Model: meteorain/Qwen__Qwen3-4B-Thinking-2507-RTN-nvfp4-g16-faked-bf16 Source: Original Platform
2537 lines
52 KiB
JSON
2537 lines
52 KiB
JSON
{
|
|
"format": "nvfp4",
|
|
"codebook": "fp4-e2m1",
|
|
"block_size": 16,
|
|
"num_modules": 252,
|
|
"num_weights": 3633315840,
|
|
"elapsed_sec": 89.115,
|
|
"quantized_modules": [
|
|
{
|
|
"name": "model.layers.0.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.542554052022751e-06,
|
|
"max_abs_err": 0.06640625
|
|
},
|
|
{
|
|
"name": "model.layers.0.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 5.047036665928317e-06,
|
|
"max_abs_err": 0.0322265625
|
|
},
|
|
{
|
|
"name": "model.layers.0.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.441323653736617e-06,
|
|
"max_abs_err": 0.01904296875
|
|
},
|
|
{
|
|
"name": "model.layers.0.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 3.981581812695367e-06,
|
|
"max_abs_err": 0.05078125
|
|
},
|
|
{
|
|
"name": "model.layers.0.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.7133685256994795e-06,
|
|
"max_abs_err": 0.037109375
|
|
},
|
|
{
|
|
"name": "model.layers.0.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.1089938349614386e-06,
|
|
"max_abs_err": 0.0390625
|
|
},
|
|
{
|
|
"name": "model.layers.0.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.540105692285579e-06,
|
|
"max_abs_err": 0.048828125
|
|
},
|
|
{
|
|
"name": "model.layers.1.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 3.9390806705341674e-06,
|
|
"max_abs_err": 0.05078125
|
|
},
|
|
{
|
|
"name": "model.layers.1.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.266446012479719e-06,
|
|
"max_abs_err": 0.0283203125
|
|
},
|
|
{
|
|
"name": "model.layers.1.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.564609298540745e-06,
|
|
"max_abs_err": 0.01904296875
|
|
},
|
|
{
|
|
"name": "model.layers.1.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.0740515032666735e-06,
|
|
"max_abs_err": 0.0546875
|
|
},
|
|
{
|
|
"name": "model.layers.1.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.442033059604e-06,
|
|
"max_abs_err": 0.02978515625
|
|
},
|
|
{
|
|
"name": "model.layers.1.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 2.0193926957290387e-06,
|
|
"max_abs_err": 0.0302734375
|
|
},
|
|
{
|
|
"name": "model.layers.1.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 2.301136873938958e-06,
|
|
"max_abs_err": 0.103515625
|
|
},
|
|
{
|
|
"name": "model.layers.2.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.2169453990936745e-06,
|
|
"max_abs_err": 0.03125
|
|
},
|
|
{
|
|
"name": "model.layers.2.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.5095621317159384e-06,
|
|
"max_abs_err": 0.0390625
|
|
},
|
|
{
|
|
"name": "model.layers.2.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.917421847494552e-06,
|
|
"max_abs_err": 0.01611328125
|
|
},
|
|
{
|
|
"name": "model.layers.2.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.2704764382506255e-06,
|
|
"max_abs_err": 0.0400390625
|
|
},
|
|
{
|
|
"name": "model.layers.2.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 2.9918223845015746e-06,
|
|
"max_abs_err": 0.03515625
|
|
},
|
|
{
|
|
"name": "model.layers.2.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 2.5080255454668077e-06,
|
|
"max_abs_err": 0.0302734375
|
|
},
|
|
{
|
|
"name": "model.layers.2.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 2.2821548100182554e-06,
|
|
"max_abs_err": 0.046875
|
|
},
|
|
{
|
|
"name": "model.layers.3.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.2362189560662955e-06,
|
|
"max_abs_err": 0.05078125
|
|
},
|
|
{
|
|
"name": "model.layers.3.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.234458629071014e-06,
|
|
"max_abs_err": 0.02880859375
|
|
},
|
|
{
|
|
"name": "model.layers.3.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.985236046195496e-06,
|
|
"max_abs_err": 0.0185546875
|
|
},
|
|
{
|
|
"name": "model.layers.3.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.320966127124848e-06,
|
|
"max_abs_err": 0.0546875
|
|
},
|
|
{
|
|
"name": "model.layers.3.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 5.353369033400668e-06,
|
|
"max_abs_err": 0.052734375
|
|
},
|
|
{
|
|
"name": "model.layers.3.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 2.8861941245850176e-06,
|
|
"max_abs_err": 0.0517578125
|
|
},
|
|
{
|
|
"name": "model.layers.3.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 3.0565827273676405e-06,
|
|
"max_abs_err": 0.0625
|
|
},
|
|
{
|
|
"name": "model.layers.4.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.399350473249797e-06,
|
|
"max_abs_err": 0.0419921875
|
|
},
|
|
{
|
|
"name": "model.layers.4.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.4655757847067434e-06,
|
|
"max_abs_err": 0.0341796875
|
|
},
|
|
{
|
|
"name": "model.layers.4.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 5.0879539230663795e-06,
|
|
"max_abs_err": 0.01708984375
|
|
},
|
|
{
|
|
"name": "model.layers.4.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.5221504478831775e-06,
|
|
"max_abs_err": 0.064453125
|
|
},
|
|
{
|
|
"name": "model.layers.4.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 8.320091183122713e-06,
|
|
"max_abs_err": 0.0341796875
|
|
},
|
|
{
|
|
"name": "model.layers.4.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 3.5067955650447402e-06,
|
|
"max_abs_err": 0.03125
|
|
},
|
|
{
|
|
"name": "model.layers.4.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 3.569574118955643e-06,
|
|
"max_abs_err": 0.0595703125
|
|
},
|
|
{
|
|
"name": "model.layers.5.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.354569227871252e-06,
|
|
"max_abs_err": 0.037109375
|
|
},
|
|
{
|
|
"name": "model.layers.5.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.3236350393272005e-06,
|
|
"max_abs_err": 0.033203125
|
|
},
|
|
{
|
|
"name": "model.layers.5.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.999436896468978e-06,
|
|
"max_abs_err": 0.01953125
|
|
},
|
|
{
|
|
"name": "model.layers.5.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.208860445942264e-06,
|
|
"max_abs_err": 0.05859375
|
|
},
|
|
{
|
|
"name": "model.layers.5.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 7.235579687403515e-06,
|
|
"max_abs_err": 0.0419921875
|
|
},
|
|
{
|
|
"name": "model.layers.5.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 3.939914222428342e-06,
|
|
"max_abs_err": 0.0302734375
|
|
},
|
|
{
|
|
"name": "model.layers.5.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 3.8194452827156056e-06,
|
|
"max_abs_err": 0.083984375
|
|
},
|
|
{
|
|
"name": "model.layers.6.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.601887667377014e-06,
|
|
"max_abs_err": 0.0400390625
|
|
},
|
|
{
|
|
"name": "model.layers.6.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.356297722551972e-06,
|
|
"max_abs_err": 0.03125
|
|
},
|
|
{
|
|
"name": "model.layers.6.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.984137831343105e-06,
|
|
"max_abs_err": 0.0263671875
|
|
},
|
|
{
|
|
"name": "model.layers.6.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.4975340642849915e-06,
|
|
"max_abs_err": 0.083984375
|
|
},
|
|
{
|
|
"name": "model.layers.6.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 6.4540076891717035e-06,
|
|
"max_abs_err": 0.0400390625
|
|
},
|
|
{
|
|
"name": "model.layers.6.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.171607542957645e-06,
|
|
"max_abs_err": 0.037109375
|
|
},
|
|
{
|
|
"name": "model.layers.6.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.125409759581089e-06,
|
|
"max_abs_err": 0.060546875
|
|
},
|
|
{
|
|
"name": "model.layers.7.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.3491049837030005e-06,
|
|
"max_abs_err": 0.0322265625
|
|
},
|
|
{
|
|
"name": "model.layers.7.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.362997515272582e-06,
|
|
"max_abs_err": 0.03515625
|
|
},
|
|
{
|
|
"name": "model.layers.7.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 5.727734333049739e-06,
|
|
"max_abs_err": 0.01904296875
|
|
},
|
|
{
|
|
"name": "model.layers.7.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.378383437142475e-06,
|
|
"max_abs_err": 0.0478515625
|
|
},
|
|
{
|
|
"name": "model.layers.7.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 6.311032848316245e-06,
|
|
"max_abs_err": 0.0546875
|
|
},
|
|
{
|
|
"name": "model.layers.7.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.225127668178175e-06,
|
|
"max_abs_err": 0.0400390625
|
|
},
|
|
{
|
|
"name": "model.layers.7.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.187504600849934e-06,
|
|
"max_abs_err": 0.046875
|
|
},
|
|
{
|
|
"name": "model.layers.8.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.398107648739824e-06,
|
|
"max_abs_err": 0.03125
|
|
},
|
|
{
|
|
"name": "model.layers.8.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.508533493208233e-06,
|
|
"max_abs_err": 0.02587890625
|
|
},
|
|
{
|
|
"name": "model.layers.8.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 5.705248440790456e-06,
|
|
"max_abs_err": 0.0166015625
|
|
},
|
|
{
|
|
"name": "model.layers.8.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.587458988680737e-06,
|
|
"max_abs_err": 0.041015625
|
|
},
|
|
{
|
|
"name": "model.layers.8.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 5.478610546560958e-06,
|
|
"max_abs_err": 0.0390625
|
|
},
|
|
{
|
|
"name": "model.layers.8.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.424239705258515e-06,
|
|
"max_abs_err": 0.04296875
|
|
},
|
|
{
|
|
"name": "model.layers.8.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.4307771531748585e-06,
|
|
"max_abs_err": 0.125
|
|
},
|
|
{
|
|
"name": "model.layers.9.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.182641532679554e-06,
|
|
"max_abs_err": 0.0439453125
|
|
},
|
|
{
|
|
"name": "model.layers.9.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.181940312264487e-06,
|
|
"max_abs_err": 0.0439453125
|
|
},
|
|
{
|
|
"name": "model.layers.9.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 5.822787898068782e-06,
|
|
"max_abs_err": 0.02099609375
|
|
},
|
|
{
|
|
"name": "model.layers.9.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.504823209572351e-06,
|
|
"max_abs_err": 0.0478515625
|
|
},
|
|
{
|
|
"name": "model.layers.9.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 6.41219185126829e-06,
|
|
"max_abs_err": 0.0380859375
|
|
},
|
|
{
|
|
"name": "model.layers.9.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.0623881432111375e-06,
|
|
"max_abs_err": 0.0302734375
|
|
},
|
|
{
|
|
"name": "model.layers.9.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 3.965729774790816e-06,
|
|
"max_abs_err": 0.037109375
|
|
},
|
|
{
|
|
"name": "model.layers.10.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.4585422074305825e-06,
|
|
"max_abs_err": 0.0341796875
|
|
},
|
|
{
|
|
"name": "model.layers.10.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.433762114786077e-06,
|
|
"max_abs_err": 0.0400390625
|
|
},
|
|
{
|
|
"name": "model.layers.10.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 5.712111033062683e-06,
|
|
"max_abs_err": 0.02099609375
|
|
},
|
|
{
|
|
"name": "model.layers.10.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.675422587752109e-06,
|
|
"max_abs_err": 0.0625
|
|
},
|
|
{
|
|
"name": "model.layers.10.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 6.128296263341326e-06,
|
|
"max_abs_err": 0.0439453125
|
|
},
|
|
{
|
|
"name": "model.layers.10.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.222978532197885e-06,
|
|
"max_abs_err": 0.0361328125
|
|
},
|
|
{
|
|
"name": "model.layers.10.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.172791250312002e-06,
|
|
"max_abs_err": 0.080078125
|
|
},
|
|
{
|
|
"name": "model.layers.11.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.435257778823143e-06,
|
|
"max_abs_err": 0.0361328125
|
|
},
|
|
{
|
|
"name": "model.layers.11.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.58741442344035e-06,
|
|
"max_abs_err": 0.0439453125
|
|
},
|
|
{
|
|
"name": "model.layers.11.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 5.841427082486916e-06,
|
|
"max_abs_err": 0.01953125
|
|
},
|
|
{
|
|
"name": "model.layers.11.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.573185833578464e-06,
|
|
"max_abs_err": 0.0322265625
|
|
},
|
|
{
|
|
"name": "model.layers.11.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 5.742780103901168e-06,
|
|
"max_abs_err": 0.0703125
|
|
},
|
|
{
|
|
"name": "model.layers.11.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.479616109165363e-06,
|
|
"max_abs_err": 0.0361328125
|
|
},
|
|
{
|
|
"name": "model.layers.11.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.366329449112527e-06,
|
|
"max_abs_err": 0.0546875
|
|
},
|
|
{
|
|
"name": "model.layers.12.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.528201770881424e-06,
|
|
"max_abs_err": 0.0390625
|
|
},
|
|
{
|
|
"name": "model.layers.12.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.417740456119645e-06,
|
|
"max_abs_err": 0.05859375
|
|
},
|
|
{
|
|
"name": "model.layers.12.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 5.7242186812800355e-06,
|
|
"max_abs_err": 0.02490234375
|
|
},
|
|
{
|
|
"name": "model.layers.12.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.483066277316539e-06,
|
|
"max_abs_err": 0.091796875
|
|
},
|
|
{
|
|
"name": "model.layers.12.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 5.440474524220917e-06,
|
|
"max_abs_err": 0.05859375
|
|
},
|
|
{
|
|
"name": "model.layers.12.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.664385869546095e-06,
|
|
"max_abs_err": 0.0380859375
|
|
},
|
|
{
|
|
"name": "model.layers.12.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.464236553758383e-06,
|
|
"max_abs_err": 0.0458984375
|
|
},
|
|
{
|
|
"name": "model.layers.13.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.484520104597323e-06,
|
|
"max_abs_err": 0.033203125
|
|
},
|
|
{
|
|
"name": "model.layers.13.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.3462850953801535e-06,
|
|
"max_abs_err": 0.0400390625
|
|
},
|
|
{
|
|
"name": "model.layers.13.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 5.498080099641811e-06,
|
|
"max_abs_err": 0.02001953125
|
|
},
|
|
{
|
|
"name": "model.layers.13.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.2810943341464736e-06,
|
|
"max_abs_err": 0.0341796875
|
|
},
|
|
{
|
|
"name": "model.layers.13.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 5.072974545328179e-06,
|
|
"max_abs_err": 0.076171875
|
|
},
|
|
{
|
|
"name": "model.layers.13.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.78162201034138e-06,
|
|
"max_abs_err": 0.0380859375
|
|
},
|
|
{
|
|
"name": "model.layers.13.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.493766937230248e-06,
|
|
"max_abs_err": 0.08203125
|
|
},
|
|
{
|
|
"name": "model.layers.14.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.489885668590432e-06,
|
|
"max_abs_err": 0.033203125
|
|
},
|
|
{
|
|
"name": "model.layers.14.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.439792519406183e-06,
|
|
"max_abs_err": 0.0322265625
|
|
},
|
|
{
|
|
"name": "model.layers.14.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 5.185019290365744e-06,
|
|
"max_abs_err": 0.01953125
|
|
},
|
|
{
|
|
"name": "model.layers.14.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.160808202868793e-06,
|
|
"max_abs_err": 0.076171875
|
|
},
|
|
{
|
|
"name": "model.layers.14.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.931144758302253e-06,
|
|
"max_abs_err": 0.0517578125
|
|
},
|
|
{
|
|
"name": "model.layers.14.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.64901950181229e-06,
|
|
"max_abs_err": 0.0390625
|
|
},
|
|
{
|
|
"name": "model.layers.14.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.3928507693635765e-06,
|
|
"max_abs_err": 0.05859375
|
|
},
|
|
{
|
|
"name": "model.layers.15.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.596051439875737e-06,
|
|
"max_abs_err": 0.037109375
|
|
},
|
|
{
|
|
"name": "model.layers.15.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.38515780842863e-06,
|
|
"max_abs_err": 0.041015625
|
|
},
|
|
{
|
|
"name": "model.layers.15.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 5.1026572691625915e-06,
|
|
"max_abs_err": 0.0205078125
|
|
},
|
|
{
|
|
"name": "model.layers.15.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 3.92215861211298e-06,
|
|
"max_abs_err": 0.041015625
|
|
},
|
|
{
|
|
"name": "model.layers.15.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.589111995301209e-06,
|
|
"max_abs_err": 0.06640625
|
|
},
|
|
{
|
|
"name": "model.layers.15.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.544973762676818e-06,
|
|
"max_abs_err": 0.048828125
|
|
},
|
|
{
|
|
"name": "model.layers.15.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.314314082876081e-06,
|
|
"max_abs_err": 0.064453125
|
|
},
|
|
{
|
|
"name": "model.layers.16.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.331095169618493e-06,
|
|
"max_abs_err": 0.0341796875
|
|
},
|
|
{
|
|
"name": "model.layers.16.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.133424681640463e-06,
|
|
"max_abs_err": 0.03125
|
|
},
|
|
{
|
|
"name": "model.layers.16.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 5.113518909638515e-06,
|
|
"max_abs_err": 0.0205078125
|
|
},
|
|
{
|
|
"name": "model.layers.16.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.153736426815158e-06,
|
|
"max_abs_err": 0.078125
|
|
},
|
|
{
|
|
"name": "model.layers.16.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.821577931579668e-06,
|
|
"max_abs_err": 0.068359375
|
|
},
|
|
{
|
|
"name": "model.layers.16.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.486418220039923e-06,
|
|
"max_abs_err": 0.107421875
|
|
},
|
|
{
|
|
"name": "model.layers.16.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.181591066299006e-06,
|
|
"max_abs_err": 0.1484375
|
|
},
|
|
{
|
|
"name": "model.layers.17.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.556003204925219e-06,
|
|
"max_abs_err": 0.03515625
|
|
},
|
|
{
|
|
"name": "model.layers.17.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 3.897896931448486e-06,
|
|
"max_abs_err": 0.033203125
|
|
},
|
|
{
|
|
"name": "model.layers.17.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.864297807216644e-06,
|
|
"max_abs_err": 0.02099609375
|
|
},
|
|
{
|
|
"name": "model.layers.17.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.013651960121933e-06,
|
|
"max_abs_err": 0.052734375
|
|
},
|
|
{
|
|
"name": "model.layers.17.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.6912750804040115e-06,
|
|
"max_abs_err": 0.05859375
|
|
},
|
|
{
|
|
"name": "model.layers.17.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.52291078545386e-06,
|
|
"max_abs_err": 0.03515625
|
|
},
|
|
{
|
|
"name": "model.layers.17.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.206545781926252e-06,
|
|
"max_abs_err": 0.064453125
|
|
},
|
|
{
|
|
"name": "model.layers.18.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.2560013753245585e-06,
|
|
"max_abs_err": 0.0537109375
|
|
},
|
|
{
|
|
"name": "model.layers.18.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 3.990614459326025e-06,
|
|
"max_abs_err": 0.03125
|
|
},
|
|
{
|
|
"name": "model.layers.18.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 5.056341706222156e-06,
|
|
"max_abs_err": 0.01904296875
|
|
},
|
|
{
|
|
"name": "model.layers.18.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.020593678433215e-06,
|
|
"max_abs_err": 0.0498046875
|
|
},
|
|
{
|
|
"name": "model.layers.18.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.573814294417389e-06,
|
|
"max_abs_err": 0.08203125
|
|
},
|
|
{
|
|
"name": "model.layers.18.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.5132910599932075e-06,
|
|
"max_abs_err": 0.041015625
|
|
},
|
|
{
|
|
"name": "model.layers.18.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.195589554001344e-06,
|
|
"max_abs_err": 0.0703125
|
|
},
|
|
{
|
|
"name": "model.layers.19.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.401343176141381e-06,
|
|
"max_abs_err": 0.03515625
|
|
},
|
|
{
|
|
"name": "model.layers.19.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 3.724827593032387e-06,
|
|
"max_abs_err": 0.052734375
|
|
},
|
|
{
|
|
"name": "model.layers.19.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.821865331905428e-06,
|
|
"max_abs_err": 0.01904296875
|
|
},
|
|
{
|
|
"name": "model.layers.19.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.074472599313594e-06,
|
|
"max_abs_err": 0.068359375
|
|
},
|
|
{
|
|
"name": "model.layers.19.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.513279691309435e-06,
|
|
"max_abs_err": 0.08203125
|
|
},
|
|
{
|
|
"name": "model.layers.19.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.537821951089427e-06,
|
|
"max_abs_err": 0.04296875
|
|
},
|
|
{
|
|
"name": "model.layers.19.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.189990249869879e-06,
|
|
"max_abs_err": 0.07421875
|
|
},
|
|
{
|
|
"name": "model.layers.20.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.71401517643244e-06,
|
|
"max_abs_err": 0.037109375
|
|
},
|
|
{
|
|
"name": "model.layers.20.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 3.6374353840074036e-06,
|
|
"max_abs_err": 0.029296875
|
|
},
|
|
{
|
|
"name": "model.layers.20.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.802538569492754e-06,
|
|
"max_abs_err": 0.02294921875
|
|
},
|
|
{
|
|
"name": "model.layers.20.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 3.893324901582673e-06,
|
|
"max_abs_err": 0.064453125
|
|
},
|
|
{
|
|
"name": "model.layers.20.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.4322032408672385e-06,
|
|
"max_abs_err": 0.06640625
|
|
},
|
|
{
|
|
"name": "model.layers.20.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.5625320126418956e-06,
|
|
"max_abs_err": 0.05078125
|
|
},
|
|
{
|
|
"name": "model.layers.20.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.26160931965569e-06,
|
|
"max_abs_err": 0.0546875
|
|
},
|
|
{
|
|
"name": "model.layers.21.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.5007700464338996e-06,
|
|
"max_abs_err": 0.046875
|
|
},
|
|
{
|
|
"name": "model.layers.21.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 3.6221692880644696e-06,
|
|
"max_abs_err": 0.029296875
|
|
},
|
|
{
|
|
"name": "model.layers.21.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.659023488784442e-06,
|
|
"max_abs_err": 0.0361328125
|
|
},
|
|
{
|
|
"name": "model.layers.21.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 3.7779172998853028e-06,
|
|
"max_abs_err": 0.0302734375
|
|
},
|
|
{
|
|
"name": "model.layers.21.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.27874283559504e-06,
|
|
"max_abs_err": 0.052734375
|
|
},
|
|
{
|
|
"name": "model.layers.21.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.5363094614003785e-06,
|
|
"max_abs_err": 0.04296875
|
|
},
|
|
{
|
|
"name": "model.layers.21.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.3228542381257284e-06,
|
|
"max_abs_err": 0.080078125
|
|
},
|
|
{
|
|
"name": "model.layers.22.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.394257302919868e-06,
|
|
"max_abs_err": 0.04296875
|
|
},
|
|
{
|
|
"name": "model.layers.22.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 3.3786041058192495e-06,
|
|
"max_abs_err": 0.10546875
|
|
},
|
|
{
|
|
"name": "model.layers.22.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.624628672900144e-06,
|
|
"max_abs_err": 0.02685546875
|
|
},
|
|
{
|
|
"name": "model.layers.22.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 3.936739176424453e-06,
|
|
"max_abs_err": 0.0546875
|
|
},
|
|
{
|
|
"name": "model.layers.22.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.576571427605813e-06,
|
|
"max_abs_err": 0.056640625
|
|
},
|
|
{
|
|
"name": "model.layers.22.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.744350917462725e-06,
|
|
"max_abs_err": 0.0361328125
|
|
},
|
|
{
|
|
"name": "model.layers.22.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.359742433734937e-06,
|
|
"max_abs_err": 0.0546875
|
|
},
|
|
{
|
|
"name": "model.layers.23.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.976443960913457e-06,
|
|
"max_abs_err": 0.0400390625
|
|
},
|
|
{
|
|
"name": "model.layers.23.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 3.7672120924980845e-06,
|
|
"max_abs_err": 0.0283203125
|
|
},
|
|
{
|
|
"name": "model.layers.23.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.793057996721473e-06,
|
|
"max_abs_err": 0.025390625
|
|
},
|
|
{
|
|
"name": "model.layers.23.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.230628746881848e-06,
|
|
"max_abs_err": 0.0478515625
|
|
},
|
|
{
|
|
"name": "model.layers.23.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.927570898871636e-06,
|
|
"max_abs_err": 0.068359375
|
|
},
|
|
{
|
|
"name": "model.layers.23.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.7841022023931146e-06,
|
|
"max_abs_err": 0.03515625
|
|
},
|
|
{
|
|
"name": "model.layers.23.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.427899511938449e-06,
|
|
"max_abs_err": 0.0859375
|
|
},
|
|
{
|
|
"name": "model.layers.24.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 5.052277629147284e-06,
|
|
"max_abs_err": 0.04296875
|
|
},
|
|
{
|
|
"name": "model.layers.24.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 3.7488789530470967e-06,
|
|
"max_abs_err": 0.02783203125
|
|
},
|
|
{
|
|
"name": "model.layers.24.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.695091774920002e-06,
|
|
"max_abs_err": 0.03515625
|
|
},
|
|
{
|
|
"name": "model.layers.24.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.3748873395088594e-06,
|
|
"max_abs_err": 0.0517578125
|
|
},
|
|
{
|
|
"name": "model.layers.24.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 5.124813469592482e-06,
|
|
"max_abs_err": 0.04296875
|
|
},
|
|
{
|
|
"name": "model.layers.24.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.8177194003073964e-06,
|
|
"max_abs_err": 0.0380859375
|
|
},
|
|
{
|
|
"name": "model.layers.24.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.5179649532656185e-06,
|
|
"max_abs_err": 0.046875
|
|
},
|
|
{
|
|
"name": "model.layers.25.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.509666723606642e-06,
|
|
"max_abs_err": 0.04296875
|
|
},
|
|
{
|
|
"name": "model.layers.25.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.015783815702889e-06,
|
|
"max_abs_err": 0.0234375
|
|
},
|
|
{
|
|
"name": "model.layers.25.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 5.1402475946815684e-06,
|
|
"max_abs_err": 0.02587890625
|
|
},
|
|
{
|
|
"name": "model.layers.25.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.262743459548801e-06,
|
|
"max_abs_err": 0.029296875
|
|
},
|
|
{
|
|
"name": "model.layers.25.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 5.243239684205037e-06,
|
|
"max_abs_err": 0.0478515625
|
|
},
|
|
{
|
|
"name": "model.layers.25.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.824907591682859e-06,
|
|
"max_abs_err": 0.05078125
|
|
},
|
|
{
|
|
"name": "model.layers.25.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.564091341308085e-06,
|
|
"max_abs_err": 0.068359375
|
|
},
|
|
{
|
|
"name": "model.layers.26.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.6665986701555084e-06,
|
|
"max_abs_err": 0.04296875
|
|
},
|
|
{
|
|
"name": "model.layers.26.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 3.902922344423132e-06,
|
|
"max_abs_err": 0.0244140625
|
|
},
|
|
{
|
|
"name": "model.layers.26.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 5.062448963144561e-06,
|
|
"max_abs_err": 0.02490234375
|
|
},
|
|
{
|
|
"name": "model.layers.26.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.452941993804416e-06,
|
|
"max_abs_err": 0.0322265625
|
|
},
|
|
{
|
|
"name": "model.layers.26.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 5.263592356641311e-06,
|
|
"max_abs_err": 0.0458984375
|
|
},
|
|
{
|
|
"name": "model.layers.26.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.938388883601874e-06,
|
|
"max_abs_err": 0.037109375
|
|
},
|
|
{
|
|
"name": "model.layers.26.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.69050974061247e-06,
|
|
"max_abs_err": 0.0703125
|
|
},
|
|
{
|
|
"name": "model.layers.27.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.592570803652052e-06,
|
|
"max_abs_err": 0.0341796875
|
|
},
|
|
{
|
|
"name": "model.layers.27.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 3.6113390251557576e-06,
|
|
"max_abs_err": 0.01953125
|
|
},
|
|
{
|
|
"name": "model.layers.27.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.759513103635982e-06,
|
|
"max_abs_err": 0.02490234375
|
|
},
|
|
{
|
|
"name": "model.layers.27.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.253867700754199e-06,
|
|
"max_abs_err": 0.0458984375
|
|
},
|
|
{
|
|
"name": "model.layers.27.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 5.2142859203740954e-06,
|
|
"max_abs_err": 0.060546875
|
|
},
|
|
{
|
|
"name": "model.layers.27.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 5.059260274720145e-06,
|
|
"max_abs_err": 0.048828125
|
|
},
|
|
{
|
|
"name": "model.layers.27.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.801653176400578e-06,
|
|
"max_abs_err": 0.0517578125
|
|
},
|
|
{
|
|
"name": "model.layers.28.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.7848834583419375e-06,
|
|
"max_abs_err": 0.0322265625
|
|
},
|
|
{
|
|
"name": "model.layers.28.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 3.8549769669771194e-06,
|
|
"max_abs_err": 0.02880859375
|
|
},
|
|
{
|
|
"name": "model.layers.28.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 5.2966233852203e-06,
|
|
"max_abs_err": 0.021484375
|
|
},
|
|
{
|
|
"name": "model.layers.28.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.2320016291341744e-06,
|
|
"max_abs_err": 0.0380859375
|
|
},
|
|
{
|
|
"name": "model.layers.28.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 5.101563147036359e-06,
|
|
"max_abs_err": 0.0703125
|
|
},
|
|
{
|
|
"name": "model.layers.28.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 5.168639290786814e-06,
|
|
"max_abs_err": 0.04296875
|
|
},
|
|
{
|
|
"name": "model.layers.28.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.9236537051911e-06,
|
|
"max_abs_err": 0.05078125
|
|
},
|
|
{
|
|
"name": "model.layers.29.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 5.249293735687388e-06,
|
|
"max_abs_err": 0.037109375
|
|
},
|
|
{
|
|
"name": "model.layers.29.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 3.6735996218340006e-06,
|
|
"max_abs_err": 0.0283203125
|
|
},
|
|
{
|
|
"name": "model.layers.29.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.809035999642219e-06,
|
|
"max_abs_err": 0.02734375
|
|
},
|
|
{
|
|
"name": "model.layers.29.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.557928150461521e-06,
|
|
"max_abs_err": 0.025390625
|
|
},
|
|
{
|
|
"name": "model.layers.29.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 5.113557108415989e-06,
|
|
"max_abs_err": 0.076171875
|
|
},
|
|
{
|
|
"name": "model.layers.29.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 5.3310809562390205e-06,
|
|
"max_abs_err": 0.056640625
|
|
},
|
|
{
|
|
"name": "model.layers.29.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 5.099047029943904e-06,
|
|
"max_abs_err": 0.056640625
|
|
},
|
|
{
|
|
"name": "model.layers.30.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.427760813996429e-06,
|
|
"max_abs_err": 0.0458984375
|
|
},
|
|
{
|
|
"name": "model.layers.30.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 3.993854534201091e-06,
|
|
"max_abs_err": 0.0302734375
|
|
},
|
|
{
|
|
"name": "model.layers.30.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 5.407737717177952e-06,
|
|
"max_abs_err": 0.03125
|
|
},
|
|
{
|
|
"name": "model.layers.30.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.478546088648727e-06,
|
|
"max_abs_err": 0.029296875
|
|
},
|
|
{
|
|
"name": "model.layers.30.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 5.053174390923232e-06,
|
|
"max_abs_err": 0.087890625
|
|
},
|
|
{
|
|
"name": "model.layers.30.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 5.4569077292399015e-06,
|
|
"max_abs_err": 0.0302734375
|
|
},
|
|
{
|
|
"name": "model.layers.30.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 5.2262294047977775e-06,
|
|
"max_abs_err": 0.064453125
|
|
},
|
|
{
|
|
"name": "model.layers.31.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.614938916347455e-06,
|
|
"max_abs_err": 0.037109375
|
|
},
|
|
{
|
|
"name": "model.layers.31.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 3.666723159767571e-06,
|
|
"max_abs_err": 0.02734375
|
|
},
|
|
{
|
|
"name": "model.layers.31.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 5.836637228640029e-06,
|
|
"max_abs_err": 0.02685546875
|
|
},
|
|
{
|
|
"name": "model.layers.31.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.746147169498727e-06,
|
|
"max_abs_err": 0.0400390625
|
|
},
|
|
{
|
|
"name": "model.layers.31.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.950442416884471e-06,
|
|
"max_abs_err": 0.08203125
|
|
},
|
|
{
|
|
"name": "model.layers.31.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 5.588029125647154e-06,
|
|
"max_abs_err": 0.033203125
|
|
},
|
|
{
|
|
"name": "model.layers.31.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 5.438460448203841e-06,
|
|
"max_abs_err": 0.060546875
|
|
},
|
|
{
|
|
"name": "model.layers.32.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.4495300244307145e-06,
|
|
"max_abs_err": 0.033203125
|
|
},
|
|
{
|
|
"name": "model.layers.32.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 3.711434374054079e-06,
|
|
"max_abs_err": 0.033203125
|
|
},
|
|
{
|
|
"name": "model.layers.32.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 5.803630301670637e-06,
|
|
"max_abs_err": 0.02783203125
|
|
},
|
|
{
|
|
"name": "model.layers.32.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.721639015770052e-06,
|
|
"max_abs_err": 0.0537109375
|
|
},
|
|
{
|
|
"name": "model.layers.32.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.801231170858955e-06,
|
|
"max_abs_err": 0.0546875
|
|
},
|
|
{
|
|
"name": "model.layers.32.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 5.616715043288423e-06,
|
|
"max_abs_err": 0.037109375
|
|
},
|
|
{
|
|
"name": "model.layers.32.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 5.446503109851619e-06,
|
|
"max_abs_err": 0.06640625
|
|
},
|
|
{
|
|
"name": "model.layers.33.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.503608124650782e-06,
|
|
"max_abs_err": 0.0390625
|
|
},
|
|
{
|
|
"name": "model.layers.33.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 2.9984480534039903e-06,
|
|
"max_abs_err": 0.0224609375
|
|
},
|
|
{
|
|
"name": "model.layers.33.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 6.625287824135739e-06,
|
|
"max_abs_err": 0.0244140625
|
|
},
|
|
{
|
|
"name": "model.layers.33.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 5.124221388541628e-06,
|
|
"max_abs_err": 0.048828125
|
|
},
|
|
{
|
|
"name": "model.layers.33.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.703628746938193e-06,
|
|
"max_abs_err": 0.06640625
|
|
},
|
|
{
|
|
"name": "model.layers.33.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 5.658939699060284e-06,
|
|
"max_abs_err": 0.0390625
|
|
},
|
|
{
|
|
"name": "model.layers.33.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 5.3191611186775845e-06,
|
|
"max_abs_err": 0.068359375
|
|
},
|
|
{
|
|
"name": "model.layers.34.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.190060280961916e-06,
|
|
"max_abs_err": 0.0390625
|
|
},
|
|
{
|
|
"name": "model.layers.34.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 3.477332938928157e-06,
|
|
"max_abs_err": 0.029296875
|
|
},
|
|
{
|
|
"name": "model.layers.34.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 6.614848189201439e-06,
|
|
"max_abs_err": 0.03125
|
|
},
|
|
{
|
|
"name": "model.layers.34.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.878462732449407e-06,
|
|
"max_abs_err": 0.048828125
|
|
},
|
|
{
|
|
"name": "model.layers.34.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.8811793931236025e-06,
|
|
"max_abs_err": 0.056640625
|
|
},
|
|
{
|
|
"name": "model.layers.34.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 5.5068662732082885e-06,
|
|
"max_abs_err": 0.060546875
|
|
},
|
|
{
|
|
"name": "model.layers.34.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.733033620141214e-06,
|
|
"max_abs_err": 0.08984375
|
|
},
|
|
{
|
|
"name": "model.layers.35.self_attn.q_proj",
|
|
"shape": [
|
|
4096,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.313304543757113e-06,
|
|
"max_abs_err": 0.037109375
|
|
},
|
|
{
|
|
"name": "model.layers.35.self_attn.k_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 3.962451046390925e-06,
|
|
"max_abs_err": 0.033203125
|
|
},
|
|
{
|
|
"name": "model.layers.35.self_attn.v_proj",
|
|
"shape": [
|
|
1024,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 5.858712938788813e-06,
|
|
"max_abs_err": 0.02392578125
|
|
},
|
|
{
|
|
"name": "model.layers.35.self_attn.o_proj",
|
|
"shape": [
|
|
2560,
|
|
4096
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.573285877995659e-06,
|
|
"max_abs_err": 0.064453125
|
|
},
|
|
{
|
|
"name": "model.layers.35.mlp.gate_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 5.399806923378492e-06,
|
|
"max_abs_err": 0.08984375
|
|
},
|
|
{
|
|
"name": "model.layers.35.mlp.up_proj",
|
|
"shape": [
|
|
9728,
|
|
2560
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 5.665603566740174e-06,
|
|
"max_abs_err": 0.0625
|
|
},
|
|
{
|
|
"name": "model.layers.35.mlp.down_proj",
|
|
"shape": [
|
|
2560,
|
|
9728
|
|
],
|
|
"dtype": "bfloat16",
|
|
"mse": 4.272512796887895e-06,
|
|
"max_abs_err": 0.076171875
|
|
}
|
|
],
|
|
"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."
|
|
}
|