_err": 0.07421875
|
|
},
|
|
{
|
|
"name": "model.layers.6.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.814850838854909e-05,
|
|
"max_abs_err": 0.0712890625
|
|
},
|
|
{
|
|
"name": "model.layers.6.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 2.0801284335902892e-05,
|
|
"max_abs_err": 0.052734375
|
|
},
|
|
{
|
|
"name": "model.layers.6.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.8788390661939047e-05,
|
|
"max_abs_err": 0.12890625
|
|
},
|
|
{
|
|
"name": "model.layers.6.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 2.6911095119430684e-05,
|
|
"max_abs_err": 0.078125
|
|
},
|
|
{
|
|
"name": "model.layers.6.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.7407210179953836e-05,
|
|
"max_abs_err": 0.1181640625
|
|
},
|
|
{
|
|
"name": "model.layers.6.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.7163194570457563e-05,
|
|
"max_abs_err": 0.12109375
|
|
},
|
|
{
|
|
"name": "model.layers.7.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.813646304071881e-05,
|
|
"max_abs_err": 0.0732421875
|
|
},
|
|
{
|
|
"name": "model.layers.7.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.816048643377144e-05,
|
|
"max_abs_err": 0.0595703125
|
|
},
|
|
{
|
|
"name": "model.layers.7.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 2.3960426915436983e-05,
|
|
"max_abs_err": 0.03564453125
|
|
},
|
|
{
|
|
"name": "model.layers.7.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.8296555936103687e-05,
|
|
"max_abs_err": 0.083984375
|
|
},
|
|
{
|
|
"name": "model.layers.7.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 2.630111157486681e-05,
|
|
"max_abs_err": 0.083984375
|
|
},
|
|
{
|
|
"name": "model.layers.7.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.7655831470619887e-05,
|
|
"max_abs_err": 0.109375
|
|
},
|
|
{
|
|
"name": "model.layers.7.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.7477472283644602e-05,
|
|
"max_abs_err": 0.119140625
|
|
},
|
|
{
|
|
"name": "model.layers.8.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.8369575627730228e-05,
|
|
"max_abs_err": 0.0703125
|
|
},
|
|
{
|
|
"name": "model.layers.8.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.883392542367801e-05,
|
|
"max_abs_err": 0.0615234375
|
|
},
|
|
{
|
|
"name": "model.layers.8.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 2.390061126789078e-05,
|
|
"max_abs_err": 0.03173828125
|
|
},
|
|
{
|
|
"name": "model.layers.8.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.9156173948431388e-05,
|
|
"max_abs_err": 0.083984375
|
|
},
|
|
{
|
|
"name": "model.layers.8.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 2.2891217668075114e-05,
|
|
"max_abs_err": 0.0869140625
|
|
},
|
|
{
|
|
"name": "model.layers.8.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.848257852543611e-05,
|
|
"max_abs_err": 0.0947265625
|
|
},
|
|
{
|
|
"name": "model.layers.8.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.847717066993937e-05,
|
|
"max_abs_err": 0.1796875
|
|
},
|
|
{
|
|
"name": "model.layers.9.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.745830559229944e-05,
|
|
"max_abs_err": 0.126953125
|
|
},
|
|
{
|
|
"name": "model.layers.9.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.739564504532609e-05,
|
|
"max_abs_err": 0.12890625
|
|
},
|
|
{
|
|
"name": "model.layers.9.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 2.4338844013982452e-05,
|
|
"max_abs_err": 0.037109375
|
|
},
|
|
{
|
|
"name": "model.layers.9.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.88392041309271e-05,
|
|
"max_abs_err": 0.0908203125
|
|
},
|
|
{
|
|
"name": "model.layers.9.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 2.6769792384584434e-05,
|
|
"max_abs_err": 0.115234375
|
|
},
|
|
{
|
|
"name": "model.layers.9.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.6967167539405636e-05,
|
|
"max_abs_err": 0.087890625
|
|
},
|
|
{
|
|
"name": "model.layers.9.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.654093648539856e-05,
|
|
"max_abs_err": 0.125
|
|
},
|
|
{
|
|
"name": "model.layers.10.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.862073622760363e-05,
|
|
"max_abs_err": 0.0673828125
|
|
},
|
|
{
|
|
"name": "model.layers.10.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.8512468159315176e-05,
|
|
"max_abs_err": 0.0556640625
|
|
},
|
|
{
|
|
"name": "model.layers.10.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 2.381021658948157e-05,
|
|
"max_abs_err": 0.041015625
|
|
},
|
|
{
|
|
"name": "model.layers.10.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.953605533344671e-05,
|
|
"max_abs_err": 0.130859375
|
|
},
|
|
{
|
|
"name": "model.layers.10.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 2.5613875550334342e-05,
|
|
"max_abs_err": 0.119140625
|
|
},
|
|
{
|
|
"name": "model.layers.10.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.7652033420745283e-05,
|
|
"max_abs_err": 0.091796875
|
|
},
|
|
{
|
|
"name": "model.layers.10.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.7395293980371207e-05,
|
|
"max_abs_err": 0.15625
|
|
},
|
|
{
|
|
"name": "model.layers.11.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.852634159149602e-05,
|
|
"max_abs_err": 0.0634765625
|
|
},
|
|
{
|
|
"name": "model.layers.11.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.912353400257416e-05,
|
|
"max_abs_err": 0.0849609375
|
|
},
|
|
{
|
|
"name": "model.layers.11.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 2.4355953428312205e-05,
|
|
"max_abs_err": 0.0322265625
|
|
},
|
|
{
|
|
"name": "model.layers.11.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.9115435861749575e-05,
|
|
"max_abs_err": 0.1171875
|
|
},
|
|
{
|
|
"name": "model.layers.11.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 2.3986754968063906e-05,
|
|
"max_abs_err": 0.1171875
|
|
},
|
|
{
|
|
"name": "model.layers.11.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.8703658497543074e-05,
|
|
"max_abs_err": 0.0849609375
|
|
},
|
|
{
|
|
"name": "model.layers.11.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.8214506781077944e-05,
|
|
"max_abs_err": 0.1796875
|
|
},
|
|
{
|
|
"name": "model.layers.12.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.8887836631620303e-05,
|
|
"max_abs_err": 0.072265625
|
|
},
|
|
{
|
|
"name": "model.layers.12.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.8410206394037232e-05,
|
|
"max_abs_err": 0.1015625
|
|
},
|
|
{
|
|
"name": "model.layers.12.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 2.3901327949715778e-05,
|
|
"max_abs_err": 0.04296875
|
|
},
|
|
{
|
|
"name": "model.layers.12.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.8733411707216874e-05,
|
|
"max_abs_err": 0.177734375
|
|
},
|
|
{
|
|
"name": "model.layers.12.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 2.274689541081898e-05,
|
|
"max_abs_err": 0.126953125
|
|
},
|
|
{
|
|
"name": "model.layers.12.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.9478269678074867e-05,
|
|
"max_abs_err": 0.083984375
|
|
},
|
|
{
|
|
"name": "model.layers.12.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.8614593500387855e-05,
|
|
"max_abs_err": 0.2109375
|
|
},
|
|
{
|
|
"name": "model.layers.13.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.8668502889340743e-05,
|
|
"max_abs_err": 0.08203125
|
|
},
|
|
{
|
|
"name": "model.layers.13.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.8202708815806545e-05,
|
|
"max_abs_err": 0.09765625
|
|
},
|
|
{
|
|
"name": "model.layers.13.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 2.296371712873224e-05,
|
|
"max_abs_err": 0.04833984375
|
|
},
|
|
{
|
|
"name": "model.layers.13.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.7901085811899975e-05,
|
|
"max_abs_err": 0.103515625
|
|
},
|
|
{
|
|
"name": "model.layers.13.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 2.1178255337872542e-05,
|
|
"max_abs_err": 0.109375
|
|
},
|
|
{
|
|
"name": "model.layers.13.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.9979464923380874e-05,
|
|
"max_abs_err": 0.125
|
|
},
|
|
{
|
|
"name": "model.layers.13.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.8744292901828885e-05,
|
|
"max_abs_err": 0.18359375
|
|
},
|
|
{
|
|
"name": "model.layers.14.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.87323985301191e-05,
|
|
"max_abs_err": 0.0576171875
|
|
},
|
|
{
|
|
"name": "model.layers.14.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.8547962099546567e-05,
|
|
"max_abs_err": 0.060546875
|
|
},
|
|
{
|
|
"name": "model.layers.14.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 2.1698268028558232e-05,
|
|
"max_abs_err": 0.03173828125
|
|
},
|
|
{
|
|
"name": "model.layers.14.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.7402418961864896e-05,
|
|
"max_abs_err": 0.123046875
|
|
},
|
|
{
|
|
"name": "model.layers.14.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 2.05986052606022e-05,
|
|
"max_abs_err": 0.12109375
|
|
},
|
|
{
|
|
"name": "model.layers.14.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.9417417206568643e-05,
|
|
"max_abs_err": 0.076171875
|
|
},
|
|
{
|
|
"name": "model.layers.14.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.8333397747483104e-05,
|
|
"max_abs_err": 0.1328125
|
|
},
|
|
{
|
|
"name": "model.layers.15.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.917348890856374e-05,
|
|
"max_abs_err": 0.0791015625
|
|
},
|
|
{
|
|
"name": "model.layers.15.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.8372924387222156e-05,
|
|
"max_abs_err": 0.0751953125
|
|
},
|
|
{
|
|
"name": "model.layers.15.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 2.1302214008755982e-05,
|
|
"max_abs_err": 0.04296875
|
|
},
|
|
{
|
|
"name": "model.layers.15.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.6400652384618297e-05,
|
|
"max_abs_err": 0.078125
|
|
},
|
|
{
|
|
"name": "model.layers.15.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.9181821699021384e-05,
|
|
"max_abs_err": 0.1484375
|
|
},
|
|
{
|
|
"name": "model.layers.15.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.898018308565952e-05,
|
|
"max_abs_err": 0.080078125
|
|
},
|
|
{
|
|
"name": "model.layers.15.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.798509583750274e-05,
|
|
"max_abs_err": 0.119140625
|
|
},
|
|
{
|
|
"name": "model.layers.16.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.808450724638533e-05,
|
|
"max_abs_err": 0.0615234375
|
|
},
|
|
{
|
|
"name": "model.layers.16.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.7200358342961408e-05,
|
|
"max_abs_err": 0.0537109375
|
|
},
|
|
{
|
|
"name": "model.layers.16.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 2.1334544726414606e-05,
|
|
"max_abs_err": 0.03564453125
|
|
},
|
|
{
|
|
"name": "model.layers.16.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.7345528249279596e-05,
|
|
"max_abs_err": 0.15234375
|
|
},
|
|
{
|
|
"name": "model.layers.16.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 2.014556412177626e-05,
|
|
"max_abs_err": 0.115234375
|
|
},
|
|
{
|
|
"name": "model.layers.16.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.8714488760451786e-05,
|
|
"max_abs_err": 0.193359375
|
|
},
|
|
{
|
|
"name": "model.layers.16.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.7413001842214726e-05,
|
|
"max_abs_err": 0.234375
|
|
},
|
|
{
|
|
"name": "model.layers.17.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.90222890523728e-05,
|
|
"max_abs_err": 0.064453125
|
|
},
|
|
{
|
|
"name": "model.layers.17.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.623477328394074e-05,
|
|
"max_abs_err": 0.0576171875
|
|
},
|
|
{
|
|
"name": "model.layers.17.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 2.0298617528169416e-05,
|
|
"max_abs_err": 0.03955078125
|
|
},
|
|
{
|
|
"name": "model.layers.17.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.676509782555513e-05,
|
|
"max_abs_err": 0.107421875
|
|
},
|
|
{
|
|
"name": "model.layers.17.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.9574454199755564e-05,
|
|
"max_abs_err": 0.1015625
|
|
},
|
|
{
|
|
"name": "model.layers.17.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.8874190573114902e-05,
|
|
"max_abs_err": 0.07421875
|
|
},
|
|
{
|
|
"name": "model.layers.17.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.7525198927614838e-05,
|
|
"max_abs_err": 0.171875
|
|
},
|
|
{
|
|
"name": "model.layers.18.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.7764516087481752e-05,
|
|
"max_abs_err": 0.076171875
|
|
},
|
|
{
|
|
"name": "model.layers.18.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.6660807887092233e-05,
|
|
"max_abs_err": 0.056640625
|
|
},
|
|
{
|
|
"name": "model.layers.18.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 2.1131247194716707e-05,
|
|
"max_abs_err": 0.03662109375
|
|
},
|
|
{
|
|
"name": "model.layers.18.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.6800162484287284e-05,
|
|
"max_abs_err": 0.095703125
|
|
},
|
|
{
|
|
"name": "model.layers.18.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.9088591216132045e-05,
|
|
"max_abs_err": 0.11328125
|
|
},
|
|
{
|
|
"name": "model.layers.18.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.883263212221209e-05,
|
|
"max_abs_err": 0.07421875
|
|
},
|
|
{
|
|
"name": "model.layers.18.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.7469137674197555e-05,
|
|
"max_abs_err": 0.1123046875
|
|
},
|
|
{
|
|
"name": "model.layers.19.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.8330898456042632e-05,
|
|
"max_abs_err": 0.0791015625
|
|
},
|
|
{
|
|
"name": "model.layers.19.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.5523120964644477e-05,
|
|
"max_abs_err": 0.09765625
|
|
},
|
|
{
|
|
"name": "model.layers.19.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 2.0182164007565007e-05,
|
|
"max_abs_err": 0.0361328125
|
|
},
|
|
{
|
|
"name": "model.layers.19.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.7039246813510545e-05,
|
|
"max_abs_err": 0.1328125
|
|
},
|
|
{
|
|
"name": "model.layers.19.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.8843915313482285e-05,
|
|
"max_abs_err": 0.11328125
|
|
},
|
|
{
|
|
"name": "model.layers.19.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.893723128887359e-05,
|
|
"max_abs_err": 0.087890625
|
|
},
|
|
{
|
|
"name": "model.layers.19.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.744461951602716e-05,
|
|
"max_abs_err": 0.1640625
|
|
},
|
|
{
|
|
"name": "model.layers.20.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.9654400603030808e-05,
|
|
"max_abs_err": 0.0751953125
|
|
},
|
|
{
|
|
"name": "model.layers.20.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.5257326595019549e-05,
|
|
"max_abs_err": 0.052734375
|
|
},
|
|
{
|
|
"name": "model.layers.20.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 2.007134389714338e-05,
|
|
"max_abs_err": 0.03857421875
|
|
},
|
|
{
|
|
"name": "model.layers.20.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.6290588973788545e-05,
|
|
"max_abs_err": 0.103515625
|
|
},
|
|
{
|
|
"name": "model.layers.20.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.8505581465433352e-05,
|
|
"max_abs_err": 0.12109375
|
|
},
|
|
{
|
|
"name": "model.layers.20.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.904497912619263e-05,
|
|
"max_abs_err": 0.0791015625
|
|
},
|
|
{
|
|
"name": "model.layers.20.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.776057069946546e-05,
|
|
"max_abs_err": 0.091796875
|
|
},
|
|
{
|
|
"name": "model.layers.21.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.8810304027283564e-05,
|
|
"max_abs_err": 0.07421875
|
|
},
|
|
{
|
|
"name": "model.layers.21.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.512696053396212e-05,
|
|
"max_abs_err": 0.04443359375
|
|
},
|
|
{
|
|
"name": "model.layers.21.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.9430408428888768e-05,
|
|
"max_abs_err": 0.056640625
|
|
},
|
|
{
|
|
"name": "model.layers.21.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.5767072909511626e-05,
|
|
"max_abs_err": 0.0595703125
|
|
},
|
|
{
|
|
"name": "model.layers.21.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.786220855137799e-05,
|
|
"max_abs_err": 0.09765625
|
|
},
|
|
{
|
|
"name": "model.layers.21.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.89486763702007e-05,
|
|
"max_abs_err": 0.091796875
|
|
},
|
|
{
|
|
"name": "model.layers.21.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.8013570297625847e-05,
|
|
"max_abs_err": 0.13671875
|
|
},
|
|
{
|
|
"name": "model.layers.22.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.82994353963295e-05,
|
|
"max_abs_err": 0.07421875
|
|
},
|
|
{
|
|
"name": "model.layers.22.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.4011558960191905e-05,
|
|
"max_abs_err": 0.162109375
|
|
},
|
|
{
|
|
"name": "model.layers.22.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.9287337636342272e-05,
|
|
"max_abs_err": 0.0478515625
|
|
},
|
|
{
|
|
"name": "model.layers.22.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.6448586393380538e-05,
|
|
"max_abs_err": 0.1162109375
|
|
},
|
|
{
|
|
"name": "model.layers.22.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.9109505956294015e-05,
|
|
"max_abs_err": 0.1083984375
|
|
},
|
|
{
|
|
"name": "model.layers.22.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.9826549760182388e-05,
|
|
"max_abs_err": 0.08984375
|
|
},
|
|
{
|
|
"name": "model.layers.22.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.814990901038982e-05,
|
|
"max_abs_err": 0.13671875
|
|
},
|
|
{
|
|
"name": "model.layers.23.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 2.0717987354146317e-05,
|
|
"max_abs_err": 0.083984375
|
|
},
|
|
{
|
|
"name": "model.layers.23.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.5667135812691413e-05,
|
|
"max_abs_err": 0.05908203125
|
|
},
|
|
{
|
|
"name": "model.layers.23.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 2.003620465984568e-05,
|
|
"max_abs_err": 0.0478515625
|
|
},
|
|
{
|
|
"name": "model.layers.23.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.769474329194054e-05,
|
|
"max_abs_err": 0.0703125
|
|
},
|
|
{
|
|
"name": "model.layers.23.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 2.0570023480104282e-05,
|
|
"max_abs_err": 0.119140625
|
|
},
|
|
{
|
|
"name": "model.layers.23.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.9975439499830827e-05,
|
|
"max_abs_err": 0.0732421875
|
|
},
|
|
{
|
|
"name": "model.layers.23.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.844682810769882e-05,
|
|
"max_abs_err": 0.130859375
|
|
},
|
|
{
|
|
"name": "model.layers.24.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 2.1032054064562544e-05,
|
|
"max_abs_err": 0.0751953125
|
|
},
|
|
{
|
|
"name": "model.layers.24.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.563769365020562e-05,
|
|
"max_abs_err": 0.05859375
|
|
},
|
|
{
|
|
"name": "model.layers.24.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.9544100723578595e-05,
|
|
"max_abs_err": 0.060546875
|
|
},
|
|
{
|
|
"name": "model.layers.24.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.8287808416062035e-05,
|
|
"max_abs_err": 0.083984375
|
|
},
|
|
{
|
|
"name": "model.layers.24.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 2.1402449419838376e-05,
|
|
"max_abs_err": 0.09765625
|
|
},
|
|
{
|
|
"name": "model.layers.24.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 2.0131476048845798e-05,
|
|
"max_abs_err": 0.07421875
|
|
},
|
|
{
|
|
"name": "model.layers.24.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.8831229681381956e-05,
|
|
"max_abs_err": 0.138671875
|
|
},
|
|
{
|
|
"name": "model.layers.25.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.8817005184246227e-05,
|
|
"max_abs_err": 0.0693359375
|
|
},
|
|
{
|
|
"name": "model.layers.25.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.677433101576753e-05,
|
|
"max_abs_err": 0.0595703125
|
|
},
|
|
{
|
|
"name": "model.layers.25.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 2.144874815712683e-05,
|
|
"max_abs_err": 0.04296875
|
|
},
|
|
{
|
|
"name": "model.layers.25.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.7827735064201988e-05,
|
|
"max_abs_err": 0.04638671875
|
|
},
|
|
{
|
|
"name": "model.layers.25.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 2.189824590459466e-05,
|
|
"max_abs_err": 0.09765625
|
|
},
|
|
{
|
|
"name": "model.layers.25.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 2.0146335373283364e-05,
|
|
"max_abs_err": 0.087890625
|
|
},
|
|
{
|
|
"name": "model.layers.25.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.903276461234782e-05,
|
|
"max_abs_err": 0.109375
|
|
},
|
|
{
|
|
"name": "model.layers.26.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.949474426510278e-05,
|
|
"max_abs_err": 0.0634765625
|
|
},
|
|
{
|
|
"name": "model.layers.26.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.6310981663991697e-05,
|
|
"max_abs_err": 0.04736328125
|
|
},
|
|
{
|
|
"name": "model.layers.26.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 2.1148482119315304e-05,
|
|
"max_abs_err": 0.041015625
|
|
},
|
|
{
|
|
"name": "model.layers.26.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.8630051272339188e-05,
|
|
"max_abs_err": 0.0546875
|
|
},
|
|
{
|
|
"name": "model.layers.26.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 2.198565925937146e-05,
|
|
"max_abs_err": 0.083984375
|
|
},
|
|
{
|
|
"name": "model.layers.26.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 2.06345666811103e-05,
|
|
"max_abs_err": 0.107421875
|
|
},
|
|
{
|
|
"name": "model.layers.26.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.9564675312722102e-05,
|
|
"max_abs_err": 0.091796875
|
|
},
|
|
{
|
|
"name": "model.layers.27.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.9162247554049827e-05,
|
|
"max_abs_err": 0.0625
|
|
},
|
|
{
|
|
"name": "model.layers.27.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.50887326526572e-05,
|
|
"max_abs_err": 0.03759765625
|
|
},
|
|
{
|
|
"name": "model.layers.27.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.99087880901061e-05,
|
|
"max_abs_err": 0.04638671875
|
|
},
|
|
{
|
|
"name": "model.layers.27.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.778020669007674e-05,
|
|
"max_abs_err": 0.0732421875
|
|
},
|
|
{
|
|
"name": "model.layers.27.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 2.1764639313914813e-05,
|
|
"max_abs_err": 0.091796875
|
|
},
|
|
{
|
|
"name": "model.layers.27.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 2.1136684154043905e-05,
|
|
"max_abs_err": 0.10546875
|
|
},
|
|
{
|
|
"name": "model.layers.27.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 2.0022198441438377e-05,
|
|
"max_abs_err": 0.0966796875
|
|
},
|
|
{
|
|
"name": "model.layers.28.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.9998567950096913e-05,
|
|
"max_abs_err": 0.0615234375
|
|
},
|
|
{
|
|
"name": "model.layers.28.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.60989638970932e-05,
|
|
"max_abs_err": 0.048828125
|
|
},
|
|
{
|
|
"name": "model.layers.28.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 2.212423351011239e-05,
|
|
"max_abs_err": 0.0400390625
|
|
},
|
|
{
|
|
"name": "model.layers.28.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.7685884813545272e-05,
|
|
"max_abs_err": 0.07421875
|
|
},
|
|
{
|
|
"name": "model.layers.28.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 2.1298999854479916e-05,
|
|
"max_abs_err": 0.09765625
|
|
},
|
|
{
|
|
"name": "model.layers.28.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 2.1588704839814454e-05,
|
|
"max_abs_err": 0.16015625
|
|
},
|
|
{
|
|
"name": "model.layers.28.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 2.0526933440123685e-05,
|
|
"max_abs_err": 0.1015625
|
|
},
|
|
{
|
|
"name": "model.layers.29.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 2.1917181584285572e-05,
|
|
"max_abs_err": 0.0751953125
|
|
},
|
|
{
|
|
"name": "model.layers.29.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.5322644685511477e-05,
|
|
"max_abs_err": 0.0615234375
|
|
},
|
|
{
|
|
"name": "model.layers.29.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 2.0084567950107157e-05,
|
|
"max_abs_err": 0.056640625
|
|
},
|
|
{
|
|
"name": "model.layers.29.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.9037179299630225e-05,
|
|
"max_abs_err": 0.0517578125
|
|
},
|
|
{
|
|
"name": "model.layers.29.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 2.135786780854687e-05,
|
|
"max_abs_err": 0.15234375
|
|
},
|
|
{
|
|
"name": "model.layers.29.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 2.2267939129960723e-05,
|
|
"max_abs_err": 0.072265625
|
|
},
|
|
{
|
|
"name": "model.layers.29.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 2.1271829609759152e-05,
|
|
"max_abs_err": 0.0830078125
|
|
},
|
|
{
|
|
"name": "model.layers.30.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.8452148651704192e-05,
|
|
"max_abs_err": 0.07421875
|
|
},
|
|
{
|
|
"name": "model.layers.30.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.6642570699332282e-05,
|
|
"max_abs_err": 0.0625
|
|
},
|
|
{
|
|
"name": "model.layers.30.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 2.257928645121865e-05,
|
|
"max_abs_err": 0.05322265625
|
|
},
|
|
{
|
|
"name": "model.layers.30.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.8730595911620185e-05,
|
|
"max_abs_err": 0.05224609375
|
|
},
|
|
{
|
|
"name": "model.layers.30.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 2.110728200932499e-05,
|
|
"max_abs_err": 0.142578125
|
|
},
|
|
{
|
|
"name": "model.layers.30.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 2.2815434931544587e-05,
|
|
"max_abs_err": 0.080078125
|
|
},
|
|
{
|
|
"name": "model.layers.30.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 2.18136228795629e-05,
|
|
"max_abs_err": 0.1015625
|
|
},
|
|
{
|
|
"name": "model.layers.31.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.9239483663113788e-05,
|
|
"max_abs_err": 0.06640625
|
|
},
|
|
{
|
|
"name": "model.layers.31.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.537404205009807e-05,
|
|
"max_abs_err": 0.0498046875
|
|
},
|
|
{
|
|
"name": "model.layers.31.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 2.4271848815260455e-05,
|
|
"max_abs_err": 0.044921875
|
|
},
|
|
{
|
|
"name": "model.layers.31.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.982345202122815e-05,
|
|
"max_abs_err": 0.08203125
|
|
},
|
|
{
|
|
"name": "model.layers.31.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 2.0661418602685444e-05,
|
|
"max_abs_err": 0.150390625
|
|
},
|
|
{
|
|
"name": "model.layers.31.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 2.3341304768109694e-05,
|
|
"max_abs_err": 0.0751953125
|
|
},
|
|
{
|
|
"name": "model.layers.31.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 2.2678856112179346e-05,
|
|
"max_abs_err": 0.0908203125
|
|
},
|
|
{
|
|
"name": "model.layers.32.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.855792652349919e-05,
|
|
"max_abs_err": 0.0595703125
|
|
},
|
|
{
|
|
"name": "model.layers.32.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.550954402773641e-05,
|
|
"max_abs_err": 0.052734375
|
|
},
|
|
{
|
|
"name": "model.layers.32.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 2.412836329312995e-05,
|
|
"max_abs_err": 0.05712890625
|
|
},
|
|
{
|
|
"name": "model.layers.32.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.973248436115682e-05,
|
|
"max_abs_err": 0.0986328125
|
|
},
|
|
{
|
|
"name": "model.layers.32.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 2.0038009097333997e-05,
|
|
"max_abs_err": 0.109375
|
|
},
|
|
{
|
|
"name": "model.layers.32.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 2.34664767049253e-05,
|
|
"max_abs_err": 0.0751953125
|
|
},
|
|
{
|
|
"name": "model.layers.32.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 2.2730804630555212e-05,
|
|
"max_abs_err": 0.134765625
|
|
},
|
|
{
|
|
"name": "model.layers.33.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.879085357359145e-05,
|
|
"max_abs_err": 0.091796875
|
|
},
|
|
{
|
|
"name": "model.layers.33.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.2517602954176255e-05,
|
|
"max_abs_err": 0.05224609375
|
|
},
|
|
{
|
|
"name": "model.layers.33.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 2.7557049179449677e-05,
|
|
"max_abs_err": 0.0439453125
|
|
},
|
|
{
|
|
"name": "model.layers.33.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 2.1428042600746267e-05,
|
|
"max_abs_err": 0.1484375
|
|
},
|
|
{
|
|
"name": "model.layers.33.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.9609004084486514e-05,
|
|
"max_abs_err": 0.13671875
|
|
},
|
|
{
|
|
"name": "model.layers.33.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 2.3617118131369352e-05,
|
|
"max_abs_err": 0.1083984375
|
|
},
|
|
{
|
|
"name": "model.layers.33.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 2.2203168555279262e-05,
|
|
"max_abs_err": 0.140625
|
|
},
|
|
{
|
|
"name": "model.layers.34.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.7506754375062883e-05,
|
|
"max_abs_err": 0.0634765625
|
|
},
|
|
{
|
|
"name": "model.layers.34.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.4562296200892888e-05,
|
|
"max_abs_err": 0.0712890625
|
|
},
|
|
{
|
|
"name": "model.layers.34.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 2.7584252165979706e-05,
|
|
"max_abs_err": 0.05078125
|
|
},
|
|
{
|
|
"name": "model.layers.34.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 2.040964136540424e-05,
|
|
"max_abs_err": 0.087890625
|
|
},
|
|
{
|
|
"name": "model.layers.34.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 2.036994555965066e-05,
|
|
"max_abs_err": 0.115234375
|
|
},
|
|
{
|
|
"name": "model.layers.34.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 2.2992129743215628e-05,
|
|
"max_abs_err": 0.08984375
|
|
},
|
|
{
|
|
"name": "model.layers.34.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.9736082322197035e-05,
|
|
"max_abs_err": 0.1484375
|
|
},
|
|
{
|
|
"name": "model.layers.35.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.801333564799279e-05,
|
|
"max_abs_err": 0.068359375
|
|
},
|
|
{
|
|
"name": "model.layers.35.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.6515648894710466e-05,
|
|
"max_abs_err": 0.05322265625
|
|
},
|
|
{
|
|
"name": "model.layers.35.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 2.445553764118813e-05,
|
|
"max_abs_err": 0.04296875
|
|
},
|
|
{
|
|
"name": "model.layers.35.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.9118611817248166e-05,
|
|
"max_abs_err": 0.12890625
|
|
},
|
|
{
|
|
"name": "model.layers.35.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 2.2545549654751085e-05,
|
|
"max_abs_err": 0.15625
|
|
},
|
|
{
|
|
"name": "model.layers.35.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 2.3660330043639988e-05,
|
|
"max_abs_err": 0.103515625
|
|
},
|
|
{
|
|
"name": "model.layers.35.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.779453486960847e-05,
|
|
"max_abs_err": 0.1171875
|
|
}
|
|
],
|
|
"source_model": "Qwen/Qwen3-4B-Thinking-2507",
|
|
"dtype": "bf16",
|
|
"target_regex": ".*",
|
|
"ignore_regex": [],
|
|
"include_lm_head": false,
|
|
"note": "Dense fake quant: rounded weights are saved as normal HF tensors."
|
|
}
|