Model: meteorain/Qwen__Qwen3-4B-Thinking-2507-RTN-nf4-g64-faked-bf16 Source: Original Platform
2537 lines
52 KiB
JSON
2537 lines
52 KiB
JSON
{
|
|
"format": "nf4",
|
|
"codebook": "nf4",
|
|
"block_size": 64,
|
|
"num_modules": 252,
|
|
"num_weights": 3633315840,
|
|
"elapsed_sec": 92.035,
|
|
"quantized_modules": [
|
|
{
|
|
"name": "model.layers.0.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.432776677276706e-06,
|
|
"max_abs_err": 0.041015625
|
|
},
|
|
{
|
|
"name": "model.layers.0.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.882299435848836e-06,
|
|
"max_abs_err": 0.0302734375
|
|
},
|
|
{
|
|
"name": "model.layers.0.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.3304157770762686e-06,
|
|
"max_abs_err": 0.01806640625
|
|
},
|
|
{
|
|
"name": "model.layers.0.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 3.884323177771876e-06,
|
|
"max_abs_err": 0.0625
|
|
},
|
|
{
|
|
"name": "model.layers.0.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.55732242699014e-06,
|
|
"max_abs_err": 0.037109375
|
|
},
|
|
{
|
|
"name": "model.layers.0.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 3.965223640989279e-06,
|
|
"max_abs_err": 0.0390625
|
|
},
|
|
{
|
|
"name": "model.layers.0.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.477888523979345e-06,
|
|
"max_abs_err": 0.060546875
|
|
},
|
|
{
|
|
"name": "model.layers.1.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 3.8442944969574455e-06,
|
|
"max_abs_err": 0.05078125
|
|
},
|
|
{
|
|
"name": "model.layers.1.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.1471043914498296e-06,
|
|
"max_abs_err": 0.025390625
|
|
},
|
|
{
|
|
"name": "model.layers.1.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.42921327703516e-06,
|
|
"max_abs_err": 0.015625
|
|
},
|
|
{
|
|
"name": "model.layers.1.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 3.93121263186913e-06,
|
|
"max_abs_err": 0.05078125
|
|
},
|
|
{
|
|
"name": "model.layers.1.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.296734005038161e-06,
|
|
"max_abs_err": 0.033203125
|
|
},
|
|
{
|
|
"name": "model.layers.1.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 1.957799668161897e-06,
|
|
"max_abs_err": 0.0283203125
|
|
},
|
|
{
|
|
"name": "model.layers.1.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 2.2658996385871433e-06,
|
|
"max_abs_err": 0.0703125
|
|
},
|
|
{
|
|
"name": "model.layers.2.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.111255293537397e-06,
|
|
"max_abs_err": 0.04296875
|
|
},
|
|
{
|
|
"name": "model.layers.2.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.46276271759416e-06,
|
|
"max_abs_err": 0.044921875
|
|
},
|
|
{
|
|
"name": "model.layers.2.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.769919996761018e-06,
|
|
"max_abs_err": 0.015625
|
|
},
|
|
{
|
|
"name": "model.layers.2.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.099542820767965e-06,
|
|
"max_abs_err": 0.0341796875
|
|
},
|
|
{
|
|
"name": "model.layers.2.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 2.9356613140407717e-06,
|
|
"max_abs_err": 0.029296875
|
|
},
|
|
{
|
|
"name": "model.layers.2.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 2.4578309876233106e-06,
|
|
"max_abs_err": 0.04296875
|
|
},
|
|
{
|
|
"name": "model.layers.2.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 2.8484221274993615e-06,
|
|
"max_abs_err": 0.060546875
|
|
},
|
|
{
|
|
"name": "model.layers.3.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.154432645009365e-06,
|
|
"max_abs_err": 0.0361328125
|
|
},
|
|
{
|
|
"name": "model.layers.3.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.152899236942176e-06,
|
|
"max_abs_err": 0.025390625
|
|
},
|
|
{
|
|
"name": "model.layers.3.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.8436895667691715e-06,
|
|
"max_abs_err": 0.0185546875
|
|
},
|
|
{
|
|
"name": "model.layers.3.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.150874701736029e-06,
|
|
"max_abs_err": 0.0537109375
|
|
},
|
|
{
|
|
"name": "model.layers.3.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 5.188927389099263e-06,
|
|
"max_abs_err": 0.03515625
|
|
},
|
|
{
|
|
"name": "model.layers.3.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 2.8084120913263177e-06,
|
|
"max_abs_err": 0.041015625
|
|
},
|
|
{
|
|
"name": "model.layers.3.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 3.132234269287437e-06,
|
|
"max_abs_err": 0.05859375
|
|
},
|
|
{
|
|
"name": "model.layers.4.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.294029622542439e-06,
|
|
"max_abs_err": 0.0400390625
|
|
},
|
|
{
|
|
"name": "model.layers.4.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.427968178788433e-06,
|
|
"max_abs_err": 0.044921875
|
|
},
|
|
{
|
|
"name": "model.layers.4.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.924415861751186e-06,
|
|
"max_abs_err": 0.015625
|
|
},
|
|
{
|
|
"name": "model.layers.4.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.3348782128305174e-06,
|
|
"max_abs_err": 0.0546875
|
|
},
|
|
{
|
|
"name": "model.layers.4.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 8.049067218962591e-06,
|
|
"max_abs_err": 0.0361328125
|
|
},
|
|
{
|
|
"name": "model.layers.4.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 3.397330601728754e-06,
|
|
"max_abs_err": 0.0263671875
|
|
},
|
|
{
|
|
"name": "model.layers.4.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 3.6408052892511478e-06,
|
|
"max_abs_err": 0.048828125
|
|
},
|
|
{
|
|
"name": "model.layers.5.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.2482711251068395e-06,
|
|
"max_abs_err": 0.037109375
|
|
},
|
|
{
|
|
"name": "model.layers.5.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.233726485836087e-06,
|
|
"max_abs_err": 0.0380859375
|
|
},
|
|
{
|
|
"name": "model.layers.5.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.861286015511723e-06,
|
|
"max_abs_err": 0.017578125
|
|
},
|
|
{
|
|
"name": "model.layers.5.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.04879529014579e-06,
|
|
"max_abs_err": 0.056640625
|
|
},
|
|
{
|
|
"name": "model.layers.5.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 7.083965556375915e-06,
|
|
"max_abs_err": 0.037109375
|
|
},
|
|
{
|
|
"name": "model.layers.5.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 3.816854587057605e-06,
|
|
"max_abs_err": 0.044921875
|
|
},
|
|
{
|
|
"name": "model.layers.5.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 3.808074097833014e-06,
|
|
"max_abs_err": 0.072265625
|
|
},
|
|
{
|
|
"name": "model.layers.6.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.508055098995101e-06,
|
|
"max_abs_err": 0.04150390625
|
|
},
|
|
{
|
|
"name": "model.layers.6.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.240621819917578e-06,
|
|
"max_abs_err": 0.03125
|
|
},
|
|
{
|
|
"name": "model.layers.6.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.904123670712579e-06,
|
|
"max_abs_err": 0.0283203125
|
|
},
|
|
{
|
|
"name": "model.layers.6.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.328182058088714e-06,
|
|
"max_abs_err": 0.078125
|
|
},
|
|
{
|
|
"name": "model.layers.6.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 6.248923909879522e-06,
|
|
"max_abs_err": 0.037109375
|
|
},
|
|
{
|
|
"name": "model.layers.6.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.035503934574081e-06,
|
|
"max_abs_err": 0.037109375
|
|
},
|
|
{
|
|
"name": "model.layers.6.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.088376954314299e-06,
|
|
"max_abs_err": 0.060546875
|
|
},
|
|
{
|
|
"name": "model.layers.7.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.244740011927206e-06,
|
|
"max_abs_err": 0.0380859375
|
|
},
|
|
{
|
|
"name": "model.layers.7.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.2962224142684136e-06,
|
|
"max_abs_err": 0.0341796875
|
|
},
|
|
{
|
|
"name": "model.layers.7.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 5.564193088503089e-06,
|
|
"max_abs_err": 0.0166015625
|
|
},
|
|
{
|
|
"name": "model.layers.7.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.218873982608784e-06,
|
|
"max_abs_err": 0.052734375
|
|
},
|
|
{
|
|
"name": "model.layers.7.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 6.10346432949882e-06,
|
|
"max_abs_err": 0.0546875
|
|
},
|
|
{
|
|
"name": "model.layers.7.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.07972856919514e-06,
|
|
"max_abs_err": 0.0322265625
|
|
},
|
|
{
|
|
"name": "model.layers.7.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.094483756489353e-06,
|
|
"max_abs_err": 0.04443359375
|
|
},
|
|
{
|
|
"name": "model.layers.8.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.29332749263267e-06,
|
|
"max_abs_err": 0.029296875
|
|
},
|
|
{
|
|
"name": "model.layers.8.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.381317921797745e-06,
|
|
"max_abs_err": 0.025390625
|
|
},
|
|
{
|
|
"name": "model.layers.8.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 5.524434072867734e-06,
|
|
"max_abs_err": 0.01708984375
|
|
},
|
|
{
|
|
"name": "model.layers.8.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.403879302117275e-06,
|
|
"max_abs_err": 0.0400390625
|
|
},
|
|
{
|
|
"name": "model.layers.8.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 5.27864813193446e-06,
|
|
"max_abs_err": 0.0390625
|
|
},
|
|
{
|
|
"name": "model.layers.8.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.272857950127218e-06,
|
|
"max_abs_err": 0.0390625
|
|
},
|
|
{
|
|
"name": "model.layers.8.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.357619673100999e-06,
|
|
"max_abs_err": 0.07421875
|
|
},
|
|
{
|
|
"name": "model.layers.9.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.131610239710426e-06,
|
|
"max_abs_err": 0.0380859375
|
|
},
|
|
{
|
|
"name": "model.layers.9.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.204764081805479e-06,
|
|
"max_abs_err": 0.048828125
|
|
},
|
|
{
|
|
"name": "model.layers.9.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 5.701755981135648e-06,
|
|
"max_abs_err": 0.0185546875
|
|
},
|
|
{
|
|
"name": "model.layers.9.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.341866770118941e-06,
|
|
"max_abs_err": 0.048828125
|
|
},
|
|
{
|
|
"name": "model.layers.9.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 6.178286639624275e-06,
|
|
"max_abs_err": 0.056640625
|
|
},
|
|
{
|
|
"name": "model.layers.9.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 3.931649189325981e-06,
|
|
"max_abs_err": 0.0341796875
|
|
},
|
|
{
|
|
"name": "model.layers.9.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 3.919267328456044e-06,
|
|
"max_abs_err": 0.0546875
|
|
},
|
|
{
|
|
"name": "model.layers.10.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.351064944785321e-06,
|
|
"max_abs_err": 0.0380859375
|
|
},
|
|
{
|
|
"name": "model.layers.10.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.355007604317507e-06,
|
|
"max_abs_err": 0.03125
|
|
},
|
|
{
|
|
"name": "model.layers.10.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 5.598906682280358e-06,
|
|
"max_abs_err": 0.02001953125
|
|
},
|
|
{
|
|
"name": "model.layers.10.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.498688213061541e-06,
|
|
"max_abs_err": 0.056640625
|
|
},
|
|
{
|
|
"name": "model.layers.10.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 5.9006410992878955e-06,
|
|
"max_abs_err": 0.07421875
|
|
},
|
|
{
|
|
"name": "model.layers.10.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.084883130417438e-06,
|
|
"max_abs_err": 0.037109375
|
|
},
|
|
{
|
|
"name": "model.layers.10.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.119125605939189e-06,
|
|
"max_abs_err": 0.0703125
|
|
},
|
|
{
|
|
"name": "model.layers.11.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.362725121609401e-06,
|
|
"max_abs_err": 0.033203125
|
|
},
|
|
{
|
|
"name": "model.layers.11.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.505641300056595e-06,
|
|
"max_abs_err": 0.046875
|
|
},
|
|
{
|
|
"name": "model.layers.11.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 5.723034064430976e-06,
|
|
"max_abs_err": 0.017578125
|
|
},
|
|
{
|
|
"name": "model.layers.11.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.3996396925649606e-06,
|
|
"max_abs_err": 0.0341796875
|
|
},
|
|
{
|
|
"name": "model.layers.11.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 5.5344107749988325e-06,
|
|
"max_abs_err": 0.046875
|
|
},
|
|
{
|
|
"name": "model.layers.11.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.3337163333490025e-06,
|
|
"max_abs_err": 0.037109375
|
|
},
|
|
{
|
|
"name": "model.layers.11.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.295663984521525e-06,
|
|
"max_abs_err": 0.064453125
|
|
},
|
|
{
|
|
"name": "model.layers.12.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.42736654804321e-06,
|
|
"max_abs_err": 0.0380859375
|
|
},
|
|
{
|
|
"name": "model.layers.12.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.359534159448231e-06,
|
|
"max_abs_err": 0.060546875
|
|
},
|
|
{
|
|
"name": "model.layers.12.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 5.615883765131002e-06,
|
|
"max_abs_err": 0.0205078125
|
|
},
|
|
{
|
|
"name": "model.layers.12.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.32626120527857e-06,
|
|
"max_abs_err": 0.076171875
|
|
},
|
|
{
|
|
"name": "model.layers.12.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 5.266308562568156e-06,
|
|
"max_abs_err": 0.05078125
|
|
},
|
|
{
|
|
"name": "model.layers.12.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.532284492597682e-06,
|
|
"max_abs_err": 0.0390625
|
|
},
|
|
{
|
|
"name": "model.layers.12.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.407591404742561e-06,
|
|
"max_abs_err": 0.078125
|
|
},
|
|
{
|
|
"name": "model.layers.13.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.41527026850963e-06,
|
|
"max_abs_err": 0.048828125
|
|
},
|
|
{
|
|
"name": "model.layers.13.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.338548478699522e-06,
|
|
"max_abs_err": 0.044921875
|
|
},
|
|
{
|
|
"name": "model.layers.13.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 5.347076694306452e-06,
|
|
"max_abs_err": 0.02099609375
|
|
},
|
|
{
|
|
"name": "model.layers.13.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.133703441766556e-06,
|
|
"max_abs_err": 0.0546875
|
|
},
|
|
{
|
|
"name": "model.layers.13.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.9299078455078416e-06,
|
|
"max_abs_err": 0.052734375
|
|
},
|
|
{
|
|
"name": "model.layers.13.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.663273557525827e-06,
|
|
"max_abs_err": 0.037109375
|
|
},
|
|
{
|
|
"name": "model.layers.13.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.506962795858271e-06,
|
|
"max_abs_err": 0.0546875
|
|
},
|
|
{
|
|
"name": "model.layers.14.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.3742661546275485e-06,
|
|
"max_abs_err": 0.03125
|
|
},
|
|
{
|
|
"name": "model.layers.14.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.337135578680318e-06,
|
|
"max_abs_err": 0.033203125
|
|
},
|
|
{
|
|
"name": "model.layers.14.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 5.114804480399471e-06,
|
|
"max_abs_err": 0.01806640625
|
|
},
|
|
{
|
|
"name": "model.layers.14.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.021666427433956e-06,
|
|
"max_abs_err": 0.0625
|
|
},
|
|
{
|
|
"name": "model.layers.14.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.78740366816055e-06,
|
|
"max_abs_err": 0.052734375
|
|
},
|
|
{
|
|
"name": "model.layers.14.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.535500011115801e-06,
|
|
"max_abs_err": 0.0361328125
|
|
},
|
|
{
|
|
"name": "model.layers.14.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.407239430292975e-06,
|
|
"max_abs_err": 0.056640625
|
|
},
|
|
{
|
|
"name": "model.layers.15.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.497607278608484e-06,
|
|
"max_abs_err": 0.037109375
|
|
},
|
|
{
|
|
"name": "model.layers.15.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.329124294599751e-06,
|
|
"max_abs_err": 0.0390625
|
|
},
|
|
{
|
|
"name": "model.layers.15.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.9755049076338764e-06,
|
|
"max_abs_err": 0.01953125
|
|
},
|
|
{
|
|
"name": "model.layers.15.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 3.7926984077785164e-06,
|
|
"max_abs_err": 0.041015625
|
|
},
|
|
{
|
|
"name": "model.layers.15.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.473585704545258e-06,
|
|
"max_abs_err": 0.060546875
|
|
},
|
|
{
|
|
"name": "model.layers.15.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.4379189603205305e-06,
|
|
"max_abs_err": 0.0400390625
|
|
},
|
|
{
|
|
"name": "model.layers.15.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.3610766624624375e-06,
|
|
"max_abs_err": 0.056640625
|
|
},
|
|
{
|
|
"name": "model.layers.16.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.228112629789393e-06,
|
|
"max_abs_err": 0.033203125
|
|
},
|
|
{
|
|
"name": "model.layers.16.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.029952378914459e-06,
|
|
"max_abs_err": 0.0341796875
|
|
},
|
|
{
|
|
"name": "model.layers.16.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.999103566660779e-06,
|
|
"max_abs_err": 0.02001953125
|
|
},
|
|
{
|
|
"name": "model.layers.16.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.00851013182546e-06,
|
|
"max_abs_err": 0.07421875
|
|
},
|
|
{
|
|
"name": "model.layers.16.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.68530834041303e-06,
|
|
"max_abs_err": 0.064453125
|
|
},
|
|
{
|
|
"name": "model.layers.16.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.385244210425299e-06,
|
|
"max_abs_err": 0.10546875
|
|
},
|
|
{
|
|
"name": "model.layers.16.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.210333372611785e-06,
|
|
"max_abs_err": 0.1484375
|
|
},
|
|
{
|
|
"name": "model.layers.17.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.4457769945438486e-06,
|
|
"max_abs_err": 0.03515625
|
|
},
|
|
{
|
|
"name": "model.layers.17.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 3.83370888812351e-06,
|
|
"max_abs_err": 0.0302734375
|
|
},
|
|
{
|
|
"name": "model.layers.17.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.7686016841907986e-06,
|
|
"max_abs_err": 0.021484375
|
|
},
|
|
{
|
|
"name": "model.layers.17.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 3.887456841766834e-06,
|
|
"max_abs_err": 0.060546875
|
|
},
|
|
{
|
|
"name": "model.layers.17.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.557432930596406e-06,
|
|
"max_abs_err": 0.076171875
|
|
},
|
|
{
|
|
"name": "model.layers.17.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.403425919008441e-06,
|
|
"max_abs_err": 0.0390625
|
|
},
|
|
{
|
|
"name": "model.layers.17.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.241866918164305e-06,
|
|
"max_abs_err": 0.083984375
|
|
},
|
|
{
|
|
"name": "model.layers.18.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.148148946114816e-06,
|
|
"max_abs_err": 0.0439453125
|
|
},
|
|
{
|
|
"name": "model.layers.18.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 3.894510882673785e-06,
|
|
"max_abs_err": 0.02734375
|
|
},
|
|
{
|
|
"name": "model.layers.18.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.939644441037672e-06,
|
|
"max_abs_err": 0.01904296875
|
|
},
|
|
{
|
|
"name": "model.layers.18.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 3.881219981849426e-06,
|
|
"max_abs_err": 0.05078125
|
|
},
|
|
{
|
|
"name": "model.layers.18.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.466749487619381e-06,
|
|
"max_abs_err": 0.09765625
|
|
},
|
|
{
|
|
"name": "model.layers.18.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.409459961607354e-06,
|
|
"max_abs_err": 0.0400390625
|
|
},
|
|
{
|
|
"name": "model.layers.18.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.252085091138724e-06,
|
|
"max_abs_err": 0.056640625
|
|
},
|
|
{
|
|
"name": "model.layers.19.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.318806077208137e-06,
|
|
"max_abs_err": 0.037109375
|
|
},
|
|
{
|
|
"name": "model.layers.19.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 3.6864505545963766e-06,
|
|
"max_abs_err": 0.052734375
|
|
},
|
|
{
|
|
"name": "model.layers.19.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.722854555438971e-06,
|
|
"max_abs_err": 0.01904296875
|
|
},
|
|
{
|
|
"name": "model.layers.19.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 3.967053999076597e-06,
|
|
"max_abs_err": 0.060546875
|
|
},
|
|
{
|
|
"name": "model.layers.19.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.417328455019742e-06,
|
|
"max_abs_err": 0.068359375
|
|
},
|
|
{
|
|
"name": "model.layers.19.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.44006309407996e-06,
|
|
"max_abs_err": 0.04296875
|
|
},
|
|
{
|
|
"name": "model.layers.19.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.264526523911627e-06,
|
|
"max_abs_err": 0.060546875
|
|
},
|
|
{
|
|
"name": "model.layers.20.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.604032255883794e-06,
|
|
"max_abs_err": 0.03515625
|
|
},
|
|
{
|
|
"name": "model.layers.20.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 3.6128153624304105e-06,
|
|
"max_abs_err": 0.033203125
|
|
},
|
|
{
|
|
"name": "model.layers.20.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.73115414934e-06,
|
|
"max_abs_err": 0.0185546875
|
|
},
|
|
{
|
|
"name": "model.layers.20.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 3.764621169466409e-06,
|
|
"max_abs_err": 0.05859375
|
|
},
|
|
{
|
|
"name": "model.layers.20.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.314027592045022e-06,
|
|
"max_abs_err": 0.060546875
|
|
},
|
|
{
|
|
"name": "model.layers.20.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.44011948275147e-06,
|
|
"max_abs_err": 0.05078125
|
|
},
|
|
{
|
|
"name": "model.layers.20.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.300146883906564e-06,
|
|
"max_abs_err": 0.052734375
|
|
},
|
|
{
|
|
"name": "model.layers.21.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.408530003274791e-06,
|
|
"max_abs_err": 0.0380859375
|
|
},
|
|
{
|
|
"name": "model.layers.21.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 3.5656448744703084e-06,
|
|
"max_abs_err": 0.025390625
|
|
},
|
|
{
|
|
"name": "model.layers.21.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.5936694732517935e-06,
|
|
"max_abs_err": 0.0341796875
|
|
},
|
|
{
|
|
"name": "model.layers.21.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 3.6658461795013864e-06,
|
|
"max_abs_err": 0.02734375
|
|
},
|
|
{
|
|
"name": "model.layers.21.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.188520506431814e-06,
|
|
"max_abs_err": 0.0498046875
|
|
},
|
|
{
|
|
"name": "model.layers.21.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.409470420796424e-06,
|
|
"max_abs_err": 0.0390625
|
|
},
|
|
{
|
|
"name": "model.layers.21.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.349980372353457e-06,
|
|
"max_abs_err": 0.06640625
|
|
},
|
|
{
|
|
"name": "model.layers.22.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.279831955500413e-06,
|
|
"max_abs_err": 0.04296875
|
|
},
|
|
{
|
|
"name": "model.layers.22.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 3.395120529603446e-06,
|
|
"max_abs_err": 0.091796875
|
|
},
|
|
{
|
|
"name": "model.layers.22.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.610105861502234e-06,
|
|
"max_abs_err": 0.025390625
|
|
},
|
|
{
|
|
"name": "model.layers.22.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 3.816996013483731e-06,
|
|
"max_abs_err": 0.0546875
|
|
},
|
|
{
|
|
"name": "model.layers.22.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.486457328312099e-06,
|
|
"max_abs_err": 0.056640625
|
|
},
|
|
{
|
|
"name": "model.layers.22.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.62035859527532e-06,
|
|
"max_abs_err": 0.06640625
|
|
},
|
|
{
|
|
"name": "model.layers.22.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.429509772307938e-06,
|
|
"max_abs_err": 0.060546875
|
|
},
|
|
{
|
|
"name": "model.layers.23.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.858595275436528e-06,
|
|
"max_abs_err": 0.0380859375
|
|
},
|
|
{
|
|
"name": "model.layers.23.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 3.701973810166237e-06,
|
|
"max_abs_err": 0.03125
|
|
},
|
|
{
|
|
"name": "model.layers.23.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.845483545068419e-06,
|
|
"max_abs_err": 0.025390625
|
|
},
|
|
{
|
|
"name": "model.layers.23.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.104576873942278e-06,
|
|
"max_abs_err": 0.0380859375
|
|
},
|
|
{
|
|
"name": "model.layers.23.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.80355765830609e-06,
|
|
"max_abs_err": 0.068359375
|
|
},
|
|
{
|
|
"name": "model.layers.23.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.631754109141184e-06,
|
|
"max_abs_err": 0.03515625
|
|
},
|
|
{
|
|
"name": "model.layers.23.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.43936005467549e-06,
|
|
"max_abs_err": 0.0703125
|
|
},
|
|
{
|
|
"name": "model.layers.24.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.948974037688458e-06,
|
|
"max_abs_err": 0.04296875
|
|
},
|
|
{
|
|
"name": "model.layers.24.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 3.714139438670827e-06,
|
|
"max_abs_err": 0.0283203125
|
|
},
|
|
{
|
|
"name": "model.layers.24.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.7844587243162096e-06,
|
|
"max_abs_err": 0.0322265625
|
|
},
|
|
{
|
|
"name": "model.layers.24.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.229699698043987e-06,
|
|
"max_abs_err": 0.046875
|
|
},
|
|
{
|
|
"name": "model.layers.24.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.97599421578343e-06,
|
|
"max_abs_err": 0.0400390625
|
|
},
|
|
{
|
|
"name": "model.layers.24.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.654815711546689e-06,
|
|
"max_abs_err": 0.03125
|
|
},
|
|
{
|
|
"name": "model.layers.24.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.4677544792648405e-06,
|
|
"max_abs_err": 0.041015625
|
|
},
|
|
{
|
|
"name": "model.layers.25.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.377608547656564e-06,
|
|
"max_abs_err": 0.0361328125
|
|
},
|
|
{
|
|
"name": "model.layers.25.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 3.913945874955971e-06,
|
|
"max_abs_err": 0.03125
|
|
},
|
|
{
|
|
"name": "model.layers.25.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 5.042701104684966e-06,
|
|
"max_abs_err": 0.021484375
|
|
},
|
|
{
|
|
"name": "model.layers.25.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.0954823816719e-06,
|
|
"max_abs_err": 0.0244140625
|
|
},
|
|
{
|
|
"name": "model.layers.25.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 5.083179075882072e-06,
|
|
"max_abs_err": 0.052734375
|
|
},
|
|
{
|
|
"name": "model.layers.25.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.656661985791288e-06,
|
|
"max_abs_err": 0.046875
|
|
},
|
|
{
|
|
"name": "model.layers.25.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.500410341279348e-06,
|
|
"max_abs_err": 0.056640625
|
|
},
|
|
{
|
|
"name": "model.layers.26.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.557187821774278e-06,
|
|
"max_abs_err": 0.033203125
|
|
},
|
|
{
|
|
"name": "model.layers.26.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 3.804395873885369e-06,
|
|
"max_abs_err": 0.0283203125
|
|
},
|
|
{
|
|
"name": "model.layers.26.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 5.004220838600304e-06,
|
|
"max_abs_err": 0.02099609375
|
|
},
|
|
{
|
|
"name": "model.layers.26.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.28900284532574e-06,
|
|
"max_abs_err": 0.0263671875
|
|
},
|
|
{
|
|
"name": "model.layers.26.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 5.0898847803182434e-06,
|
|
"max_abs_err": 0.044921875
|
|
},
|
|
{
|
|
"name": "model.layers.26.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.752734184876317e-06,
|
|
"max_abs_err": 0.037109375
|
|
},
|
|
{
|
|
"name": "model.layers.26.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.570770670397906e-06,
|
|
"max_abs_err": 0.056640625
|
|
},
|
|
{
|
|
"name": "model.layers.27.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.443979378265794e-06,
|
|
"max_abs_err": 0.0322265625
|
|
},
|
|
{
|
|
"name": "model.layers.27.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 3.522326778693241e-06,
|
|
"max_abs_err": 0.0205078125
|
|
},
|
|
{
|
|
"name": "model.layers.27.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.739751148008509e-06,
|
|
"max_abs_err": 0.02734375
|
|
},
|
|
{
|
|
"name": "model.layers.27.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.103467290406115e-06,
|
|
"max_abs_err": 0.037109375
|
|
},
|
|
{
|
|
"name": "model.layers.27.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 5.037284154241206e-06,
|
|
"max_abs_err": 0.0546875
|
|
},
|
|
{
|
|
"name": "model.layers.27.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.8673609853722155e-06,
|
|
"max_abs_err": 0.0341796875
|
|
},
|
|
{
|
|
"name": "model.layers.27.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.680827714764746e-06,
|
|
"max_abs_err": 0.04296875
|
|
},
|
|
{
|
|
"name": "model.layers.28.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.625020665116608e-06,
|
|
"max_abs_err": 0.033203125
|
|
},
|
|
{
|
|
"name": "model.layers.28.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 3.759183073270833e-06,
|
|
"max_abs_err": 0.02734375
|
|
},
|
|
{
|
|
"name": "model.layers.28.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 5.18985461894772e-06,
|
|
"max_abs_err": 0.01904296875
|
|
},
|
|
{
|
|
"name": "model.layers.28.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.063952928845538e-06,
|
|
"max_abs_err": 0.0341796875
|
|
},
|
|
{
|
|
"name": "model.layers.28.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.932152023684466e-06,
|
|
"max_abs_err": 0.05859375
|
|
},
|
|
{
|
|
"name": "model.layers.28.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.970076133758994e-06,
|
|
"max_abs_err": 0.068359375
|
|
},
|
|
{
|
|
"name": "model.layers.28.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.789372724189889e-06,
|
|
"max_abs_err": 0.060546875
|
|
},
|
|
{
|
|
"name": "model.layers.29.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 5.1148495003872085e-06,
|
|
"max_abs_err": 0.0341796875
|
|
},
|
|
{
|
|
"name": "model.layers.29.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 3.635831035353476e-06,
|
|
"max_abs_err": 0.033203125
|
|
},
|
|
{
|
|
"name": "model.layers.29.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.815466581931105e-06,
|
|
"max_abs_err": 0.02734375
|
|
},
|
|
{
|
|
"name": "model.layers.29.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.3866789383173455e-06,
|
|
"max_abs_err": 0.02685546875
|
|
},
|
|
{
|
|
"name": "model.layers.29.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.919627372146351e-06,
|
|
"max_abs_err": 0.064453125
|
|
},
|
|
{
|
|
"name": "model.layers.29.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 5.120305559103144e-06,
|
|
"max_abs_err": 0.044921875
|
|
},
|
|
{
|
|
"name": "model.layers.29.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.91496894028387e-06,
|
|
"max_abs_err": 0.056640625
|
|
},
|
|
{
|
|
"name": "model.layers.30.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.318022092775209e-06,
|
|
"max_abs_err": 0.041015625
|
|
},
|
|
{
|
|
"name": "model.layers.30.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 3.9061606003087945e-06,
|
|
"max_abs_err": 0.0283203125
|
|
},
|
|
{
|
|
"name": "model.layers.30.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 5.369913196773268e-06,
|
|
"max_abs_err": 0.03125
|
|
},
|
|
{
|
|
"name": "model.layers.30.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.30595719080884e-06,
|
|
"max_abs_err": 0.02734375
|
|
},
|
|
{
|
|
"name": "model.layers.30.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.865625669481233e-06,
|
|
"max_abs_err": 0.060546875
|
|
},
|
|
{
|
|
"name": "model.layers.30.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 5.2353257160575595e-06,
|
|
"max_abs_err": 0.04296875
|
|
},
|
|
{
|
|
"name": "model.layers.30.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 5.042840712121688e-06,
|
|
"max_abs_err": 0.052734375
|
|
},
|
|
{
|
|
"name": "model.layers.31.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.488820195547305e-06,
|
|
"max_abs_err": 0.0341796875
|
|
},
|
|
{
|
|
"name": "model.layers.31.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 3.6213489238434704e-06,
|
|
"max_abs_err": 0.02734375
|
|
},
|
|
{
|
|
"name": "model.layers.31.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 5.777186743216589e-06,
|
|
"max_abs_err": 0.02490234375
|
|
},
|
|
{
|
|
"name": "model.layers.31.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.571093995764386e-06,
|
|
"max_abs_err": 0.046875
|
|
},
|
|
{
|
|
"name": "model.layers.31.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.759885996463709e-06,
|
|
"max_abs_err": 0.068359375
|
|
},
|
|
{
|
|
"name": "model.layers.31.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 5.364705884858267e-06,
|
|
"max_abs_err": 0.0361328125
|
|
},
|
|
{
|
|
"name": "model.layers.31.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 5.2394671001820825e-06,
|
|
"max_abs_err": 0.052734375
|
|
},
|
|
{
|
|
"name": "model.layers.32.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.357744728622492e-06,
|
|
"max_abs_err": 0.037109375
|
|
},
|
|
{
|
|
"name": "model.layers.32.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 3.6557835301209707e-06,
|
|
"max_abs_err": 0.0302734375
|
|
},
|
|
{
|
|
"name": "model.layers.32.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 5.786484052805463e-06,
|
|
"max_abs_err": 0.025390625
|
|
},
|
|
{
|
|
"name": "model.layers.32.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.5518822844314855e-06,
|
|
"max_abs_err": 0.044921875
|
|
},
|
|
{
|
|
"name": "model.layers.32.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.632854597730329e-06,
|
|
"max_abs_err": 0.05078125
|
|
},
|
|
{
|
|
"name": "model.layers.32.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 5.398054781835526e-06,
|
|
"max_abs_err": 0.037109375
|
|
},
|
|
{
|
|
"name": "model.layers.32.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 5.252433766145259e-06,
|
|
"max_abs_err": 0.0703125
|
|
},
|
|
{
|
|
"name": "model.layers.33.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.438436008058488e-06,
|
|
"max_abs_err": 0.0380859375
|
|
},
|
|
{
|
|
"name": "model.layers.33.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 2.9729155812674435e-06,
|
|
"max_abs_err": 0.0224609375
|
|
},
|
|
{
|
|
"name": "model.layers.33.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 6.5771591835073195e-06,
|
|
"max_abs_err": 0.02587890625
|
|
},
|
|
{
|
|
"name": "model.layers.33.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.9389359446649905e-06,
|
|
"max_abs_err": 0.060546875
|
|
},
|
|
{
|
|
"name": "model.layers.33.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.552728569251485e-06,
|
|
"max_abs_err": 0.0546875
|
|
},
|
|
{
|
|
"name": "model.layers.33.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 5.443682766781421e-06,
|
|
"max_abs_err": 0.044921875
|
|
},
|
|
{
|
|
"name": "model.layers.33.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 5.14817429575487e-06,
|
|
"max_abs_err": 0.07421875
|
|
},
|
|
{
|
|
"name": "model.layers.34.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.09982249038876e-06,
|
|
"max_abs_err": 0.03125
|
|
},
|
|
{
|
|
"name": "model.layers.34.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 3.415584842514363e-06,
|
|
"max_abs_err": 0.03515625
|
|
},
|
|
{
|
|
"name": "model.layers.34.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 6.533237865369301e-06,
|
|
"max_abs_err": 0.025390625
|
|
},
|
|
{
|
|
"name": "model.layers.34.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.707650987256784e-06,
|
|
"max_abs_err": 0.0380859375
|
|
},
|
|
{
|
|
"name": "model.layers.34.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.729331521957647e-06,
|
|
"max_abs_err": 0.076171875
|
|
},
|
|
{
|
|
"name": "model.layers.34.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 5.304823844198836e-06,
|
|
"max_abs_err": 0.05078125
|
|
},
|
|
{
|
|
"name": "model.layers.34.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.668886958825169e-06,
|
|
"max_abs_err": 0.078125
|
|
},
|
|
{
|
|
"name": "model.layers.35.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.15862632507924e-06,
|
|
"max_abs_err": 0.037109375
|
|
},
|
|
{
|
|
"name": "model.layers.35.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 3.8521648093592376e-06,
|
|
"max_abs_err": 0.033203125
|
|
},
|
|
{
|
|
"name": "model.layers.35.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 5.735487320635002e-06,
|
|
"max_abs_err": 0.0234375
|
|
},
|
|
{
|
|
"name": "model.layers.35.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.418748630996561e-06,
|
|
"max_abs_err": 0.06640625
|
|
},
|
|
{
|
|
"name": "model.layers.35.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 5.227528163231909e-06,
|
|
"max_abs_err": 0.087890625
|
|
},
|
|
{
|
|
"name": "model.layers.35.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 5.463487923407229e-06,
|
|
"max_abs_err": 0.08203125
|
|
},
|
|
{
|
|
"name": "model.layers.35.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.352178621047642e-06,
|
|
"max_abs_err": 0.06640625
|
|
}
|
|
],
|
|
"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."
|
|
}
|