{ "method": "HQQ-style-FP4", "format": "fp4-e3m0", "nbits": 4, "group_size": 128, "axis": 1, "round_zero": false, "codebook": "fp4-e3m0", "block_size": 128, "scale_opt_steps": 8, "optimize": true, "optimize_device": "cuda", "num_modules": 252, "num_weights": 3633315840, "elapsed_sec": 0.914, "quantized_modules": [ { "name": "model.layers.0.self_attn.q_proj", "shape": [ 4096, 2560 ], "dtype": "bfloat16", "mse": 1.719890678941738e-05, "max_abs_err": 0.1171875 }, { "name": "model.layers.0.self_attn.k_proj", "shape": [ 1024, 2560 ], "dtype": "bfloat16", "mse": 1.9118719137622975e-05, "max_abs_err": 0.0625 }, { "name": "model.layers.0.self_attn.v_proj", "shape": [ 1024, 2560 ], "dtype": "bfloat16", "mse": 1.6793477698229253e-05, "max_abs_err": 0.0322265625 }, { "name": "model.layers.0.self_attn.o_proj", "shape": [ 2560, 4096 ], "dtype": "bfloat16", "mse": 1.5022829757072031e-05, "max_abs_err": 0.107421875 }, { "name": "model.layers.0.mlp.gate_proj", "shape": [ 9728, 2560 ], "dtype": "bfloat16", "mse": 1.7813934391597286e-05, "max_abs_err": 0.087890625 }, { "name": "model.layers.0.mlp.up_proj", "shape": [ 9728, 2560 ], "dtype": "bfloat16", "mse": 1.5537463696091436e-05, "max_abs_err": 0.05517578125 }, { "name": "model.layers.0.mlp.down_proj", "shape": [ 2560, 9728 ], "dtype": "bfloat16", "mse": 1.716708720778115e-05, "max_abs_err": 0.1123046875 }, { "name": "model.layers.1.self_attn.q_proj", "shape": [ 4096, 2560 ], "dtype": "bfloat16", "mse": 1.4880002709105611e-05, "max_abs_err": 0.0791015625 }, { "name": "model.layers.1.self_attn.k_proj", "shape": [ 1024, 2560 ], "dtype": "bfloat16", "mse": 1.6146706911968067e-05, "max_abs_err": 0.0400390625 }, { "name": "model.layers.1.self_attn.v_proj", "shape": [ 1024, 2560 ], "dtype": "bfloat16", "mse": 1.729782707116101e-05, "max_abs_err": 0.033203125 }, { "name": "model.layers.1.self_attn.o_proj", "shape": [ 2560, 4096 ], "dtype": "bfloat16", "mse": 1.5409294064738788e-05, "max_abs_err": 0.091796875 }, { "name": "model.layers.1.mlp.gate_proj", "shape": [ 9728, 2560 ], "dtype": "bfloat16", "mse": 1.6731833966332488e-05, "max_abs_err": 0.0517578125 }, { "name": "model.layers.1.mlp.up_proj", "shape": [ 9728, 2560 ], "dtype": "bfloat16", "mse": 7.637258022441529e-06, "max_abs_err": 0.05859375 }, { "name": "model.layers.1.mlp.down_proj", "shape": [ 2560, 9728 ], "dtype": "bfloat16", "mse": 8.50557626108639e-06, "max_abs_err": 0.15234375 }, { "name": "model.layers.2.self_attn.q_proj", "shape": [ 4096, 2560 ], "dtype": "bfloat16", "mse": 1.5947609426802956e-05, "max_abs_err": 0.06640625 }, { "name": "model.layers.2.self_attn.k_proj", "shape": [ 1024, 2560 ], "dtype": "bfloat16", "mse": 1.7001228115987033e-05, "max_abs_err": 0.056640625 }, { "name": "model.layers.2.self_attn.v_proj", "shape": [ 1024, 2560 ], "dtype": "bfloat16", "mse": 1.8633050785865635e-05, "max_abs_err": 0.03515625 }, { "name": "model.layers.2.self_attn.o_proj", "shape": [ 2560, 4096 ], "dtype": "bfloat16", "mse": 1.614650318515487e-05, "max_abs_err": 0.06640625 }, { "name": "model.layers.2.mlp.gate_proj", "shape": [ 9728, 2560 ], "dtype": "bfloat16", "mse": 1.1282760169706307e-05, "max_abs_err": 0.04931640625 }, { "name": "model.layers.2.mlp.up_proj", "shape": [ 9728, 2560 ], "dtype": "bfloat16", "mse": 9.478022548137233e-06, "max_abs_err": 0.09375 }, { "name": "model.layers.2.mlp.down_proj", "shape": [ 2560, 9728 ], "dtype": "bfloat16", "mse": 8.1053976828116e-06, "max_abs_err": 0.1015625 }, { "name": "model.layers.3.self_attn.q_proj", "shape": [ 4096, 2560 ], "dtype": "bfloat16", "mse": 1.6014668290154077e-05, "max_abs_err": 0.078125 }, { "name": "model.layers.3.self_attn.k_proj", "shape": [ 1024, 2560 ], "dtype": "bfloat16", "mse": 1.598071503394749e-05, "max_abs_err": 0.044921875 }, { "name": "model.layers.3.self_attn.v_proj", "shape": [ 1024, 2560 ], "dtype": "bfloat16", "mse": 1.8832359273801558e-05, "max_abs_err": 0.029541015625 }, { "name": "model.layers.3.self_attn.o_proj", "shape": [ 2560, 4096 ], "dtype": "bfloat16", "mse": 1.634876934986096e-05, "max_abs_err": 0.0859375 }, { "name": "model.layers.3.mlp.gate_proj", "shape": [ 9728, 2560 ], "dtype": "bfloat16", "mse": 2.0244984625605866e-05, "max_abs_err": 0.08203125 }, { "name": "model.layers.3.mlp.up_proj", "shape": [ 9728, 2560 ], "dtype": "bfloat16", "mse": 1.0916028259089217e-05, "max_abs_err": 0.0654296875 }, { "name": "model.layers.3.mlp.down_proj", "shape": [ 2560, 9728 ], "dtype": "bfloat16", "mse": 1.1440579328336753e-05, "max_abs_err": 0.1171875 }, { "name": "model.layers.4.self_attn.q_proj", "shape": [ 4096, 2560 ], "dtype": "bfloat16", "mse": 1.659669214859605e-05, "max_abs_err": 0.0625 }, { "name": "model.layers.4.self_attn.k_proj", "shape": [ 1024, 2560 ], "dtype": "bfloat16", "mse": 1.6826257706270553e-05, "max_abs_err": 0.099609375 }, { "name": "model.layers.4.self_attn.v_proj", "shape": [ 1024, 2560 ], "dtype": "bfloat16", "mse": 1.9200519091100432e-05, "max_abs_err": 0.0322265625 }, { "name": "model.layers.4.self_attn.o_proj", "shape": [ 2560, 4096 ], "dtype": "bfloat16", "mse": 1.710886681394186e-05, "max_abs_err": 0.0888671875 }, { "name": "model.layers.4.mlp.gate_proj", "shape": [ 9728, 2560 ], "dtype": "bfloat16", "mse": 3.14745084324386e-05, "max_abs_err": 0.0595703125 }, { "name": "model.layers.4.mlp.up_proj", "shape": [ 9728, 2560 ], "dtype": "bfloat16", "mse": 1.3255451449367683e-05, "max_abs_err": 0.0576171875 }, { "name": "model.layers.4.mlp.down_proj", "shape": [ 2560, 9728 ], "dtype": "bfloat16", "mse": 1.3456040505843703e-05, "max_abs_err": 0.11328125 }, { "name": "model.layers.5.self_attn.q_proj", "shape": [ 4096, 2560 ], "dtype": "bfloat16", "mse": 1.6459682228742167e-05, "max_abs_err": 0.076171875 }, { "name": "model.layers.5.self_attn.k_proj", "shape": [ 1024, 2560 ], "dtype": "bfloat16", "mse": 1.6336600310751237e-05, "max_abs_err": 0.0634765625 }, { "name": "model.layers.5.self_attn.v_proj", "shape": [ 1024, 2560 ], "dtype": "bfloat16", "mse": 1.894819433800876e-05, "max_abs_err": 0.03515625 }, { "name": "model.layers.5.self_attn.o_proj", "shape": [ 2560, 4096 ], "dtype": "bfloat16", "mse": 1.592212720424868e-05, "max_abs_err": 0.09765625 }, { "name": "model.layers.5.mlp.gate_proj", "shape": [ 9728, 2560 ], "dtype": "bfloat16", "mse": 2.733454675762914e-05, "max_abs_err": 0.08203125 }, { "name": "model.layers.5.mlp.up_proj", "shape": [ 9728, 2560 ], "dtype": "bfloat16", "mse": 1.4910734535078518e-05, "max_abs_err": 0.078125 }, { "name": "model.layers.5.mlp.down_proj", "shape": [ 2560, 9728 ], "dtype": "bfloat16", "mse": 1.4434205695579294e-05, "max_abs_err": 0.12109375 }, { "name": "model.layers.6.self_attn.q_proj", "shape": [ 4096, 2560 ], "dtype": "bfloat16", "mse": 1.7389325876138173e-05, "max_abs_err": 0.0576171875 }, { "name": "model.layers.6.self_attn.k_proj", "shape": [ 1024, 2560 ], "dtype": "bfloat16", "mse": 1.645899646973703e-05, "max_abs_err": 0.0576171875 }, { "name": "model.layers.6.self_attn.v_proj", "shape": [ 1024, 2560 ], "dtype": "bfloat16", "mse": 1.8835136870620772e-05, "max_abs_err": 0.05078125 }, { "name": "model.layers.6.self_attn.o_proj", "shape": [ 2560, 4096 ], "dtype": "bfloat16", "mse": 1.6883930584299378e-05, "max_abs_err": 0.119140625 }, { "name": "model.layers.6.mlp.gate_proj", "shape": [ 9728, 2560 ], "dtype": "bfloat16", "mse": 2.4390483304159716e-05, "max_abs_err": 0.068359375 }, { "name": "model.layers.6.mlp.up_proj", "shape": [ 9728, 2560 ], "dtype": "bfloat16", "mse": 1.577007060404867e-05, "max_abs_err": 0.11328125 }, { "name": "model.layers.6.mlp.down_proj", "shape": [ 2560, 9728 ], "dtype": "bfloat16", "mse": 1.5560388419544324e-05, "max_abs_err": 0.1025390625 }, { "name": "model.layers.7.self_attn.q_proj", "shape": [ 4096, 2560 ], "dtype": "bfloat16", "mse": 1.6432099073426798e-05, "max_abs_err": 0.087890625 }, { "name": "model.layers.7.self_attn.k_proj", "shape": [ 1024, 2560 ], "dtype": "bfloat16", "mse": 1.6453714124509133e-05, "max_abs_err": 0.0595703125 }, { "name": "model.layers.7.self_attn.v_proj", "shape": [ 1024, 2560 ], "dtype": "bfloat16", "mse": 2.170585503336042e-05, "max_abs_err": 0.03173828125 }, { "name": "model.layers.7.self_attn.o_proj", "shape": [ 2560, 4096 ], "dtype": "bfloat16", "mse": 1.6572452295804396e-05, "max_abs_err": 0.109375 }, { "name": "model.layers.7.mlp.gate_proj", "shape": [ 9728, 2560 ], "dtype": "bfloat16", "mse": 2.3831778889871202e-05, "max_abs_err": 0.083984375 }, { "name": "model.layers.7.mlp.up_proj", "shape": [ 9728, 2560 ], "dtype": "bfloat16", "mse": 1.5985622667358257e-05, "max_abs_err": 0.087890625 }, { "name": "model.layers.7.mlp.down_proj", "shape": [ 2560, 9728 ], "dtype": "bfloat16", "mse": 1.5843241271795705e-05, "max_abs_err": 0.109375 }, { "name": "model.layers.8.self_attn.q_proj", "shape": [ 4096, 2560 ], "dtype": "bfloat16", "mse": 1.6639822206343524e-05, "max_abs_err": 0.0625 }, { "name": "model.layers.8.self_attn.k_proj", "shape": [ 1024, 2560 ], "dtype": "bfloat16", "mse": 1.7067113731172867e-05, "max_abs_err": 0.060546875 }, { "name": "model.layers.8.self_attn.v_proj", "shape": [ 1024, 2560 ], "dtype": "bfloat16", "mse": 2.1615083824144676e-05, "max_abs_err": 0.02978515625 }, { "name": "model.layers.8.self_attn.o_proj", "shape": [ 2560, 4096 ], "dtype": "bfloat16", "mse": 1.735561090754345e-05, "max_abs_err": 0.064453125 }, { "name": "model.layers.8.mlp.gate_proj", "shape": [ 9728, 2560 ], "dtype": "bfloat16", "mse": 2.0730678443214856e-05, "max_abs_err": 0.087890625 }, { "name": "model.layers.8.mlp.up_proj", "shape": [ 9728, 2560 ], "dtype": "bfloat16", "mse": 1.673734914220404e-05, "max_abs_err": 0.0830078125 }, { "name": "model.layers.8.mlp.down_proj", "shape": [ 2560, 9728 ], "dtype": "bfloat16", "mse": 1.6755271644797176e-05, "max_abs_err": 0.09765625 }, { "name": "model.layers.9.self_attn.q_proj", "shape": [ 4096, 2560 ], "dtype": "bfloat16", "mse": 1.5801730114617385e-05, "max_abs_err": 0.11328125 }, { "name": "model.layers.9.self_attn.k_proj", "shape": [ 1024, 2560 ], "dtype": "bfloat16", "mse": 1.5767296645208262e-05, "max_abs_err": 0.09375 }, { "name": "model.layers.9.self_attn.v_proj", "shape": [ 1024, 2560 ], "dtype": "bfloat16", "mse": 2.2033251298125833e-05, "max_abs_err": 0.03662109375 }, { "name": "model.layers.9.self_attn.o_proj", "shape": [ 2560, 4096 ], "dtype": "bfloat16", "mse": 1.706122566247359e-05, "max_abs_err": 0.091796875 }, { "name": "model.layers.9.mlp.gate_proj", "shape": [ 9728, 2560 ], "dtype": "bfloat16", "mse": 2.425502498226706e-05, "max_abs_err": 0.1015625 }, { "name": "model.layers.9.mlp.up_proj", "shape": [ 9728, 2560 ], "dtype": "bfloat16", "mse": 1.5369872926385142e-05, "max_abs_err": 0.06640625 }, { "name": "model.layers.9.mlp.down_proj", "shape": [ 2560, 9728 ], "dtype": "bfloat16", "mse": 1.5004680790298153e-05, "max_abs_err": 0.12109375 }, { "name": "model.layers.10.self_attn.q_proj", "shape": [ 4096, 2560 ], "dtype": "bfloat16", "mse": 1.6874904758878984e-05, "max_abs_err": 0.05859375 }, { "name": "model.layers.10.self_attn.k_proj", "shape": [ 1024, 2560 ], "dtype": "bfloat16", "mse": 1.6799225704744458e-05, "max_abs_err": 0.0546875 }, { "name": "model.layers.10.self_attn.v_proj", "shape": [ 1024, 2560 ], "dtype": "bfloat16", "mse": 2.1573048798018135e-05, "max_abs_err": 0.03857421875 }, { "name": "model.layers.10.self_attn.o_proj", "shape": [ 2560, 4096 ], "dtype": "bfloat16", "mse": 1.768007132341154e-05, "max_abs_err": 0.125 }, { "name": "model.layers.10.mlp.gate_proj", "shape": [ 9728, 2560 ], "dtype": "bfloat16", "mse": 2.3180533389677294e-05, "max_abs_err": 0.07421875 }, { "name": "model.layers.10.mlp.up_proj", "shape": [ 9728, 2560 ], "dtype": "bfloat16", "mse": 1.598827657289803e-05, "max_abs_err": 0.0810546875 }, { "name": "model.layers.10.mlp.down_proj", "shape": [ 2560, 9728 ], "dtype": "bfloat16", "mse": 1.5775680367369205e-05, "max_abs_err": 0.14453125 }, { "name": "model.layers.11.self_attn.q_proj", "shape": [ 4096, 2560 ], "dtype": "bfloat16", "mse": 1.6782150851213373e-05, "max_abs_err": 0.068359375 }, { "name": "model.layers.11.self_attn.k_proj", "shape": [ 1024, 2560 ], "dtype": "bfloat16", "mse": 1.7268477677134797e-05, "max_abs_err": 0.083984375 }, { "name": "model.layers.11.self_attn.v_proj", "shape": [ 1024, 2560 ], "dtype": "bfloat16", "mse": 2.2072868887335062e-05, "max_abs_err": 0.03076171875 }, { "name": "model.layers.11.self_attn.o_proj", "shape": [ 2560, 4096 ], "dtype": "bfloat16", "mse": 1.7301174011663534e-05, "max_abs_err": 0.07421875 }, { "name": "model.layers.11.mlp.gate_proj", "shape": [ 9728, 2560 ], "dtype": "bfloat16", "mse": 2.172264430555515e-05, "max_abs_err": 0.109375 }, { "name": "model.layers.11.mlp.up_proj", "shape": [ 9728, 2560 ], "dtype": "bfloat16", "mse": 1.6930407582549378e-05, "max_abs_err": 0.0771484375 }, { "name": "model.layers.11.mlp.down_proj", "shape": [ 2560, 9728 ], "dtype": "bfloat16", "mse": 1.6510774003108963e-05, "max_abs_err": 0.142578125 }, { "name": "model.layers.12.self_attn.q_proj", "shape": [ 4096, 2560 ], "dtype": "bfloat16", "mse": 1.7128559193224646e-05, "max_abs_err": 0.060546875 }, { "name": "model.layers.12.self_attn.k_proj", "shape": [ 1024, 2560 ], "dtype": "bfloat16", "mse": 1.6684743968653493e-05, "max_abs_err": 0.1025390625 }, { "name": "model.layers.12.self_attn.v_proj", "shape": [ 1024, 2560 ], "dtype": "bfloat16", "mse": 2.1708199710701592e-05, "max_abs_err": 0.04296875 }, { "name": "model.layers.12.self_attn.o_proj", "shape": [ 2560, 4096 ], "dtype": "bfloat16", "mse": 1.6969233911368065e-05, "max_abs_err": 0.19140625 }, { "name": "model.layers.12.mlp.gate_proj", "shape": [ 9728, 2560 ], "dtype": "bfloat16", "mse": 2.0589332052622922e-05, "max_abs_err": 0.091796875 }, { "name": "model.layers.12.mlp.up_proj", "shape": [ 9728, 2560 ], "dtype": "bfloat16", "mse": 1.7654972907621413e-05, "max_abs_err": 0.0830078125 }, { "name": "model.layers.12.mlp.down_proj", "shape": [ 2560, 9728 ], "dtype": "bfloat16", "mse": 1.6883001080714166e-05, "max_abs_err": 0.13671875 }, { "name": "model.layers.13.self_attn.q_proj", "shape": [ 4096, 2560 ], "dtype": "bfloat16", "mse": 1.693702688498888e-05, "max_abs_err": 0.05859375 }, { "name": "model.layers.13.self_attn.k_proj", "shape": [ 1024, 2560 ], "dtype": "bfloat16", "mse": 1.6465170119772665e-05, "max_abs_err": 0.080078125 }, { "name": "model.layers.13.self_attn.v_proj", "shape": [ 1024, 2560 ], "dtype": "bfloat16", "mse": 2.0833480448345654e-05, "max_abs_err": 0.04345703125 }, { "name": "model.layers.13.self_attn.o_proj", "shape": [ 2560, 4096 ], "dtype": "bfloat16", "mse": 1.6220859833993018e-05, "max_abs_err": 0.0771484375 }, { "name": "model.layers.13.mlp.gate_proj", "shape": [ 9728, 2560 ], "dtype": "bfloat16", "mse": 1.918758061947301e-05, "max_abs_err": 0.1064453125 }, { "name": "model.layers.13.mlp.up_proj", "shape": [ 9728, 2560 ], "dtype": "bfloat16", "mse": 1.8108754375134595e-05, "max_abs_err": 0.109375 }, { "name": "model.layers.13.mlp.down_proj", "shape": [ 2560, 9728 ], "dtype": "bfloat16", "mse": 1.7016274796333164e-05, "max_abs_err": 0.15234375 }, { "name": "model.layers.14.self_attn.q_proj", "shape": [ 4096, 2560 ], "dtype": "bfloat16", "mse": 1.698016785667278e-05, "max_abs_err": 0.0576171875 }, { "name": "model.layers.14.self_attn.k_proj", "shape": [ 1024, 2560 ], "dtype": "bfloat16", "mse": 1.679834349488374e-05, "max_abs_err": 0.0537109375 }, { "name": "model.layers.14.self_attn.v_proj", "shape": [ 1024, 2560 ], "dtype": "bfloat16", "mse": 1.969084951269906e-05, "max_abs_err": 0.03173828125 }, { "name": "model.layers.14.self_attn.o_proj", "shape": [ 2560, 4096 ], "dtype": "bfloat16", "mse": 1.5758592780912295e-05, "max_abs_err": 0.103515625 }, { "name": "model.layers.14.mlp.gate_proj", "shape": [ 9728, 2560 ], "dtype": "bfloat16", "mse": 1.8653519873623736e-05, "max_abs_err": 0.099609375 }, { "name": "model.layers.14.mlp.up_proj", "shape": [ 9728, 2560 ], "dtype": "bfloat16", "mse": 1.759439692250453e-05, "max_abs_err": 0.064453125 }, { "name": "model.layers.14.mlp.down_proj", "shape": [ 2560, 9728 ], "dtype": "bfloat16", "mse": 1.663642979110591e-05, "max_abs_err": 0.123046875 }, { "name": "model.layers.15.self_attn.q_proj", "shape": [ 4096, 2560 ], "dtype": "bfloat16", "mse": 1.7361626305500977e-05, "max_abs_err": 0.072265625 }, { "name": "model.layers.15.self_attn.k_proj", "shape": [ 1024, 2560 ], "dtype": "bfloat16", "mse": 1.6651087207719684e-05, "max_abs_err": 0.08203125 }, { "name": "model.layers.15.self_attn.v_proj", "shape": [ 1024, 2560 ], "dtype": "bfloat16", "mse": 1.9272689314675517e-05, "max_abs_err": 0.03955078125 }, { "name": "model.layers.15.self_attn.o_proj", "shape": [ 2560, 4096 ], "dtype": "bfloat16", "mse": 1.4866357560094912e-05, "max_abs_err": 0.06640625 }, { "name": "model.layers.15.mlp.gate_proj", "shape": [ 9728, 2560 ], "dtype": "bfloat16", "mse": 1.737766979204025e-05, "max_abs_err": 0.125 }, { "name": "model.layers.15.mlp.up_proj", "shape": [ 9728, 2560 ], "dtype": "bfloat16", "mse": 1.720309410302434e-05, "max_abs_err": 0.0810546875 }, { "name": "model.layers.15.mlp.down_proj", "shape": [ 2560, 9728 ], "dtype": "bfloat16", "mse": 1.6329977370332927e-05, "max_abs_err": 0.115234375 }, { "name": "model.layers.16.self_attn.q_proj", "shape": [ 4096, 2560 ], "dtype": "bfloat16", "mse": 1.6376508938265033e-05, "max_abs_err": 0.0673828125 }, { "name": "model.layers.16.self_attn.k_proj", "shape": [ 1024, 2560 ], "dtype": "bfloat16", "mse": 1.5598416212014854e-05, "max_abs_err": 0.05126953125 }, { "name": "model.layers.16.self_attn.v_proj", "shape": [ 1024, 2560 ], "dtype": "bfloat16", "mse": 1.931911174324341e-05, "max_abs_err": 0.033203125 }, { "name": "model.layers.16.self_attn.o_proj", "shape": [ 2560, 4096 ], "dtype": "bfloat16", "mse": 1.5727329810033552e-05, "max_abs_err": 0.15625 }, { "name": "model.layers.16.mlp.gate_proj", "shape": [ 9728, 2560 ], "dtype": "bfloat16", "mse": 1.8253442249260843e-05, "max_abs_err": 0.1328125 }, { "name": "model.layers.16.mlp.up_proj", "shape": [ 9728, 2560 ], "dtype": "bfloat16", "mse": 1.6971369404927827e-05, "max_abs_err": 0.169921875 }, { "name": "model.layers.16.mlp.down_proj", "shape": [ 2560, 9728 ], "dtype": "bfloat16", "mse": 1.5798255844856612e-05, "max_abs_err": 0.15625 }, { "name": "model.layers.17.self_attn.q_proj", "shape": [ 4096, 2560 ], "dtype": "bfloat16", "mse": 1.7241725799976848e-05, "max_abs_err": 0.068359375 }, { "name": "model.layers.17.self_attn.k_proj", "shape": [ 1024, 2560 ], "dtype": "bfloat16", "mse": 1.4724444554303773e-05, "max_abs_err": 0.05078125 }, { "name": "model.layers.17.self_attn.v_proj", "shape": [ 1024, 2560 ], "dtype": "bfloat16", "mse": 1.84071122930618e-05, "max_abs_err": 0.04296875 }, { "name": "model.layers.17.self_attn.o_proj", "shape": [ 2560, 4096 ], "dtype": "bfloat16", "mse": 1.5188844372460153e-05, "max_abs_err": 0.11328125 }, { "name": "model.layers.17.mlp.gate_proj", "shape": [ 9728, 2560 ], "dtype": "bfloat16", "mse": 1.7732727428665385e-05, "max_abs_err": 0.076171875 }, { "name": "model.layers.17.mlp.up_proj", "shape": [ 9728, 2560 ], "dtype": "bfloat16", "mse": 1.7098542230087332e-05, "max_abs_err": 0.05859375 }, { "name": "model.layers.17.mlp.down_proj", "shape": [ 2560, 9728 ], "dtype": "bfloat16", "mse": 1.5923718819976784e-05, "max_abs_err": 0.12890625 }, { "name": "model.layers.18.self_attn.q_proj", "shape": [ 4096, 2560 ], "dtype": "bfloat16", "mse": 1.6095396858872846e-05, "max_abs_err": 0.0654296875 }, { "name": "model.layers.18.self_attn.k_proj", "shape": [ 1024, 2560 ], "dtype": "bfloat16", "mse": 1.509094909124542e-05, "max_abs_err": 0.0537109375 }, { "name": "model.layers.18.self_attn.v_proj", "shape": [ 1024, 2560 ], "dtype": "bfloat16", "mse": 1.915367101901211e-05, "max_abs_err": 0.03466796875 }, { "name": "model.layers.18.self_attn.o_proj", "shape": [ 2560, 4096 ], "dtype": "bfloat16", "mse": 1.522593720437726e-05, "max_abs_err": 0.09765625 }, { "name": "model.layers.18.mlp.gate_proj", "shape": [ 9728, 2560 ], "dtype": "bfloat16", "mse": 1.7290623873122968e-05, "max_abs_err": 0.103515625 }, { "name": "model.layers.18.mlp.up_proj", "shape": [ 9728, 2560 ], "dtype": "bfloat16", "mse": 1.7067102817236446e-05, "max_abs_err": 0.06640625 }, { "name": "model.layers.18.mlp.down_proj", "shape": [ 2560, 9728 ], "dtype": "bfloat16", "mse": 1.5870682545937598e-05, "max_abs_err": 0.1083984375 }, { "name": "model.layers.19.self_attn.q_proj", "shape": [ 4096, 2560 ], "dtype": "bfloat16", "mse": 1.660223279031925e-05, "max_abs_err": 0.0673828125 }, { "name": "model.layers.19.self_attn.k_proj", "shape": [ 1024, 2560 ], "dtype": "bfloat16", "mse": 1.399809025315335e-05, "max_abs_err": 0.095703125 }, { "name": "model.layers.19.self_attn.v_proj", "shape": [ 1024, 2560 ], "dtype": "bfloat16", "mse": 1.827698179113213e-05, "max_abs_err": 0.0341796875 }, { "name": "model.layers.19.self_attn.o_proj", "shape": [ 2560, 4096 ], "dtype": "bfloat16", "mse": 1.543254984426312e-05, "max_abs_err": 0.111328125 }, { "name": "model.layers.19.mlp.gate_proj", "shape": [ 9728, 2560 ], "dtype": "bfloat16", "mse": 1.707305091258604e-05, "max_abs_err": 0.109375 }, { "name": "model.layers.19.mlp.up_proj", "shape": [ 9728, 2560 ], "dtype": "bfloat16", "mse": 1.7165415556519292e-05, "max_abs_err": 0.0673828125 }, { "name": "model.layers.19.mlp.down_proj", "shape": [ 2560, 9728 ], "dtype": "bfloat16", "mse": 1.584030360390898e-05, "max_abs_err": 0.162109375 }, { "name": "model.layers.20.self_attn.q_proj", "shape": [ 4096, 2560 ], "dtype": "bfloat16", "mse": 1.781398350431118e-05, "max_abs_err": 0.0751953125 }, { "name": "model.layers.20.self_attn.k_proj", "shape": [ 1024, 2560 ], "dtype": "bfloat16", "mse": 1.3817766557622235e-05, "max_abs_err": 0.04833984375 }, { "name": "model.layers.20.self_attn.v_proj", "shape": [ 1024, 2560 ], "dtype": "bfloat16", "mse": 1.817273187043611e-05, "max_abs_err": 0.0341796875 }, { "name": "model.layers.20.self_attn.o_proj", "shape": [ 2560, 4096 ], "dtype": "bfloat16", "mse": 1.4749757610843517e-05, "max_abs_err": 0.078125 }, { "name": "model.layers.20.mlp.gate_proj", "shape": [ 9728, 2560 ], "dtype": "bfloat16", "mse": 1.675738167250529e-05, "max_abs_err": 0.109375 }, { "name": "model.layers.20.mlp.up_proj", "shape": [ 9728, 2560 ], "dtype": "bfloat16", "mse": 1.7268621377297677e-05, "max_abs_err": 0.080078125 }, { "name": "model.layers.20.mlp.down_proj", "shape": [ 2560, 9728 ], "dtype": "bfloat16", "mse": 1.6121419321279973e-05, "max_abs_err": 0.0849609375 }, { "name": "model.layers.21.self_attn.q_proj", "shape": [ 4096, 2560 ], "dtype": "bfloat16", "mse": 1.703898851701524e-05, "max_abs_err": 0.0654296875 }, { "name": "model.layers.21.self_attn.k_proj", "shape": [ 1024, 2560 ], "dtype": "bfloat16", "mse": 1.370652171317488e-05, "max_abs_err": 0.04296875 }, { "name": "model.layers.21.self_attn.v_proj", "shape": [ 1024, 2560 ], "dtype": "bfloat16", "mse": 1.761260864441283e-05, "max_abs_err": 0.0546875 }, { "name": "model.layers.21.self_attn.o_proj", "shape": [ 2560, 4096 ], "dtype": "bfloat16", "mse": 1.4290511899162084e-05, "max_abs_err": 0.0615234375 }, { "name": "model.layers.21.mlp.gate_proj", "shape": [ 9728, 2560 ], "dtype": "bfloat16", "mse": 1.618803435121663e-05, "max_abs_err": 0.08984375 }, { "name": "model.layers.21.mlp.up_proj", "shape": [ 9728, 2560 ], "dtype": "bfloat16", "mse": 1.717154918878805e-05, "max_abs_err": 0.060546875 }, { "name": "model.layers.21.mlp.down_proj", "shape": [ 2560, 9728 ], "dtype": "bfloat16", "mse": 1.6351204976672307e-05, "max_abs_err": 0.13671875 }, { "name": "model.layers.22.self_attn.q_proj", "shape": [ 4096, 2560 ], "dtype": "bfloat16", "mse": 1.6588825019425713e-05, "max_abs_err": 0.0654296875 }, { "name": "model.layers.22.self_attn.k_proj", "shape": [ 1024, 2560 ], "dtype": "bfloat16", "mse": 1.2678183338721283e-05, "max_abs_err": 0.107421875 }, { "name": "model.layers.22.self_attn.v_proj", "shape": [ 1024, 2560 ], "dtype": "bfloat16", "mse": 1.7473985280958004e-05, "max_abs_err": 0.04736328125 }, { "name": "model.layers.22.self_attn.o_proj", "shape": [ 2560, 4096 ], "dtype": "bfloat16", "mse": 1.4902734619681723e-05, "max_abs_err": 0.109375 }, { "name": "model.layers.22.mlp.gate_proj", "shape": [ 9728, 2560 ], "dtype": "bfloat16", "mse": 1.7299809769610874e-05, "max_abs_err": 0.1103515625 }, { "name": "model.layers.22.mlp.up_proj", "shape": [ 9728, 2560 ], "dtype": "bfloat16", "mse": 1.795753814803902e-05, "max_abs_err": 0.0703125 }, { "name": "model.layers.22.mlp.down_proj", "shape": [ 2560, 9728 ], "dtype": "bfloat16", "mse": 1.6482328646816313e-05, "max_abs_err": 0.0986328125 }, { "name": "model.layers.23.self_attn.q_proj", "shape": [ 4096, 2560 ], "dtype": "bfloat16", "mse": 1.8804832507157698e-05, "max_abs_err": 0.087890625 }, { "name": "model.layers.23.self_attn.k_proj", "shape": [ 1024, 2560 ], "dtype": "bfloat16", "mse": 1.4226406165107619e-05, "max_abs_err": 0.0537109375 }, { "name": "model.layers.23.self_attn.v_proj", "shape": [ 1024, 2560 ], "dtype": "bfloat16", "mse": 1.8124350390280597e-05, "max_abs_err": 0.0419921875 }, { "name": "model.layers.23.self_attn.o_proj", "shape": [ 2560, 4096 ], "dtype": "bfloat16", "mse": 1.60254530783277e-05, "max_abs_err": 0.076171875 }, { "name": "model.layers.23.mlp.gate_proj", "shape": [ 9728, 2560 ], "dtype": "bfloat16", "mse": 1.8611648556543514e-05, "max_abs_err": 0.09375 }, { "name": "model.layers.23.mlp.up_proj", "shape": [ 9728, 2560 ], "dtype": "bfloat16", "mse": 1.8098797227139585e-05, "max_abs_err": 0.0576171875 }, { "name": "model.layers.23.mlp.down_proj", "shape": [ 2560, 9728 ], "dtype": "bfloat16", "mse": 1.674131999607198e-05, "max_abs_err": 0.12890625 }, { "name": "model.layers.24.self_attn.q_proj", "shape": [ 4096, 2560 ], "dtype": "bfloat16", "mse": 1.909290767798666e-05, "max_abs_err": 0.0732421875 }, { "name": "model.layers.24.self_attn.k_proj", "shape": [ 1024, 2560 ], "dtype": "bfloat16", "mse": 1.4168143934512045e-05, "max_abs_err": 0.05615234375 }, { "name": "model.layers.24.self_attn.v_proj", "shape": [ 1024, 2560 ], "dtype": "bfloat16", "mse": 1.7670316083240323e-05, "max_abs_err": 0.0498046875 }, { "name": "model.layers.24.self_attn.o_proj", "shape": [ 2560, 4096 ], "dtype": "bfloat16", "mse": 1.656645690673031e-05, "max_abs_err": 0.083984375 }, { "name": "model.layers.24.mlp.gate_proj", "shape": [ 9728, 2560 ], "dtype": "bfloat16", "mse": 1.9394497940083966e-05, "max_abs_err": 0.091796875 }, { "name": "model.layers.24.mlp.up_proj", "shape": [ 9728, 2560 ], "dtype": "bfloat16", "mse": 1.8227949112770148e-05, "max_abs_err": 0.0703125 }, { "name": "model.layers.24.mlp.down_proj", "shape": [ 2560, 9728 ], "dtype": "bfloat16", "mse": 1.7075506548280828e-05, "max_abs_err": 0.12109375 }, { "name": "model.layers.25.self_attn.q_proj", "shape": [ 4096, 2560 ], "dtype": "bfloat16", "mse": 1.7038068108377047e-05, "max_abs_err": 0.0615234375 }, { "name": "model.layers.25.self_attn.k_proj", "shape": [ 1024, 2560 ], "dtype": "bfloat16", "mse": 1.5202610484266188e-05, "max_abs_err": 0.0498046875 }, { "name": "model.layers.25.self_attn.v_proj", "shape": [ 1024, 2560 ], "dtype": "bfloat16", "mse": 1.941895970958285e-05, "max_abs_err": 0.044921875 }, { "name": "model.layers.25.self_attn.o_proj", "shape": [ 2560, 4096 ], "dtype": "bfloat16", "mse": 1.614114808035083e-05, "max_abs_err": 0.0458984375 }, { "name": "model.layers.25.mlp.gate_proj", "shape": [ 9728, 2560 ], "dtype": "bfloat16", "mse": 1.982031244551763e-05, "max_abs_err": 0.091796875 }, { "name": "model.layers.25.mlp.up_proj", "shape": [ 9728, 2560 ], "dtype": "bfloat16", "mse": 1.8250872017233633e-05, "max_abs_err": 0.0732421875 }, { "name": "model.layers.25.mlp.down_proj", "shape": [ 2560, 9728 ], "dtype": "bfloat16", "mse": 1.7261329048778862e-05, "max_abs_err": 0.09765625 }, { "name": "model.layers.26.self_attn.q_proj", "shape": [ 4096, 2560 ], "dtype": "bfloat16", "mse": 1.7668819054961205e-05, "max_abs_err": 0.0634765625 }, { "name": "model.layers.26.self_attn.k_proj", "shape": [ 1024, 2560 ], "dtype": "bfloat16", "mse": 1.4761316379008349e-05, "max_abs_err": 0.041015625 }, { "name": "model.layers.26.self_attn.v_proj", "shape": [ 1024, 2560 ], "dtype": "bfloat16", "mse": 1.9145734768244438e-05, "max_abs_err": 0.037109375 }, { "name": "model.layers.26.self_attn.o_proj", "shape": [ 2560, 4096 ], "dtype": "bfloat16", "mse": 1.6871248590177856e-05, "max_abs_err": 0.05859375 }, { "name": "model.layers.26.mlp.gate_proj", "shape": [ 9728, 2560 ], "dtype": "bfloat16", "mse": 1.991515455301851e-05, "max_abs_err": 0.087890625 }, { "name": "model.layers.26.mlp.up_proj", "shape": [ 9728, 2560 ], "dtype": "bfloat16", "mse": 1.8679349523154087e-05, "max_abs_err": 0.0625 }, { "name": "model.layers.26.mlp.down_proj", "shape": [ 2560, 9728 ], "dtype": "bfloat16", "mse": 1.77289912244305e-05, "max_abs_err": 0.0859375 }, { "name": "model.layers.27.self_attn.q_proj", "shape": [ 4096, 2560 ], "dtype": "bfloat16", "mse": 1.7379041310050525e-05, "max_abs_err": 0.052734375 }, { "name": "model.layers.27.self_attn.k_proj", "shape": [ 1024, 2560 ], "dtype": "bfloat16", "mse": 1.3688032595382538e-05, "max_abs_err": 0.03662109375 }, { "name": "model.layers.27.self_attn.v_proj", "shape": [ 1024, 2560 ], "dtype": "bfloat16", "mse": 1.8019320123130456e-05, "max_abs_err": 0.046875 }, { "name": "model.layers.27.self_attn.o_proj", "shape": [ 2560, 4096 ], "dtype": "bfloat16", "mse": 1.611689913261216e-05, "max_abs_err": 0.0556640625 }, { "name": "model.layers.27.mlp.gate_proj", "shape": [ 9728, 2560 ], "dtype": "bfloat16", "mse": 1.9707769752130844e-05, "max_abs_err": 0.087890625 }, { "name": "model.layers.27.mlp.up_proj", "shape": [ 9728, 2560 ], "dtype": "bfloat16", "mse": 1.9140639778925106e-05, "max_abs_err": 0.095703125 }, { "name": "model.layers.27.mlp.down_proj", "shape": [ 2560, 9728 ], "dtype": "bfloat16", "mse": 1.8150087271351367e-05, "max_abs_err": 0.0888671875 }, { "name": "model.layers.28.self_attn.q_proj", "shape": [ 4096, 2560 ], "dtype": "bfloat16", "mse": 1.8118531443178654e-05, "max_abs_err": 0.060546875 }, { "name": "model.layers.28.self_attn.k_proj", "shape": [ 1024, 2560 ], "dtype": "bfloat16", "mse": 1.4594187632610556e-05, "max_abs_err": 0.046875 }, { "name": "model.layers.28.self_attn.v_proj", "shape": [ 1024, 2560 ], "dtype": "bfloat16", "mse": 2.0017729184473865e-05, "max_abs_err": 0.033203125 }, { "name": "model.layers.28.self_attn.o_proj", "shape": [ 2560, 4096 ], "dtype": "bfloat16", "mse": 1.602481643203646e-05, "max_abs_err": 0.0693359375 }, { "name": "model.layers.28.mlp.gate_proj", "shape": [ 9728, 2560 ], "dtype": "bfloat16", "mse": 1.9288887415314093e-05, "max_abs_err": 0.10546875 }, { "name": "model.layers.28.mlp.up_proj", "shape": [ 9728, 2560 ], "dtype": "bfloat16", "mse": 1.9545614122762345e-05, "max_abs_err": 0.109375 }, { "name": "model.layers.28.mlp.down_proj", "shape": [ 2560, 9728 ], "dtype": "bfloat16", "mse": 1.8596885638544336e-05, "max_abs_err": 0.09765625 }, { "name": "model.layers.29.self_attn.q_proj", "shape": [ 4096, 2560 ], "dtype": "bfloat16", "mse": 1.9868792151100934e-05, "max_abs_err": 0.064453125 }, { "name": "model.layers.29.self_attn.k_proj", "shape": [ 1024, 2560 ], "dtype": "bfloat16", "mse": 1.3874183423467912e-05, "max_abs_err": 0.0546875 }, { "name": "model.layers.29.self_attn.v_proj", "shape": [ 1024, 2560 ], "dtype": "bfloat16", "mse": 1.815088398871012e-05, "max_abs_err": 0.0537109375 }, { "name": "model.layers.29.self_attn.o_proj", "shape": [ 2560, 4096 ], "dtype": "bfloat16", "mse": 1.7245971321244724e-05, "max_abs_err": 0.05078125 }, { "name": "model.layers.29.mlp.gate_proj", "shape": [ 9728, 2560 ], "dtype": "bfloat16", "mse": 1.9345470718690194e-05, "max_abs_err": 0.10546875 }, { "name": "model.layers.29.mlp.up_proj", "shape": [ 9728, 2560 ], "dtype": "bfloat16", "mse": 2.0173825760139152e-05, "max_abs_err": 0.06103515625 }, { "name": "model.layers.29.mlp.down_proj", "shape": [ 2560, 9728 ], "dtype": "bfloat16", "mse": 1.9278711988590658e-05, "max_abs_err": 0.0888671875 }, { "name": "model.layers.30.self_attn.q_proj", "shape": [ 4096, 2560 ], "dtype": "bfloat16", "mse": 1.6717636754037812e-05, "max_abs_err": 0.0703125 }, { "name": "model.layers.30.self_attn.k_proj", "shape": [ 1024, 2560 ], "dtype": "bfloat16", "mse": 1.5045689906401094e-05, "max_abs_err": 0.0595703125 }, { "name": "model.layers.30.self_attn.v_proj", "shape": [ 1024, 2560 ], "dtype": "bfloat16", "mse": 2.039750324911438e-05, "max_abs_err": 0.05078125 }, { "name": "model.layers.30.self_attn.o_proj", "shape": [ 2560, 4096 ], "dtype": "bfloat16", "mse": 1.6942367437877692e-05, "max_abs_err": 0.05859375 }, { "name": "model.layers.30.mlp.gate_proj", "shape": [ 9728, 2560 ], "dtype": "bfloat16", "mse": 1.911395156639628e-05, "max_abs_err": 0.1328125 }, { "name": "model.layers.30.mlp.up_proj", "shape": [ 9728, 2560 ], "dtype": "bfloat16", "mse": 2.0659232177422382e-05, "max_abs_err": 0.08203125 }, { "name": "model.layers.30.mlp.down_proj", "shape": [ 2560, 9728 ], "dtype": "bfloat16", "mse": 1.975557279365603e-05, "max_abs_err": 0.0966796875 }, { "name": "model.layers.31.self_attn.q_proj", "shape": [ 4096, 2560 ], "dtype": "bfloat16", "mse": 1.7437650967622176e-05, "max_abs_err": 0.06640625 }, { "name": "model.layers.31.self_attn.k_proj", "shape": [ 1024, 2560 ], "dtype": "bfloat16", "mse": 1.3905271771363914e-05, "max_abs_err": 0.0498046875 }, { "name": "model.layers.31.self_attn.v_proj", "shape": [ 1024, 2560 ], "dtype": "bfloat16", "mse": 2.1965097403153777e-05, "max_abs_err": 0.0439453125 }, { "name": "model.layers.31.self_attn.o_proj", "shape": [ 2560, 4096 ], "dtype": "bfloat16", "mse": 1.795162279449869e-05, "max_abs_err": 0.0810546875 }, { "name": "model.layers.31.mlp.gate_proj", "shape": [ 9728, 2560 ], "dtype": "bfloat16", "mse": 1.8703731257119216e-05, "max_abs_err": 0.126953125 }, { "name": "model.layers.31.mlp.up_proj", "shape": [ 9728, 2560 ], "dtype": "bfloat16", "mse": 2.1140878743608482e-05, "max_abs_err": 0.0703125 }, { "name": "model.layers.31.mlp.down_proj", "shape": [ 2560, 9728 ], "dtype": "bfloat16", "mse": 2.0536434021778405e-05, "max_abs_err": 0.083984375 }, { "name": "model.layers.32.self_attn.q_proj", "shape": [ 4096, 2560 ], "dtype": "bfloat16", "mse": 1.6798245269455947e-05, "max_abs_err": 0.0615234375 }, { "name": "model.layers.32.self_attn.k_proj", "shape": [ 1024, 2560 ], "dtype": "bfloat16", "mse": 1.4044644558453001e-05, "max_abs_err": 0.056640625 }, { "name": "model.layers.32.self_attn.v_proj", "shape": [ 1024, 2560 ], "dtype": "bfloat16", "mse": 2.1764282791991718e-05, "max_abs_err": 0.04931640625 }, { "name": "model.layers.32.self_attn.o_proj", "shape": [ 2560, 4096 ], "dtype": "bfloat16", "mse": 1.7867676433525048e-05, "max_abs_err": 0.0888671875 }, { "name": "model.layers.32.mlp.gate_proj", "shape": [ 9728, 2560 ], "dtype": "bfloat16", "mse": 1.8150054529542103e-05, "max_abs_err": 0.09765625 }, { "name": "model.layers.32.mlp.up_proj", "shape": [ 9728, 2560 ], "dtype": "bfloat16", "mse": 2.1252080841804855e-05, "max_abs_err": 0.0751953125 }, { "name": "model.layers.32.mlp.down_proj", "shape": [ 2560, 9728 ], "dtype": "bfloat16", "mse": 2.0602758013410494e-05, "max_abs_err": 0.130859375 }, { "name": "model.layers.33.self_attn.q_proj", "shape": [ 4096, 2560 ], "dtype": "bfloat16", "mse": 1.7009329894790426e-05, "max_abs_err": 0.091796875 }, { "name": "model.layers.33.self_attn.k_proj", "shape": [ 1024, 2560 ], "dtype": "bfloat16", "mse": 1.1332662325003184e-05, "max_abs_err": 0.0419921875 }, { "name": "model.layers.33.self_attn.v_proj", "shape": [ 1024, 2560 ], "dtype": "bfloat16", "mse": 2.493846113793552e-05, "max_abs_err": 0.0419921875 }, { "name": "model.layers.33.self_attn.o_proj", "shape": [ 2560, 4096 ], "dtype": "bfloat16", "mse": 1.9399094526306726e-05, "max_abs_err": 0.107421875 }, { "name": "model.layers.33.mlp.gate_proj", "shape": [ 9728, 2560 ], "dtype": "bfloat16", "mse": 1.7770391423255205e-05, "max_abs_err": 0.109375 }, { "name": "model.layers.33.mlp.up_proj", "shape": [ 9728, 2560 ], "dtype": "bfloat16", "mse": 2.1407859094324522e-05, "max_abs_err": 0.0888671875 }, { "name": "model.layers.33.mlp.down_proj", "shape": [ 2560, 9728 ], "dtype": "bfloat16", "mse": 2.0121487978030927e-05, "max_abs_err": 0.1171875 }, { "name": "model.layers.34.self_attn.q_proj", "shape": [ 4096, 2560 ], "dtype": "bfloat16", "mse": 1.5853764125495218e-05, "max_abs_err": 0.0546875 }, { "name": "model.layers.34.self_attn.k_proj", "shape": [ 1024, 2560 ], "dtype": "bfloat16", "mse": 1.3166883036319632e-05, "max_abs_err": 0.05517578125 }, { "name": "model.layers.34.self_attn.v_proj", "shape": [ 1024, 2560 ], "dtype": "bfloat16", "mse": 2.4885564926080406e-05, "max_abs_err": 0.0498046875 }, { "name": "model.layers.34.self_attn.o_proj", "shape": [ 2560, 4096 ], "dtype": "bfloat16", "mse": 1.8469421775080264e-05, "max_abs_err": 0.08984375 }, { "name": "model.layers.34.mlp.gate_proj", "shape": [ 9728, 2560 ], "dtype": "bfloat16", "mse": 1.8445696696289815e-05, "max_abs_err": 0.109375 }, { "name": "model.layers.34.mlp.up_proj", "shape": [ 9728, 2560 ], "dtype": "bfloat16", "mse": 2.0818995835725218e-05, "max_abs_err": 0.07421875 }, { "name": "model.layers.34.mlp.down_proj", "shape": [ 2560, 9728 ], "dtype": "bfloat16", "mse": 1.7882033716887236e-05, "max_abs_err": 0.1220703125 }, { "name": "model.layers.35.self_attn.q_proj", "shape": [ 4096, 2560 ], "dtype": "bfloat16", "mse": 1.6315227185259573e-05, "max_abs_err": 0.056640625 }, { "name": "model.layers.35.self_attn.k_proj", "shape": [ 1024, 2560 ], "dtype": "bfloat16", "mse": 1.4981170352257323e-05, "max_abs_err": 0.05126953125 }, { "name": "model.layers.35.self_attn.v_proj", "shape": [ 1024, 2560 ], "dtype": "bfloat16", "mse": 2.2129837816464715e-05, "max_abs_err": 0.04248046875 }, { "name": "model.layers.35.self_attn.o_proj", "shape": [ 2560, 4096 ], "dtype": "bfloat16", "mse": 1.7290023606619798e-05, "max_abs_err": 0.08984375 }, { "name": "model.layers.35.mlp.gate_proj", "shape": [ 9728, 2560 ], "dtype": "bfloat16", "mse": 2.042667256318964e-05, "max_abs_err": 0.14453125 }, { "name": "model.layers.35.mlp.up_proj", "shape": [ 9728, 2560 ], "dtype": "bfloat16", "mse": 2.1418125470518135e-05, "max_abs_err": 0.103515625 }, { "name": "model.layers.35.mlp.down_proj", "shape": [ 2560, 9728 ], "dtype": "bfloat16", "mse": 1.615256223885808e-05, "max_abs_err": 0.1064453125 } ], "source_model": "Qwen/Qwen3-4B-Thinking-2507", "dtype": "bf16", "target_regex": "(?:q_proj|k_proj|v_proj|o_proj|gate_proj|up_proj|down_proj)$", "ignore_regex": [], "include_lm_head": false, "note": "Dense fake quant: quantized weights are dequantized and saved as normal HF tensors. Codebook formats are HQQ-style optimized fake-quant variants, not packed runtime checkpoints." }