diff --git a/results/metrics/qwen25-7b_awq_w4_gpqa_run0_metrics.json b/results/metrics/qwen25-7b_awq_w4_gpqa_run0_metrics.json new file mode 100644 index 0000000000000000000000000000000000000000..e63a716b8a34f4089c517810bb11dd8609db3346 --- /dev/null +++ b/results/metrics/qwen25-7b_awq_w4_gpqa_run0_metrics.json @@ -0,0 +1,47 @@ +{ + "model": "qwen25-7b", + "quantization": "awq_w4", + "accuracy": 0.26262626262626265, + "accuracy_delta": 0.0, + "avg_ffs": 0.2808219178082192, + "median_ffs": 0.0, + "ecr": 0.9636923276847184, + "ssr_curve": [ + 0.4292929292929293, + 0.29381443298969073, + 0.26136363636363635, + 0.2611464968152866, + 0.2734375, + 0.24271844660194175, + 0.2602739726027397, + 0.2711864406779661, + 0.29411764705882354, + 0.3111111111111111, + 0.28205128205128205, + 0.3333333333333333, + 0.32142857142857145, + 0.32, + 0.34782608695652173, + 0.3333333333333333, + 0.3333333333333333, + 0.35294117647058826, + 0.3125, + 0.35714285714285715, + 0.45454545454545453, + 0.4, + 0.42857142857142855, + 0.5, + 0.5, + 0.5, + 0.5, + 0.3333333333333333, + 0.5, + 0.5 + ], + "error_type_dist": { + "conceptual": 0.7818361303060217, + "executional": 0.003948667324777887, + "logical": 0.21224086870681144, + "methodological": 0.0019743336623889436 + } +} \ No newline at end of file diff --git a/results/metrics/qwen25-7b_awq_w4_gpqa_run0_sig.json b/results/metrics/qwen25-7b_awq_w4_gpqa_run0_sig.json new file mode 100644 index 0000000000000000000000000000000000000000..230e803fa6585f901093f8ebff67bd6267524218 --- /dev/null +++ b/results/metrics/qwen25-7b_awq_w4_gpqa_run0_sig.json @@ -0,0 +1,7 @@ +{ + "n_pairs": 198, + "delta_acc_observed": 0.015151515151515138, + "delta_acc_ci_lo": -0.06060606060606061, + "delta_acc_ci_hi": 0.08598484848484667, + "p_value": 0.7588 +} \ No newline at end of file diff --git a/results/metrics/qwen25-7b_awq_w4_gsm8k_run0_ci.json b/results/metrics/qwen25-7b_awq_w4_gsm8k_run0_ci.json new file mode 100644 index 0000000000000000000000000000000000000000..213bb4cee1e89e9d5cffc560a79e9e3217a5b04e --- /dev/null +++ b/results/metrics/qwen25-7b_awq_w4_gsm8k_run0_ci.json @@ -0,0 +1,116 @@ +{ + "n": 1319, + "accuracy": { + "mean": 0.8663734647460198, + "ci_lo": 0.8476118271417741, + "ci_hi": 0.8847611827141774 + }, + "avg_ffs": { + "mean": 1.1130837434525587, + "ci_lo": 0.917098445595855, + "ci_hi": 1.3142989417989417 + }, + "ecr": { + "mean": 0.8587336219360299, + "ci_lo": 0.8182060164014389, + "ci_hi": 0.8973816909975669 + }, + "ssr_curve_ci": { + "mean": [ + 0.9452821834723265, + 0.9019858434901767, + 0.8817650255702835, + 0.873284771722402, + 0.855534350872108, + 0.8541805681614897, + 0.8624117378281537, + 0.8624123505602438, + 0.8486782654972008, + 0.7811723178775143, + 0.7668866786060701, + 0.7089726195822806, + 0.49842731495099374, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "ci_lo": [ + 0.9325246398786959, + 0.8857589984350548, + 0.8628281117696867, + 0.8529117199227677, + 0.831280406190028, + 0.8198584583056626, + 0.8178137651821862, + 0.7889882772680938, + 0.7446808510638298, + 0.6284821428571429, + 0.5, + 0.3333333333333333, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "ci_hi": [ + 0.9567854435178166, + 0.9182242990654206, + 0.9000853970964987, + 0.8931862212660185, + 0.8795620437956204, + 0.8861630945328368, + 0.9042178910544727, + 0.9285894660894659, + 0.94, + 0.918918918918919, + 1.0, + 1.0, + 1.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ] + } +} \ No newline at end of file diff --git a/results/metrics/qwen25-7b_awq_w4_gsm8k_run0_metrics.json b/results/metrics/qwen25-7b_awq_w4_gsm8k_run0_metrics.json new file mode 100644 index 0000000000000000000000000000000000000000..5289831c6a7ea8f1320cc071bc3e534430c6643e --- /dev/null +++ b/results/metrics/qwen25-7b_awq_w4_gsm8k_run0_metrics.json @@ -0,0 +1,47 @@ +{ + "model": "qwen25-7b", + "quantization": "awq_w4", + "accuracy": 0.866565579984837, + "accuracy_delta": 0.0, + "avg_ffs": 1.1136363636363635, + "median_ffs": 1.0, + "ecr": 0.8588141025641022, + "ssr_curve": [ + 0.9454131918119788, + 0.9021909233176839, + 0.8819031435853866, + 0.8733459357277883, + 0.8555691554467564, + 0.8539325842696629, + 0.8623481781376519, + 0.8631578947368421, + 0.8490566037735849, + 0.78125, + 0.7692307692307693, + 0.7142857142857143, + 0.5, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "error_type_dist": { + "conceptual": 0.6656626506024096, + "executional": 0.0572289156626506, + "logical": 0.2756024096385542, + "methodological": 0.0015060240963855422 + } +} \ No newline at end of file diff --git a/results/metrics/qwen25-7b_awq_w4_gsm8k_run0_sig.json b/results/metrics/qwen25-7b_awq_w4_gsm8k_run0_sig.json new file mode 100644 index 0000000000000000000000000000000000000000..0df0c9a93234f341d1d81b043b1071c41b616cd2 --- /dev/null +++ b/results/metrics/qwen25-7b_awq_w4_gsm8k_run0_sig.json @@ -0,0 +1,7 @@ +{ + "n_pairs": 1319, + "delta_acc_observed": 0.015921152388172932, + "delta_acc_ci_lo": -0.0030326004548900665, + "delta_acc_ci_hi": 0.03487490523123571, + "p_value": 0.1028 +} \ No newline at end of file diff --git a/results/metrics/qwen25-7b_awq_w4_restored_gpqa_run0_ci.json b/results/metrics/qwen25-7b_awq_w4_restored_gpqa_run0_ci.json new file mode 100644 index 0000000000000000000000000000000000000000..ecddd3eeffce75f1a5d6342916d93b0b9472f114 --- /dev/null +++ b/results/metrics/qwen25-7b_awq_w4_restored_gpqa_run0_ci.json @@ -0,0 +1,116 @@ +{ + "n": 198, + "accuracy": { + "mean": 0.2772131313131313, + "ci_lo": 0.21717171717171718, + "ci_hi": 0.3383838383838384 + }, + "avg_ffs": { + "mean": 0.6147496415468082, + "ci_lo": 0.421043557168784, + "ci_hi": 0.8417277570348783 + }, + "ecr": { + "mean": 0.9419172171354703, + "ci_lo": 0.9124032598260806, + "ci_hi": 0.9672991054226294 + }, + "ssr_curve_ci": { + "mean": [ + 0.5249171717171652, + 0.366978095144372, + 0.31249728264796606, + 0.32457386836581104, + 0.3237086401011688, + 0.3148892522221575, + 0.30917209798111134, + 0.26172052932777223, + 0.2947778691115077, + 0.3188645542102334, + 0.26502168063609344, + 0.2590141631999209, + 0.2688982409812325, + 0.2721552529014958, + 0.278255126345129, + 0.278255126345129, + 0.278255126345129, + 0.26543828701624356, + 0.26543828701624356, + 0.26543828701624356, + 0.214657971869537, + 0.214657971869537, + 0.214657971869537, + 0.15415841790794613, + 0.1817478234610836, + 0.1817478234610836, + 0.10988116381296713, + 0.10988116381296713, + 0.14052041498366147, + 0.14052041498366147 + ], + "ci_lo": [ + 0.45454545454545453, + 0.29896907216494845, + 0.24581005586592178, + 0.2546583850931677, + 0.24666438356164383, + 0.22880750605326877, + 0.21516726943942135, + 0.15789473684210525, + 0.17307692307692307, + 0.1836734693877551, + 0.12193274205469327, + 0.10344827586206896, + 0.10714285714285714, + 0.09523809523809523, + 0.07692307692307693, + 0.07692307692307693, + 0.07692307692307693, + 0.0625, + 0.0625, + 0.0625, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "ci_hi": [ + 0.5959595959595959, + 0.43455497382198954, + 0.3791208791208791, + 0.39622641509433965, + 0.40131578947368424, + 0.4056603773584906, + 0.4117647058823529, + 0.3709677419354839, + 0.4218836805555554, + 0.46153846153846156, + 0.42105263157894735, + 0.43333333333333335, + 0.4482758620689655, + 0.47058823529411764, + 0.5, + 0.5, + 0.5, + 0.5, + 0.5, + 0.5, + 0.45454545454545453, + 0.45454545454545453, + 0.45454545454545453, + 0.375, + 0.4444444444444444, + 0.4444444444444444, + 0.375, + 0.375, + 0.5, + 0.5 + ] + } +} \ No newline at end of file diff --git a/results/metrics/qwen25-7b_awq_w4_restored_gpqa_run0_metrics.json b/results/metrics/qwen25-7b_awq_w4_restored_gpqa_run0_metrics.json new file mode 100644 index 0000000000000000000000000000000000000000..8c6cfbce33eda995541939280e0499d8e1643ac5 --- /dev/null +++ b/results/metrics/qwen25-7b_awq_w4_restored_gpqa_run0_metrics.json @@ -0,0 +1,47 @@ +{ + "model": "qwen25-7b", + "quantization": "awq_w4_restored", + "accuracy": 0.2777777777777778, + "accuracy_delta": 0.0, + "avg_ffs": 0.6153846153846154, + "median_ffs": 0.0, + "ecr": 0.9416252592550498, + "ssr_curve": [ + 0.5252525252525253, + 0.36787564766839376, + 0.3131868131868132, + 0.325, + 0.323943661971831, + 0.3148148148148148, + 0.30952380952380953, + 0.26153846153846155, + 0.29411764705882354, + 0.3181818181818182, + 0.2647058823529412, + 0.25925925925925924, + 0.2692307692307692, + 0.2727272727272727, + 0.2777777777777778, + 0.2777777777777778, + 0.2777777777777778, + 0.26666666666666666, + 0.26666666666666666, + 0.26666666666666666, + 0.21428571428571427, + 0.21428571428571427, + 0.21428571428571427, + 0.15384615384615385, + 0.18181818181818182, + 0.18181818181818182, + 0.1111111111111111, + 0.1111111111111111, + 0.14285714285714285, + 0.14285714285714285 + ], + "error_type_dist": { + "conceptual": 0.7881679389312977, + "executional": 0.0028625954198473282, + "logical": 0.2032442748091603, + "methodological": 0.0057251908396946565 + } +} \ No newline at end of file diff --git a/results/metrics/qwen25-7b_awq_w4_restored_gsm8k_run0_ci.json b/results/metrics/qwen25-7b_awq_w4_restored_gsm8k_run0_ci.json new file mode 100644 index 0000000000000000000000000000000000000000..365ffa91509aa62014c18f424242c1f83cc325f6 --- /dev/null +++ b/results/metrics/qwen25-7b_awq_w4_restored_gsm8k_run0_ci.json @@ -0,0 +1,116 @@ +{ + "n": 1319, + "accuracy": { + "mean": 0.8823316148597421, + "ci_lo": 0.865049279757392, + "ci_hi": 0.8999241849886277 + }, + "avg_ffs": { + "mean": 1.6693245232453067, + "ci_lo": 1.4013583777460028, + "ci_hi": 1.967121179966044 + }, + "ecr": { + "mean": 0.7620505063766115, + "ci_lo": 0.699723606557377, + "ci_hi": 0.8213879456220987 + }, + "ssr_curve_ci": { + "mean": [ + 0.9710574677786185, + 0.9264562814336372, + 0.911648353827803, + 0.8967685702574044, + 0.887217981305208, + 0.8725629643420875, + 0.8838926660672402, + 0.8585736275292504, + 0.8770366703182916, + 0.8846775662035041, + 0.9246207462271337, + 1.0, + 0.7519935406386951, + 1.0, + null, + null, + null, + null, + null, + null, + null, + null, + null, + null, + null, + null, + null, + null, + null, + null + ], + "ci_lo": [ + 0.9620924943138741, + 0.912333441952691, + 0.8958149735592532, + 0.8782606675286893, + 0.8639445719116048, + 0.838148236337362, + 0.8369536613272311, + 0.7816002155172415, + 0.7758620689655172, + 0.75, + 0.75, + 1.0, + 0.0, + 1.0, + null, + null, + null, + null, + null, + null, + null, + null, + null, + null, + null, + null, + null, + null, + null, + null + ], + "ci_hi": [ + 0.979529946929492, + 0.9400480758311288, + 0.9275150627592124, + 0.9153039457398522, + 0.9103260869565217, + 0.9054080655458607, + 0.9269662921348315, + 0.9285714285714286, + 0.9607843137254902, + 1.0, + 1.0, + 1.0, + 1.0, + 1.0, + null, + null, + null, + null, + null, + null, + null, + null, + null, + null, + null, + null, + null, + null, + null, + null + ] + } +} \ No newline at end of file diff --git a/results/metrics/qwen25-7b_awq_w4_restored_gsm8k_run0_metrics.json b/results/metrics/qwen25-7b_awq_w4_restored_gsm8k_run0_metrics.json new file mode 100644 index 0000000000000000000000000000000000000000..144b05e1fcdab66ed3e8879708146482844fed31 --- /dev/null +++ b/results/metrics/qwen25-7b_awq_w4_restored_gsm8k_run0_metrics.json @@ -0,0 +1,46 @@ +{ + "model": "qwen25-7b", + "quantization": "awq_w4_restored", + "accuracy": 0.8824867323730099, + "accuracy_delta": 0.0, + "avg_ffs": 1.6709677419354838, + "median_ffs": 1.0, + "ecr": 0.7624, + "ssr_curve": [ + 0.9711902956785443, + 0.9265274555297757, + 0.9117889530090684, + 0.8969754253308129, + 0.8876560332871013, + 0.873015873015873, + 0.8842105263157894, + 0.8588235294117647, + 0.8775510204081632, + 0.8846153846153846, + 0.9230769230769231, + 1.0, + 0.75, + 1.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "error_type_dist": { + "conceptual": 0.5310344827586206, + "executional": 0.08045977011494253, + "logical": 0.38850574712643676 + } +} \ No newline at end of file diff --git a/results/metrics/qwen25-7b_bnb_nf4_w4_gpqa_run0_ci.json b/results/metrics/qwen25-7b_bnb_nf4_w4_gpqa_run0_ci.json new file mode 100644 index 0000000000000000000000000000000000000000..18c3df476d0fea88235101326fdd722d70a2e6d2 --- /dev/null +++ b/results/metrics/qwen25-7b_bnb_nf4_w4_gpqa_run0_ci.json @@ -0,0 +1,116 @@ +{ + "n": 198, + "accuracy": { + "mean": 0.3129959595959596, + "ci_lo": 0.25239898989898996, + "ci_hi": 0.3787878787878788 + }, + "avg_ffs": { + "mean": 0.3009145275190325, + "ci_lo": 0.21258649486995945, + "ci_hi": 0.4 + }, + "ecr": { + "mean": 0.9638939283823105, + "ci_lo": 0.9449814170262291, + "ci_hi": 0.9803937107491645 + }, + "ssr_curve_ci": { + "mean": [ + 0.489749494949499, + 0.3364034661564047, + 0.3293664632535101, + 0.3263619589788414, + 0.3071514879759616, + 0.34210047238648145, + 0.36421499669597057, + 0.3437965333776358, + 0.38880981958954136, + 0.3225928581084274, + 0.3327097132032657, + 0.3477350527832801, + 0.36340379162861436, + 0.3505085875786862, + 0.27846531626539306, + 0.3084013104285189, + 0.40082066639219416, + 0.40082066639219416, + 0.502421864192322, + 0.502421864192322, + 0.502421864192322, + 0.502421864192322, + 0.6035371843902793, + 0.6035371843902793, + 0.5073448319029379, + 0.6702496298951736, + 0.6702496298951736, + 0.6702496298951736, + 0.6702496298951736, + 0.6702496298951736 + ], + "ci_lo": [ + 0.42424242424242425, + 0.2708333333333333, + 0.26011560693641617, + 0.25157136242620676, + 0.22900763358778625, + 0.25225225225225223, + 0.2537313432835821, + 0.21664215686274513, + 0.2391304347826087, + 0.16, + 0.16, + 0.15789473684210525, + 0.16666666666666666, + 0.14285714285714285, + 0.07400793650793652, + 0.07142857142857142, + 0.09090909090909091, + 0.09090909090909091, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "ci_hi": [ + 0.5606060606060606, + 0.40414507772020725, + 0.4, + 0.4025177652536143, + 0.3897058823529412, + 0.4361750483558994, + 0.47692307692307695, + 0.4772727272727273, + 0.5454545454545454, + 0.5, + 0.5217391304347826, + 0.5555555555555556, + 0.5789473684210527, + 0.5714285714285714, + 0.5, + 0.5833333333333334, + 0.7272727272727273, + 0.7272727272727273, + 1.0, + 1.0, + 1.0, + 1.0, + 1.0, + 1.0, + 1.0, + 1.0, + 1.0, + 1.0, + 1.0, + 1.0 + ] + } +} \ No newline at end of file diff --git a/results/metrics/qwen25-7b_bnb_nf4_w4_gpqa_run0_sig.json b/results/metrics/qwen25-7b_bnb_nf4_w4_gpqa_run0_sig.json new file mode 100644 index 0000000000000000000000000000000000000000..b654d8412bc76171376d979e31f46e511f3b1ed5 --- /dev/null +++ b/results/metrics/qwen25-7b_bnb_nf4_w4_gpqa_run0_sig.json @@ -0,0 +1,7 @@ +{ + "n_pairs": 198, + "delta_acc_observed": 0.0, + "delta_acc_ci_lo": -0.07070707070707072, + "delta_acc_ci_hi": 0.07070707070707069, + "p_value": 1.0496 +} \ No newline at end of file diff --git a/results/metrics/qwen25-7b_bnb_nf4_w4_gsm8k_run0_ci.json b/results/metrics/qwen25-7b_bnb_nf4_w4_gsm8k_run0_ci.json new file mode 100644 index 0000000000000000000000000000000000000000..ab227d11a6dbf201b9a13e2d9695bc8958547316 --- /dev/null +++ b/results/metrics/qwen25-7b_bnb_nf4_w4_gsm8k_run0_ci.json @@ -0,0 +1,116 @@ +{ + "n": 1319, + "accuracy": { + "mean": 0.8635822592873388, + "ci_lo": 0.844579226686884, + "ci_hi": 0.8817285822592873 + }, + "avg_ffs": { + "mean": 1.3238990129406885, + "ci_lo": 1.109274768353528, + "ci_hi": 1.5537747765254564 + }, + "ecr": { + "mean": 0.866001662890616, + "ci_lo": 0.826406728915853, + "ci_hi": 0.9034009502923976 + }, + "ssr_curve_ci": { + "mean": [ + 0.9531190295678487, + 0.905036025273055, + 0.8862429185538575, + 0.8695986568425902, + 0.8523569073842906, + 0.851208965904284, + 0.8261093172125309, + 0.8016798866558521, + 0.7993970020708107, + 0.8049932226867372, + 0.7674034216710551, + 0.6981488531480916, + 0.498328133473569, + 0.4966229181057636, + 0.4966229181057636, + 0.4966229181057636, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "ci_lo": [ + 0.9408642911296436, + 0.888715953307393, + 0.8674386726580352, + 0.8480298078610283, + 0.8269948630890075, + 0.8155844155844156, + 0.7731481481481481, + 0.72, + 0.6851851851851852, + 0.6538461538461539, + 0.5, + 0.38461538461538464, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "ci_hi": [ + 0.9643669446550417, + 0.9207082371054658, + 0.9040233627290127, + 0.8894927536231884, + 0.8778640824824031, + 0.8854186545801526, + 0.8755986812349734, + 0.8737864077669902, + 0.8974489795918366, + 0.935501029512697, + 1.0, + 1.0, + 1.0, + 1.0, + 1.0, + 1.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ] + } +} \ No newline at end of file diff --git a/results/metrics/qwen25-7b_bnb_nf4_w4_gsm8k_run0_metrics.json b/results/metrics/qwen25-7b_bnb_nf4_w4_gsm8k_run0_metrics.json new file mode 100644 index 0000000000000000000000000000000000000000..9da91890a9b3bcfb7e93568f9e293ea3514f3e47 --- /dev/null +++ b/results/metrics/qwen25-7b_bnb_nf4_w4_gsm8k_run0_metrics.json @@ -0,0 +1,46 @@ +{ + "model": "qwen25-7b", + "quantization": "bnb_nf4_w4", + "accuracy": 0.8635329795299469, + "accuracy_delta": 0.0, + "avg_ffs": 1.3222222222222222, + "median_ffs": 1.0, + "ecr": 0.8659139784946237, + "ssr_curve": [ + 0.9529946929492039, + 0.9049459041731066, + 0.8861386138613861, + 0.8694444444444445, + 0.8522427440633246, + 0.851010101010101, + 0.8262910798122066, + 0.8018867924528302, + 0.8, + 0.8064516129032258, + 0.7692307692307693, + 0.7, + 0.5, + 0.5, + 0.5, + 0.5, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "error_type_dist": { + "conceptual": 0.5975975975975976, + "executional": 0.057057057057057055, + "logical": 0.34534534534534533 + } +} \ No newline at end of file diff --git a/results/metrics/qwen25-7b_bnb_nf4_w4_gsm8k_run0_sig.json b/results/metrics/qwen25-7b_bnb_nf4_w4_gsm8k_run0_sig.json new file mode 100644 index 0000000000000000000000000000000000000000..039a8e06b7fa0a3729b485b54feb68587940723f --- /dev/null +++ b/results/metrics/qwen25-7b_bnb_nf4_w4_gsm8k_run0_sig.json @@ -0,0 +1,7 @@ +{ + "n_pairs": 1319, + "delta_acc_observed": -0.019711902956785377, + "delta_acc_ci_lo": -0.04094010614101595, + "delta_acc_ci_hi": 0.0015163002274450887, + "p_value": 0.0764 +} \ No newline at end of file diff --git a/results/metrics/qwen25-7b_bnb_nf4_w4_math500_run0_sig.json b/results/metrics/qwen25-7b_bnb_nf4_w4_math500_run0_sig.json new file mode 100644 index 0000000000000000000000000000000000000000..99bb2c3b6bf1b2b91eb14a9dd56cb860beb0284e --- /dev/null +++ b/results/metrics/qwen25-7b_bnb_nf4_w4_math500_run0_sig.json @@ -0,0 +1,7 @@ +{ + "n_pairs": 500, + "delta_acc_observed": -0.010000000000000009, + "delta_acc_ci_lo": -0.04400000000000004, + "delta_acc_ci_hi": 0.02200000000000002, + "p_value": 0.5772 +} \ No newline at end of file diff --git a/results/metrics/qwen25-7b_bnb_nf4_w4_restored_gpqa_run0_ci.json b/results/metrics/qwen25-7b_bnb_nf4_w4_restored_gpqa_run0_ci.json new file mode 100644 index 0000000000000000000000000000000000000000..4e7b7699d521a95d6dc4fbdf603a3ec4adfd3adf --- /dev/null +++ b/results/metrics/qwen25-7b_bnb_nf4_w4_restored_gpqa_run0_ci.json @@ -0,0 +1,116 @@ +{ + "n": 198, + "accuracy": { + "mean": 0.312679797979798, + "ci_lo": 0.25252525252525254, + "ci_hi": 0.3787878787878788 + }, + "avg_ffs": { + "mean": 0.500689736642367, + "ci_lo": 0.36169803600654665, + "ci_hi": 0.6439482916412446 + }, + "ecr": { + "mean": 0.9428519616966791, + "ci_lo": 0.9137926286116983, + "ci_hi": 0.9681568225392683 + }, + "ssr_curve_ci": { + "mean": [ + 0.545783838383829, + 0.3851573833288215, + 0.3191100594123924, + 0.28599551255747724, + 0.2624343320300904, + 0.2706478030484445, + 0.27820266929115944, + 0.2722553950238139, + 0.2804353622414716, + 0.26671821028988857, + 0.2232377708341524, + 0.22680859686213362, + 0.25122085968266505, + 0.2812482316490329, + 0.23582645710644323, + 0.2669804964563325, + 0.2669804964563325, + 0.2669804964563325, + 0.23223423904953241, + 0.22458671926886167, + 0.2535624184819287, + 0.2535624184819287, + 0.1473327652543453, + 0.1706716317365487, + 0.2030012812890366, + 0.2030012812890366, + 0.2549428893923098, + 0.0, + 0.0, + 0.0 + ], + "ci_lo": [ + 0.47474747474747475, + 0.3177083333333333, + 0.2533253012048193, + 0.21476510067114093, + 0.19148936170212766, + 0.19, + 0.18461538461538463, + 0.1694915254237288, + 0.1691666666666667, + 0.14, + 0.09297040169133194, + 0.08333333333333333, + 0.09375, + 0.1111111111111111, + 0.05263157894736842, + 0.058823529411764705, + 0.058823529411764705, + 0.058823529411764705, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "ci_hi": [ + 0.6161616161616161, + 0.45549738219895286, + 0.3905393457117594, + 0.36, + 0.33813214134574693, + 0.35715923984272585, + 0.3815789473684211, + 0.3835616438356164, + 0.4, + 0.4, + 0.36585365853658536, + 0.38235294117647056, + 0.4166666666666667, + 0.47058823529411764, + 0.4547202797202772, + 0.5, + 0.5, + 0.5, + 0.5, + 0.5555555555555556, + 0.625, + 0.625, + 0.5, + 0.5595238095238095, + 0.6666666666666666, + 0.6666666666666666, + 1.0, + 0.0, + 0.0, + 0.0 + ] + } +} \ No newline at end of file diff --git a/results/metrics/qwen25-7b_bnb_nf4_w4_restored_gpqa_run0_metrics.json b/results/metrics/qwen25-7b_bnb_nf4_w4_restored_gpqa_run0_metrics.json new file mode 100644 index 0000000000000000000000000000000000000000..3b569f0bc8454d850098bdf6e719e31def1c4b83 --- /dev/null +++ b/results/metrics/qwen25-7b_bnb_nf4_w4_restored_gpqa_run0_metrics.json @@ -0,0 +1,47 @@ +{ + "model": "qwen25-7b", + "quantization": "bnb_nf4_w4_restored", + "accuracy": 0.31313131313131315, + "accuracy_delta": 0.0, + "avg_ffs": 0.5, + "median_ffs": 0.0, + "ecr": 0.943037920789117, + "ssr_curve": [ + 0.5454545454545454, + 0.3854166666666667, + 0.31976744186046513, + 0.2866666666666667, + 0.26277372262773724, + 0.27102803738317754, + 0.27848101265822783, + 0.2727272727272727, + 0.2807017543859649, + 0.26666666666666666, + 0.2222222222222222, + 0.22580645161290322, + 0.25, + 0.28, + 0.23529411764705882, + 0.26666666666666666, + 0.26666666666666666, + 0.26666666666666666, + 0.23076923076923078, + 0.2222222222222222, + 0.25, + 0.25, + 0.14285714285714285, + 0.16666666666666666, + 0.2, + 0.2, + 0.25, + 0.0, + 0.0, + 0.0 + ], + "error_type_dist": { + "conceptual": 0.7429406037000974, + "executional": 0.0029211295034079843, + "logical": 0.25121713729308665, + "methodological": 0.0029211295034079843 + } +} \ No newline at end of file diff --git a/results/metrics/qwen25-7b_bnb_nf4_w4_restored_gsm8k_run0_ci.json b/results/metrics/qwen25-7b_bnb_nf4_w4_restored_gsm8k_run0_ci.json new file mode 100644 index 0000000000000000000000000000000000000000..edfbc06b7371e6443723d552bd13bd7d1fb885d6 --- /dev/null +++ b/results/metrics/qwen25-7b_bnb_nf4_w4_restored_gsm8k_run0_ci.json @@ -0,0 +1,116 @@ +{ + "n": 1319, + "accuracy": { + "mean": 0.8438679302501896, + "ci_lo": 0.824109173616376, + "ci_hi": 0.8627748294162244 + }, + "avg_ffs": { + "mean": 1.368657154587038, + "ci_lo": 1.1693967386590338, + "ci_hi": 1.5808149058149057 + }, + "ecr": { + "mean": 0.803701460009669, + "ci_lo": 0.7564720567898913, + "ci_hi": 0.849325791107041 + }, + "ssr_curve_ci": { + "mean": [ + 0.9417137225170591, + 0.9017153921793984, + 0.8737602512662026, + 0.8672202560295622, + 0.8429157709734909, + 0.8044121847391416, + 0.8008640785229358, + 0.7973444398423337, + 0.7879257526993056, + 0.7503623309545595, + 0.8003742335145025, + 1.0, + 1.0, + 1.0, + 1.0, + null, + null, + null, + null, + null, + null, + null, + null, + null, + null, + null, + null, + null, + null, + null + ], + "ci_lo": [ + 0.9287338893100834, + 0.8854466981460121, + 0.8546511627906976, + 0.8462291633908426, + 0.8150884943648101, + 0.7634947618652735, + 0.7424203373945643, + 0.7159090909090909, + 0.6739019520851819, + 0.5384615384615384, + 0.5, + 1.0, + 1.0, + 1.0, + 1.0, + null, + null, + null, + null, + null, + null, + null, + null, + null, + null, + null, + null, + null, + null, + null + ], + "ci_hi": [ + 0.9537528430629265, + 0.9173745173745174, + 0.8915567919458268, + 0.8873920442439515, + 0.8689047688784748, + 0.8445072401205677, + 0.8556149732620321, + 0.8787878787878788, + 0.8936170212765957, + 0.9375, + 1.0, + 1.0, + 1.0, + 1.0, + 1.0, + null, + null, + null, + null, + null, + null, + null, + null, + null, + null, + null, + null, + null, + null, + null + ] + } +} \ No newline at end of file diff --git a/results/metrics/qwen25-7b_bnb_nf4_w4_restored_gsm8k_run0_metrics.json b/results/metrics/qwen25-7b_bnb_nf4_w4_restored_gsm8k_run0_metrics.json new file mode 100644 index 0000000000000000000000000000000000000000..c444c5809042b091ab762d80de7dc88d992a71eb --- /dev/null +++ b/results/metrics/qwen25-7b_bnb_nf4_w4_restored_gsm8k_run0_metrics.json @@ -0,0 +1,46 @@ +{ + "model": "qwen25-7b", + "quantization": "bnb_nf4_w4_restored", + "accuracy": 0.8438210765731615, + "accuracy_delta": 0.0, + "avg_ffs": 1.3689320388349515, + "median_ffs": 1.0, + "ecr": 0.8035092751960223, + "ssr_curve": [ + 0.9416224412433661, + 0.901702786377709, + 0.8737541528239202, + 0.8671799807507219, + 0.8429394812680115, + 0.8048128342245989, + 0.8010471204188482, + 0.7978723404255319, + 0.7884615384615384, + 0.75, + 0.8, + 1.0, + 1.0, + 1.0, + 1.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "error_type_dist": { + "conceptual": 0.5707692307692308, + "executional": 0.07230769230769231, + "logical": 0.3569230769230769 + } +} \ No newline at end of file diff --git a/results/metrics/qwen25-7b_bnb_nf4_w4_restored_math500_run0_metrics.json b/results/metrics/qwen25-7b_bnb_nf4_w4_restored_math500_run0_metrics.json new file mode 100644 index 0000000000000000000000000000000000000000..3cff56709fe1766fe3ecbb485bce37539f0df563 --- /dev/null +++ b/results/metrics/qwen25-7b_bnb_nf4_w4_restored_math500_run0_metrics.json @@ -0,0 +1,47 @@ +{ + "model": "qwen25-7b", + "quantization": "bnb_nf4_w4_restored", + "accuracy": 0.722, + "accuracy_delta": 0.0, + "avg_ffs": 0.7122302158273381, + "median_ffs": 0.0, + "ecr": 0.9219371897943327, + "ssr_curve": [ + 0.846, + 0.757085020242915, + 0.7443609022556391, + 0.7557755775577558, + 0.7373271889400922, + 0.6842105263157895, + 0.7064220183486238, + 0.6904761904761905, + 0.625, + 0.5686274509803921, + 0.5476190476190477, + 0.47058823529411764, + 0.4666666666666667, + 0.4583333333333333, + 0.42857142857142855, + 0.375, + 0.3076923076923077, + 0.3333333333333333, + 0.3333333333333333, + 0.3333333333333333, + 0.3333333333333333, + 0.3333333333333333, + 0.3333333333333333, + 0.25, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "error_type_dist": { + "conceptual": 0.5786259541984733, + "executional": 0.03206106870229008, + "logical": 0.3755725190839695, + "methodological": 0.013740458015267175 + } +} \ No newline at end of file diff --git a/results/metrics/qwen25-7b_gptq_w4_gpqa_run0_ci.json b/results/metrics/qwen25-7b_gptq_w4_gpqa_run0_ci.json new file mode 100644 index 0000000000000000000000000000000000000000..18c10d026fbc26b7053e72080daff9e1c7eb382e --- /dev/null +++ b/results/metrics/qwen25-7b_gptq_w4_gpqa_run0_ci.json @@ -0,0 +1,116 @@ +{ + "n": 198, + "accuracy": { + "mean": 0.2680050505050505, + "ci_lo": 0.20707070707070707, + "ci_hi": 0.3282828282828283 + }, + "avg_ffs": { + "mean": 0.25536829442679415, + "ci_lo": 0.176056338028169, + "ci_hi": 0.3424657534246575 + }, + "ecr": { + "mean": 0.9646514354939871, + "ci_lo": 0.9444374361461334, + "ci_hi": 0.9813586454093275 + }, + "ssr_curve_ci": { + "mean": [ + 0.43467171717172187, + 0.26731580247105496, + 0.2594558383131973, + 0.254530249594509, + 0.28573324414462486, + 0.27190091150824486, + 0.2499570292903606, + 0.21925476563714502, + 0.24996332413636702, + 0.26486986249756567, + 0.19975775904316348, + 0.21832032368264181, + 0.2688362842268626, + 0.28527397947213384, + 0.24923405747296543, + 0.2772244204936014, + 0.2662999091018617, + 0.15338855904839024, + 0.16605316195766276, + 0.0998368358736652, + 0.1110954877088751, + 0.1110954877088751, + 0.12474080168981627, + 0.1430075759043663, + 0.16762255462345774, + 0.16762255462345774, + 0.16762255462345774, + 0.0, + 0.0, + 0.0 + ], + "ci_lo": [ + 0.36363636363636365, + 0.20618556701030927, + 0.1988691759978728, + 0.19074456372144233, + 0.2127659574468085, + 0.19327731092436976, + 0.1625, + 0.12903225806451613, + 0.14754098360655737, + 0.14285714285714285, + 0.08108108108108109, + 0.08327702702702704, + 0.10344827586206896, + 0.09523809523809523, + 0.0625, + 0.07142857142857142, + 0.058741830065359504, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "ci_hi": [ + 0.5050505050505051, + 0.33157894736842103, + 0.3241758241758242, + 0.3214337928153717, + 0.3617146925783212, + 0.35398230088495575, + 0.345679012345679, + 0.3194444444444444, + 0.36065573770491804, + 0.4, + 0.3333333333333333, + 0.375, + 0.45454545454545453, + 0.5, + 0.45, + 0.5, + 0.5, + 0.38461538461538464, + 0.4166666666666667, + 0.3333333333333333, + 0.375, + 0.375, + 0.42857142857142855, + 0.5, + 0.5, + 0.5, + 0.5, + 0.0, + 0.0, + 0.0 + ] + } +} \ No newline at end of file diff --git a/results/metrics/qwen25-7b_gptq_w4_gsm8k_run0_ci.json b/results/metrics/qwen25-7b_gptq_w4_gsm8k_run0_ci.json new file mode 100644 index 0000000000000000000000000000000000000000..aa0f81c650b7f5ab6fb6144840d97922d768bb25 --- /dev/null +++ b/results/metrics/qwen25-7b_gptq_w4_gsm8k_run0_ci.json @@ -0,0 +1,116 @@ +{ + "n": 1319, + "accuracy": { + "mean": 0.8594833965125095, + "ci_lo": 0.8407884761182715, + "ci_hi": 0.8779378316906747 + }, + "avg_ffs": { + "mean": 1.103108781002656, + "ci_lo": 0.9285714285714286, + "ci_hi": 1.2823562645397142 + }, + "ecr": { + "mean": 0.8439696431495372, + "ci_lo": 0.8011137868498416, + "ci_hi": 0.8835817547531832 + }, + "ssr_curve_ci": { + "mean": [ + 0.9476667172100081, + 0.8950164880289548, + 0.8754005068273732, + 0.8601291080690123, + 0.8493054237570207, + 0.8271800824675436, + 0.8078368059567558, + 0.7879495025999508, + 0.7496246045406726, + 0.8133342673985481, + 0.7984957928446947, + 0.6644169829135926, + 0.7127630388715678, + 0.799239787793852, + 1.0, + 1.0, + null, + null, + null, + null, + null, + null, + null, + null, + null, + null, + null, + null, + null, + null + ], + "ci_lo": [ + 0.935557240333586, + 0.8778606884897701, + 0.856687735947401, + 0.8398940790764143, + 0.8242564588490099, + 0.7899755887577511, + 0.7543840744101633, + 0.7129606710671067, + 0.6461538461538462, + 0.6875, + 0.6, + 0.375, + 0.3294871794871797, + 0.3333333333333333, + 1.0, + 1.0, + null, + null, + null, + null, + null, + null, + null, + null, + null, + null, + null, + null, + null, + null + ], + "ci_hi": [ + 0.9598180439727066, + 0.9115384615384615, + 0.8936, + 0.880558288148421, + 0.8734495125624158, + 0.861583418570292, + 0.85897577092511, + 0.859657894736842, + 0.8461538461538461, + 0.9230769230769231, + 0.9545454545454546, + 0.9168269230769207, + 1.0, + 1.0, + 1.0, + 1.0, + null, + null, + null, + null, + null, + null, + null, + null, + null, + null, + null, + null, + null, + null + ] + } +} \ No newline at end of file diff --git a/results/metrics/qwen25-7b_gptq_w4_math500_run0_ci.json b/results/metrics/qwen25-7b_gptq_w4_math500_run0_ci.json new file mode 100644 index 0000000000000000000000000000000000000000..ed0250b871c95a168e1b26d8f6a6b8f8669dbf20 --- /dev/null +++ b/results/metrics/qwen25-7b_gptq_w4_math500_run0_ci.json @@ -0,0 +1,116 @@ +{ + "n": 500, + "accuracy": { + "mean": 0.6939124, + "ci_lo": 0.652, + "ci_hi": 0.734 + }, + "avg_ffs": { + "mean": 0.5029682660717776, + "ci_lo": 0.38848920863309355, + "ci_hi": 0.6293706293706294 + }, + "ecr": { + "mean": 0.9353212372758597, + "ci_lo": 0.9033852338344526, + "ci_hi": 0.9632433515246014 + }, + "ssr_curve_ci": { + "mean": [ + 0.8219631999999976, + 0.7062865072159568, + 0.7342478783267923, + 0.7261701513345061, + 0.7125028203628714, + 0.7459832347350905, + 0.7369743896746226, + 0.7387592296422759, + 0.7651301920790297, + 0.7424094693340728, + 0.7076346961610959, + 0.7221481980856038, + 0.5830786895289611, + 0.5830786895289611, + 0.6263220016309803, + 0.5032376316261248, + 0.3357115299823631, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "ci_lo": [ + 0.788, + 0.6653225806451613, + 0.6894692180843754, + 0.6745742805141622, + 0.6482388795831007, + 0.6736111111111112, + 0.6499893162393162, + 0.6307692307692307, + 0.6363636363636364, + 0.5882352941176471, + 0.519962962962963, + 0.5, + 0.2727272727272727, + 0.2727272727272727, + 0.25, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "ci_hi": [ + 0.854, + 0.7454545454545455, + 0.7783505154639175, + 0.7760313853392502, + 0.7734061446720248, + 0.8151325145442792, + 0.8191657082614527, + 0.8412698412698413, + 0.8837209302325582, + 0.8888888888888888, + 0.8928571428571429, + 0.9333333333333333, + 0.8825163398692787, + 0.8825163398692787, + 1.0, + 1.0, + 1.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ] + } +} \ No newline at end of file diff --git a/results/metrics/qwen25-7b_gptq_w4_math500_run0_metrics.json b/results/metrics/qwen25-7b_gptq_w4_math500_run0_metrics.json new file mode 100644 index 0000000000000000000000000000000000000000..61d10a1ba77fae8b71e330c0b9056d4424d4eabc --- /dev/null +++ b/results/metrics/qwen25-7b_gptq_w4_math500_run0_metrics.json @@ -0,0 +1,47 @@ +{ + "model": "qwen25-7b", + "quantization": "gptq_w4", + "accuracy": 0.694, + "accuracy_delta": 0.0, + "avg_ffs": 0.5032679738562091, + "median_ffs": 0.0, + "ecr": 0.9351899335232667, + "ssr_curve": [ + 0.822, + 0.7064777327935222, + 0.734375, + 0.7263513513513513, + 0.7128712871287128, + 0.7464788732394366, + 0.7378640776699029, + 0.7391304347826086, + 0.7659574468085106, + 0.7428571428571429, + 0.7083333333333334, + 0.7222222222222222, + 0.5833333333333334, + 0.5833333333333334, + 0.625, + 0.5, + 0.3333333333333333, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "error_type_dist": { + "conceptual": 0.5618556701030928, + "executional": 0.030927835051546393, + "logical": 0.3917525773195876, + "methodological": 0.015463917525773196 + } +} \ No newline at end of file diff --git a/results/metrics/qwen25-7b_gptq_w4_math500_run0_sig.json b/results/metrics/qwen25-7b_gptq_w4_math500_run0_sig.json new file mode 100644 index 0000000000000000000000000000000000000000..95e02ab4f36b645798605314ce20fbfd41f86451 --- /dev/null +++ b/results/metrics/qwen25-7b_gptq_w4_math500_run0_sig.json @@ -0,0 +1,7 @@ +{ + "n_pairs": 500, + "delta_acc_observed": 0.05400000000000005, + "delta_acc_ci_lo": 0.018000000000000016, + "delta_acc_ci_hi": 0.09000000000000008, + "p_value": 0.004 +} \ No newline at end of file diff --git a/results/metrics/qwen25-7b_gptq_w4_restored_gpqa_run0_ci.json b/results/metrics/qwen25-7b_gptq_w4_restored_gpqa_run0_ci.json new file mode 100644 index 0000000000000000000000000000000000000000..78721eebce9b148b6a43c809a4478d3efb3bf0d9 --- /dev/null +++ b/results/metrics/qwen25-7b_gptq_w4_restored_gpqa_run0_ci.json @@ -0,0 +1,116 @@ +{ + "n": 198, + "accuracy": { + "mean": 0.3083767676767677, + "ci_lo": 0.24242424242424243, + "ci_hi": 0.37373737373737376 + }, + "avg_ffs": { + "mean": 0.632478567591307, + "ci_lo": 0.46575081041514166, + "ci_hi": 0.8263888888888888 + }, + "ecr": { + "mean": 0.9292991960989246, + "ci_lo": 0.8967125985486121, + "ci_hi": 0.9574497108888449 + }, + "ssr_curve_ci": { + "mean": [ + 0.5652464646464578, + 0.41707541689491373, + 0.3386583931101849, + 0.30855665764431267, + 0.29743185616992013, + 0.30135501506761697, + 0.36558240678269327, + 0.3339046310271485, + 0.2500619848514425, + 0.23077251585900593, + 0.2354488865737635, + 0.23999433654271024, + 0.2851867656516091, + 0.3329895105568544, + 0.3329895105568544, + 0.3329895105568544, + 0.3329895105568544, + 0.29447775464580633, + 0.3836259820942796, + 0.49922833201542044, + 0.49922833201542044, + 0.4271122350076677, + 0.49525749083394616, + 0.49525749083394616, + 0.49525749083394616, + 0.49525749083394616, + 0.3303112979694067, + 0.3303112979694067, + 0.0, + 0.0 + ], + "ci_lo": [ + 0.494949494949495, + 0.3471502590673575, + 0.269652117545376, + 0.2356687898089172, + 0.2206871196754564, + 0.21495327102803738, + 0.25859456635318706, + 0.21052631578947367, + 0.13043478260869565, + 0.10808397683397684, + 0.10256410256410256, + 0.08, + 0.1, + 0.11764705882352941, + 0.11764705882352941, + 0.11764705882352941, + 0.11764705882352941, + 0.08333333333333333, + 0.1111111111111111, + 0.16666666666666666, + 0.16666666666666666, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "ci_hi": [ + 0.6363636363636364, + 0.48717948717948717, + 0.40782122905027934, + 0.38064516129032255, + 0.37594135338345863, + 0.39090909090909093, + 0.48, + 0.4642857142857143, + 0.38235294117647056, + 0.370396825396825, + 0.3870967741935484, + 0.4166666666666667, + 0.5, + 0.5652173913043478, + 0.5652173913043478, + 0.5652173913043478, + 0.5652173913043478, + 0.5333333333333333, + 0.6666666666666666, + 0.8181818181818182, + 0.8181818181818182, + 0.8333333333333334, + 1.0, + 1.0, + 1.0, + 1.0, + 1.0, + 1.0, + 0.0, + 0.0 + ] + } +} \ No newline at end of file diff --git a/results/metrics/qwen25-7b_gptq_w4_restored_gpqa_run0_metrics.json b/results/metrics/qwen25-7b_gptq_w4_restored_gpqa_run0_metrics.json new file mode 100644 index 0000000000000000000000000000000000000000..de8a08e2ea8f2b1035f82a3b117ca4c072cad14b --- /dev/null +++ b/results/metrics/qwen25-7b_gptq_w4_restored_gpqa_run0_metrics.json @@ -0,0 +1,47 @@ +{ + "model": "qwen25-7b", + "quantization": "gptq_w4_restored", + "accuracy": 0.30808080808080807, + "accuracy_delta": 0.0, + "avg_ffs": 0.635036496350365, + "median_ffs": 0.0, + "ecr": 0.9292356336473983, + "ssr_curve": [ + 0.5656565656565656, + 0.4175257731958763, + 0.3388888888888889, + 0.3081761006289308, + 0.2971014492753623, + 0.30097087378640774, + 0.36486486486486486, + 0.3333333333333333, + 0.25, + 0.23076923076923078, + 0.23529411764705882, + 0.24, + 0.2857142857142857, + 0.3333333333333333, + 0.3333333333333333, + 0.3333333333333333, + 0.3333333333333333, + 0.29411764705882354, + 0.38461538461538464, + 0.5, + 0.5, + 0.42857142857142855, + 0.5, + 0.5, + 0.5, + 0.5, + 0.3333333333333333, + 0.3333333333333333, + 0.0, + 0.0 + ], + "error_type_dist": { + "conceptual": 0.7843137254901961, + "executional": 0.0034602076124567475, + "logical": 0.20876585928489041, + "methodological": 0.0034602076124567475 + } +} \ No newline at end of file diff --git a/results/metrics/qwen25-7b_gptq_w4_restored_gsm8k_run0_metrics.json b/results/metrics/qwen25-7b_gptq_w4_restored_gsm8k_run0_metrics.json new file mode 100644 index 0000000000000000000000000000000000000000..9c9c7c9ddedc065a54411357c8fa016112a0ba5e --- /dev/null +++ b/results/metrics/qwen25-7b_gptq_w4_restored_gsm8k_run0_metrics.json @@ -0,0 +1,47 @@ +{ + "model": "qwen25-7b", + "quantization": "gptq_w4_restored", + "accuracy": 0.8786959818043972, + "accuracy_delta": 0.0, + "avg_ffs": 1.6, + "median_ffs": 1.0, + "ecr": 0.7970959595959594, + "ssr_curve": [ + 0.9704321455648218, + 0.9240898528272656, + 0.9040867389491243, + 0.89207258834766, + 0.8802308802308803, + 0.8728323699421965, + 0.864516129032258, + 0.7971014492753623, + 0.8055555555555556, + 0.7619047619047619, + 0.8, + 1.0, + 1.0, + 1.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "error_type_dist": { + "conceptual": 0.5150214592274678, + "executional": 0.07510729613733906, + "logical": 0.40772532188841204, + "methodological": 0.002145922746781116 + } +} \ No newline at end of file diff --git a/results/metrics/qwen25-7b_gptq_w4_restored_math500_run0_ci.json b/results/metrics/qwen25-7b_gptq_w4_restored_math500_run0_ci.json new file mode 100644 index 0000000000000000000000000000000000000000..aa741095ecabb58b09ece2c86b9be602c3d31529 --- /dev/null +++ b/results/metrics/qwen25-7b_gptq_w4_restored_math500_run0_ci.json @@ -0,0 +1,116 @@ +{ + "n": 500, + "accuracy": { + "mean": 0.7479752000000001, + "ci_lo": 0.71, + "ci_hi": 0.786 + }, + "avg_ffs": { + "mean": 0.8503586775464093, + "ci_lo": 0.6438266825491364, + "ci_hi": 1.0769379844961238 + }, + "ecr": { + "mean": 0.8973266816237024, + "ci_lo": 0.8527643977594014, + "ci_hi": 0.9366657150306703 + }, + "ssr_curve_ci": { + "mean": [ + 0.876084400000003, + 0.7875854194374848, + 0.7665338940508576, + 0.7955800113202738, + 0.7789613520209591, + 0.7671658814614322, + 0.7319310124587006, + 0.695910429638771, + 0.6570859332026908, + 0.5839268463131989, + 0.5143886485055151, + 0.4694822926495244, + 0.434066254849938, + 0.4168874853964888, + 0.47024156138578993, + 0.4997251980919913, + 0.38440408921360014, + 0.38440408921360014, + 0.2988101699324044, + 0.37557150998710254, + 0.37557150998710254, + 0.3360400555553099, + 0.4035511765384647, + 0.2544260519026025, + 0.2544260519026025, + 0.33898764808953846, + 0.33898764808953846, + 0.33898764808953846, + 0.5073889893510153, + 1.0 + ], + "ci_lo": [ + 0.846, + 0.751004016064257, + 0.7251140644505574, + 0.7492063492063492, + 0.7232988037447989, + 0.6986276188557615, + 0.6464646464646465, + 0.5954989299090423, + 0.5357142857142857, + 0.4444444444444444, + 0.3548387096774194, + 0.29411764705882354, + 0.25806451612903225, + 0.21739130434782608, + 0.2222222222222222, + 0.25, + 0.1111111111111111, + 0.1111111111111111, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 1.0 + ], + "ci_hi": [ + 0.904, + 0.8236472945891784, + 0.8060765614988621, + 0.8394680219617153, + 0.8325369676762866, + 0.8344370860927153, + 0.815225306036302, + 0.7976190476190477, + 0.7755190007037296, + 0.7254901960784313, + 0.6808510638297872, + 0.6521739130434783, + 0.6206896551724138, + 0.625, + 0.7142857142857143, + 0.75, + 0.6666666666666666, + 0.6666666666666666, + 0.625, + 0.75, + 0.75, + 0.8, + 1.0, + 1.0, + 1.0, + 1.0, + 1.0, + 1.0, + 1.0, + 1.0 + ] + } +} \ No newline at end of file diff --git a/results/metrics/qwen25-7b_gptq_w4_restored_math500_run0_metrics.json b/results/metrics/qwen25-7b_gptq_w4_restored_math500_run0_metrics.json new file mode 100644 index 0000000000000000000000000000000000000000..8ed2177539b840dc1ee240c647d9cc395a563ec4 --- /dev/null +++ b/results/metrics/qwen25-7b_gptq_w4_restored_math500_run0_metrics.json @@ -0,0 +1,47 @@ +{ + "model": "qwen25-7b", + "quantization": "gptq_w4_restored", + "accuracy": 0.748, + "accuracy_delta": 0.0, + "avg_ffs": 0.8492063492063492, + "median_ffs": 1.0, + "ecr": 0.8975203244176458, + "ssr_curve": [ + 0.876, + 0.7875751503006012, + 0.7665094339622641, + 0.7954545454545454, + 0.7788018433179723, + 0.7666666666666667, + 0.7314814814814815, + 0.6951219512195121, + 0.65625, + 0.5833333333333334, + 0.5135135135135135, + 0.46875, + 0.43333333333333335, + 0.4166666666666667, + 0.47058823529411764, + 0.5, + 0.38461538461538464, + 0.38461538461538464, + 0.3, + 0.375, + 0.375, + 0.3333333333333333, + 0.4, + 0.25, + 0.25, + 0.3333333333333333, + 0.3333333333333333, + 0.3333333333333333, + 0.5, + 1.0 + ], + "error_type_dist": { + "conceptual": 0.5255731922398589, + "executional": 0.021164021164021163, + "logical": 0.42857142857142855, + "methodological": 0.024691358024691357 + } +} \ No newline at end of file diff --git a/results/metrics/r1-llama-8b_awq_w4_gpqa_run0_ci.json b/results/metrics/r1-llama-8b_awq_w4_gpqa_run0_ci.json new file mode 100644 index 0000000000000000000000000000000000000000..6a0d8933dadb93332f1d0ca596e5f7d30256f745 --- /dev/null +++ b/results/metrics/r1-llama-8b_awq_w4_gpqa_run0_ci.json @@ -0,0 +1,116 @@ +{ + "n": 198, + "accuracy": { + "mean": 0.24758585858585858, + "ci_lo": 0.1919191919191919, + "ci_hi": 0.30808080808080807 + }, + "avg_ffs": { + "mean": 0.5039832552951411, + "ci_lo": 0.40425531914893614, + "ci_hi": 0.6129091123145749 + }, + "ecr": { + "mean": 0.9880485903354089, + "ci_lo": 0.9817796917455708, + "ci_hi": 0.9930130475051788 + }, + "ssr_curve_ci": { + "mean": [ + 0.5861090909090856, + 0.2728090909090915, + 0.2527272727272724, + 0.25658116679828236, + 0.25524673363740064, + 0.24224529098068912, + 0.24224529098068912, + 0.2394312918981755, + 0.2419762823817417, + 0.2392484539299488, + 0.24461583325589903, + 0.243126008169238, + 0.24451044290029264, + 0.24451044290029264, + 0.2295914478293636, + 0.22631588137564668, + 0.22766286744235495, + 0.22901071952206867, + 0.23175663061879054, + 0.23130664417572602, + 0.23422991139775923, + 0.24035191501028985, + 0.24035191501028985, + 0.23534072463242978, + 0.23844364771263654, + 0.2432486654772189, + 0.24311389990426338, + 0.24827837495543945, + 0.25372175059845675, + 0.26326579181413007 + ], + "ci_lo": [ + 0.5202020202020202, + 0.21212121212121213, + 0.19696969696969696, + 0.19796954314720813, + 0.19576719576719576, + 0.18324607329842932, + 0.18324607329842932, + 0.18031702662850205, + 0.18181818181818182, + 0.17894736842105263, + 0.1839080459770115, + 0.18285560675883256, + 0.1839080459770115, + 0.1839080459770115, + 0.16766467065868262, + 0.1638393299633264, + 0.1656441717791411, + 0.16666666666666666, + 0.1685830832708177, + 0.16770186335403728, + 0.16967751842751844, + 0.17419354838709677, + 0.17419354838709677, + 0.1697913898485251, + 0.17218543046357615, + 0.17567567567567569, + 0.174496644295302, + 0.17856450719822814, + 0.18248175182481752, + 0.19046780684104628 + ], + "ci_hi": [ + 0.6515151515151515, + 0.3333333333333333, + 0.31313131313131315, + 0.31794871794871793, + 0.3172043010752688, + 0.3027149223691775, + 0.3027149223691775, + 0.3010752688172043, + 0.3036649214659686, + 0.3010752688172043, + 0.3064607771260996, + 0.3063583815028902, + 0.3081395348837209, + 0.3081395348837209, + 0.29239766081871343, + 0.28917623155343236, + 0.2909148446490218, + 0.2926829268292683, + 0.2962962962962963, + 0.29677908113391976, + 0.3, + 0.3077044025157231, + 0.3077044025157231, + 0.3021660449418242, + 0.30714285714285716, + 0.3129251700680272, + 0.3137254901960784, + 0.32116788321167883, + 0.32639243197278905, + 0.3361793682990431 + ] + } +} \ No newline at end of file diff --git a/results/metrics/r1-llama-8b_awq_w4_gpqa_run0_metrics.json b/results/metrics/r1-llama-8b_awq_w4_gpqa_run0_metrics.json new file mode 100644 index 0000000000000000000000000000000000000000..6368a304c15df4b537c16aa207d03aff32c36e1d --- /dev/null +++ b/results/metrics/r1-llama-8b_awq_w4_gpqa_run0_metrics.json @@ -0,0 +1,47 @@ +{ + "model": "r1-llama-8b", + "quantization": "awq_w4", + "accuracy": 0.2474747474747475, + "accuracy_delta": 0.0, + "avg_ffs": 0.5033557046979866, + "median_ffs": 0.0, + "ecr": 0.9880374821760477, + "ssr_curve": [ + 0.5858585858585859, + 0.2727272727272727, + 0.25252525252525254, + 0.2564102564102564, + 0.2552083333333333, + 0.24210526315789474, + 0.24210526315789474, + 0.2393617021276596, + 0.24193548387096775, + 0.2391304347826087, + 0.24444444444444444, + 0.24293785310734464, + 0.24431818181818182, + 0.24431818181818182, + 0.22941176470588234, + 0.2261904761904762, + 0.2275449101796407, + 0.2289156626506024, + 0.23170731707317074, + 0.23125, + 0.23417721518987342, + 0.24025974025974026, + 0.24025974025974026, + 0.23529411764705882, + 0.23841059602649006, + 0.24324324324324326, + 0.24305555555555555, + 0.24822695035460993, + 0.2536231884057971, + 0.2631578947368421 + ], + "error_type_dist": { + "conceptual": 0.7406291203616501, + "executional": 0.006592578640045206, + "logical": 0.25221322282915803, + "methodological": 0.000565078169146732 + } +} \ No newline at end of file diff --git a/results/metrics/r1-llama-8b_awq_w4_gpqa_run0_sig.json b/results/metrics/r1-llama-8b_awq_w4_gpqa_run0_sig.json new file mode 100644 index 0000000000000000000000000000000000000000..7ee347cf311f58f8f45636866a3c3c7381c59569 --- /dev/null +++ b/results/metrics/r1-llama-8b_awq_w4_gpqa_run0_sig.json @@ -0,0 +1,7 @@ +{ + "n_pairs": 198, + "delta_acc_observed": -0.015151515151515166, + "delta_acc_ci_lo": -0.08585858585858586, + "delta_acc_ci_hi": 0.06060606060606058, + "p_value": 0.7484 +} \ No newline at end of file diff --git a/results/metrics/r1-llama-8b_awq_w4_gsm8k_run0_metrics.json b/results/metrics/r1-llama-8b_awq_w4_gsm8k_run0_metrics.json new file mode 100644 index 0000000000000000000000000000000000000000..26258ee3db7a0505be08c28a8287228c7323d87a --- /dev/null +++ b/results/metrics/r1-llama-8b_awq_w4_gsm8k_run0_metrics.json @@ -0,0 +1,47 @@ +{ + "model": "r1-llama-8b", + "quantization": "awq_w4", + "accuracy": 0.6482183472327521, + "accuracy_delta": 0.0, + "avg_ffs": 0.6702586206896551, + "median_ffs": 0.0, + "ecr": 0.8176335113895783, + "ssr_curve": [ + 0.7907505686125853, + 0.6960486322188449, + 0.6671949286846276, + 0.6506437768240343, + 0.634, + 0.6307870370370371, + 0.6008902077151336, + 0.5818965517241379, + 0.5768025078369906, + 0.5339366515837104, + 0.5321637426900585, + 0.5337837837837838, + 0.5161290322580645, + 0.4954128440366973, + 0.47474747474747475, + 0.43956043956043955, + 0.4235294117647059, + 0.3924050632911392, + 0.4, + 0.4027777777777778, + 0.38571428571428573, + 0.3787878787878788, + 0.38461538461538464, + 0.36507936507936506, + 0.3275862068965517, + 0.3076923076923077, + 0.29411764705882354, + 0.28, + 0.28, + 0.2708333333333333 + ], + "error_type_dist": { + "conceptual": 0.7345342498222328, + "executional": 0.040293908509125384, + "logical": 0.22422374970372125, + "methodological": 0.0009480919649205973 + } +} \ No newline at end of file diff --git a/results/metrics/r1-llama-8b_awq_w4_math500_run0_ci.json b/results/metrics/r1-llama-8b_awq_w4_math500_run0_ci.json new file mode 100644 index 0000000000000000000000000000000000000000..7a8951cafe277e959ea6f0e5e922413188e4836d --- /dev/null +++ b/results/metrics/r1-llama-8b_awq_w4_math500_run0_ci.json @@ -0,0 +1,116 @@ +{ + "n": 500, + "accuracy": { + "mean": 0.6602784, + "ci_lo": 0.618, + "ci_hi": 0.702 + }, + "avg_ffs": { + "mean": 0.6591961851990448, + "ci_lo": 0.5297217806041336, + "ci_hi": 0.7965164196268847 + }, + "ecr": { + "mean": 0.9658575797768403, + "ci_lo": 0.9509716256044652, + "ci_hi": 0.9783187716740227 + }, + "ssr_curve_ci": { + "mean": [ + 0.826188399999999, + 0.6942600000000043, + 0.6716195861535702, + 0.6608637333342388, + 0.6558892443861063, + 0.6377408763128497, + 0.6283455995295512, + 0.6219058111647392, + 0.6070476437498861, + 0.5958269130818367, + 0.5838752750027844, + 0.5675799726530462, + 0.54569484975349, + 0.5294285394848295, + 0.5083196468162798, + 0.48291257722168457, + 0.4650133288756974, + 0.451775307997617, + 0.4390933993946535, + 0.4350286605783337, + 0.40855589135647136, + 0.40684950692066263, + 0.3846891480786377, + 0.3762370580627854, + 0.35564559893898595, + 0.3475507782026942, + 0.3558703828284054, + 0.336709433010237, + 0.33702590817271794, + 0.3349098937543365 + ], + "ci_lo": [ + 0.792, + 0.652, + 0.6285140562248996, + 0.6178816803258577, + 0.6116850459589038, + 0.592425843700793, + 0.5821917808219178, + 0.5744174072694375, + 0.5583111493892892, + 0.5459314237506026, + 0.5321637426900585, + 0.5133531157270029, + 0.49006456953642386, + 0.47241126586543863, + 0.450179889298893, + 0.4232209737827715, + 0.403770154373928, + 0.3891209318984745, + 0.37548171936758895, + 0.3706896551724138, + 0.34374655837004403, + 0.3409090909090909, + 0.3188405797101449, + 0.3097791602295299, + 0.28888591800356506, + 0.2808885996163333, + 0.2864967897271268, + 0.2670807453416149, + 0.26514923329682366, + 0.2614306232046959 + ], + "ci_hi": [ + 0.8580499999999992, + 0.734, + 0.712, + 0.703030303030303, + 0.6983471074380165, + 0.6823336737671432, + 0.6735668609357293, + 0.6682692307692307, + 0.6546799108936201, + 0.6451635429807172, + 0.6353649686300215, + 0.6208031796683571, + 0.6019417475728155, + 0.5874629988772072, + 0.5667870036101083, + 0.5421245421245421, + 0.5263207735247208, + 0.5137282027435479, + 0.5021645021645021, + 0.49804034257381113, + 0.47342995169082125, + 0.4724770642201835, + 0.45128640776699025, + 0.4417503951230526, + 0.4247409225494778, + 0.4189985949887933, + 0.4293785310734463, + 0.41250726744186034, + 0.41305072463768105, + 0.4110429447852761 + ] + } +} \ No newline at end of file diff --git a/results/metrics/r1-llama-8b_awq_w4_math500_run0_metrics.json b/results/metrics/r1-llama-8b_awq_w4_math500_run0_metrics.json new file mode 100644 index 0000000000000000000000000000000000000000..fea10ff9e5d9d38849b2eec0fa68c386ee258573 --- /dev/null +++ b/results/metrics/r1-llama-8b_awq_w4_math500_run0_metrics.json @@ -0,0 +1,47 @@ +{ + "model": "r1-llama-8b", + "quantization": "awq_w4", + "accuracy": 0.66, + "accuracy_delta": 0.0, + "avg_ffs": 0.6588235294117647, + "median_ffs": 0.0, + "ecr": 0.9657817970235313, + "ssr_curve": [ + 0.826, + 0.694, + 0.6713426853707415, + 0.6605691056910569, + 0.6556016597510373, + 0.6375545851528385, + 0.6281179138321995, + 0.6217494089834515, + 0.6069651741293532, + 0.5957446808510638, + 0.5837988826815642, + 0.5676470588235294, + 0.546031746031746, + 0.5298013245033113, + 0.5087719298245614, + 0.4833948339483395, + 0.4653846153846154, + 0.452, + 0.4393305439330544, + 0.4353448275862069, + 0.4090909090909091, + 0.4074074074074074, + 0.3853658536585366, + 0.3768844221105528, + 0.35638297872340424, + 0.34831460674157305, + 0.3567251461988304, + 0.3374233128834356, + 0.3375796178343949, + 0.33548387096774196 + ], + "error_type_dist": { + "conceptual": 0.7354963851172632, + "executional": 0.018867924528301886, + "logical": 0.23752424616469758, + "methodological": 0.00811144418973726 + } +} \ No newline at end of file diff --git a/results/metrics/r1-llama-8b_awq_w4_math500_run0_sig.json b/results/metrics/r1-llama-8b_awq_w4_math500_run0_sig.json new file mode 100644 index 0000000000000000000000000000000000000000..f2626db0c147be0c894d4ed6ce9e9e6bc09c2d0e --- /dev/null +++ b/results/metrics/r1-llama-8b_awq_w4_math500_run0_sig.json @@ -0,0 +1,7 @@ +{ + "n_pairs": 500, + "delta_acc_observed": -0.18800000000000006, + "delta_acc_ci_lo": -0.23600000000000004, + "delta_acc_ci_hi": -0.14200000000000002, + "p_value": 0.0 +} \ No newline at end of file diff --git a/results/metrics/r1-llama-8b_awq_w4_restored_gsm8k_run0_ci.json b/results/metrics/r1-llama-8b_awq_w4_restored_gsm8k_run0_ci.json new file mode 100644 index 0000000000000000000000000000000000000000..d2480655570164489dd96cfe50bf92c941d37f73 --- /dev/null +++ b/results/metrics/r1-llama-8b_awq_w4_restored_gsm8k_run0_ci.json @@ -0,0 +1,116 @@ +{ + "n": 1319, + "accuracy": { + "mean": 0.5538495830174375, + "ci_lo": 0.5269143290371494, + "ci_hi": 0.5807429871114481 + }, + "avg_ffs": { + "mean": 0.3997386875002321, + "ci_lo": 0.3355926515754749, + "ci_hi": 0.4700858250276855 + }, + "ecr": { + "mean": 0.8607660736659057, + "ci_lo": 0.8417515689502542, + "ci_hi": 0.878554719275398 + }, + "ssr_curve_ci": { + "mean": [ + 0.6790782410917386, + 0.5852604819684138, + 0.5638594405651735, + 0.5555082479062055, + 0.5473610544509252, + 0.5291402975287399, + 0.5080742263501905, + 0.4744868815791439, + 0.46411927858759483, + 0.47545476283668386, + 0.4627992567576668, + 0.46277648620631334, + 0.44964759722978204, + 0.4516054970877596, + 0.4385866808218154, + 0.42972532977768196, + 0.4244819141750499, + 0.4250725668222648, + 0.4209584604633516, + 0.4154998072991936, + 0.41396304295527453, + 0.40756395625388103, + 0.41176555199890175, + 0.39936659138068437, + 0.3853955784167955, + 0.384088643993255, + 0.38018224619577584, + 0.3840114346353488, + 0.366695914856007, + 0.34862000251563124 + ], + "ci_lo": [ + 0.6535253980288097, + 0.5584398128752968, + 0.5369645981694932, + 0.5279106858054227, + 0.518355044345898, + 0.4985699984430526, + 0.47417713259097305, + 0.4360196795541623, + 0.4194881082558775, + 0.425414364640884, + 0.4048579493716991, + 0.40310077519379844, + 0.388, + 0.38822768762677484, + 0.375, + 0.36363636363636365, + 0.35930426716141, + 0.3598021320093458, + 0.35406698564593303, + 0.34895530523255813, + 0.3467276073349202, + 0.3387894073139975, + 0.3428531300160514, + 0.3301851937912445, + 0.31471796041749844, + 0.3136094674556213, + 0.3089005235602094, + 0.31033937145351376, + 0.29012345679012347, + 0.27333229813664595 + ], + "ci_hi": [ + 0.7050796057619408, + 0.6119206239868452, + 0.5909818552736079, + 0.5834670947030498, + 0.5761531766753699, + 0.5599641745460401, + 0.5415728444957659, + 0.5125638697690131, + 0.5087736531899494, + 0.5274151436031331, + 0.5202708935873143, + 0.5246696905657783, + 0.5124026613360383, + 0.5150375939849624, + 0.5023474178403756, + 0.49590644211065565, + 0.4905725289942876, + 0.49099497727816305, + 0.4876906784702886, + 0.48246370235934655, + 0.48128404717876955, + 0.47593582887700536, + 0.48023151216740834, + 0.4692737430167598, + 0.45641618497109815, + 0.45454545454545453, + 0.45223309068850465, + 0.4565248806232721, + 0.44171779141104295, + 0.4264731591937474 + ] + } +} \ No newline at end of file diff --git a/results/metrics/r1-llama-8b_awq_w4_restored_math500_run0_ci.json b/results/metrics/r1-llama-8b_awq_w4_restored_math500_run0_ci.json new file mode 100644 index 0000000000000000000000000000000000000000..e1c0e20e3dc5df846d3de4080aba12e91abc4cab --- /dev/null +++ b/results/metrics/r1-llama-8b_awq_w4_restored_math500_run0_ci.json @@ -0,0 +1,116 @@ +{ + "n": 500, + "accuracy": { + "mean": 0.47172200000000003, + "ci_lo": 0.43, + "ci_hi": 0.514 + }, + "avg_ffs": { + "mean": 0.5911770646334961, + "ci_lo": 0.503968253968254, + "ci_hi": 0.6844106463878327 + }, + "ecr": { + "mean": 0.9653287854278844, + "ci_lo": 0.9538673482569967, + "ci_hi": 0.9754733208089089 + }, + "ssr_curve_ci": { + "mean": [ + 0.7318864000000024, + 0.5008836575859381, + 0.484544128681263, + 0.4742524829228788, + 0.46561200170315054, + 0.45865167934146894, + 0.4525804204203134, + 0.44442761406149284, + 0.4465405707627004, + 0.4245072119650639, + 0.4086739800219384, + 0.4030833821965393, + 0.39818164285218965, + 0.3839792770829044, + 0.37660264196289595, + 0.35644646435573346, + 0.3394308303090971, + 0.33312006896475327, + 0.32327036698509837, + 0.3105060877718699, + 0.30625845752241965, + 0.2983399121305476, + 0.2894130118087134, + 0.2936592952161559, + 0.2922819579657003, + 0.2974918862600696, + 0.2833832565614938, + 0.2721683239041574, + 0.2679918911555058, + 0.2751186825594855 + ], + "ci_lo": [ + 0.694, + 0.458, + 0.4424184030979163, + 0.4323163685536911, + 0.42323055032504875, + 0.4159622372857667, + 0.4075702981662627, + 0.3985595708281412, + 0.4, + 0.3777721088435374, + 0.36061381074168797, + 0.3549158894645941, + 0.34844192634560905, + 0.3342694673093174, + 0.32732596232596234, + 0.30695759120271837, + 0.2879245238765363, + 0.28205128205128205, + 0.27152204091850435, + 0.25838926174496646, + 0.2534722222222222, + 0.24460234731922492, + 0.23505515305326738, + 0.23694514901712113, + 0.23599622641509432, + 0.2396279479533749, + 0.2242152466367713, + 0.21296067106710673, + 0.2074468085106383, + 0.21142636229749634 + ], + "ci_hi": [ + 0.77, + 0.544, + 0.5272727272727272, + 0.5182186234817814, + 0.5091659065635425, + 0.501094917624437, + 0.49777802359882006, + 0.49074074074074076, + 0.49394746054435656, + 0.4720245336577452, + 0.45776889430661116, + 0.45355822344975044, + 0.4492007911508314, + 0.43515850144092216, + 0.4281711822660097, + 0.40785799677497075, + 0.3904761904761905, + 0.3850936702245271, + 0.3754358600899979, + 0.36200716845878134, + 0.35906040268456374, + 0.3507495417648599, + 0.34328639819769075, + 0.3493975903614458, + 0.34831575796647635, + 0.3562231759656652, + 0.34299872122762143, + 0.3333333333333333, + 0.3298429319371728, + 0.34000253807106595 + ] + } +} \ No newline at end of file diff --git a/results/metrics/r1-llama-8b_awq_w4_restored_math500_run0_metrics.json b/results/metrics/r1-llama-8b_awq_w4_restored_math500_run0_metrics.json new file mode 100644 index 0000000000000000000000000000000000000000..de4cc5864a575c2ebabc9b6808738b6c1cbbd493 --- /dev/null +++ b/results/metrics/r1-llama-8b_awq_w4_restored_math500_run0_metrics.json @@ -0,0 +1,47 @@ +{ + "model": "r1-llama-8b", + "quantization": "awq_w4_restored", + "accuracy": 0.472, + "accuracy_delta": 0.0, + "avg_ffs": 0.5909090909090909, + "median_ffs": 0.0, + "ecr": 0.9654005170191907, + "ssr_curve": [ + 0.732, + 0.501002004008016, + 0.48484848484848486, + 0.4746450304259635, + 0.4658385093167702, + 0.4589473684210526, + 0.45274725274725275, + 0.4447004608294931, + 0.44680851063829785, + 0.4246913580246914, + 0.4088541666666667, + 0.4032258064516129, + 0.3983516483516483, + 0.384180790960452, + 0.37681159420289856, + 0.3564954682779456, + 0.33962264150943394, + 0.3333333333333333, + 0.3234323432343234, + 0.3108108108108108, + 0.30662020905923343, + 0.29856115107913667, + 0.2894736842105263, + 0.29365079365079366, + 0.29218106995884774, + 0.2974137931034483, + 0.2831050228310502, + 0.27184466019417475, + 0.2676767676767677, + 0.27472527472527475 + ], + "error_type_dist": { + "conceptual": 0.7536068247396813, + "executional": 0.017563668297578723, + "logical": 0.22218040396437083, + "methodological": 0.0066491029983690875 + } +} \ No newline at end of file diff --git a/results/metrics/r1-llama-8b_bnb_nf4_w4_gpqa_run0_metrics.json b/results/metrics/r1-llama-8b_bnb_nf4_w4_gpqa_run0_metrics.json new file mode 100644 index 0000000000000000000000000000000000000000..d60d7709198088edc3afb8e0a096582e5ef0cf15 --- /dev/null +++ b/results/metrics/r1-llama-8b_bnb_nf4_w4_gpqa_run0_metrics.json @@ -0,0 +1,47 @@ +{ + "model": "r1-llama-8b", + "quantization": "bnb_nf4_w4", + "accuracy": 0.1919191919191919, + "accuracy_delta": 0.0, + "avg_ffs": 0.63125, + "median_ffs": 1.0, + "ecr": 0.9820504261007766, + "ssr_curve": [ + 0.6161616161616161, + 0.2222222222222222, + 0.20812182741116753, + 0.20918367346938777, + 0.211340206185567, + 0.20526315789473684, + 0.19576719576719576, + 0.1891891891891892, + 0.18478260869565216, + 0.1878453038674033, + 0.1878453038674033, + 0.18888888888888888, + 0.18435754189944134, + 0.1864406779661017, + 0.1839080459770115, + 0.1871345029239766, + 0.18452380952380953, + 0.18452380952380953, + 0.18181818181818182, + 0.18518518518518517, + 0.18633540372670807, + 0.1910828025477707, + 0.19480519480519481, + 0.19594594594594594, + 0.19310344827586207, + 0.19718309859154928, + 0.1956521739130435, + 0.20300751879699247, + 0.2076923076923077, + 0.203125 + ], + "error_type_dist": { + "conceptual": 0.7444424030865332, + "executional": 0.00551166636046298, + "logical": 0.24949476391695757, + "methodological": 0.000551166636046298 + } +} \ No newline at end of file diff --git a/results/metrics/r1-llama-8b_bnb_nf4_w4_gpqa_run0_sig.json b/results/metrics/r1-llama-8b_bnb_nf4_w4_gpqa_run0_sig.json new file mode 100644 index 0000000000000000000000000000000000000000..66a7b0e581e294215a72abada3e53bef3fd49e46 --- /dev/null +++ b/results/metrics/r1-llama-8b_bnb_nf4_w4_gpqa_run0_sig.json @@ -0,0 +1,7 @@ +{ + "n_pairs": 198, + "delta_acc_observed": 0.03535353535353536, + "delta_acc_ci_lo": -0.030303030303030304, + "delta_acc_ci_hi": 0.10606060606060605, + "p_value": 0.3552 +} \ No newline at end of file diff --git a/results/metrics/r1-llama-8b_bnb_nf4_w4_gsm8k_run0_ci.json b/results/metrics/r1-llama-8b_bnb_nf4_w4_gsm8k_run0_ci.json new file mode 100644 index 0000000000000000000000000000000000000000..a9a8ef8159b13762204f34785a3719696d709f1f --- /dev/null +++ b/results/metrics/r1-llama-8b_bnb_nf4_w4_gsm8k_run0_ci.json @@ -0,0 +1,116 @@ +{ + "n": 1319, + "accuracy": { + "mean": 0.6539213040181956, + "ci_lo": 0.6284874905231236, + "ci_hi": 0.6800606520090978 + }, + "avg_ffs": { + "mean": 0.7586694553687747, + "ci_lo": 0.6491587424443462, + "ci_hi": 0.8792401146328994 + }, + "ecr": { + "mean": 0.8058983827046329, + "ci_lo": 0.7791931475647129, + "ci_hi": 0.8316099490360587 + }, + "ssr_curve_ci": { + "mean": [ + 0.8092761182714178, + 0.7043951309409076, + 0.6794631803841585, + 0.6682864653484831, + 0.6637185133038286, + 0.6432595124753404, + 0.6298323641861275, + 0.6160398321841276, + 0.6108502890660309, + 0.6189766299626962, + 0.6166621500361136, + 0.6063850203462586, + 0.5830470391004281, + 0.5838964750105949, + 0.5413188535280379, + 0.5453260071905861, + 0.5473506265456675, + 0.5383287732178385, + 0.5305445009434799, + 0.5145909009298857, + 0.5058676078608482, + 0.4758591269450997, + 0.45258131348364694, + 0.44650341285960865, + 0.4311492270218363, + 0.4314467995890249, + 0.4296359547922957, + 0.4177652013744095, + 0.41519334649845996, + 0.41519334649845996 + ], + "ci_lo": [ + 0.7877179681576952, + 0.680365296803653, + 0.6536312849162011, + 0.6411863978474213, + 0.6332245163296166, + 0.6097232930622972, + 0.5927719589543431, + 0.5720809466709345, + 0.5585585585585585, + 0.5572504999859159, + 0.5480718085106383, + 0.5340894217039505, + 0.5058122828120823, + 0.503448275862069, + 0.456787693205016, + 0.45925576519916145, + 0.4596669653524492, + 0.44826688218390803, + 0.44093043415444627, + 0.42342342342342343, + 0.4117647058823529, + 0.37999074074074074, + 0.3522727272727273, + 0.345679012345679, + 0.3285602678571429, + 0.3235294117647059, + 0.32, + 0.30434782608695654, + 0.3, + 0.3 + ], + "ci_hi": [ + 0.8309325246398787, + 0.7290076335877863, + 0.7054585677403938, + 0.6963170824202596, + 0.6947051575696844, + 0.6779677067282999, + 0.6682042349274947, + 0.6600032679738562, + 0.662170608108108, + 0.6787185326672543, + 0.6837256081261693, + 0.6767676767676768, + 0.6588235294117647, + 0.6629483146067411, + 0.6240626734621789, + 0.6285768398268398, + 0.6335896757614692, + 0.6290322580645161, + 0.6203720600162206, + 0.6083423913043476, + 0.6018518518518519, + 0.5754716981132075, + 0.5566062050184788, + 0.5500140449438201, + 0.5408253301320527, + 0.5421686746987951, + 0.5428659611992943, + 0.532258064516129, + 0.5312598892405062, + 0.5312598892405062 + ] + } +} \ No newline at end of file diff --git a/results/metrics/r1-llama-8b_bnb_nf4_w4_gsm8k_run0_sig.json b/results/metrics/r1-llama-8b_bnb_nf4_w4_gsm8k_run0_sig.json new file mode 100644 index 0000000000000000000000000000000000000000..40965f8339a538d73a61682af27630746fca9923 --- /dev/null +++ b/results/metrics/r1-llama-8b_bnb_nf4_w4_gsm8k_run0_sig.json @@ -0,0 +1,7 @@ +{ + "n_pairs": 1319, + "delta_acc_observed": -0.09552691432903715, + "delta_acc_ci_lo": -0.12357846853677024, + "delta_acc_ci_hi": -0.06823351023502655, + "p_value": 0.0 +} \ No newline at end of file diff --git a/results/metrics/r1-llama-8b_bnb_nf4_w4_math500_run0_metrics.json b/results/metrics/r1-llama-8b_bnb_nf4_w4_math500_run0_metrics.json new file mode 100644 index 0000000000000000000000000000000000000000..105fe520bad8a4e5a6d6882ff1e8a194978dd27d --- /dev/null +++ b/results/metrics/r1-llama-8b_bnb_nf4_w4_math500_run0_metrics.json @@ -0,0 +1,47 @@ +{ + "model": "r1-llama-8b", + "quantization": "bnb_nf4_w4", + "accuracy": 0.634, + "accuracy_delta": 0.0, + "avg_ffs": 0.7213114754098361, + "median_ffs": 1.0, + "ecr": 0.9478199237742374, + "ssr_curve": [ + 0.848, + 0.666, + 0.6484848484848484, + 0.6369168356997972, + 0.629399585921325, + 0.6175213675213675, + 0.6062639821029083, + 0.5990566037735849, + 0.5833333333333334, + 0.5785340314136126, + 0.5619834710743802, + 0.5384615384615384, + 0.514018691588785, + 0.4900662251655629, + 0.46503496503496505, + 0.45323741007194246, + 0.43820224719101125, + 0.43137254901960786, + 0.4098360655737705, + 0.4092827004219409, + 0.39473684210526316, + 0.37209302325581395, + 0.36538461538461536, + 0.3434343434343434, + 0.3471502590673575, + 0.34054054054054056, + 0.33707865168539325, + 0.33727810650887574, + 0.33540372670807456, + 0.33544303797468356 + ], + "error_type_dist": { + "conceptual": 0.7483314341526941, + "executional": 0.018394921048347713, + "logical": 0.22497151229041185, + "methodological": 0.008302132508546313 + } +} \ No newline at end of file diff --git a/results/metrics/r1-llama-8b_bnb_nf4_w4_math500_run0_sig.json b/results/metrics/r1-llama-8b_bnb_nf4_w4_math500_run0_sig.json new file mode 100644 index 0000000000000000000000000000000000000000..28de8804a86bdc5d2ed89a96cdc3389cdcd5a5d3 --- /dev/null +++ b/results/metrics/r1-llama-8b_bnb_nf4_w4_math500_run0_sig.json @@ -0,0 +1,7 @@ +{ + "n_pairs": 500, + "delta_acc_observed": -0.136, + "delta_acc_ci_lo": -0.18000000000000005, + "delta_acc_ci_hi": -0.09199999999999997, + "p_value": 0.0 +} \ No newline at end of file diff --git a/results/metrics/r1-llama-8b_bnb_nf4_w4_restored_gpqa_run0_ci.json b/results/metrics/r1-llama-8b_bnb_nf4_w4_restored_gpqa_run0_ci.json new file mode 100644 index 0000000000000000000000000000000000000000..449b8b4c57af67980a014cc615931a962a35e825 --- /dev/null +++ b/results/metrics/r1-llama-8b_bnb_nf4_w4_restored_gpqa_run0_ci.json @@ -0,0 +1,116 @@ +{ + "n": 198, + "accuracy": { + "mean": 0.2279111111111111, + "ci_lo": 0.1717171717171717, + "ci_hi": 0.2878787878787879 + }, + "avg_ffs": { + "mean": 0.5808505127927905, + "ci_lo": 0.4266451612903226, + "ci_hi": 0.7638910455486542 + }, + "ecr": { + "mean": 0.9935639166094618, + "ci_lo": 0.9905204205589955, + "ci_hi": 0.9961170545595286 + }, + "ssr_curve_ci": { + "mean": [ + 0.5501616161616039, + 0.2783606060606056, + 0.2441882101675159, + 0.24543208905781044, + 0.2467287012571267, + 0.23984870837764347, + 0.24370891597465016, + 0.23545624960706021, + 0.2380495184754989, + 0.23384290410688133, + 0.23086931198381389, + 0.23086931198381389, + 0.23618648836665634, + 0.23618648836665634, + 0.23752133743326975, + 0.23752133743326975, + 0.23582545731727156, + 0.2372236080149986, + 0.23405285523074745, + 0.2321777540198068, + 0.23164430988039938, + 0.2313585256150353, + 0.22943889839158566, + 0.22871318245604694, + 0.22804073714128958, + 0.2284417435506782, + 0.23187153003809857, + 0.23479760700135743, + 0.23456125565597466, + 0.23197806108023714 + ], + "ci_lo": [ + 0.4847222222222223, + 0.21717171717171718, + 0.18652849740932642, + 0.1875, + 0.18947368421052632, + 0.1827956989247312, + 0.18478260869565216, + 0.17679558011049723, + 0.1793478260869565, + 0.1751412429378531, + 0.17142857142857143, + 0.17142857142857143, + 0.1746987951807229, + 0.1746987951807229, + 0.17613636363636365, + 0.17613636363636365, + 0.1744186046511628, + 0.17575757575757575, + 0.17261904761904762, + 0.17058823529411765, + 0.16981132075471697, + 0.16770186335403728, + 0.16560509554140126, + 0.1643835616438356, + 0.16249155405405405, + 0.1619718309859155, + 0.16417910447761194, + 0.16428571428571428, + 0.1623931623931624, + 0.15966386554621848 + ], + "ci_hi": [ + 0.6212121212121212, + 0.3383838383838384, + 0.30571005604314244, + 0.30851063829787234, + 0.31016042780748665, + 0.30158730158730157, + 0.3064516129032258, + 0.29891304347826086, + 0.30113636363636365, + 0.2967032967032967, + 0.29480479484506766, + 0.29480479484506766, + 0.30057803468208094, + 0.30057803468208094, + 0.3016759776536313, + 0.3016759776536313, + 0.3005789212383418, + 0.3028571428571429, + 0.30057803468208094, + 0.2981366459627329, + 0.29870327842888816, + 0.3, + 0.29931972789115646, + 0.29931972789115646, + 0.3, + 0.30233786439567617, + 0.3059785531800456, + 0.3114754098360656, + 0.312067819148936, + 0.30952380952380953 + ] + } +} \ No newline at end of file diff --git a/results/metrics/r1-llama-8b_bnb_nf4_w4_restored_gpqa_run0_metrics.json b/results/metrics/r1-llama-8b_bnb_nf4_w4_restored_gpqa_run0_metrics.json new file mode 100644 index 0000000000000000000000000000000000000000..d6214a06cde4dd22259eaf372f4d040a93f8d0ab --- /dev/null +++ b/results/metrics/r1-llama-8b_bnb_nf4_w4_restored_gpqa_run0_metrics.json @@ -0,0 +1,47 @@ +{ + "model": "r1-llama-8b", + "quantization": "bnb_nf4_w4_restored", + "accuracy": 0.22727272727272727, + "accuracy_delta": 0.0, + "avg_ffs": 0.5816993464052288, + "median_ffs": 0.0, + "ecr": 0.9935599099683323, + "ssr_curve": [ + 0.5505050505050505, + 0.2777777777777778, + 0.24352331606217617, + 0.24479166666666666, + 0.24607329842931938, + 0.2393617021276596, + 0.24324324324324326, + 0.23497267759562843, + 0.23756906077348067, + 0.23333333333333334, + 0.2303370786516854, + 0.2303370786516854, + 0.23563218390804597, + 0.23563218390804597, + 0.23699421965317918, + 0.23699421965317918, + 0.23529411764705882, + 0.23668639053254437, + 0.23353293413173654, + 0.23170731707317074, + 0.23125, + 0.23076923076923078, + 0.22875816993464052, + 0.22818791946308725, + 0.22758620689655173, + 0.22794117647058823, + 0.23134328358208955, + 0.234375, + 0.23387096774193547, + 0.23140495867768596 + ], + "error_type_dist": { + "conceptual": 0.7450509898020395, + "executional": 0.003799240151969606, + "logical": 0.24875024995001, + "methodological": 0.0023995200959808036 + } +} \ No newline at end of file diff --git a/results/metrics/r1-llama-8b_bnb_nf4_w4_restored_gsm8k_run0_metrics.json b/results/metrics/r1-llama-8b_bnb_nf4_w4_restored_gsm8k_run0_metrics.json new file mode 100644 index 0000000000000000000000000000000000000000..851861d80bd95c750769051da911ab005b7805a2 --- /dev/null +++ b/results/metrics/r1-llama-8b_bnb_nf4_w4_restored_gsm8k_run0_metrics.json @@ -0,0 +1,47 @@ +{ + "model": "r1-llama-8b", + "quantization": "bnb_nf4_w4_restored", + "accuracy": 0.5579984836997726, + "accuracy_delta": 0.0, + "avg_ffs": 0.3687821612349914, + "median_ffs": 0.0, + "ecr": 0.8678137655057995, + "ssr_curve": [ + 0.6800606520090978, + 0.583969465648855, + 0.5687645687645687, + 0.5660828025477707, + 0.5616319444444444, + 0.5461465271170314, + 0.5151869158878505, + 0.4867807153965785, + 0.47912524850894633, + 0.4690721649484536, + 0.47112462006079026, + 0.46, + 0.45652173913043476, + 0.4517374517374517, + 0.4534412955465587, + 0.45267489711934156, + 0.4435146443514644, + 0.4334763948497854, + 0.42731277533039647, + 0.4200913242009132, + 0.41203703703703703, + 0.4, + 0.3961352657004831, + 0.400990099009901, + 0.4020100502512563, + 0.4020618556701031, + 0.4052631578947368, + 0.39344262295081966, + 0.39664804469273746, + 0.38922155688622756 + ], + "error_type_dist": { + "conceptual": 0.8049426301853486, + "executional": 0.023452275879460346, + "logical": 0.1709746564115496, + "methodological": 0.0006304375236414071 + } +} \ No newline at end of file diff --git a/results/metrics/r1-llama-8b_gptq_w4_gpqa_run0_sig.json b/results/metrics/r1-llama-8b_gptq_w4_gpqa_run0_sig.json new file mode 100644 index 0000000000000000000000000000000000000000..d18d2bc66d106e4008b642caf7769829b2caea54 --- /dev/null +++ b/results/metrics/r1-llama-8b_gptq_w4_gpqa_run0_sig.json @@ -0,0 +1,7 @@ +{ + "n_pairs": 198, + "delta_acc_observed": 0.02020202020202022, + "delta_acc_ci_lo": -0.04545454545454544, + "delta_acc_ci_hi": 0.08585858585858586, + "p_value": 0.5944 +} \ No newline at end of file diff --git a/results/metrics/r1-llama-8b_gptq_w4_gsm8k_run0_metrics.json b/results/metrics/r1-llama-8b_gptq_w4_gsm8k_run0_metrics.json new file mode 100644 index 0000000000000000000000000000000000000000..13daae07792d1540641d31340b7ae020bf1a6f5b --- /dev/null +++ b/results/metrics/r1-llama-8b_gptq_w4_gsm8k_run0_metrics.json @@ -0,0 +1,47 @@ +{ + "model": "r1-llama-8b", + "quantization": "gptq_w4", + "accuracy": 0.6269901440485216, + "accuracy_delta": 0.0, + "avg_ffs": 0.5630081300813008, + "median_ffs": 0.0, + "ecr": 0.8508756090274743, + "ssr_curve": [ + 0.7611827141774071, + 0.6572948328267477, + 0.6426307448494454, + 0.6348936170212766, + 0.6217038539553753, + 0.6070559610705596, + 0.5771144278606966, + 0.5649202733485194, + 0.5570469798657718, + 0.526829268292683, + 0.5256410256410257, + 0.5298507462686567, + 0.5040650406504065, + 0.4956521739130435, + 0.48214285714285715, + 0.46601941747572817, + 0.4387755102040816, + 0.43617021276595747, + 0.4157303370786517, + 0.4069767441860465, + 0.4117647058823529, + 0.4125, + 0.41025641025641024, + 0.4155844155844156, + 0.41333333333333333, + 0.41333333333333333, + 0.41333333333333333, + 0.4057971014492754, + 0.40298507462686567, + 0.40298507462686567 + ], + "error_type_dist": { + "conceptual": 0.7759391771019678, + "executional": 0.02817531305903399, + "logical": 0.19521466905187834, + "methodological": 0.0006708407871198568 + } +} \ No newline at end of file diff --git a/results/metrics/r1-llama-8b_gptq_w4_math500_run0_ci.json b/results/metrics/r1-llama-8b_gptq_w4_math500_run0_ci.json new file mode 100644 index 0000000000000000000000000000000000000000..0d6b59249c631ec2b3487d62afd91b464c31b44f --- /dev/null +++ b/results/metrics/r1-llama-8b_gptq_w4_math500_run0_ci.json @@ -0,0 +1,116 @@ +{ + "n": 500, + "accuracy": { + "mean": 0.5877952, + "ci_lo": 0.544, + "ci_hi": 0.632 + }, + "avg_ffs": { + "mean": 0.4999723980462574, + "ci_lo": 0.40608760114590137, + "ci_hi": 0.6067429366204865 + }, + "ecr": { + "mean": 0.9654562950697054, + "ci_lo": 0.9535297179651581, + "ci_hi": 0.9754792017461859 + }, + "ssr_curve_ci": { + "mean": [ + 0.7577927999999989, + 0.6091008530329064, + 0.5965990103990365, + 0.5953468126937098, + 0.5828240972715899, + 0.5703084804397279, + 0.5593345744880749, + 0.5540533983349477, + 0.5404803583200183, + 0.5244766438386844, + 0.5013360959928438, + 0.4767940645808539, + 0.4690158631057886, + 0.4606387007077694, + 0.4500002922570554, + 0.4223578714204785, + 0.406783073459411, + 0.3897171217705542, + 0.38063274579332984, + 0.37087958544308225, + 0.35803094192776563, + 0.3505199677165461, + 0.33877697641290466, + 0.3128308630405522, + 0.3079075318541842, + 0.2942165215574197, + 0.2862656354715286, + 0.27394640524992936, + 0.27406765053525783, + 0.2666344336334726 + ], + "ci_lo": [ + 0.71995, + 0.566, + 0.5533199195171026, + 0.5521472392638037, + 0.5386221294363257, + 0.5254237288135594, + 0.512642231515821, + 0.5078980377764483, + 0.4936370415432014, + 0.4756756756756757, + 0.4508167859616717, + 0.42469386091374045, + 0.4156967988231998, + 0.4070761564445437, + 0.39549655339091694, + 0.36824324324324326, + 0.35087262426900584, + 0.33454545454545453, + 0.32592384519350814, + 0.31517117771981584, + 0.30071725135286437, + 0.29302189316860466, + 0.2808473310936917, + 0.2544604371921182, + 0.24873389462941703, + 0.23499354460093896, + 0.22579854522454143, + 0.21287128712871287, + 0.21264008620689653, + 0.20441988950276244 + ], + "ci_hi": [ + 0.796, + 0.652, + 0.6398390342052314, + 0.6383838383838384, + 0.6265878246117708, + 0.6147229897761811, + 0.6051432611903591, + 0.600478029001299, + 0.5889423076923077, + 0.5744153629525851, + 0.5530741479305253, + 0.529252007209569, + 0.5220186439729018, + 0.5150625228185468, + 0.5047658079625292, + 0.4781144781144781, + 0.4638211141613025, + 0.4473709809543313, + 0.43866354673288566, + 0.43044033640114243, + 0.41729397312178085, + 0.4108527131782946, + 0.398406374501992, + 0.37333476394849785, + 0.3686445559634461, + 0.3557713963963964, + 0.3484936575052853, + 0.336734693877551, + 0.3383084577114428, + 0.3314924633677636 + ] + } +} \ No newline at end of file diff --git a/results/metrics/r1-llama-8b_gptq_w4_math500_run0_metrics.json b/results/metrics/r1-llama-8b_gptq_w4_math500_run0_metrics.json new file mode 100644 index 0000000000000000000000000000000000000000..4862352f6169175e71012ef7ddbe5ac0a7953b13 --- /dev/null +++ b/results/metrics/r1-llama-8b_gptq_w4_math500_run0_metrics.json @@ -0,0 +1,47 @@ +{ + "model": "r1-llama-8b", + "quantization": "gptq_w4", + "accuracy": 0.588, + "accuracy_delta": 0.0, + "avg_ffs": 0.5, + "median_ffs": 0.0, + "ecr": 0.9653983912160197, + "ssr_curve": [ + 0.758, + 0.6092184368737475, + 0.5967741935483871, + 0.5955284552845529, + 0.58298755186722, + 0.5705128205128205, + 0.5595505617977528, + 0.5542725173210161, + 0.5407407407407407, + 0.5246753246753246, + 0.5013698630136987, + 0.476878612716763, + 0.4690265486725664, + 0.46060606060606063, + 0.45, + 0.42244224422442245, + 0.4068965517241379, + 0.3898916967509025, + 0.38095238095238093, + 0.3712121212121212, + 0.35826771653543305, + 0.35080645161290325, + 0.3389121338912134, + 0.3130434782608696, + 0.3080357142857143, + 0.29439252336448596, + 0.28640776699029125, + 0.27411167512690354, + 0.27419354838709675, + 0.26666666666666666 + ], + "error_type_dist": { + "conceptual": 0.7283163265306123, + "executional": 0.015589569160997732, + "logical": 0.24914965986394558, + "methodological": 0.006944444444444444 + } +} \ No newline at end of file diff --git a/results/metrics/r1-llama-8b_gptq_w4_math500_run0_sig.json b/results/metrics/r1-llama-8b_gptq_w4_math500_run0_sig.json new file mode 100644 index 0000000000000000000000000000000000000000..c76d28a712e7ad4931fabfa86866e65c92e61adb --- /dev/null +++ b/results/metrics/r1-llama-8b_gptq_w4_math500_run0_sig.json @@ -0,0 +1,7 @@ +{ + "n_pairs": 500, + "delta_acc_observed": -0.08599999999999997, + "delta_acc_ci_lo": -0.128, + "delta_acc_ci_hi": -0.04400000000000004, + "p_value": 0.0 +} \ No newline at end of file diff --git a/results/metrics/r1-llama-8b_gptq_w4_restored_gpqa_run0_ci.json b/results/metrics/r1-llama-8b_gptq_w4_restored_gpqa_run0_ci.json new file mode 100644 index 0000000000000000000000000000000000000000..4abce2c7baa520b3b125596f832c634257592831 --- /dev/null +++ b/results/metrics/r1-llama-8b_gptq_w4_restored_gpqa_run0_ci.json @@ -0,0 +1,116 @@ +{ + "n": 198, + "accuracy": { + "mean": 0.21206666666666665, + "ci_lo": 0.15656565656565657, + "ci_hi": 0.2727272727272727 + }, + "avg_ffs": { + "mean": 0.48622771790835456, + "ci_lo": 0.3356631382316314, + "ci_hi": 0.6687116564417178 + }, + "ecr": { + "mean": 0.9907833344563225, + "ci_lo": 0.9858076595858701, + "ci_hi": 0.9950417053240144 + }, + "ssr_curve_ci": { + "mean": [ + 0.4895363636363675, + 0.24746868686868678, + 0.22830165950191222, + 0.22660624597933346, + 0.23136443553270192, + 0.22203830313815248, + 0.22015921770995087, + 0.22504253037822022, + 0.22077288154314537, + 0.21535488151423424, + 0.21654386223624392, + 0.21772953905284634, + 0.21772953905284634, + 0.2170191591924341, + 0.2170191591924341, + 0.20805512283848215, + 0.20582114851252817, + 0.20582114851252817, + 0.20582114851252817, + 0.21338433796079398, + 0.2027362992841377, + 0.20532708883560455, + 0.20277096431961633, + 0.1935564589936935, + 0.19192118439328273, + 0.19734151025586422, + 0.19877717625245822, + 0.21233281256464764, + 0.20330790830028567, + 0.2103177243130986 + ], + "ci_lo": [ + 0.41919191919191917, + 0.18686868686868688, + 0.1683673469387755, + 0.16751269035532995, + 0.17096516370852166, + 0.1630249829273845, + 0.16042780748663102, + 0.16393442622950818, + 0.16021546961325966, + 0.15516050676319298, + 0.15555555555555556, + 0.15728970543577286, + 0.15728970543577286, + 0.15555555555555556, + 0.15555555555555556, + 0.14792899408284024, + 0.14533146848032763, + 0.14533146848032763, + 0.14533146848032763, + 0.15, + 0.13923631486293905, + 0.1411042944785276, + 0.1393939393939394, + 0.13071895424836602, + 0.12751677852348994, + 0.13138686131386862, + 0.13234272875816994, + 0.14283181357649444, + 0.13385826771653545, + 0.139107458465889 + ], + "ci_hi": [ + 0.5555555555555556, + 0.30808080808080807, + 0.2878787878787879, + 0.28717948717948716, + 0.29320923517029174, + 0.28421052631578947, + 0.2826086956521739, + 0.2887700534759358, + 0.28415300546448086, + 0.27717391304347827, + 0.27869805566145617, + 0.2802367576243978, + 0.2802367576243978, + 0.27976785714285707, + 0.27976785714285707, + 0.2711864406779661, + 0.26947301715587213, + 0.26947301715587213, + 0.26947301715587213, + 0.2800063694267515, + 0.2692307692307692, + 0.2721528349936334, + 0.26973684210526316, + 0.2605633802816901, + 0.2589928057553957, + 0.26573426573426573, + 0.2677265205979686, + 0.2857142857142857, + 0.2764227642276423, + 0.288135593220339 + ] + } +} \ No newline at end of file diff --git a/results/metrics/r1-llama-8b_gptq_w4_restored_gpqa_run0_metrics.json b/results/metrics/r1-llama-8b_gptq_w4_restored_gpqa_run0_metrics.json new file mode 100644 index 0000000000000000000000000000000000000000..a9a24f7ef9291099a9c55c59469ca526d830d5f1 --- /dev/null +++ b/results/metrics/r1-llama-8b_gptq_w4_restored_gpqa_run0_metrics.json @@ -0,0 +1,47 @@ +{ + "model": "r1-llama-8b", + "quantization": "gptq_w4_restored", + "accuracy": 0.21212121212121213, + "accuracy_delta": 0.0, + "avg_ffs": 0.48717948717948717, + "median_ffs": 0.0, + "ecr": 0.9907699867083272, + "ssr_curve": [ + 0.4898989898989899, + 0.2474747474747475, + 0.22842639593908629, + 0.2268041237113402, + 0.23157894736842105, + 0.2222222222222222, + 0.22043010752688172, + 0.22527472527472528, + 0.22099447513812154, + 0.2154696132596685, + 0.21666666666666667, + 0.21787709497206703, + 0.21787709497206703, + 0.21714285714285714, + 0.21714285714285714, + 0.20809248554913296, + 0.20588235294117646, + 0.20588235294117646, + 0.20588235294117646, + 0.21341463414634146, + 0.20253164556962025, + 0.20512820512820512, + 0.20261437908496732, + 0.19333333333333333, + 0.1917808219178082, + 0.19718309859154928, + 0.19858156028368795, + 0.21212121212121213, + 0.203125, + 0.21008403361344538 + ], + "error_type_dist": { + "conceptual": 0.7895234493192134, + "executional": 0.0045385779122541605, + "logical": 0.2046142208774584, + "methodological": 0.00132375189107413 + } +} \ No newline at end of file diff --git a/results/metrics/r1-llama-8b_gptq_w4_restored_math500_run0_ci.json b/results/metrics/r1-llama-8b_gptq_w4_restored_math500_run0_ci.json new file mode 100644 index 0000000000000000000000000000000000000000..f9546ed58c939a593cb9a8381ede3cb984daeed3 --- /dev/null +++ b/results/metrics/r1-llama-8b_gptq_w4_restored_math500_run0_ci.json @@ -0,0 +1,116 @@ +{ + "n": 500, + "accuracy": { + "mean": 0.5022399999999999, + "ci_lo": 0.46, + "ci_hi": 0.546 + }, + "avg_ffs": { + "mean": 0.5142997798067076, + "ci_lo": 0.43253968253968256, + "ci_hi": 0.601593625498008 + }, + "ecr": { + "mean": 0.9725986894837013, + "ci_lo": 0.965380068066655, + "ci_hi": 0.9793339250290694 + }, + "ssr_curve_ci": { + "mean": [ + 0.7180836000000053, + 0.5304422821472117, + 0.5144011704477766, + 0.5033364829580478, + 0.4951358550577128, + 0.4886024404662638, + 0.4815397452499762, + 0.47141289213021353, + 0.4615798880790233, + 0.4579980754711315, + 0.4437652647742633, + 0.41838024087808134, + 0.40659486659156585, + 0.39011261317642665, + 0.3724475864607543, + 0.36583876951215755, + 0.36148264483828096, + 0.351238370449042, + 0.335694120982898, + 0.3185667880630699, + 0.31943105703892444, + 0.3164926579991887, + 0.3073110650104345, + 0.3080087790581214, + 0.2963550513827293, + 0.3017950644315696, + 0.3083500212218115, + 0.31750262915253785, + 0.32414050143652035, + 0.31611085379568676 + ], + "ci_lo": [ + 0.68, + 0.48697394789579157, + 0.4717741935483871, + 0.4609053497942387, + 0.45188284518828453, + 0.44468535099460005, + 0.4365256124721604, + 0.42562204251529867, + 0.41445783132530123, + 0.4105793450881612, + 0.39546306601839376, + 0.36885245901639346, + 0.3563209787321908, + 0.338368580060423, + 0.32038762281262717, + 0.3137898907689013, + 0.30792096622889303, + 0.2971222281972606, + 0.28214285714285714, + 0.2629582319545823, + 0.26377418665690044, + 0.260865335814583, + 0.2489054130485411, + 0.24997489959839359, + 0.23766816143497757, + 0.24056603773584906, + 0.24468085106382978, + 0.2513089005235602, + 0.25742284754791267, + 0.24855316167454894 + ], + "ci_hi": [ + 0.756, + 0.5734406438631791, + 0.5575757575757576, + 0.5467479674796748, + 0.5390986889092949, + 0.5326127171615038, + 0.5265486725663717, + 0.5186937826334956, + 0.5110037966394934, + 0.5073179698708752, + 0.49326145552560646, + 0.4684951076320939, + 0.45763306989144914, + 0.4414434171013118, + 0.4237829905351292, + 0.4174757281553398, + 0.41463618565951543, + 0.40468426501035193, + 0.39016593362654933, + 0.374556818181818, + 0.3763848161295301, + 0.3741348671904226, + 0.36666666666666664, + 0.3679662337662337, + 0.35656363041637795, + 0.36448713226831175, + 0.3731361136899236, + 0.38503521586017986, + 0.393446982908964, + 0.3863694939081536 + ] + } +} \ No newline at end of file diff --git a/results/metrics/r1-llama-8b_gptq_w4_restored_math500_run0_metrics.json b/results/metrics/r1-llama-8b_gptq_w4_restored_math500_run0_metrics.json new file mode 100644 index 0000000000000000000000000000000000000000..22f2503ce1c152f21754bce2e8f8d677496ca9d9 --- /dev/null +++ b/results/metrics/r1-llama-8b_gptq_w4_restored_math500_run0_metrics.json @@ -0,0 +1,47 @@ +{ + "model": "r1-llama-8b", + "quantization": "gptq_w4_restored", + "accuracy": 0.502, + "accuracy_delta": 0.0, + "avg_ffs": 0.5140562248995983, + "median_ffs": 0.0, + "ecr": 0.972656194994821, + "ssr_curve": [ + 0.718, + 0.5301204819277109, + 0.5141129032258065, + 0.5030549898167006, + 0.4948453608247423, + 0.488272921108742, + 0.4811529933481153, + 0.4709976798143852, + 0.46116504854368934, + 0.4575, + 0.44329896907216493, + 0.4180327868852459, + 0.40625, + 0.3898809523809524, + 0.3723076923076923, + 0.365625, + 0.36129032258064514, + 0.3509933774834437, + 0.3356643356643357, + 0.31868131868131866, + 0.31954887218045114, + 0.3166023166023166, + 0.3073770491803279, + 0.3080168776371308, + 0.29646017699115046, + 0.3018867924528302, + 0.30845771144278605, + 0.3177083333333333, + 0.32432432432432434, + 0.3163841807909605 + ], + "error_type_dist": { + "conceptual": 0.7429426860564585, + "executional": 0.021243227830054178, + "logical": 0.2256914741944682, + "methodological": 0.010122611919019104 + } +} \ No newline at end of file diff --git a/results/metrics/r1-qwen-1.5b_awq_w4_gpqa_run0_ci.json b/results/metrics/r1-qwen-1.5b_awq_w4_gpqa_run0_ci.json new file mode 100644 index 0000000000000000000000000000000000000000..9c12e76b8ad3c57863d8419fcbed2ef589ac9029 --- /dev/null +++ b/results/metrics/r1-qwen-1.5b_awq_w4_gpqa_run0_ci.json @@ -0,0 +1,116 @@ +{ + "n": 198, + "accuracy": { + "mean": 0.2173949494949495, + "ci_lo": 0.16161616161616163, + "ci_hi": 0.2777777777777778 + }, + "avg_ffs": { + "mean": 0.45222331570317353, + "ci_lo": 0.36075949367088606, + "ci_hi": 0.5454545454545454 + }, + "ecr": { + "mean": 0.9966490493395519, + "ci_lo": 0.994459238967503, + "ci_hi": 0.9984790677323808 + }, + "ssr_curve_ci": { + "mean": [ + 0.5457747474747364, + 0.23779494949494961, + 0.22831635970793424, + 0.22899426967646172, + 0.2323265788738771, + 0.2320088947525965, + 0.23157247698594718, + 0.22968091957597012, + 0.22509087022995636, + 0.2218776141219355, + 0.2186528404182737, + 0.22550816093716586, + 0.22207316926687176, + 0.2263921538778211, + 0.2243162458220298, + 0.22579908671282584, + 0.22879198471846707, + 0.2197858239585601, + 0.21589447176887647, + 0.21589447176887647, + 0.21327436369903566, + 0.21632227251910727, + 0.21789079606855022, + 0.215236704184383, + 0.21093960516662327, + 0.20813412838679715, + 0.2052174031364751, + 0.21022453371862057, + 0.21259431064243708, + 0.21293425148424713 + ], + "ci_lo": [ + 0.47474747474747475, + 0.17676767676767677, + 0.16923076923076924, + 0.1693121693121693, + 0.1712707182320442, + 0.1694915254237288, + 0.16874811746987953, + 0.16666666666666666, + 0.16201117318435754, + 0.15975696083058047, + 0.15606373395390738, + 0.1610738255033557, + 0.15757575757575756, + 0.16129032258064516, + 0.15893240465859787, + 0.16, + 0.16216216216216217, + 0.15384615384615385, + 0.1503186274509804, + 0.1503186274509804, + 0.145985401459854, + 0.14814814814814814, + 0.14959748501586556, + 0.14705882352941177, + 0.14285714285714285, + 0.1397016073871409, + 0.1349206349206349, + 0.13821138211382114, + 0.13792312559316672, + 0.13636363636363635 + ], + "ci_hi": [ + 0.6161616161616161, + 0.29797979797979796, + 0.2894736842105263, + 0.291005291005291, + 0.29508196721311475, + 0.2947976878612717, + 0.2958579881656805, + 0.2958579881656805, + 0.2911464737793851, + 0.28823529411764703, + 0.2840433300876336, + 0.2929936305732484, + 0.28948992852501604, + 0.29448835928897266, + 0.29333333333333333, + 0.2948825503355703, + 0.2980132450331126, + 0.2887323943661972, + 0.2857142857142857, + 0.2857142857142857, + 0.28368794326241137, + 0.28888888888888886, + 0.29078675770085727, + 0.28671457219251334, + 0.28170895602623347, + 0.2800075757575757, + 0.2781954887218045, + 0.2845528455284553, + 0.2894736842105263, + 0.29060606060606053 + ] + } +} \ No newline at end of file diff --git a/results/metrics/r1-qwen-1.5b_awq_w4_gpqa_run0_metrics.json b/results/metrics/r1-qwen-1.5b_awq_w4_gpqa_run0_metrics.json new file mode 100644 index 0000000000000000000000000000000000000000..3241b87b05100bdb8ad1b4808f427a921983452a --- /dev/null +++ b/results/metrics/r1-qwen-1.5b_awq_w4_gpqa_run0_metrics.json @@ -0,0 +1,47 @@ +{ + "model": "r1-qwen-1.5b", + "quantization": "awq_w4", + "accuracy": 0.21717171717171718, + "accuracy_delta": 0.0, + "avg_ffs": 0.45161290322580644, + "median_ffs": 0.0, + "ecr": 0.9966560285264121, + "ssr_curve": [ + 0.5454545454545454, + 0.23737373737373738, + 0.22797927461139897, + 0.22872340425531915, + 0.23204419889502761, + 0.23163841807909605, + 0.23121387283236994, + 0.22941176470588234, + 0.22485207100591717, + 0.2215568862275449, + 0.21818181818181817, + 0.225, + 0.22151898734177214, + 0.22580645161290322, + 0.2236842105263158, + 0.2251655629139073, + 0.22818791946308725, + 0.2191780821917808, + 0.2152777777777778, + 0.2152777777777778, + 0.2127659574468085, + 0.2158273381294964, + 0.21739130434782608, + 0.21481481481481482, + 0.21052631578947367, + 0.2076923076923077, + 0.2047244094488189, + 0.20967741935483872, + 0.211864406779661, + 0.21238938053097345 + ], + "error_type_dist": { + "conceptual": 0.7239985299522235, + "executional": 0.0011025358324145535, + "logical": 0.27342888643880925, + "methodological": 0.001470047776552738 + } +} \ No newline at end of file diff --git a/results/metrics/r1-qwen-1.5b_awq_w4_gpqa_run0_sig.json b/results/metrics/r1-qwen-1.5b_awq_w4_gpqa_run0_sig.json new file mode 100644 index 0000000000000000000000000000000000000000..94dc93ca1d8527fb6c42b0c354b6c5bdd43ed9de --- /dev/null +++ b/results/metrics/r1-qwen-1.5b_awq_w4_gpqa_run0_sig.json @@ -0,0 +1,7 @@ +{ + "n_pairs": 198, + "delta_acc_observed": -0.06565656565656566, + "delta_acc_ci_lo": -0.13131313131313133, + "delta_acc_ci_hi": 0.0, + "p_value": 0.0564 +} \ No newline at end of file diff --git a/results/metrics/r1-qwen-1.5b_awq_w4_gsm8k_run0_ci.json b/results/metrics/r1-qwen-1.5b_awq_w4_gsm8k_run0_ci.json new file mode 100644 index 0000000000000000000000000000000000000000..09f1e158842e2ab38f15d51ec0ab6460b4957f40 --- /dev/null +++ b/results/metrics/r1-qwen-1.5b_awq_w4_gsm8k_run0_ci.json @@ -0,0 +1,116 @@ +{ + "n": 1319, + "accuracy": { + "mean": 0.6305951478392723, + "ci_lo": 0.6049848369977255, + "ci_hi": 0.6565769522365426 + }, + "avg_ffs": { + "mean": 0.5884457530362415, + "ci_lo": 0.48803923649663644, + "ci_hi": 0.6953165187757201 + }, + "ecr": { + "mean": 0.857286042901041, + "ci_lo": 0.8347355926604786, + "ci_hi": 0.8780093377454411 + }, + "ssr_curve_ci": { + "mean": [ + 0.7626949203942394, + 0.6673850956014561, + 0.6572803970622774, + 0.6602573829800197, + 0.6694846276096863, + 0.6525889807548994, + 0.6407083398369283, + 0.5940829979807558, + 0.49797191857470596, + 0.47342817748292354, + 0.42174747294123943, + 0.4027312508087868, + 0.36118328725162757, + 0.3083475385561924, + 0.3083475385561924, + 0.31641828266661204, + 0.3247729592914834, + 0.30030137137629453, + 0.283857250566295, + 0.25022256153190303, + 0.23881620622294913, + 0.23022284646168037, + 0.20772655174966873, + 0.2011195881133659, + 0.20476352908430573, + 0.20476352908430573, + 0.20867357870282194, + 0.21681675484256613, + 0.20099189275793608, + 0.18437133492197155 + ], + "ci_lo": [ + 0.7391963608794542, + 0.6414373088685015, + 0.6307392804180271, + 0.6321921407827866, + 0.6382508055093298, + 0.6182994354554161, + 0.6, + 0.5442349991853421, + 0.42986326109391126, + 0.3916083916083916, + 0.3333333333333333, + 0.30487514518002323, + 0.26136363636363635, + 0.20588235294117646, + 0.20588235294117646, + 0.21212121212121213, + 0.2191780821917808, + 0.1935483870967742, + 0.17857142857142858, + 0.14752892960462874, + 0.13793103448275862, + 0.13041465378421901, + 0.10909090909090909, + 0.1016949152542373, + 0.10344827586206896, + 0.10344827586206896, + 0.10526315789473684, + 0.10908102766798419, + 0.09433962264150944, + 0.08163265306122448 + ], + "ci_hi": [ + 0.7854435178165277, + 0.6932530663057676, + 0.6835353945520385, + 0.688373268638851, + 0.6996790681465976, + 0.6864064602960969, + 0.6813823996210716, + 0.6432967168285889, + 0.5669642857142857, + 0.5542168674698795, + 0.5126160029229081, + 0.5048543689320388, + 0.4666666666666667, + 0.4166666666666667, + 0.4166666666666667, + 0.4264705882352941, + 0.4361750483558994, + 0.41464565117349267, + 0.4, + 0.36363636363636365, + 0.3508771929824561, + 0.34615384615384615, + 0.32142857142857145, + 0.31666666666666665, + 0.32558139534883723, + 0.32558139534883723, + 0.3333333333333333, + 0.34146341463414637, + 0.3225806451612903, + 0.3023255813953488 + ] + } +} \ No newline at end of file diff --git a/results/metrics/r1-qwen-1.5b_awq_w4_gsm8k_run0_metrics.json b/results/metrics/r1-qwen-1.5b_awq_w4_gsm8k_run0_metrics.json new file mode 100644 index 0000000000000000000000000000000000000000..853c00c549fea1097e7edb8b49570e147207756c --- /dev/null +++ b/results/metrics/r1-qwen-1.5b_awq_w4_gsm8k_run0_metrics.json @@ -0,0 +1,46 @@ +{ + "model": "r1-qwen-1.5b", + "quantization": "awq_w4", + "accuracy": 0.6307808946171342, + "accuracy_delta": 0.0, + "avg_ffs": 0.5872689938398358, + "median_ffs": 0.0, + "ecr": 0.8577052220864806, + "ssr_curve": [ + 0.7626990144048522, + 0.6674347158218126, + 0.6573033707865169, + 0.6602914389799636, + 0.6695371367061357, + 0.6526867627785059, + 0.6404494382022472, + 0.5939226519337016, + 0.49760765550239233, + 0.4726027397260274, + 0.42105263157894735, + 0.4020618556701031, + 0.36046511627906974, + 0.3076923076923077, + 0.3076923076923077, + 0.3157894736842105, + 0.32432432432432434, + 0.3, + 0.2835820895522388, + 0.25, + 0.23809523809523808, + 0.22950819672131148, + 0.20689655172413793, + 0.2, + 0.2037037037037037, + 0.2037037037037037, + 0.20754716981132076, + 0.21568627450980393, + 0.2, + 0.1836734693877551 + ], + "error_type_dist": { + "conceptual": 0.7262609519299077, + "executional": 0.04546530902202226, + "logical": 0.22827373904807008 + } +} \ No newline at end of file diff --git a/results/metrics/r1-qwen-1.5b_awq_w4_gsm8k_run0_sig.json b/results/metrics/r1-qwen-1.5b_awq_w4_gsm8k_run0_sig.json new file mode 100644 index 0000000000000000000000000000000000000000..8e098c89d0c5a4ada8d5d0158b048b59858b166d --- /dev/null +++ b/results/metrics/r1-qwen-1.5b_awq_w4_gsm8k_run0_sig.json @@ -0,0 +1,7 @@ +{ + "n_pairs": 1319, + "delta_acc_observed": 0.00833965125094771, + "delta_acc_ci_lo": -0.016698256254738475, + "delta_acc_ci_hi": 0.03411675511751333, + "p_value": 0.5252 +} \ No newline at end of file diff --git a/results/metrics/r1-qwen-1.5b_awq_w4_math500_run0_metrics.json b/results/metrics/r1-qwen-1.5b_awq_w4_math500_run0_metrics.json new file mode 100644 index 0000000000000000000000000000000000000000..d62e573b6237c2e3096875d54e3f47e6d8d45a42 --- /dev/null +++ b/results/metrics/r1-qwen-1.5b_awq_w4_math500_run0_metrics.json @@ -0,0 +1,47 @@ +{ + "model": "r1-qwen-1.5b", + "quantization": "awq_w4", + "accuracy": 0.586, + "accuracy_delta": 0.0, + "avg_ffs": 0.6763285024154589, + "median_ffs": 1.0, + "ecr": 0.9628518628372117, + "ssr_curve": [ + 0.814, + 0.6192384769539078, + 0.5983935742971888, + 0.5875251509054326, + 0.5811088295687885, + 0.5777310924369747, + 0.5594713656387665, + 0.5448275862068965, + 0.5322195704057279, + 0.5113350125944585, + 0.5078125, + 0.48633879781420764, + 0.4828571428571429, + 0.4658753709198813, + 0.4472049689440994, + 0.42857142857142855, + 0.4135593220338983, + 0.40070921985815605, + 0.38783269961977185, + 0.3709677419354839, + 0.3516949152542373, + 0.32286995515695066, + 0.3113207547169811, + 0.3088235294117647, + 0.29949238578680204, + 0.2727272727272727, + 0.27624309392265195, + 0.27011494252873564, + 0.2603550295857988, + 0.2641509433962264 + ], + "error_type_dist": { + "conceptual": 0.6755516066933215, + "executional": 0.02310084406930253, + "logical": 0.29172219754183326, + "methodological": 0.009625351695542721 + } +} \ No newline at end of file diff --git a/results/metrics/r1-qwen-1.5b_awq_w4_math500_run0_sig.json b/results/metrics/r1-qwen-1.5b_awq_w4_math500_run0_sig.json new file mode 100644 index 0000000000000000000000000000000000000000..bd8bb347e0633ef884bf8e121517caf10851fbee --- /dev/null +++ b/results/metrics/r1-qwen-1.5b_awq_w4_math500_run0_sig.json @@ -0,0 +1,7 @@ +{ + "n_pairs": 500, + "delta_acc_observed": -0.014000000000000012, + "delta_acc_ci_lo": -0.052000000000000046, + "delta_acc_ci_hi": 0.026000000000000023, + "p_value": 0.534 +} \ No newline at end of file diff --git a/results/metrics/r1-qwen-1.5b_awq_w4_restored_gpqa_run0_ci.json b/results/metrics/r1-qwen-1.5b_awq_w4_restored_gpqa_run0_ci.json new file mode 100644 index 0000000000000000000000000000000000000000..414fd20c9828a87706e9a82001a949b590dfb65e --- /dev/null +++ b/results/metrics/r1-qwen-1.5b_awq_w4_restored_gpqa_run0_ci.json @@ -0,0 +1,116 @@ +{ + "n": 198, + "accuracy": { + "mean": 0.15164949494949498, + "ci_lo": 0.10606060606060606, + "ci_hi": 0.20202020202020202 + }, + "avg_ffs": { + "mean": 0.32083529714810716, + "ci_lo": 0.2485207100591716, + "ci_hi": 0.3974378881987577 + }, + "ecr": { + "mean": 0.9909123296728712, + "ci_lo": 0.984739513446361, + "ci_hi": 0.9959489740172274 + }, + "ssr_curve_ci": { + "mean": [ + 0.4138606060606101, + 0.1574173016231522, + 0.1296299681707118, + 0.13213073723380667, + 0.13366956202084182, + 0.13321599082446195, + 0.14089387935108666, + 0.1427168697599522, + 0.14365364663815305, + 0.14165694585011046, + 0.13855575204945558, + 0.1405204154782022, + 0.14149289311468025, + 0.13610542550211932, + 0.13710130305090185, + 0.13808328386997357, + 0.14118612879763542, + 0.1368761398946769, + 0.13131701112535552, + 0.134477118632158, + 0.13669153090139566, + 0.13781036165659113, + 0.1389252197507564, + 0.1389252197507564, + 0.1340390027134538, + 0.13378765708113674, + 0.13748933664693191, + 0.14137644053677417, + 0.1371601911328755, + 0.12988997777603042 + ], + "ci_lo": [ + 0.3484848484848485, + 0.10767857142857144, + 0.08333333333333333, + 0.08383233532934131, + 0.08484848484848485, + 0.08333333333333333, + 0.08804500254971953, + 0.08916870255649594, + 0.090308105872622, + 0.08759124087591241, + 0.0851063829787234, + 0.08609271523178808, + 0.08666666666666667, + 0.08208955223880597, + 0.08270676691729323, + 0.08333333333333333, + 0.08509410801963993, + 0.08088235294117647, + 0.07575757575757576, + 0.078125, + 0.07936507936507936, + 0.08, + 0.08064516129032258, + 0.08064516129032258, + 0.07563025210084033, + 0.07406648451730419, + 0.07618788819875777, + 0.07766990291262135, + 0.07368421052631578, + 0.06796116504854369 + ], + "ci_hi": [ + 0.48484848484848486, + 0.20812182741116753, + 0.1793478260869565, + 0.18285714285714286, + 0.18497109826589594, + 0.18562874251497005, + 0.1962025316455696, + 0.1986754966887417, + 0.2, + 0.1986754966887417, + 0.19594594594594594, + 0.19858156028368795, + 0.2, + 0.19531690554511272, + 0.19565597446032224, + 0.1984142130134496, + 0.20155038759689922, + 0.19834710743801653, + 0.19230769230769232, + 0.19658469945355186, + 0.2, + 0.20149253731343283, + 0.20178045638571915, + 0.20178045638571915, + 0.19827586206896552, + 0.2, + 0.20535714285714285, + 0.21, + 0.20952380952380953, + 0.2 + ] + } +} \ No newline at end of file diff --git a/results/metrics/r1-qwen-1.5b_awq_w4_restored_gsm8k_run0_metrics.json b/results/metrics/r1-qwen-1.5b_awq_w4_restored_gsm8k_run0_metrics.json new file mode 100644 index 0000000000000000000000000000000000000000..b9142a4d2f4621668a53f2c2d986f0ac4551e0e7 --- /dev/null +++ b/results/metrics/r1-qwen-1.5b_awq_w4_restored_gsm8k_run0_metrics.json @@ -0,0 +1,46 @@ +{ + "model": "r1-qwen-1.5b", + "quantization": "awq_w4_restored", + "accuracy": 0.6391205458680819, + "accuracy_delta": 0.0, + "avg_ffs": 0.5189075630252101, + "median_ffs": 0.0, + "ecr": 0.8266529985463869, + "ssr_curve": [ + 0.7482941622441244, + 0.6750380517503806, + 0.6523076923076923, + 0.6482649842271293, + 0.6371527777777778, + 0.6322645290581163, + 0.5915875169606513, + 0.5621181262729125, + 0.5353535353535354, + 0.4632768361581921, + 0.39473684210526316, + 0.32941176470588235, + 0.33783783783783783, + 0.3382352941176471, + 0.31746031746031744, + 0.3, + 0.30357142857142855, + 0.30357142857142855, + 0.30357142857142855, + 0.30357142857142855, + 0.30357142857142855, + 0.30357142857142855, + 0.2962962962962963, + 0.2830188679245283, + 0.2830188679245283, + 0.27450980392156865, + 0.2857142857142857, + 0.2708333333333333, + 0.2708333333333333, + 0.26666666666666666 + ], + "error_type_dist": { + "conceptual": 0.7673627974745022, + "executional": 0.03351141330743079, + "logical": 0.19912578921806703 + } +} \ No newline at end of file diff --git a/results/metrics/r1-qwen-1.5b_awq_w4_restored_math500_run0_ci.json b/results/metrics/r1-qwen-1.5b_awq_w4_restored_math500_run0_ci.json new file mode 100644 index 0000000000000000000000000000000000000000..c2de1bdb54841e3c7e970857c8bd72f62bfcfe08 --- /dev/null +++ b/results/metrics/r1-qwen-1.5b_awq_w4_restored_math500_run0_ci.json @@ -0,0 +1,116 @@ +{ + "n": 500, + "accuracy": { + "mean": 0.5719059999999999, + "ci_lo": 0.53, + "ci_hi": 0.614 + }, + "avg_ffs": { + "mean": 0.49956350456948695, + "ci_lo": 0.41070586253369273, + "ci_hi": 0.593607305936073 + }, + "ecr": { + "mean": 0.9582862745943225, + "ci_lo": 0.9443319473501485, + "ci_hi": 0.9701962945884653 + }, + "ssr_curve_ci": { + "mean": [ + 0.7498268000000007, + 0.5975269082934563, + 0.5796718531151988, + 0.5729949457593873, + 0.565155912176914, + 0.5481938499415124, + 0.540320180141196, + 0.532710977310042, + 0.5203023714107521, + 0.5078037653969383, + 0.49416716473504696, + 0.4852255206919914, + 0.4711564404127727, + 0.4503359662390782, + 0.4403143318118929, + 0.42506842620422697, + 0.4081777058473464, + 0.38784471110573554, + 0.37626722221547304, + 0.35484488848965545, + 0.3401424715486591, + 0.3176825392125704, + 0.292310952544336, + 0.2735154174055176, + 0.2624979728928327, + 0.25355913703360117, + 0.23564563385957568, + 0.2350298503872149, + 0.2209050187571102, + 0.22319687074968833 + ], + "ci_lo": [ + 0.712, + 0.5544354838709677, + 0.5373699774404, + 0.5297680633330126, + 0.5219206680584552, + 0.5021929824561403, + 0.49321189839572194, + 0.48529230657780353, + 0.472077657205014, + 0.45801474999482816, + 0.44382896487254164, + 0.4329587893165991, + 0.41907445637214424, + 0.3972956123737374, + 0.3867572590011614, + 0.3698630136986301, + 0.3539518900343643, + 0.33203125, + 0.31983390854184646, + 0.29706622646443515, + 0.2809917355371901, + 0.2570030310684516, + 0.23039215686274508, + 0.21104997880971121, + 0.19892398859383353, + 0.19017805120471606, + 0.17307692307692307, + 0.16993155752867187, + 0.15624445921985816, + 0.15671641791044777 + ], + "ci_hi": [ + 0.788, + 0.6405622489959839, + 0.6222222222222222, + 0.6163265306122448, + 0.6082495704467353, + 0.5929222904281481, + 0.5868594378772636, + 0.5804898277655761, + 0.5695538057742782, + 0.5587467362924282, + 0.5459830483917705, + 0.5391608569107892, + 0.5261538461538462, + 0.5064102564102564, + 0.49673255665902727, + 0.48197015005072885, + 0.4659507351327628, + 0.4485624090503685, + 0.43795620437956206, + 0.4163269276582228, + 0.40336134453781514, + 0.3807344247821581, + 0.35609756097560974, + 0.3368421052631579, + 0.32663565351510027, + 0.31958762886597936, + 0.30357142857142855, + 0.3055555555555556, + 0.2922102897102897, + 0.29457364341085274 + ] + } +} \ No newline at end of file diff --git a/results/metrics/r1-qwen-1.5b_awq_w4_restored_math500_run0_metrics.json b/results/metrics/r1-qwen-1.5b_awq_w4_restored_math500_run0_metrics.json new file mode 100644 index 0000000000000000000000000000000000000000..132f7a916208958883c223a2d1ae78ca6014fa1f --- /dev/null +++ b/results/metrics/r1-qwen-1.5b_awq_w4_restored_math500_run0_metrics.json @@ -0,0 +1,47 @@ +{ + "model": "r1-qwen-1.5b", + "quantization": "awq_w4_restored", + "accuracy": 0.572, + "accuracy_delta": 0.0, + "avg_ffs": 0.5, + "median_ffs": 0.0, + "ecr": 0.9582267538104763, + "ssr_curve": [ + 0.75, + 0.5975855130784709, + 0.5797979797979798, + 0.573170731707317, + 0.5652173913043478, + 0.5482456140350878, + 0.5404157043879908, + 0.5328467153284672, + 0.5205128205128206, + 0.5080213903743316, + 0.49441340782122906, + 0.4853801169590643, + 0.47112462006079026, + 0.4504792332268371, + 0.44039735099337746, + 0.42517006802721086, + 0.4084507042253521, + 0.3880597014925373, + 0.3764705882352941, + 0.3551020408163265, + 0.3404255319148936, + 0.31797235023041476, + 0.2926829268292683, + 0.27411167512690354, + 0.2631578947368421, + 0.2541436464088398, + 0.23636363636363636, + 0.2356687898089172, + 0.2214765100671141, + 0.22377622377622378 + ], + "error_type_dist": { + "conceptual": 0.6965190390586697, + "executional": 0.027945742768426214, + "logical": 0.2660565451871221, + "methodological": 0.009478672985781991 + } +} \ No newline at end of file diff --git a/results/metrics/r1-qwen-1.5b_bnb_nf4_w4_gpqa_run0_ci.json b/results/metrics/r1-qwen-1.5b_bnb_nf4_w4_gpqa_run0_ci.json new file mode 100644 index 0000000000000000000000000000000000000000..89360d1d18c096871a5713635fdae3511309d77b --- /dev/null +++ b/results/metrics/r1-qwen-1.5b_bnb_nf4_w4_gpqa_run0_ci.json @@ -0,0 +1,116 @@ +{ + "n": 198, + "accuracy": { + "mean": 0.1560161616161616, + "ci_lo": 0.10606060606060606, + "ci_hi": 0.20707070707070707 + }, + "avg_ffs": { + "mean": 0.5266689962536266, + "ci_lo": 0.4294478527607362, + "ci_hi": 0.6385542168674698 + }, + "ecr": { + "mean": 0.9898341925058673, + "ci_lo": 0.9830532198322214, + "ci_hi": 0.9949870415061589 + }, + "ssr_curve_ci": { + "mean": [ + 0.5596303030302925, + 0.17634343434343333, + 0.1644924755433365, + 0.16452649406021486, + 0.17001630296520176, + 0.16264467065537355, + 0.1596393759539758, + 0.16054157795039167, + 0.163374391492233, + 0.15450662437420382, + 0.156376555664571, + 0.15828728134445333, + 0.1541401274085991, + 0.1541401274085991, + 0.15508204414052146, + 0.15703655321495058, + 0.15703655321495058, + 0.159017091764921, + 0.159017091764921, + 0.159017091764921, + 0.16003761367557875, + 0.16211364297037487, + 0.16533891186080502, + 0.16192814487076063, + 0.1630381760663023, + 0.16758959443560423, + 0.1640243299412807, + 0.16278462720891423, + 0.14583894381250723, + 0.14813709144883694 + ], + "ci_lo": [ + 0.4898989898989899, + 0.12626262626262627, + 0.11282051282051282, + 0.11290115798180314, + 0.11666666666666667, + 0.11049643453681099, + 0.10795454545454546, + 0.10857142857142857, + 0.11046511627906977, + 0.10240963855421686, + 0.10365853658536585, + 0.10456814627691569, + 0.10119047619047619, + 0.10119047619047619, + 0.10126582278481013, + 0.10256410256410256, + 0.10256410256410256, + 0.1038961038961039, + 0.1038961038961039, + 0.1038961038961039, + 0.10457516339869281, + 0.10625, + 0.10810810810810811, + 0.10457516339869281, + 0.10562453669384729, + 0.1094729132528924, + 0.10605809753160747, + 0.10236220472440945, + 0.08823529411764706, + 0.09016393442622951 + ], + "ci_hi": [ + 0.6262626262626263, + 0.23232323232323232, + 0.21649484536082475, + 0.21808510638297873, + 0.2247191011235955, + 0.2178899377127077, + 0.21429580306698934, + 0.2159090909090909, + 0.21965317919075145, + 0.21084337349397592, + 0.2138364779874214, + 0.21686746987951808, + 0.21052631578947367, + 0.21052631578947367, + 0.21192554685932163, + 0.21429666958808047, + 0.21429666958808047, + 0.21739130434782608, + 0.21739130434782608, + 0.21739130434782608, + 0.21854822019867542, + 0.22151898734177214, + 0.22560975609756098, + 0.22153656821378315, + 0.2236842105263158, + 0.23076923076923078, + 0.22627737226277372, + 0.22627737226277372, + 0.2071494708994708, + 0.2101544622425628 + ] + } +} \ No newline at end of file diff --git a/results/metrics/r1-qwen-1.5b_bnb_nf4_w4_gpqa_run0_metrics.json b/results/metrics/r1-qwen-1.5b_bnb_nf4_w4_gpqa_run0_metrics.json new file mode 100644 index 0000000000000000000000000000000000000000..d7423d859e13175765b4ca27c927f6c17a48313e --- /dev/null +++ b/results/metrics/r1-qwen-1.5b_bnb_nf4_w4_gpqa_run0_metrics.json @@ -0,0 +1,47 @@ +{ + "model": "r1-qwen-1.5b", + "quantization": "bnb_nf4_w4", + "accuracy": 0.15656565656565657, + "accuracy_delta": 0.0, + "avg_ffs": 0.5269461077844312, + "median_ffs": 0.0, + "ecr": 0.9898016581394141, + "ssr_curve": [ + 0.5606060606060606, + 0.17676767676767677, + 0.16494845360824742, + 0.16489361702127658, + 0.17032967032967034, + 0.16292134831460675, + 0.16, + 0.16091954022988506, + 0.16374269005847952, + 0.15476190476190477, + 0.1566265060240964, + 0.15853658536585366, + 0.15432098765432098, + 0.15432098765432098, + 0.15527950310559005, + 0.15723270440251572, + 0.15723270440251572, + 0.1592356687898089, + 0.1592356687898089, + 0.1592356687898089, + 0.16025641025641027, + 0.16233766233766234, + 0.16556291390728478, + 0.16216216216216217, + 0.16326530612244897, + 0.16783216783216784, + 0.16428571428571428, + 0.16296296296296298, + 0.14615384615384616, + 0.1484375 + ], + "error_type_dist": { + "conceptual": 0.7836821451929366, + "executional": 0.0009810333551340746, + "logical": 0.2146827992151733, + "methodological": 0.0006540222367560497 + } +} \ No newline at end of file diff --git a/results/metrics/r1-qwen-1.5b_bnb_nf4_w4_gpqa_run0_sig.json b/results/metrics/r1-qwen-1.5b_bnb_nf4_w4_gpqa_run0_sig.json new file mode 100644 index 0000000000000000000000000000000000000000..1100fb9378e0fa047d15466d6c43086a14746c34 --- /dev/null +++ b/results/metrics/r1-qwen-1.5b_bnb_nf4_w4_gpqa_run0_sig.json @@ -0,0 +1,7 @@ +{ + "n_pairs": 198, + "delta_acc_observed": 0.06565656565656564, + "delta_acc_ci_lo": -0.005050505050505055, + "delta_acc_ci_hi": 0.14141414141414144, + "p_value": 0.0908 +} \ No newline at end of file diff --git a/results/metrics/r1-qwen-1.5b_bnb_nf4_w4_gsm8k_run0_sig.json b/results/metrics/r1-qwen-1.5b_bnb_nf4_w4_gsm8k_run0_sig.json new file mode 100644 index 0000000000000000000000000000000000000000..239e958dbf21f0dc019e1b651075c74ac60ccd2c --- /dev/null +++ b/results/metrics/r1-qwen-1.5b_bnb_nf4_w4_gsm8k_run0_sig.json @@ -0,0 +1,7 @@ +{ + "n_pairs": 1319, + "delta_acc_observed": 0.012888551933282866, + "delta_acc_ci_lo": -0.0015163002274449777, + "delta_acc_ci_hi": 0.027293404094010598, + "p_value": 0.0852 +} \ No newline at end of file diff --git a/results/metrics/r1-qwen-1.5b_bnb_nf4_w4_math500_run0_metrics.json b/results/metrics/r1-qwen-1.5b_bnb_nf4_w4_math500_run0_metrics.json new file mode 100644 index 0000000000000000000000000000000000000000..cbd218e89d88ed673f81b96b008b15ef2d792780 --- /dev/null +++ b/results/metrics/r1-qwen-1.5b_bnb_nf4_w4_math500_run0_metrics.json @@ -0,0 +1,47 @@ +{ + "model": "r1-qwen-1.5b", + "quantization": "bnb_nf4_w4", + "accuracy": 0.574, + "accuracy_delta": 0.0, + "avg_ffs": 0.6901408450704225, + "median_ffs": 1.0, + "ecr": 0.9619094105674102, + "ssr_curve": [ + 0.806, + 0.61, + 0.5895372233400402, + 0.5801217038539553, + 0.5681818181818182, + 0.5536480686695279, + 0.5436241610738255, + 0.5225653206650831, + 0.5098522167487685, + 0.5062972292191436, + 0.4973821989528796, + 0.488, + 0.4720670391061452, + 0.45345345345345345, + 0.4423676012461059, + 0.42394822006472493, + 0.40604026845637586, + 0.37857142857142856, + 0.35714285714285715, + 0.33725490196078434, + 0.33064516129032256, + 0.31512605042016806, + 0.30131004366812225, + 0.2831050228310502, + 0.2857142857142857, + 0.2814070351758794, + 0.25925925925925924, + 0.2541436464088398, + 0.24561403508771928, + 0.23809523809523808 + ], + "error_type_dist": { + "conceptual": 0.7072143997677457, + "executional": 0.022209319204528958, + "logical": 0.26418928726956015, + "methodological": 0.006386993758165191 + } +} \ No newline at end of file diff --git a/results/metrics/r1-qwen-1.5b_bnb_nf4_w4_restored_gpqa_run0_ci.json b/results/metrics/r1-qwen-1.5b_bnb_nf4_w4_restored_gpqa_run0_ci.json new file mode 100644 index 0000000000000000000000000000000000000000..c47fbbe715d3d81a4d4bf1d0a0e05a361f44ecf2 --- /dev/null +++ b/results/metrics/r1-qwen-1.5b_bnb_nf4_w4_restored_gpqa_run0_ci.json @@ -0,0 +1,116 @@ +{ + "n": 198, + "accuracy": { + "mean": 0.22208282828282833, + "ci_lo": 0.16666666666666666, + "ci_hi": 0.2828282828282828 + }, + "avg_ffs": { + "mean": 0.35686202867066463, + "ci_lo": 0.2733333333333333, + "ci_hi": 0.44666666666666666 + }, + "ecr": { + "mean": 0.9867687396024336, + "ci_lo": 0.9754536652290905, + "ci_hi": 0.9954085997043257 + }, + "ssr_curve_ci": { + "mean": [ + 0.4799828282828351, + 0.23195252525252477, + 0.21837541857435075, + 0.21312063098814146, + 0.20691768703437816, + 0.2092827000705692, + 0.19267412259150798, + 0.19004003779949752, + 0.19236171184820136, + 0.1909108406322232, + 0.18942677044846837, + 0.18535038601604953, + 0.18099483718289133, + 0.18062424676845115, + 0.18186639614205605, + 0.18186639614205605, + 0.18318271764662922, + 0.18578927584818478, + 0.18717507373118344, + 0.18995834942009587, + 0.19136351982940508, + 0.19563953475551402, + 0.19563953475551402, + 0.1832533090747385, + 0.1769413424315768, + 0.1783465939581599, + 0.18115464070603282, + 0.17095998101796236, + 0.18129991900738934, + 0.1729976730844036 + ], + "ci_lo": [ + 0.4090909090909091, + 0.17676767676767677, + 0.16313986107400483, + 0.15591397849462366, + 0.14857142857142858, + 0.15028901734104047, + 0.13579450418160094, + 0.13291139240506328, + 0.13414634146341464, + 0.1320754716981132, + 0.1310344827586207, + 0.1267582159624413, + 0.12326999558108705, + 0.12080536912751678, + 0.12179054054054055, + 0.12179054054054055, + 0.12244530905887535, + 0.12413667253964258, + 0.125, + 0.1276595744680851, + 0.12878787878787878, + 0.13178294573643412, + 0.13178294573643412, + 0.12096774193548387, + 0.11450381679389313, + 0.11538461538461539, + 0.11811023622047244, + 0.10768595041322315, + 0.11403508771929824, + 0.10576923076923077 + ], + "ci_hi": [ + 0.5505050505050505, + 0.29292929292929293, + 0.2804232804232804, + 0.2752808988764045, + 0.27058823529411763, + 0.27380952380952384, + 0.25610137195121946, + 0.25316455696202533, + 0.2564102564102564, + 0.25625, + 0.2553191489361702, + 0.25161397137363806, + 0.24647887323943662, + 0.24817647667857326, + 0.2483221476510067, + 0.2483221476510067, + 0.2500413907284762, + 0.2553191489361702, + 0.25694940476190464, + 0.2608695652173913, + 0.2619088669950738, + 0.2676056338028169, + 0.2676056338028169, + 0.25396825396825395, + 0.2462713605883625, + 0.24812030075187969, + 0.25203252032520324, + 0.24193548387096775, + 0.2564102564102564, + 0.2477084517333766 + ] + } +} \ No newline at end of file diff --git a/results/metrics/r1-qwen-1.5b_bnb_nf4_w4_restored_gpqa_run0_metrics.json b/results/metrics/r1-qwen-1.5b_bnb_nf4_w4_restored_gpqa_run0_metrics.json new file mode 100644 index 0000000000000000000000000000000000000000..0573afe1eda1afdab5a154615f45e33d25c7c61d --- /dev/null +++ b/results/metrics/r1-qwen-1.5b_bnb_nf4_w4_restored_gpqa_run0_metrics.json @@ -0,0 +1,47 @@ +{ + "model": "r1-qwen-1.5b", + "quantization": "bnb_nf4_w4_restored", + "accuracy": 0.2222222222222222, + "accuracy_delta": 0.0, + "avg_ffs": 0.35714285714285715, + "median_ffs": 0.0, + "ecr": 0.986845600449691, + "ssr_curve": [ + 0.4797979797979798, + 0.23232323232323232, + 0.21875, + 0.21311475409836064, + 0.20689655172413793, + 0.20930232558139536, + 0.1927710843373494, + 0.1901840490797546, + 0.19254658385093168, + 0.1910828025477707, + 0.1895424836601307, + 0.18543046357615894, + 0.18120805369127516, + 0.18055555555555555, + 0.18181818181818182, + 0.18181818181818182, + 0.18309859154929578, + 0.18571428571428572, + 0.18705035971223022, + 0.1897810218978102, + 0.19117647058823528, + 0.19548872180451127, + 0.19548872180451127, + 0.183206106870229, + 0.17692307692307693, + 0.17829457364341086, + 0.18110236220472442, + 0.17073170731707318, + 0.1810344827586207, + 0.17272727272727273 + ], + "error_type_dist": { + "conceptual": 0.7398496240601504, + "executional": 0.00037593984962406017, + "logical": 0.2580827067669173, + "methodological": 0.0016917293233082707 + } +} \ No newline at end of file diff --git a/results/metrics/r1-qwen-1.5b_bnb_nf4_w4_restored_gsm8k_run0_ci.json b/results/metrics/r1-qwen-1.5b_bnb_nf4_w4_restored_gsm8k_run0_ci.json new file mode 100644 index 0000000000000000000000000000000000000000..4791aba0a184eeabedfd56cd6ab9783946cd5e64 --- /dev/null +++ b/results/metrics/r1-qwen-1.5b_bnb_nf4_w4_restored_gsm8k_run0_ci.json @@ -0,0 +1,116 @@ +{ + "n": 1319, + "accuracy": { + "mean": 0.6405987869598181, + "ci_lo": 0.6148597422289613, + "ci_hi": 0.6671721000758151 + }, + "avg_ffs": { + "mean": 0.8116932945681711, + "ci_lo": 0.6943317628131463, + "ci_hi": 0.9329209363478638 + }, + "ecr": { + "mean": 0.8214401864175835, + "ci_lo": 0.7956108328490901, + "ci_hi": 0.8466872515541066 + }, + "ssr_curve_ci": { + "mean": [ + 0.7991552691432913, + 0.7049795288454366, + 0.6808275651379063, + 0.672038337141794, + 0.6702837744264747, + 0.653405064121486, + 0.6326520502411499, + 0.595402690116086, + 0.5780742504310692, + 0.4799119233477354, + 0.45856591700535826, + 0.4175027445735618, + 0.4016060815261986, + 0.3875621195086363, + 0.3805229106755658, + 0.37386534471752386, + 0.3711131635479947, + 0.3611223123071096, + 0.3463130027808741, + 0.3426715395448019, + 0.3349973263126924, + 0.3270537225629823, + 0.3265873288216204, + 0.3255569519516153, + 0.3346118305256126, + 0.3492144200715925, + 0.3648327133097085, + 0.3704305616535991, + 0.3704305616535991, + 0.3761913557075657 + ], + "ci_lo": [ + 0.77710386656558, + 0.6796338672768879, + 0.6543000287290314, + 0.6445197021355362, + 0.6409437819145498, + 0.6199482065553494, + 0.590997951263571, + 0.5454545454545454, + 0.5166666666666667, + 0.40717265469061875, + 0.37593984962406013, + 0.3302752293577982, + 0.3113207547169811, + 0.2954491341991342, + 0.2856766917293233, + 0.2794032147742818, + 0.2765957446808511, + 0.26373626373626374, + 0.2476993284781992, + 0.24468085106382978, + 0.23684210526315788, + 0.22727272727272727, + 0.2249929775280899, + 0.2191780821917808, + 0.22618087557603686, + 0.23635732323232322, + 0.25, + 0.25332911392405066, + 0.25332911392405066, + 0.25757575757575757 + ], + "ci_hi": [ + 0.821076573161486, + 0.7297090352220521, + 0.7064449997734208, + 0.6993185689948893, + 0.7, + 0.6875800256081946, + 0.6742006933744222, + 0.6450619457485655, + 0.6367191485969388, + 0.554358993448481, + 0.5447761194029851, + 0.5083368644067796, + 0.49523809523809526, + 0.4819330895008605, + 0.47663823081938705, + 0.47058823529411764, + 0.4673913043478261, + 0.46153846153846156, + 0.4473684210526316, + 0.4457831325301205, + 0.4375, + 0.42857142857142855, + 0.4318251964085296, + 0.43478260869565216, + 0.44286166365280283, + 0.4626865671641791, + 0.4807753164556961, + 0.4852941176470588, + 0.4852941176470588, + 0.4927587262706674 + ] + } +} \ No newline at end of file diff --git a/results/metrics/r1-qwen-1.5b_bnb_nf4_w4_restored_gsm8k_run0_metrics.json b/results/metrics/r1-qwen-1.5b_bnb_nf4_w4_restored_gsm8k_run0_metrics.json new file mode 100644 index 0000000000000000000000000000000000000000..46eaa16565dbf505fcbd41a392627ac2c99eb634 --- /dev/null +++ b/results/metrics/r1-qwen-1.5b_bnb_nf4_w4_restored_gsm8k_run0_metrics.json @@ -0,0 +1,47 @@ +{ + "model": "r1-qwen-1.5b", + "quantization": "bnb_nf4_w4_restored", + "accuracy": 0.640636846095527, + "accuracy_delta": 0.0, + "avg_ffs": 0.8122362869198312, + "median_ffs": 0.0, + "ecr": 0.8215136509173403, + "ssr_curve": [ + 0.799090219863533, + 0.7050691244239631, + 0.6809864757358791, + 0.6721453287197232, + 0.670442842430484, + 0.6535031847133758, + 0.6325757575757576, + 0.5951742627345844, + 0.5776892430278885, + 0.47928994082840237, + 0.4573643410852713, + 0.415929203539823, + 0.4, + 0.38613861386138615, + 0.37894736842105264, + 0.3723404255319149, + 0.3695652173913043, + 0.3595505617977528, + 0.3448275862068966, + 0.3411764705882353, + 0.3333333333333333, + 0.3253012048192771, + 0.325, + 0.32432432432432434, + 0.3333333333333333, + 0.34782608695652173, + 0.36363636363636365, + 0.36923076923076925, + 0.36923076923076925, + 0.375 + ], + "error_type_dist": { + "conceptual": 0.7167979002624671, + "executional": 0.05380577427821522, + "logical": 0.22913385826771654, + "methodological": 0.00026246719160104987 + } +} \ No newline at end of file diff --git a/results/metrics/r1-qwen-1.5b_bnb_nf4_w4_restored_math500_run0_metrics.json b/results/metrics/r1-qwen-1.5b_bnb_nf4_w4_restored_math500_run0_metrics.json new file mode 100644 index 0000000000000000000000000000000000000000..f448664bdc9adb8bc7aefdbf574f17b97699d890 --- /dev/null +++ b/results/metrics/r1-qwen-1.5b_bnb_nf4_w4_restored_math500_run0_metrics.json @@ -0,0 +1,47 @@ +{ + "model": "r1-qwen-1.5b", + "quantization": "bnb_nf4_w4_restored", + "accuracy": 0.578, + "accuracy_delta": 0.0, + "avg_ffs": 0.5308056872037915, + "median_ffs": 0.0, + "ecr": 0.9542110386033714, + "ssr_curve": [ + 0.74, + 0.5971943887775552, + 0.5885947046843177, + 0.5847107438016529, + 0.5806451612903226, + 0.5727272727272728, + 0.5700712589073634, + 0.5433673469387755, + 0.5333333333333333, + 0.5140449438202247, + 0.5014492753623189, + 0.4909090909090909, + 0.4779874213836478, + 0.46405228758169936, + 0.45791245791245794, + 0.4444444444444444, + 0.43272727272727274, + 0.4153846153846154, + 0.4063745019920319, + 0.3891213389121339, + 0.36607142857142855, + 0.3568075117370892, + 0.33663366336633666, + 0.33157894736842103, + 0.31843575418994413, + 0.32, + 0.3067484662576687, + 0.3087248322147651, + 0.3006993006993007, + 0.29927007299270075 + ], + "error_type_dist": { + "conceptual": 0.7059951202509586, + "executional": 0.030498431509236666, + "logical": 0.2558382711746253, + "methodological": 0.007668177065179505 + } +} \ No newline at end of file diff --git a/results/metrics/r1-qwen-1.5b_gptq_w4_gpqa_run0_ci.json b/results/metrics/r1-qwen-1.5b_gptq_w4_gpqa_run0_ci.json new file mode 100644 index 0000000000000000000000000000000000000000..7892de816090206bd982dbaa5e1c0efa4d67c684 --- /dev/null +++ b/results/metrics/r1-qwen-1.5b_gptq_w4_gpqa_run0_ci.json @@ -0,0 +1,116 @@ +{ + "n": 198, + "accuracy": { + "mean": 0.21175151515151516, + "ci_lo": 0.15656565656565657, + "ci_hi": 0.2676767676767677 + }, + "avg_ffs": { + "mean": 0.40900575803745143, + "ci_lo": 0.29411764705882354, + "ci_hi": 0.5534591194968553 + }, + "ecr": { + "mean": 0.9908308700634533, + "ci_lo": 0.9836417153064911, + "ci_hi": 0.9965791153084091 + }, + "ssr_curve_ci": { + "mean": [ + 0.4792818181818249, + 0.22173131313131267, + 0.2101670233323784, + 0.21816394649940798, + 0.21978955312543366, + 0.223564445654136, + 0.21452422488012932, + 0.1994515682924241, + 0.20183791999091671, + 0.20674872200420522, + 0.19704368546158138, + 0.20334925824229333, + 0.20999780262166493, + 0.20892516394662453, + 0.2117898447967446, + 0.2117898447967446, + 0.20773434780424022, + 0.2091664311389541, + 0.21519167775170595, + 0.22161883360310872, + 0.2232828939120333, + 0.22669618818133536, + 0.22669618818133536, + 0.22258887458935137, + 0.22964483913040484, + 0.22530697929873908, + 0.23478924349956656, + 0.2282073859404565, + 0.22268250599447015, + 0.22670886769957183 + ], + "ci_lo": [ + 0.4090909090909091, + 0.16666666666666666, + 0.1542349150061916, + 0.16022099447513813, + 0.16, + 0.16265060240963855, + 0.15517241379310345, + 0.1419753086419753, + 0.14374906437125748, + 0.14743098159509202, + 0.1377245508982036, + 0.1419753086419753, + 0.14743589743589744, + 0.1456953642384106, + 0.14743217054263566, + 0.14743217054263566, + 0.14285714285714285, + 0.14471842105263158, + 0.14893617021276595, + 0.15441053099091986, + 0.1555518018018018, + 0.15789473684210525, + 0.15789473684210525, + 0.15384615384615385, + 0.15965777615393983, + 0.15502642706131078, + 0.1612834882449426, + 0.15447005021520804, + 0.14781895937277265, + 0.15 + ], + "ci_hi": [ + 0.5454545454545454, + 0.2777777777777778, + 0.26842105263157895, + 0.2786885245901639, + 0.28160919540229884, + 0.2865176752743281, + 0.27684296508066153, + 0.26137024312896395, + 0.2641563652136195, + 0.26993865030674846, + 0.2597402597402597, + 0.2676148393629751, + 0.27564102564102566, + 0.2751677852348993, + 0.27972027972027974, + 0.27972027972027974, + 0.27517264857504126, + 0.2777777777777778, + 0.2857142857142857, + 0.29365218115218117, + 0.2962962962962963, + 0.3007518796992481, + 0.3007518796992481, + 0.2962962962962963, + 0.3050847457627119, + 0.3000187969924809, + 0.3130532121998701, + 0.30579815352542605, + 0.30097087378640774, + 0.30632191597538105 + ] + } +} \ No newline at end of file diff --git a/results/metrics/r1-qwen-1.5b_gptq_w4_gpqa_run0_metrics.json b/results/metrics/r1-qwen-1.5b_gptq_w4_gpqa_run0_metrics.json new file mode 100644 index 0000000000000000000000000000000000000000..630efd7ce1e1cedd2dcc393658d13c0041a4e2a2 --- /dev/null +++ b/results/metrics/r1-qwen-1.5b_gptq_w4_gpqa_run0_metrics.json @@ -0,0 +1,47 @@ +{ + "model": "r1-qwen-1.5b", + "quantization": "gptq_w4", + "accuracy": 0.21212121212121213, + "accuracy_delta": 0.0, + "avg_ffs": 0.41025641025641024, + "median_ffs": 0.0, + "ecr": 0.9908687280157069, + "ssr_curve": [ + 0.4797979797979798, + 0.2222222222222222, + 0.21052631578947367, + 0.2185792349726776, + 0.22033898305084745, + 0.22413793103448276, + 0.21511627906976744, + 0.2, + 0.20238095238095238, + 0.2073170731707317, + 0.19753086419753085, + 0.20382165605095542, + 0.21052631578947367, + 0.20945945945945946, + 0.21232876712328766, + 0.21232876712328766, + 0.20833333333333334, + 0.2097902097902098, + 0.2158273381294964, + 0.2222222222222222, + 0.22388059701492538, + 0.22727272727272727, + 0.22727272727272727, + 0.2230769230769231, + 0.23015873015873015, + 0.22580645161290322, + 0.23529411764705882, + 0.2288135593220339, + 0.22321428571428573, + 0.22727272727272727 + ], + "error_type_dist": { + "conceptual": 0.7860657350915174, + "executional": 0.001574493209998032, + "logical": 0.21137571344223577, + "methodological": 0.00098405825624877 + } +} \ No newline at end of file diff --git a/results/metrics/r1-qwen-1.5b_gptq_w4_gsm8k_run0_ci.json b/results/metrics/r1-qwen-1.5b_gptq_w4_gsm8k_run0_ci.json new file mode 100644 index 0000000000000000000000000000000000000000..baa86c0cdea422105053a66efe8331b9e0663622 --- /dev/null +++ b/results/metrics/r1-qwen-1.5b_gptq_w4_gsm8k_run0_ci.json @@ -0,0 +1,116 @@ +{ + "n": 1319, + "accuracy": { + "mean": 0.6256168309325246, + "ci_lo": 0.599696739954511, + "ci_hi": 0.6512509476876421 + }, + "avg_ffs": { + "mean": 0.5425309807310789, + "ci_lo": 0.4532796109521732, + "ci_hi": 0.6401691258074044 + }, + "ecr": { + "mean": 0.8276576113940671, + "ci_lo": 0.8038325248523832, + "ci_hi": 0.8511329393550913 + }, + "ssr_curve_ci": { + "mean": [ + 0.7453240333586066, + 0.6675377928445432, + 0.6489622463656435, + 0.6317905472094585, + 0.6160369614347678, + 0.6001847156806599, + 0.5728050945888401, + 0.5310939597150932, + 0.49589861848226685, + 0.4737728438176532, + 0.4589456584413745, + 0.44600302045062773, + 0.3654071914193886, + 0.35075622970839293, + 0.33922903120225173, + 0.33316159228647496, + 0.320875965484163, + 0.3272537688516193, + 0.3142707505075969, + 0.30641844514403205, + 0.2985402305555058, + 0.26339109496486274, + 0.24471642436856086, + 0.24471642436856086, + 0.25081758512003005, + 0.25081758512003005, + 0.25081758512003005, + 0.2570364210582993, + 0.2504861545675132, + 0.257575998625482 + ], + "ci_lo": [ + 0.7224981046247156, + 0.642473782635705, + 0.6227289950424231, + 0.6045746980647425, + 0.5862361268597739, + 0.5678573445181235, + 0.5340896197221839, + 0.48192552291490415, + 0.4317157084227569, + 0.39473236663086286, + 0.36081575161831697, + 0.3332417582417583, + 0.25, + 0.22807017543859648, + 0.21739130434782608, + 0.20833333333333334, + 0.2, + 0.20369585687382297, + 0.19148936170212766, + 0.18, + 0.1730603448275862, + 0.13513513513513514, + 0.11997619047619049, + 0.11997619047619049, + 0.12195121951219512, + 0.12195121951219512, + 0.12195121951219512, + 0.125, + 0.11428571428571428, + 0.11901260504201681 + ], + "ci_hi": [ + 0.7687642153146323, + 0.6923678869504781, + 0.6742015326627047, + 0.6580441755548022, + 0.6437111124395359, + 0.6304579637864971, + 0.6110281606261698, + 0.5775290284519053, + 0.559056012884753, + 0.5526356652500817, + 0.5568181818181818, + 0.5588235294117647, + 0.48530462184873935, + 0.4791874999999997, + 0.4642857142857143, + 0.46153846153846156, + 0.4500245098039212, + 0.4583333333333333, + 0.4444444444444444, + 0.4386071887034658, + 0.43138368983957204, + 0.40476190476190477, + 0.3829787234042553, + 0.3829787234042553, + 0.39132566069906194, + 0.39132566069906194, + 0.39132566069906194, + 0.4, + 0.4, + 0.4090909090909091 + ] + } +} \ No newline at end of file diff --git a/results/metrics/r1-qwen-1.5b_gptq_w4_gsm8k_run0_sig.json b/results/metrics/r1-qwen-1.5b_gptq_w4_gsm8k_run0_sig.json new file mode 100644 index 0000000000000000000000000000000000000000..3c2e04b4b8170ac51f06b2f2d28d9de78ed28aa6 --- /dev/null +++ b/results/metrics/r1-qwen-1.5b_gptq_w4_gsm8k_run0_sig.json @@ -0,0 +1,7 @@ +{ + "n_pairs": 1319, + "delta_acc_observed": 0.014404852160727732, + "delta_acc_ci_lo": -0.011372251705837777, + "delta_acc_ci_hi": 0.039423805913570864, + "p_value": 0.302 +} \ No newline at end of file diff --git a/results/metrics/r1-qwen-1.5b_gptq_w4_math500_run0_ci.json b/results/metrics/r1-qwen-1.5b_gptq_w4_math500_run0_ci.json new file mode 100644 index 0000000000000000000000000000000000000000..a9353d1600f11c6af9fddf6c1dd0c10111d1a0b0 --- /dev/null +++ b/results/metrics/r1-qwen-1.5b_gptq_w4_math500_run0_ci.json @@ -0,0 +1,116 @@ +{ + "n": 500, + "accuracy": { + "mean": 0.5003084, + "ci_lo": 0.458, + "ci_hi": 0.544 + }, + "avg_ffs": { + "mean": 0.47643101643114405, + "ci_lo": 0.3946360153256705, + "ci_hi": 0.5632783236751489 + }, + "ecr": { + "mean": 0.9660786757277452, + "ci_lo": 0.9563440602256191, + "ci_hi": 0.9746200486526497 + }, + "ssr_curve_ci": { + "mean": [ + 0.7064728000000046, + 0.5183590181779478, + 0.5083933131054454, + 0.4952069833453463, + 0.4878191313836927, + 0.4733332261472824, + 0.45204205304030093, + 0.44000428142962994, + 0.42859805066840706, + 0.4114719731046476, + 0.4015323896982401, + 0.3960874480136543, + 0.3758286404926524, + 0.3673631866450891, + 0.3563330180933734, + 0.34419206422871595, + 0.331420058508511, + 0.32815364875789355, + 0.31136239696449036, + 0.303997046991391, + 0.2914460018749638, + 0.2910496476171269, + 0.2771572503338463, + 0.28031552221825173, + 0.27729961022316046, + 0.2767136498804996, + 0.2713690365920019, + 0.2592130725169884, + 0.257008760066108, + 0.2536571976740673 + ], + "ci_lo": [ + 0.666, + 0.47484909456740443, + 0.46464646464646464, + 0.45081967213114754, + 0.44375, + 0.4279475982532751, + 0.40582536593620444, + 0.3936390487220496, + 0.38130411255411256, + 0.3631713554987212, + 0.3516483516483517, + 0.34687900389032544, + 0.3257774180493727, + 0.3181818181818182, + 0.30563176317987173, + 0.29283489096573206, + 0.2812475039936102, + 0.27743853265145557, + 0.2611683848797251, + 0.25333116883116885, + 0.2405475328559062, + 0.2392839635854342, + 0.2248972634724946, + 0.22763749402199904, + 0.22413614250965805, + 0.22270742358078602, + 0.21701743142783902, + 0.20276342010212978, + 0.2, + 0.1949967948717949 + ], + "ci_hi": [ + 0.746, + 0.56312625250501, + 0.5524193548387096, + 0.5398773006134969, + 0.5326315789473685, + 0.5185342789598106, + 0.4977277846027846, + 0.48668441319758454, + 0.4763092269326683, + 0.4596774193548387, + 0.4507042253521127, + 0.44628215903254426, + 0.4265139859001286, + 0.41945507109667046, + 0.40880854783290127, + 0.39743690851735014, + 0.38368961518508343, + 0.3811074918566775, + 0.3633217993079585, + 0.3566666666666667, + 0.34408602150537637, + 0.34432632584806494, + 0.3320316270511583, + 0.33461575875486377, + 0.3320928753180657, + 0.3333653846153841, + 0.33043571561858553, + 0.319047619047619, + 0.31731150793650786, + 0.3157894736842105 + ] + } +} \ No newline at end of file diff --git a/results/metrics/r1-qwen-1.5b_gptq_w4_math500_run0_sig.json b/results/metrics/r1-qwen-1.5b_gptq_w4_math500_run0_sig.json new file mode 100644 index 0000000000000000000000000000000000000000..d823f9805784ec95c35a47bed49b814680e585ac --- /dev/null +++ b/results/metrics/r1-qwen-1.5b_gptq_w4_math500_run0_sig.json @@ -0,0 +1,7 @@ +{ + "n_pairs": 500, + "delta_acc_observed": 0.07599999999999996, + "delta_acc_ci_lo": 0.03199999999999992, + "delta_acc_ci_hi": 0.12, + "p_value": 0.0012 +} \ No newline at end of file diff --git a/results/metrics/r1-qwen-1.5b_gptq_w4_restored_gpqa_run0_ci.json b/results/metrics/r1-qwen-1.5b_gptq_w4_restored_gpqa_run0_ci.json new file mode 100644 index 0000000000000000000000000000000000000000..46b2458fb140877b57b748d3bcb54024cd2ebc86 --- /dev/null +++ b/results/metrics/r1-qwen-1.5b_gptq_w4_restored_gpqa_run0_ci.json @@ -0,0 +1,116 @@ +{ + "n": 198, + "accuracy": { + "mean": 0.21697272727272726, + "ci_lo": 0.16161616161616163, + "ci_hi": 0.2727272727272727 + }, + "avg_ffs": { + "mean": 0.2517106774719704, + "ci_lo": 0.18237564322469982, + "ci_hi": 0.32741560665361985 + }, + "ecr": { + "mean": 0.9917941311721037, + "ci_lo": 0.9793100353775753, + "ci_hi": 0.998696109182373 + }, + "ssr_curve_ci": { + "mean": [ + 0.4039515151515197, + 0.22710202020202058, + 0.21449314012955506, + 0.21520809264778026, + 0.21621289714565922, + 0.22275703988968532, + 0.2137492206294774, + 0.21284517823222981, + 0.21696647101010558, + 0.21838494630365454, + 0.21452286686566907, + 0.21452286686566907, + 0.21744344045399472, + 0.21744344045399472, + 0.21660171449947416, + 0.21814449399966399, + 0.22124002739424545, + 0.22615933621338277, + 0.22779894035503467, + 0.22375355383503448, + 0.22375355383503448, + 0.22299064913507996, + 0.2300669333898155, + 0.2458048723731743, + 0.2414624883111661, + 0.25014557561362444, + 0.25544684920642186, + 0.2605639573637483, + 0.2658582919364365, + 0.26928529536698076 + ], + "ci_lo": [ + 0.3383838383838384, + 0.1717171717171717, + 0.15625, + 0.15728970543577286, + 0.15624548410404623, + 0.16149068322981366, + 0.1518891446106636, + 0.1503267973856209, + 0.15383333333333335, + 0.15436241610738255, + 0.14893617021276595, + 0.14893617021276595, + 0.15106928156105254, + 0.15106928156105254, + 0.1503186274509804, + 0.15106928156105254, + 0.15384615384615385, + 0.15714285714285714, + 0.15827338129496402, + 0.15516907244052394, + 0.15516907244052394, + 0.15267175572519084, + 0.15908189033189032, + 0.1705201411960133, + 0.1650485436893204, + 0.1717171717171717, + 0.1717171717171717, + 0.17647058823529413, + 0.17976807836358397, + 0.18072289156626506 + ], + "ci_hi": [ + 0.4696969696969697, + 0.2878787878787879, + 0.2736842105263158, + 0.2752808988764045, + 0.2784090909090909, + 0.2874251497005988, + 0.27879097093382804, + 0.2781456953642384, + 0.2830188679245283, + 0.2847233719646799, + 0.28205128205128205, + 0.28205128205128205, + 0.2857142857142857, + 0.2857142857142857, + 0.2846715328467153, + 0.2866666666666667, + 0.29008794099442936, + 0.2962962962962963, + 0.2978723404255319, + 0.29452187533184665, + 0.29452187533184665, + 0.2949669772378818, + 0.3043563179347825, + 0.3238095238095238, + 0.32000970873786394, + 0.3305785123966942, + 0.34021166922570734, + 0.3473684210526316, + 0.3541666666666667, + 0.36001162790697655 + ] + } +} \ No newline at end of file diff --git a/results/metrics/r1-qwen-1.5b_gptq_w4_restored_math500_run0_metrics.json b/results/metrics/r1-qwen-1.5b_gptq_w4_restored_math500_run0_metrics.json new file mode 100644 index 0000000000000000000000000000000000000000..de3a0e8938b4710b5f9033871631d8242f58f044 --- /dev/null +++ b/results/metrics/r1-qwen-1.5b_gptq_w4_restored_math500_run0_metrics.json @@ -0,0 +1,47 @@ +{ + "model": "r1-qwen-1.5b", + "quantization": "gptq_w4_restored", + "accuracy": 0.576, + "accuracy_delta": 0.0, + "avg_ffs": 0.47641509433962265, + "median_ffs": 0.0, + "ecr": 0.9607474555787682, + "ssr_curve": [ + 0.74, + 0.5967741935483871, + 0.5926680244399185, + 0.5833333333333334, + 0.5711206896551724, + 0.5605381165919282, + 0.5521327014218009, + 0.5355329949238579, + 0.5225464190981433, + 0.5111731843575419, + 0.4956268221574344, + 0.47560975609756095, + 0.4588607594936709, + 0.4473684210526316, + 0.4334470989761092, + 0.4234875444839858, + 0.41544117647058826, + 0.40384615384615385, + 0.38645418326693226, + 0.37916666666666665, + 0.3695652173913043, + 0.3623853211009174, + 0.33816425120772947, + 0.31794871794871793, + 0.31693989071038253, + 0.3202247191011236, + 0.32335329341317365, + 0.32515337423312884, + 0.33112582781456956, + 0.3076923076923077 + ], + "error_type_dist": { + "conceptual": 0.7344057780695995, + "executional": 0.02757715036112935, + "logical": 0.2304661851608667, + "methodological": 0.0075508864084044645 + } +} \ No newline at end of file diff --git a/results/metrics/r1-qwen-14b_gptq_w4_gpqa_run0_ci.json b/results/metrics/r1-qwen-14b_gptq_w4_gpqa_run0_ci.json new file mode 100644 index 0000000000000000000000000000000000000000..a5c5e43f5a687b2120f7e46a261e740be5d82e9c --- /dev/null +++ b/results/metrics/r1-qwen-14b_gptq_w4_gpqa_run0_ci.json @@ -0,0 +1,116 @@ +{ + "n": 198, + "accuracy": { + "mean": 0.18196363636363636, + "ci_lo": 0.13131313131313133, + "ci_hi": 0.23737373737373738 + }, + "avg_ffs": { + "mean": 0.2589544674829587, + "ci_lo": 0.1910828025477707, + "ci_hi": 0.33116986516668045 + }, + "ecr": { + "mean": 0.9929294421918492, + "ci_lo": 0.9888839391280477, + "ci_hi": 0.9963452601965833 + }, + "ssr_curve_ci": { + "mean": [ + 0.38873131313131765, + 0.18703333333333136, + 0.18383462003145135, + 0.18383462003145135, + 0.18674383500369, + 0.18971819495375944, + 0.19072725025715362, + 0.18740318474651935, + 0.18609594795041048, + 0.18357649681071908, + 0.18560698962702035, + 0.1829953052046644, + 0.17854068035462867, + 0.17463496949709778, + 0.1800807126583898, + 0.17593477758648415, + 0.18049191348324287, + 0.18285959425373902, + 0.1853302218936871, + 0.1779964507693018, + 0.1747348299916307, + 0.1837417272878272, + 0.1817340667116796, + 0.1859442197182928, + 0.17578485318731052, + 0.17073876731693274, + 0.1767921340816941, + 0.16658801591510825, + 0.14688810605103408, + 0.15442225214164884 + ], + "ci_lo": [ + 0.32323232323232326, + 0.13636363636363635, + 0.1319796954314721, + 0.1319796954314721, + 0.13333333333333333, + 0.13541666666666666, + 0.13612565445026178, + 0.13368983957219252, + 0.13114754098360656, + 0.12777777777777777, + 0.12921348314606743, + 0.1264352027283062, + 0.12195121951219512, + 0.11801242236024845, + 0.12195121951219512, + 0.11801242236024845, + 0.12160655878808746, + 0.12258064516129032, + 0.125, + 0.11805555555555555, + 0.11486486486486487, + 0.12030075187969924, + 0.11764705882352941, + 0.12095106718408927, + 0.11191956882255391, + 0.10434782608695652, + 0.10810810810810811, + 0.09899366985878917, + 0.08163265306122448, + 0.0865354938271605 + ], + "ci_hi": [ + 0.45454545454545453, + 0.24242424242424243, + 0.23857868020304568, + 0.23857868020304568, + 0.24210526315789474, + 0.24607329842931938, + 0.24736842105263157, + 0.24468085106382978, + 0.2430939226519337, + 0.24138415342890346, + 0.2430939226519337, + 0.2413793103448276, + 0.23809523809523808, + 0.2331288343558282, + 0.23952095808383234, + 0.23530348445110513, + 0.24114665143509906, + 0.2447621914509331, + 0.24817647667857326, + 0.24113475177304963, + 0.23880597014925373, + 0.25, + 0.24817518248175183, + 0.2554785686844849, + 0.24561775200713643, + 0.2413793103448276, + 0.25, + 0.2403846153846154, + 0.21782178217821782, + 0.22826086956521738 + ] + } +} \ No newline at end of file diff --git a/results/metrics/r1-qwen-14b_gptq_w4_gsm8k_run0_ci.json b/results/metrics/r1-qwen-14b_gptq_w4_gsm8k_run0_ci.json new file mode 100644 index 0000000000000000000000000000000000000000..e4e1ca8904acc0e40618c0ed6b491496cc01134a --- /dev/null +++ b/results/metrics/r1-qwen-14b_gptq_w4_gsm8k_run0_ci.json @@ -0,0 +1,116 @@ +{ + "n": 1319, + "accuracy": { + "mean": 0.7923034116755118, + "ci_lo": 0.7702805155420773, + "ci_hi": 0.8142532221379833 + }, + "avg_ffs": { + "mean": 0.3644158477215975, + "ci_lo": 0.27898298111900227, + "ci_hi": 0.46071519990248655 + }, + "ecr": { + "mean": 0.856261920517607, + "ci_lo": 0.8261066770410117, + "ci_hi": 0.8839986620203054 + }, + "ssr_curve_ci": { + "mean": [ + 0.8483974222896129, + 0.8035381435908585, + 0.795225958805212, + 0.7998573613563424, + 0.8004650850711102, + 0.7972226005976814, + 0.7739388228759887, + 0.7558702765291522, + 0.7387239462314338, + 0.7177210340830896, + 0.6539182847568281, + 0.6040923612603517, + 0.5510754399009182, + 0.4946309577619734, + 0.4941891165090382, + 0.46702190696553364, + 0.47414079838462947, + 0.43833657201201104, + 0.43572873869089945, + 0.4110566508582051, + 0.4110566508582051, + 0.40575728040818204, + 0.41182588456938446, + 0.3940503045682649, + 0.31053246181288974, + 0.31053246181288974, + 0.2862579886508331, + 0.2917058677921078, + 0.3044737296082288, + 0.33327885990131695 + ], + "ci_lo": [ + 0.8286580742987112, + 0.7818459191456903, + 0.7725449435823477, + 0.7764078995439543, + 0.7749469214437368, + 0.768700662671278, + 0.7378585403669387, + 0.7134901043932145, + 0.6859154203197524, + 0.6505950181092468, + 0.5714285714285714, + 0.5113573232323232, + 0.4482758620689655, + 0.38461538461538464, + 0.3793103448275862, + 0.33999056603773586, + 0.34424948770491803, + 0.3, + 0.29545454545454547, + 0.2580414746543779, + 0.2580414746543779, + 0.25, + 0.25, + 0.23076923076923078, + 0.14695378151260508, + 0.14695378151260508, + 0.125, + 0.11538461538461539, + 0.12, + 0.13333333333333333 + ], + "ci_hi": [ + 0.8673426838514023, + 0.8252882769141744, + 0.817741935483871, + 0.8226381461675579, + 0.8250803270458295, + 0.8251943595561618, + 0.808349476690755, + 0.7989825079596455, + 0.791342762897881, + 0.7833359872611465, + 0.734697742733457, + 0.696969696969697, + 0.6521831245394251, + 0.6086956521739131, + 0.6125100806451612, + 0.5964912280701754, + 0.6065720140515221, + 0.5814059196617335, + 0.5813953488372093, + 0.5682629870129858, + 0.5682629870129858, + 0.5666891891891889, + 0.5833333333333334, + 0.5666666666666667, + 0.48781905842314216, + 0.48781905842314216, + 0.4666666666666667, + 0.5, + 0.5, + 0.5555555555555556 + ] + } +} \ No newline at end of file diff --git a/results/metrics/r1-qwen-14b_gptq_w4_gsm8k_run0_metrics.json b/results/metrics/r1-qwen-14b_gptq_w4_gsm8k_run0_metrics.json new file mode 100644 index 0000000000000000000000000000000000000000..3616eb196ed435955d0c065ec121862ddb370abb --- /dev/null +++ b/results/metrics/r1-qwen-14b_gptq_w4_gsm8k_run0_metrics.json @@ -0,0 +1,47 @@ +{ + "model": "r1-qwen-14b", + "quantization": "gptq_w4", + "accuracy": 0.7922668688400303, + "accuracy_delta": 0.0, + "avg_ffs": 0.36496350364963503, + "median_ffs": 0.0, + "ecr": 0.8563846181664173, + "ssr_curve": [ + 0.8483699772554966, + 0.8035168195718655, + 0.7951807228915663, + 0.7998236331569665, + 0.8004246284501062, + 0.7971391417425228, + 0.7740112994350282, + 0.7559055118110236, + 0.7386363636363636, + 0.7175141242937854, + 0.6538461538461539, + 0.6037735849056604, + 0.550561797752809, + 0.49333333333333335, + 0.49295774647887325, + 0.46551724137931033, + 0.4727272727272727, + 0.4375, + 0.43478260869565216, + 0.41025641025641024, + 0.41025641025641024, + 0.40540540540540543, + 0.4117647058823529, + 0.3939393939393939, + 0.3103448275862069, + 0.3103448275862069, + 0.2857142857142857, + 0.2916666666666667, + 0.30434782608695654, + 0.3333333333333333 + ], + "error_type_dist": { + "conceptual": 0.7451352634076887, + "executional": 0.027052681537731372, + "logical": 0.2273374466065496, + "methodological": 0.00047460844803037496 + } +} \ No newline at end of file diff --git a/results/metrics/r1-qwen-14b_gptq_w4_math500_run0_ci.json b/results/metrics/r1-qwen-14b_gptq_w4_math500_run0_ci.json new file mode 100644 index 0000000000000000000000000000000000000000..dfed4f9389dcff96f1ee912f3d1d22c9797b8bd2 --- /dev/null +++ b/results/metrics/r1-qwen-14b_gptq_w4_math500_run0_ci.json @@ -0,0 +1,116 @@ +{ + "n": 500, + "accuracy": { + "mean": 0.7464472, + "ci_lo": 0.708, + "ci_hi": 0.784 + }, + "avg_ffs": { + "mean": 0.479491936130859, + "ci_lo": 0.3582064340459863, + "ci_hi": 0.6220472440944882 + }, + "ecr": { + "mean": 0.9714926312071605, + "ci_lo": 0.9615374554909905, + "ci_hi": 0.9804365286773215 + }, + "ssr_curve_ci": { + "mean": [ + 0.8480528000000007, + 0.7585423999999994, + 0.7479253984662718, + 0.7463533917283045, + 0.7447970644505899, + 0.7404977502230047, + 0.7257746321585083, + 0.710144433150815, + 0.6986402065187858, + 0.6865024331965024, + 0.6791821798930227, + 0.6594852709041463, + 0.6531608065209481, + 0.6433228750788781, + 0.6273978833216235, + 0.6126793597760439, + 0.598524499470218, + 0.5733574742008521, + 0.5524778003019118, + 0.5380472206847102, + 0.5135591250416869, + 0.49813591112786254, + 0.47131098101572905, + 0.4608132437578638, + 0.4413170938468075, + 0.42995976834686445, + 0.41773013421962435, + 0.42516772434556577, + 0.42474550402033284, + 0.4171749146278622 + ], + "ci_lo": [ + 0.816, + 0.72, + 0.708243259557344, + 0.7073170731707317, + 0.7055214723926381, + 0.70042194092827, + 0.6843267108167771, + 0.6666666666666666, + 0.6543163464216096, + 0.639888123622169, + 0.6312814818206794, + 0.6090875169606513, + 0.6013061553735418, + 0.589741641337386, + 0.5719063545150501, + 0.5551635353419636, + 0.5384615384615384, + 0.5098794853909944, + 0.4877023777265953, + 0.4703196347031963, + 0.44387755102040816, + 0.42602564102564106, + 0.3953488372093023, + 0.38460451977401133, + 0.3652593461725198, + 0.34965034965034963, + 0.3357142857142857, + 0.34166158536585367, + 0.33673216811477064, + 0.32710280373831774 + ], + "ci_hi": [ + 0.878, + 0.796, + 0.7855711422845691, + 0.7842741935483871, + 0.7829614604462475, + 0.7794171150788797, + 0.7660079794687237, + 0.7522935779816514, + 0.7418546365914787, + 0.7320454811905563, + 0.726292474463206, + 0.7078313253012049, + 0.7033665673024049, + 0.6952062948467058, + 0.6801346801346801, + 0.667808515428734, + 0.6561430921052631, + 0.6334906802227063, + 0.6141139311166105, + 0.6030170180800911, + 0.5803153039302412, + 0.5683060109289617, + 0.544828691423519, + 0.5354896313364055, + 0.517488716956802, + 0.5096805819101834, + 0.5, + 0.511815748031496, + 0.5130434782608696, + 0.5086244706594071 + ] + } +} \ No newline at end of file diff --git a/results/metrics/r1-qwen-14b_gptq_w4_math500_run0_metrics.json b/results/metrics/r1-qwen-14b_gptq_w4_math500_run0_metrics.json new file mode 100644 index 0000000000000000000000000000000000000000..ab1ceb27efab87acc2fc5b24d66a3eaf7ce52f09 --- /dev/null +++ b/results/metrics/r1-qwen-14b_gptq_w4_math500_run0_metrics.json @@ -0,0 +1,47 @@ +{ + "model": "r1-qwen-14b", + "quantization": "gptq_w4", + "accuracy": 0.746, + "accuracy_delta": 0.0, + "avg_ffs": 0.48031496062992124, + "median_ffs": 0.0, + "ecr": 0.9715710841753716, + "ssr_curve": [ + 0.848, + 0.758, + 0.7474949899799599, + 0.7459349593495935, + 0.7443762781186094, + 0.740041928721174, + 0.7252747252747253, + 0.7096018735362998, + 0.698019801980198, + 0.6858638743455497, + 0.6785714285714286, + 0.6588921282798834, + 0.6525679758308157, + 0.6426332288401254, + 0.6266666666666667, + 0.6118881118881119, + 0.5977443609022557, + 0.5725806451612904, + 0.5517241379310345, + 0.5370370370370371, + 0.5125628140703518, + 0.4972375690607735, + 0.47023809523809523, + 0.45962732919254656, + 0.44, + 0.42857142857142855, + 0.4166666666666667, + 0.424, + 0.423728813559322, + 0.415929203539823 + ], + "error_type_dist": { + "conceptual": 0.6606817590424148, + "executional": 0.017434296122820714, + "logical": 0.31043455633619566, + "methodological": 0.011449388498568826 + } +} \ No newline at end of file diff --git a/results/metrics/r1-qwen-7b_awq_w4_gpqa_run0_ci.json b/results/metrics/r1-qwen-7b_awq_w4_gpqa_run0_ci.json new file mode 100644 index 0000000000000000000000000000000000000000..d09991486a27b66ff1c616f30f1ce5938514c5bd --- /dev/null +++ b/results/metrics/r1-qwen-7b_awq_w4_gpqa_run0_ci.json @@ -0,0 +1,116 @@ +{ + "n": 198, + "accuracy": { + "mean": 0.22693434343434346, + "ci_lo": 0.1717171717171717, + "ci_hi": 0.2828282828282828 + }, + "avg_ffs": { + "mean": 0.5227857274032723, + "ci_lo": 0.4374896523178808, + "ci_hi": 0.6111199575371549 + }, + "ecr": { + "mean": 0.9887358860624578, + "ci_lo": 0.9835269701367509, + "ci_hi": 0.9934432757703112 + }, + "ssr_curve_ci": { + "mean": [ + 0.610848484848485, + 0.24712323232323274, + 0.22693434343434296, + 0.22412647922799198, + 0.2247120198804914, + 0.22411207786417023, + 0.2152076022422531, + 0.21761199114716906, + 0.22005565843842065, + 0.21810084690928774, + 0.22063721439678183, + 0.22000056256767866, + 0.21324896689898734, + 0.20835583245803113, + 0.2047344432628236, + 0.2021252189960006, + 0.20476313498074122, + 0.20376768649189259, + 0.20508233786862343, + 0.205345968942258, + 0.21567754785422677, + 0.23065192281852337, + 0.21862757592350512, + 0.2113285872955606, + 0.1844185290818805, + 0.18377955747736335, + 0.1812701636005921, + 0.17655842483042922, + 0.1681309199568954, + 0.17218297732899973 + ], + "ci_lo": [ + 0.5404040404040404, + 0.18686868686868688, + 0.1717171717171717, + 0.1683673469387755, + 0.16666666666666666, + 0.16574585635359115, + 0.15819209039548024, + 0.16, + 0.16201117318435754, + 0.1599940828402367, + 0.16184971098265896, + 0.15976331360946747, + 0.1524390243902439, + 0.14880357142857142, + 0.14465045261206239, + 0.14193548387096774, + 0.1437908496732026, + 0.14283509700176367, + 0.14285714285714285, + 0.14184397163120568, + 0.14893617021276595, + 0.16101694915254236, + 0.14925373134328357, + 0.14148732311320755, + 0.1166626213592233, + 0.11537610619469027, + 0.1111111111111111, + 0.10752688172043011, + 0.0945945945945946, + 0.0958904109589041 + ], + "ci_hi": [ + 0.6767676767676768, + 0.30808080808080807, + 0.2828282828282828, + 0.28205128205128205, + 0.28272251308900526, + 0.28342245989304815, + 0.2737430167597765, + 0.27661019888991656, + 0.27932960893854747, + 0.2774566473988439, + 0.28070668243642805, + 0.28235626453971413, + 0.2754491017964072, + 0.27044395116537173, + 0.26666666666666666, + 0.26506024096385544, + 0.2684563758389262, + 0.2679738562091503, + 0.26923758865248215, + 0.272108843537415, + 0.28368794326241137, + 0.3023255813953488, + 0.2890625, + 0.2818240093240092, + 0.25663716814159293, + 0.257149929278642, + 0.25663716814159293, + 0.25, + 0.24719101123595505, + 0.25716589861751116 + ] + } +} \ No newline at end of file diff --git a/results/metrics/r1-qwen-7b_awq_w4_gpqa_run0_metrics.json b/results/metrics/r1-qwen-7b_awq_w4_gpqa_run0_metrics.json new file mode 100644 index 0000000000000000000000000000000000000000..022369387b1647047fbe1dce282de967125a4785 --- /dev/null +++ b/results/metrics/r1-qwen-7b_awq_w4_gpqa_run0_metrics.json @@ -0,0 +1,47 @@ +{ + "model": "r1-qwen-7b", + "quantization": "awq_w4", + "accuracy": 0.22727272727272727, + "accuracy_delta": 0.0, + "avg_ffs": 0.5228758169934641, + "median_ffs": 0.0, + "ecr": 0.9887266730707228, + "ssr_curve": [ + 0.6111111111111112, + 0.2474747474747475, + 0.22727272727272727, + 0.22448979591836735, + 0.225130890052356, + 0.22459893048128343, + 0.2154696132596685, + 0.21787709497206703, + 0.22033898305084745, + 0.21839080459770116, + 0.22093023255813954, + 0.22023809523809523, + 0.21341463414634146, + 0.2085889570552147, + 0.20496894409937888, + 0.20253164556962025, + 0.20512820512820512, + 0.20394736842105263, + 0.2052980132450331, + 0.2054794520547945, + 0.2158273381294964, + 0.23076923076923078, + 0.21875, + 0.21138211382113822, + 0.18421052631578946, + 0.1834862385321101, + 0.18095238095238095, + 0.17647058823529413, + 0.16853932584269662, + 0.1728395061728395 + ], + "error_type_dist": { + "conceptual": 0.7214397496087637, + "executional": 0.01006036217303823, + "logical": 0.266934942991281, + "methodological": 0.001564945226917058 + } +} \ No newline at end of file diff --git a/results/metrics/r1-qwen-7b_awq_w4_gsm8k_run0_metrics.json b/results/metrics/r1-qwen-7b_awq_w4_gsm8k_run0_metrics.json new file mode 100644 index 0000000000000000000000000000000000000000..61a20fc47a2717ac4e2ceecfec5262f36b8251df --- /dev/null +++ b/results/metrics/r1-qwen-7b_awq_w4_gsm8k_run0_metrics.json @@ -0,0 +1,47 @@ +{ + "model": "r1-qwen-7b", + "quantization": "awq_w4", + "accuracy": 0.7573919636087946, + "accuracy_delta": 0.0, + "avg_ffs": 0.88125, + "median_ffs": 0.0, + "ecr": 0.7271533697750157, + "ssr_curve": [ + 0.865049279757392, + 0.8067226890756303, + 0.7849802371541502, + 0.7773109243697479, + 0.77734375, + 0.7598097502972652, + 0.7495881383855024, + 0.7112299465240641, + 0.7, + 0.6504065040650406, + 0.6351351351351351, + 0.6, + 0.5681818181818182, + 0.6, + 0.5714285714285714, + 0.5333333333333333, + 0.5172413793103449, + 0.52, + 0.47619047619047616, + 0.45, + 0.35294117647058826, + 0.375, + 0.375, + 0.375, + 0.4, + 0.35714285714285715, + 0.3076923076923077, + 0.3076923076923077, + 0.25, + 0.3 + ], + "error_type_dist": { + "conceptual": 0.7411556603773585, + "executional": 0.04304245283018868, + "logical": 0.21521226415094338, + "methodological": 0.0005896226415094339 + } +} \ No newline at end of file diff --git a/results/metrics/r1-qwen-7b_awq_w4_gsm8k_run0_sig.json b/results/metrics/r1-qwen-7b_awq_w4_gsm8k_run0_sig.json new file mode 100644 index 0000000000000000000000000000000000000000..39883be142a288ec8205f9764677cdfa1b5a08e6 --- /dev/null +++ b/results/metrics/r1-qwen-7b_awq_w4_gsm8k_run0_sig.json @@ -0,0 +1,7 @@ +{ + "n_pairs": 1319, + "delta_acc_observed": 0.011372251705837777, + "delta_acc_ci_lo": -0.012888551933282755, + "delta_acc_ci_hi": 0.0363912054586808, + "p_value": 0.3508 +} \ No newline at end of file diff --git a/results/metrics/r1-qwen-7b_awq_w4_math500_run0_metrics.json b/results/metrics/r1-qwen-7b_awq_w4_math500_run0_metrics.json new file mode 100644 index 0000000000000000000000000000000000000000..c9d88b2d071c7baa3d6c82c6fb506e5c13a6af90 --- /dev/null +++ b/results/metrics/r1-qwen-7b_awq_w4_math500_run0_metrics.json @@ -0,0 +1,47 @@ +{ + "model": "r1-qwen-7b", + "quantization": "awq_w4", + "accuracy": 0.72, + "accuracy_delta": 0.0, + "avg_ffs": 0.65, + "median_ffs": 0.5, + "ecr": 0.9584342785463587, + "ssr_curve": [ + 0.86, + 0.75, + 0.7254509018036072, + 0.7248995983935743, + 0.7154471544715447, + 0.7002141327623126, + 0.6872146118721462, + 0.6739130434782609, + 0.654639175257732, + 0.6432432432432432, + 0.6278409090909091, + 0.6142433234421365, + 0.5911949685534591, + 0.5784313725490197, + 0.565068493150685, + 0.5467625899280576, + 0.5318352059925093, + 0.508, + 0.48739495798319327, + 0.48, + 0.4669811320754717, + 0.4461538461538462, + 0.42857142857142855, + 0.40119760479041916, + 0.4025974025974026, + 0.3732394366197183, + 0.3230769230769231, + 0.319672131147541, + 0.3275862068965517, + 0.3181818181818182 + ], + "error_type_dist": { + "conceptual": 0.09387853859182192, + "methodological": 0.5381079119283814, + "executional": 0.05976288410355674, + "logical": 0.30825066537624 + } +} \ No newline at end of file diff --git a/results/metrics/r1-qwen-7b_awq_w4_math500_run0_sig.json b/results/metrics/r1-qwen-7b_awq_w4_math500_run0_sig.json new file mode 100644 index 0000000000000000000000000000000000000000..e8bd206fb055dd24b963f93f33166fa1e3c5be45 --- /dev/null +++ b/results/metrics/r1-qwen-7b_awq_w4_math500_run0_sig.json @@ -0,0 +1,7 @@ +{ + "n_pairs": 500, + "delta_acc_observed": 0.0020000000000000018, + "delta_acc_ci_lo": -0.03400000000000003, + "delta_acc_ci_hi": 0.038000000000000034, + "p_value": 0.9884 +} \ No newline at end of file diff --git a/results/metrics/r1-qwen-7b_awq_w4_restored_gpqa_run0_ci.json b/results/metrics/r1-qwen-7b_awq_w4_restored_gpqa_run0_ci.json new file mode 100644 index 0000000000000000000000000000000000000000..1d8608baacb18b952eeca13bd7c6e61d883e71e4 --- /dev/null +++ b/results/metrics/r1-qwen-7b_awq_w4_restored_gpqa_run0_ci.json @@ -0,0 +1,116 @@ +{ + "n": 198, + "accuracy": { + "mean": 0.23806767676767676, + "ci_lo": 0.17676767676767677, + "ci_hi": 0.29797979797979796 + }, + "avg_ffs": { + "mean": 0.515817385128816, + "ci_lo": 0.41875, + "ci_hi": 0.6126817355747386 + }, + "ecr": { + "mean": 0.9875228572833148, + "ci_lo": 0.9809477915665348, + "ci_hi": 0.9928666431082718 + }, + "ssr_curve_ci": { + "mean": [ + 0.5909020202020184, + 0.27826464646464655, + 0.23665578128449943, + 0.2306470822069668, + 0.22892051687260578, + 0.22793171900167816, + 0.23054513865198384, + 0.22676791303964236, + 0.22133432034951453, + 0.22413174185391255, + 0.20942411249589604, + 0.21344016981563874, + 0.2162118672203188, + 0.2162118672203188, + 0.21810906052313353, + 0.21035228655063776, + 0.20488609115117223, + 0.20928421404113004, + 0.21393597579664014, + 0.21008297992809916, + 0.21171469909689217, + 0.20238131363720974, + 0.20406467292769864, + 0.2092503494911014, + 0.2078784205653403, + 0.22844042745272267, + 0.23796721271919596, + 0.23249094422093008, + 0.24272270832023454, + 0.23593392583579867 + ], + "ci_lo": [ + 0.5202020202020202, + 0.21717171717171718, + 0.1785487672226251, + 0.17203881048387099, + 0.16939890710382513, + 0.16756756756756758, + 0.1694915254237288, + 0.16463221673355974, + 0.1592356687898089, + 0.16025641025641027, + 0.1464968152866242, + 0.149652133580705, + 0.1518891446106636, + 0.1518891446106636, + 0.15172413793103448, + 0.14582988410596026, + 0.14084507042253522, + 0.14388489208633093, + 0.147281121751026, + 0.14285714285714285, + 0.144, + 0.1333203125, + 0.13333333333333333, + 0.13636363636363635, + 0.13392857142857142, + 0.14814814814814814, + 0.1553398058252427, + 0.14942528735632185, + 0.15555555555555556, + 0.14606741573033707 + ], + "ci_hi": [ + 0.6565656565656566, + 0.3383838383838384, + 0.29743589743589743, + 0.2925531914893617, + 0.2918918918918919, + 0.29214140219462503, + 0.2954606549364613, + 0.2924047924689772, + 0.28750204248366007, + 0.2909090909090909, + 0.27672955974842767, + 0.281437125748503, + 0.2848101265822785, + 0.2848101265822785, + 0.28846153846153844, + 0.27972727272727266, + 0.27345087756849296, + 0.27972727272727266, + 0.2846715328467153, + 0.28360460627894973, + 0.2857142857142857, + 0.27643183797909393, + 0.28, + 0.2881394105970377, + 0.288135593220339, + 0.31521739130434784, + 0.32632116004296446, + 0.32222222222222224, + 0.3333333333333333, + 0.3333333333333333 + ] + } +} \ No newline at end of file diff --git a/results/metrics/r1-qwen-7b_awq_w4_restored_gpqa_run0_metrics.json b/results/metrics/r1-qwen-7b_awq_w4_restored_gpqa_run0_metrics.json new file mode 100644 index 0000000000000000000000000000000000000000..482c847e2554514f20ae71d64fdf1e55f81f7486 --- /dev/null +++ b/results/metrics/r1-qwen-7b_awq_w4_restored_gpqa_run0_metrics.json @@ -0,0 +1,47 @@ +{ + "model": "r1-qwen-7b", + "quantization": "awq_w4_restored", + "accuracy": 0.23737373737373738, + "accuracy_delta": 0.0, + "avg_ffs": 0.5165562913907285, + "median_ffs": 0.0, + "ecr": 0.9875674595122634, + "ssr_curve": [ + 0.5909090909090909, + 0.2777777777777778, + 0.2358974358974359, + 0.22994652406417113, + 0.22826086956521738, + 0.22727272727272727, + 0.22988505747126436, + 0.2261904761904762, + 0.22085889570552147, + 0.2236024844720497, + 0.2088607594936709, + 0.2129032258064516, + 0.21568627450980393, + 0.21568627450980393, + 0.21768707482993196, + 0.2097902097902098, + 0.20422535211267606, + 0.20863309352517986, + 0.21323529411764705, + 0.20930232558139536, + 0.2109375, + 0.20161290322580644, + 0.2033898305084746, + 0.20869565217391303, + 0.2072072072072072, + 0.22772277227722773, + 0.23711340206185566, + 0.23157894736842105, + 0.24175824175824176, + 0.2345679012345679 + ], + "error_type_dist": { + "conceptual": 0.7197353914002205, + "executional": 0.003969128996692392, + "logical": 0.2747519294377067, + "methodological": 0.0015435501653803748 + } +} \ No newline at end of file diff --git a/results/metrics/r1-qwen-7b_awq_w4_restored_gsm8k_run0_metrics.json b/results/metrics/r1-qwen-7b_awq_w4_restored_gsm8k_run0_metrics.json new file mode 100644 index 0000000000000000000000000000000000000000..95e7322d16c25eae08ad2629951c6417c559cd43 --- /dev/null +++ b/results/metrics/r1-qwen-7b_awq_w4_restored_gsm8k_run0_metrics.json @@ -0,0 +1,47 @@ +{ + "model": "r1-qwen-7b", + "quantization": "awq_w4_restored", + "accuracy": 0.7687642153146323, + "accuracy_delta": 0.0, + "avg_ffs": 0.46229508196721314, + "median_ffs": 0.0, + "ecr": 0.8130970083200567, + "ssr_curve": [ + 0.8369977255496588, + 0.795758051846033, + 0.7944397641112048, + 0.8106343283582089, + 0.8085339168490153, + 0.8013071895424837, + 0.7922606924643585, + 0.7889273356401384, + 0.782608695652174, + 0.7875, + 0.7619047619047619, + 0.7666666666666667, + 0.7727272727272727, + 0.7777777777777778, + 0.75, + 0.7333333333333333, + 0.7333333333333333, + 0.75, + 0.75, + 0.7272727272727273, + 0.625, + 0.5714285714285714, + 0.5, + 0.5, + 0.5, + 0.5, + 0.4, + 0.5, + 0.5, + 0.5 + ], + "error_type_dist": { + "conceptual": 0.7603686635944701, + "executional": 0.02534562211981567, + "logical": 0.21274961597542244, + "methodological": 0.0015360983102918587 + } +} \ No newline at end of file diff --git a/results/metrics/r1-qwen-7b_awq_w4_restored_math500_run0_ci.json b/results/metrics/r1-qwen-7b_awq_w4_restored_math500_run0_ci.json new file mode 100644 index 0000000000000000000000000000000000000000..18c61c9474255652ba9e98b63c1b660af667f9e3 --- /dev/null +++ b/results/metrics/r1-qwen-7b_awq_w4_restored_math500_run0_ci.json @@ -0,0 +1,116 @@ +{ + "n": 500, + "accuracy": { + "mean": 0.7222088, + "ci_lo": 0.682, + "ci_hi": 0.762 + }, + "avg_ffs": { + "mean": 0.6840308293686401, + "ci_lo": 0.5364061021759697, + "ci_hi": 0.8518518518518519 + }, + "ecr": { + "mean": 0.9517845050512754, + "ci_lo": 0.9343111670373138, + "ci_hi": 0.967953716904225 + }, + "ssr_curve_ci": { + "mean": [ + 0.8601628000000057, + 0.747214960877652, + 0.7426067686974549, + 0.737174710258403, + 0.722173917992703, + 0.7064297880486186, + 0.6993028504148582, + 0.6761869509138455, + 0.6610785306068966, + 0.6423823934930181, + 0.6314657172203342, + 0.6032223957918844, + 0.5864899020407732, + 0.5637113297456453, + 0.5489627436738554, + 0.5316757905397895, + 0.5127956093338204, + 0.5064781995967823, + 0.5019003090037112, + 0.4737329351372889, + 0.43459030105040886, + 0.41648577185016195, + 0.39932557292620574, + 0.3984777943343872, + 0.40158298677641313, + 0.401868775024193, + 0.38913614237987826, + 0.3848816194384001, + 0.38034091897888944, + 0.3657096373813009 + ], + "ci_lo": [ + 0.828, + 0.7088353413654619, + 0.7032520325203252, + 0.6969072164948453, + 0.6804325523523335, + 0.6635801006362876, + 0.6543735733666395, + 0.6286472148541115, + 0.6115702479338843, + 0.5911760643483913, + 0.5786163522012578, + 0.5477690126263195, + 0.52856784519828, + 0.5037030228758169, + 0.48738720580483114, + 0.46863430035973996, + 0.4481322907695921, + 0.44075628564543334, + 0.4349945652173913, + 0.4043659220249724, + 0.36416184971098264, + 0.3428531300160514, + 0.3235204447632712, + 0.31976645309286506, + 0.31858243198951164, + 0.3125, + 0.29896907216494845, + 0.2908863636363636, + 0.28125, + 0.2615340909090909 + ], + "ci_hi": [ + 0.89, + 0.7847082494969819, + 0.7811908546732569, + 0.7761506276150628, + 0.7629310344827587, + 0.7488789237668162, + 0.7433430011438791, + 0.723582434034059, + 0.7100271002710027, + 0.6941176470588235, + 0.684716753522486, + 0.6601957306447598, + 0.6450539469338324, + 0.625, + 0.6127819548872181, + 0.5982152604491578, + 0.5809555984555984, + 0.5754716981132075, + 0.5714285714285714, + 0.544973544973545, + 0.5082013004853924, + 0.49122807017543857, + 0.4768211920529801, + 0.48091603053435117, + 0.4878048780487805, + 0.49107142857142855, + 0.4818181818181818, + 0.4789959733893557, + 0.48148148148148145, + 0.4731324278438029 + ] + } +} \ No newline at end of file diff --git a/results/metrics/r1-qwen-7b_awq_w4_restored_math500_run0_metrics.json b/results/metrics/r1-qwen-7b_awq_w4_restored_math500_run0_metrics.json new file mode 100644 index 0000000000000000000000000000000000000000..99ab9b6be7b092c97a9827e46ac69f5375af0db8 --- /dev/null +++ b/results/metrics/r1-qwen-7b_awq_w4_restored_math500_run0_metrics.json @@ -0,0 +1,47 @@ +{ + "model": "r1-qwen-7b", + "quantization": "awq_w4_restored", + "accuracy": 0.722, + "accuracy_delta": 0.0, + "avg_ffs": 0.6834532374100719, + "median_ffs": 0.0, + "ecr": 0.951738439908583, + "ssr_curve": [ + 0.86, + 0.7469879518072289, + 0.7423312883435583, + 0.7369519832985386, + 0.7219827586206896, + 0.7062780269058296, + 0.6992840095465394, + 0.6762402088772846, + 0.6611570247933884, + 0.6424418604651163, + 0.631578947368421, + 0.6033333333333334, + 0.5865724381625441, + 0.5639097744360902, + 0.549407114624506, + 0.5319148936170213, + 0.5132743362831859, + 0.5069124423963134, + 0.5023923444976076, + 0.4742268041237113, + 0.4350282485875706, + 0.4171779141104294, + 0.4, + 0.39855072463768115, + 0.4015748031496063, + 0.4017094017094017, + 0.3888888888888889, + 0.38461538461538464, + 0.3804347826086957, + 0.36585365853658536 + ], + "error_type_dist": { + "conceptual": 0.08236626096801584, + "methodological": 0.5771299179167846, + "executional": 0.0993489951882253, + "logical": 0.24115482592697424 + } +} \ No newline at end of file diff --git a/results/metrics/r1-qwen-7b_bnb_nf4_w4_gsm8k_run0_ci.json b/results/metrics/r1-qwen-7b_bnb_nf4_w4_gsm8k_run0_ci.json new file mode 100644 index 0000000000000000000000000000000000000000..3adbba682effc45b6246ee4f5d683c281686cf11 --- /dev/null +++ b/results/metrics/r1-qwen-7b_bnb_nf4_w4_gsm8k_run0_ci.json @@ -0,0 +1,116 @@ +{ + "n": 1319, + "accuracy": { + "mean": 0.78089613343442, + "ci_lo": 0.7581501137225171, + "ci_hi": 0.8028809704321456 + }, + "avg_ffs": { + "mean": 1.018229661083791, + "ci_lo": 0.8285691169671753, + "ci_hi": 1.2200784342293776 + }, + "ecr": { + "mean": 0.7769183478502724, + "ci_lo": 0.7389889374411535, + "ci_hi": 0.8149680941797441 + }, + "ssr_curve_ci": { + "mean": [ + 0.8770336618650505, + 0.8336123891856734, + 0.8083589822008282, + 0.7990670023475851, + 0.8006342087763959, + 0.7850475177450101, + 0.778243716421561, + 0.7632724146102728, + 0.765512361900783, + 0.7267793150634787, + 0.7003669309975097, + 0.6919433458237457, + 0.6508032856516415, + 0.6244726500093991, + 0.6060367140307433, + 0.5765808685084325, + 0.576960313283127, + 0.4881523464316931, + 0.46247998537320584, + 0.4727454882423706, + 0.39356640809245297, + 0.3995813275188912, + 0.41323385182928607, + 0.4069161290507015, + 0.38386314693085766, + 0.37380899873159695, + 0.4073956568736798, + 0.3988151540445296, + 0.4195495043804598, + 0.44271481746543395 + ], + "ci_lo": [ + 0.8589840788476119, + 0.8129735273706984, + 0.7866651090342679, + 0.7767692055469363, + 0.7764372790282494, + 0.7575380149405088, + 0.7455138662316476, + 0.7216678307967551, + 0.7136276223776223, + 0.6613756613756614, + 0.6220422814795209, + 0.60503693412849, + 0.5541912330171751, + 0.5211122901794327, + 0.5, + 0.45588235294117646, + 0.4489619985925405, + 0.3469387755102041, + 0.31032430213464696, + 0.3142857142857143, + 0.23071428571428573, + 0.22727272727272727, + 0.23684210526315788, + 0.2222222222222222, + 0.2, + 0.18181818181818182, + 0.2, + 0.18181818181818182, + 0.19047619047619047, + 0.2 + ], + "ci_hi": [ + 0.8946171341925702, + 0.8535469107551488, + 0.8296875, + 0.821546052631579, + 0.8246268656716418, + 0.8120730840378529, + 0.8107667837871753, + 0.8035714285714286, + 0.8149606299212598, + 0.7897499999999998, + 0.7769244604316546, + 0.7758655172413793, + 0.7431223654847507, + 0.726031474233529, + 0.7160493827160493, + 0.6949152542372882, + 0.7021276595744681, + 0.627906976744186, + 0.615425531914893, + 0.631578947368421, + 0.5666666666666667, + 0.5789473684210527, + 0.5937711148648646, + 0.6, + 0.5789473684210527, + 0.5714285714285714, + 0.6190476190476191, + 0.6190476190476191, + 0.6428571428571429, + 0.6666666666666666 + ] + } +} \ No newline at end of file diff --git a/results/metrics/r1-qwen-7b_bnb_nf4_w4_gsm8k_run0_metrics.json b/results/metrics/r1-qwen-7b_bnb_nf4_w4_gsm8k_run0_metrics.json new file mode 100644 index 0000000000000000000000000000000000000000..8e39cafeee3760624655024aaefaaeb20cf60b74 --- /dev/null +++ b/results/metrics/r1-qwen-7b_bnb_nf4_w4_gsm8k_run0_metrics.json @@ -0,0 +1,47 @@ +{ + "model": "r1-qwen-7b", + "quantization": "bnb_nf4_w4", + "accuracy": 0.7808946171341926, + "accuracy_delta": 0.0, + "avg_ffs": 1.0207612456747406, + "median_ffs": 0.0, + "ecr": 0.7767809796849329, + "ssr_curve": [ + 0.8771796815769523, + 0.8337147215865751, + 0.8084112149532711, + 0.7991836734693878, + 0.8007554296506137, + 0.7852272727272728, + 0.7783171521035599, + 0.7635467980295566, + 0.7657992565055762, + 0.7272727272727273, + 0.7014925373134329, + 0.6936936936936937, + 0.6526315789473685, + 0.6265060240963856, + 0.6081081081081081, + 0.578125, + 0.5789473684210527, + 0.48936170212765956, + 0.4634146341463415, + 0.47368421052631576, + 0.3939393939393939, + 0.4, + 0.41379310344827586, + 0.4074074074074074, + 0.38461538461538464, + 0.375, + 0.4090909090909091, + 0.4, + 0.42105263157894735, + 0.4444444444444444 + ], + "error_type_dist": { + "conceptual": 0.6953992596509783, + "executional": 0.03807509254362771, + "logical": 0.2654680063458488, + "methodological": 0.0010576414595452142 + } +} \ No newline at end of file diff --git a/results/metrics/r1-qwen-7b_bnb_nf4_w4_gsm8k_run0_sig.json b/results/metrics/r1-qwen-7b_bnb_nf4_w4_gsm8k_run0_sig.json new file mode 100644 index 0000000000000000000000000000000000000000..89e0163aeca27ef479c91db7a26cb1186b4f3573 --- /dev/null +++ b/results/metrics/r1-qwen-7b_bnb_nf4_w4_gsm8k_run0_sig.json @@ -0,0 +1,7 @@ +{ + "n_pairs": 1319, + "delta_acc_observed": -0.05307050796057622, + "delta_acc_ci_lo": -0.07884761182714173, + "delta_acc_ci_hi": -0.028809704321455687, + "p_value": 0.0 +} \ No newline at end of file diff --git a/results/metrics/r1-qwen-7b_bnb_nf4_w4_math500_run0_ci.json b/results/metrics/r1-qwen-7b_bnb_nf4_w4_math500_run0_ci.json new file mode 100644 index 0000000000000000000000000000000000000000..b14960f0a09b1efd4c21169595dbf76f307ddc71 --- /dev/null +++ b/results/metrics/r1-qwen-7b_bnb_nf4_w4_math500_run0_ci.json @@ -0,0 +1,116 @@ +{ + "n": 500, + "accuracy": { + "mean": 0.752178, + "ci_lo": 0.714, + "ci_hi": 0.79 + }, + "avg_ffs": { + "mean": 0.7090075260570627, + "ci_lo": 0.5507162116616111, + "ci_hi": 0.8833430232558138 + }, + "ecr": { + "mean": 0.9377221465035007, + "ci_lo": 0.9115887775322801, + "ci_hi": 0.9590061747082081 + }, + "ssr_curve_ci": { + "mean": [ + 0.8761724000000015, + 0.7841667999999943, + 0.7621794030708953, + 0.7542212121849576, + 0.7470495040065157, + 0.7386095779975843, + 0.7301834821161122, + 0.7192549931393029, + 0.6999623391016064, + 0.6786748108615326, + 0.6581047457811889, + 0.6441952749021422, + 0.6258537031539894, + 0.6161921339650808, + 0.5975472231712137, + 0.5897578221138375, + 0.5650461871306558, + 0.5391275404578226, + 0.49548086571491456, + 0.4899941211005184, + 0.4891209018973703, + 0.47257257859384183, + 0.4393494634219672, + 0.418639425203228, + 0.41476077459019517, + 0.4006970815722064, + 0.373526818498084, + 0.35442863582015666, + 0.3378110335599332, + 0.32527935060247426 + ], + "ci_lo": [ + 0.848, + 0.748, + 0.7243460764587525, + 0.7163265306122449, + 0.7080745341614907, + 0.6986899563318777, + 0.6889362694956184, + 0.6769596199524941, + 0.6553479787521383, + 0.6321525885558583, + 0.6083055685702883, + 0.5914634146341463, + 0.5723900186071238, + 0.5618692968330649, + 0.539568345323741, + 0.5298245614035088, + 0.5038076036866359, + 0.4729713087471708, + 0.42702702702702705, + 0.4191574950657344, + 0.4161490683229814, + 0.3931034482758621, + 0.35848352576738945, + 0.3363579067990833, + 0.33070866141732286, + 0.31531286721504115, + 0.2845510933557611, + 0.2616717184474194, + 0.23880172998643148, + 0.2222222222222222 + ], + "ci_hi": [ + 0.904, + 0.82, + 0.7987927565392354, + 0.7918367346938775, + 0.7848393810534241, + 0.7777777777777778, + 0.7711670480549199, + 0.7612293144208038, + 0.7449507575757576, + 0.727019864160434, + 0.7093048101769032, + 0.6974541951718172, + 0.6802721088435374, + 0.6714298607529654, + 0.6550390973178836, + 0.6494512080157298, + 0.6289644607843137, + 0.6055078175474867, + 0.5659340659340659, + 0.5618640047186746, + 0.5645208728652751, + 0.5517241379310345, + 0.5217391304347826, + 0.5040033057851239, + 0.5, + 0.49074494949494946, + 0.464923137050547, + 0.4486076673660117, + 0.4367816091954023, + 0.43037974683544306 + ] + } +} \ No newline at end of file diff --git a/results/metrics/r1-qwen-7b_bnb_nf4_w4_math500_run0_metrics.json b/results/metrics/r1-qwen-7b_bnb_nf4_w4_math500_run0_metrics.json new file mode 100644 index 0000000000000000000000000000000000000000..c38b78ff1367abb17c20908c4aac5aef9cac733c --- /dev/null +++ b/results/metrics/r1-qwen-7b_bnb_nf4_w4_math500_run0_metrics.json @@ -0,0 +1,47 @@ +{ + "model": "r1-qwen-7b", + "quantization": "bnb_nf4_w4", + "accuracy": 0.752, + "accuracy_delta": 0.0, + "avg_ffs": 0.7096774193548387, + "median_ffs": 0.5, + "ecr": 0.9376135303258154, + "ssr_curve": [ + 0.876, + 0.784, + 0.7620967741935484, + 0.7540983606557377, + 0.7468879668049793, + 0.7384615384615385, + 0.7299771167048055, + 0.719047619047619, + 0.6997455470737913, + 0.6784741144414169, + 0.6578171091445427, + 0.6439628482972136, + 0.6254071661237784, + 0.6156462585034014, + 0.5971223021582733, + 0.5893536121673004, + 0.5648535564853556, + 0.5388127853881278, + 0.49489795918367346, + 0.48936170212765956, + 0.4883720930232558, + 0.4716981132075472, + 0.4383561643835616, + 0.417910447761194, + 0.4140625, + 0.4, + 0.37272727272727274, + 0.35353535353535354, + 0.3372093023255814, + 0.3246753246753247 + ], + "error_type_dist": { + "conceptual": 0.08772987639433223, + "methodological": 0.5942116370214049, + "executional": 0.08230328610189931, + "logical": 0.2357552004823636 + } +} \ No newline at end of file diff --git a/results/metrics/r1-qwen-7b_bnb_nf4_w4_math500_run0_sig.json b/results/metrics/r1-qwen-7b_bnb_nf4_w4_math500_run0_sig.json new file mode 100644 index 0000000000000000000000000000000000000000..f92f075346799a2ec63b29bc4e4b87c60a07f52d --- /dev/null +++ b/results/metrics/r1-qwen-7b_bnb_nf4_w4_math500_run0_sig.json @@ -0,0 +1,7 @@ +{ + "n_pairs": 500, + "delta_acc_observed": -0.020000000000000018, + "delta_acc_ci_lo": -0.05400000000000005, + "delta_acc_ci_hi": 0.014000000000000012, + "p_value": 0.2684 +} \ No newline at end of file diff --git a/results/metrics/r1-qwen-7b_bnb_nf4_w4_restored_gpqa_run0_ci.json b/results/metrics/r1-qwen-7b_bnb_nf4_w4_restored_gpqa_run0_ci.json new file mode 100644 index 0000000000000000000000000000000000000000..67e3ddd0dfd3884425f089c989f0745653e2ae0a --- /dev/null +++ b/results/metrics/r1-qwen-7b_bnb_nf4_w4_restored_gpqa_run0_ci.json @@ -0,0 +1,116 @@ +{ + "n": 198, + "accuracy": { + "mean": 0.21143737373737376, + "ci_lo": 0.15656565656565657, + "ci_hi": 0.2676767676767677 + }, + "avg_ffs": { + "mean": 0.5124720807302965, + "ci_lo": 0.40789473684210525, + "ci_hi": 0.6291390728476821 + }, + "ecr": { + "mean": 0.9905354987381726, + "ci_lo": 0.9858085927260353, + "ci_hi": 0.994505102866936 + }, + "ssr_curve_ci": { + "mean": [ + 0.5701353535353454, + 0.2368949494949501, + 0.216248712727504, + 0.2143500695798783, + 0.20803831649189983, + 0.20594692956096974, + 0.20860162566953494, + 0.20643075780420844, + 0.2101721786130087, + 0.20797492708443227, + 0.20551878624564301, + 0.2017781627791015, + 0.19282457001035463, + 0.18885661577563762, + 0.1913499928039663, + 0.1938391373117962, + 0.1950317889621688, + 0.19923853698138183, + 0.20514658476115816, + 0.21144531688531848, + 0.1961749380150442, + 0.20089850616400085, + 0.2075665899070856, + 0.19752783206722815, + 0.19228842164619409, + 0.19025954355041907, + 0.19917365726753353, + 0.20097984588769877, + 0.2050822804023988, + 0.19668290708719657 + ], + "ci_lo": [ + 0.5050505050505051, + 0.18181818181818182, + 0.15873015873015872, + 0.15675675675675677, + 0.15053763440860216, + 0.14752892960462874, + 0.14772727272727273, + 0.14545190274841438, + 0.14792899408284024, + 0.14634146341463414, + 0.14455862666351682, + 0.14110232814220544, + 0.1317365269461078, + 0.12751677852348994, + 0.1292517006802721, + 0.1310344827586207, + 0.1310344827586207, + 0.13380281690140844, + 0.13768115942028986, + 0.14283181357649444, + 0.12781954887218044, + 0.13114754098360656, + 0.13445378151260504, + 0.12605042016806722, + 0.12149532710280374, + 0.11818181818181818, + 0.125, + 0.12380952380952381, + 0.12631578947368421, + 0.11494252873563218 + ], + "ci_hi": [ + 0.6414141414141414, + 0.29797979797979796, + 0.2751322751322751, + 0.27419354838709675, + 0.2670538101604277, + 0.26667613636363624, + 0.27071823204419887, + 0.2694610778443114, + 0.27388535031847133, + 0.2710843373493976, + 0.26875, + 0.26416968636761184, + 0.25609756097560976, + 0.25157232704402516, + 0.25477707006369427, + 0.2582781456953642, + 0.2605633802816901, + 0.2656277316433566, + 0.2733812949640288, + 0.28125578703703696, + 0.2672413793103448, + 0.27352150537634384, + 0.2820610687022899, + 0.27274502840909065, + 0.2672413793103448, + 0.26605504587155965, + 0.27722772277227725, + 0.2828282828282828, + 0.28865979381443296, + 0.2826086956521739 + ] + } +} \ No newline at end of file diff --git a/results/metrics/r1-qwen-7b_bnb_nf4_w4_restored_gsm8k_run0_metrics.json b/results/metrics/r1-qwen-7b_bnb_nf4_w4_restored_gsm8k_run0_metrics.json new file mode 100644 index 0000000000000000000000000000000000000000..c9deef66cb9ee6c1fec9cb65173037b0d6758e2a --- /dev/null +++ b/results/metrics/r1-qwen-7b_bnb_nf4_w4_restored_gsm8k_run0_metrics.json @@ -0,0 +1,46 @@ +{ + "model": "r1-qwen-7b", + "quantization": "bnb_nf4_w4_restored", + "accuracy": 0.7278241091736164, + "accuracy_delta": 0.0, + "avg_ffs": 0.4206128133704735, + "median_ffs": 0.0, + "ecr": 0.822003749658252, + "ssr_curve": [ + 0.8066717210007581, + 0.7577741407528642, + 0.7562097516099356, + 0.7932584269662921, + 0.7969401947148818, + 0.7873754152823921, + 0.7688442211055276, + 0.7543103448275862, + 0.7388059701492538, + 0.7083333333333334, + 0.7, + 0.6129032258064516, + 0.5769230769230769, + 0.5769230769230769, + 0.5769230769230769, + 0.5454545454545454, + 0.5263157894736842, + 0.5, + 0.4375, + 0.4, + 0.4, + 0.42857142857142855, + 0.38461538461538464, + 0.38461538461538464, + 0.38461538461538464, + 0.38461538461538464, + 0.38461538461538464, + 0.38461538461538464, + 0.38461538461538464, + 0.3333333333333333 + ], + "error_type_dist": { + "conceptual": 0.6964047936085219, + "executional": 0.041278295605858856, + "logical": 0.2623169107856192 + } +} \ No newline at end of file diff --git a/results/metrics/r1-qwen-7b_bnb_nf4_w4_restored_math500_run0_ci.json b/results/metrics/r1-qwen-7b_bnb_nf4_w4_restored_math500_run0_ci.json new file mode 100644 index 0000000000000000000000000000000000000000..157c4735ad66129e068c4f4f7033f2672eedf986 --- /dev/null +++ b/results/metrics/r1-qwen-7b_bnb_nf4_w4_restored_math500_run0_ci.json @@ -0,0 +1,116 @@ +{ + "n": 500, + "accuracy": { + "mean": 0.7318868000000001, + "ci_lo": 0.692, + "ci_hi": 0.772 + }, + "avg_ffs": { + "mean": 0.6877363717054457, + "ci_lo": 0.526296580868229, + "ci_hi": 0.8651190476190471 + }, + "ecr": { + "mean": 0.9377102371430005, + "ci_lo": 0.9113965292223641, + "ci_hi": 0.9597171063242401 + }, + "ssr_curve_ci": { + "mean": [ + 0.8638348000000009, + 0.7597361650529383, + 0.757296993181495, + 0.7532561428817206, + 0.7443951701305589, + 0.7389674795363513, + 0.7303852354584914, + 0.7195662721284821, + 0.7005875496026258, + 0.6897568937975481, + 0.6714866548839157, + 0.6517327608935684, + 0.6326893674742878, + 0.615550341036012, + 0.5923443616602126, + 0.5722668850736228, + 0.5378983560424434, + 0.5299822749734359, + 0.516001803964798, + 0.5054358957090324, + 0.46699452265839325, + 0.4589659395466884, + 0.4381467994588024, + 0.43883850139966696, + 0.42685655787437254, + 0.4086688609035219, + 0.4315919046123687, + 0.44021810942052536, + 0.44103119814001884, + 0.4323132664822112 + ], + "ci_lo": [ + 0.832, + 0.7209775967413442, + 0.7178947368421053, + 0.7133265524625269, + 0.7022727272727273, + 0.696551724137931, + 0.6871846453680234, + 0.6741293532338308, + 0.6530026419184392, + 0.6404469741770156, + 0.6190476190476191, + 0.5978154997176737, + 0.5750892857142856, + 0.556420233463035, + 0.5308578987150416, + 0.5084710451977401, + 0.4714285714285714, + 0.45933014354066987, + 0.44385026737967914, + 0.4327471322537112, + 0.39226248510453904, + 0.38271604938271603, + 0.35810602268935604, + 0.35664335664335667, + 0.3423408307636496, + 0.32037864077669903, + 0.3364485981308411, + 0.34285714285714286, + 0.3404255319148936, + 0.32608695652173914 + ], + "ci_hi": [ + 0.894, + 0.7991886409736308, + 0.7958333333333333, + 0.7943122614074474, + 0.7857142857142857, + 0.7808895957315821, + 0.7734939759036145, + 0.7647058823529411, + 0.7479457623343357, + 0.7390139751552794, + 0.7234073842302878, + 0.7058823529411765, + 0.6900009225092251, + 0.6751848385272173, + 0.6543802217976097, + 0.6367794372920583, + 0.6067961165048543, + 0.6, + 0.5889035087719297, + 0.5795454545454546, + 0.5448727544910179, + 0.5384615384615384, + 0.5211267605633803, + 0.5220588235294118, + 0.5138921339563862, + 0.49638888888888755, + 0.5267857142857143, + 0.5376369446135117, + 0.5416764018691587, + 0.5349044850498336 + ] + } +} \ No newline at end of file diff --git a/results/metrics/r1-qwen-7b_bnb_nf4_w4_restored_math500_run0_metrics.json b/results/metrics/r1-qwen-7b_bnb_nf4_w4_restored_math500_run0_metrics.json new file mode 100644 index 0000000000000000000000000000000000000000..1715f739e0774bb781d3d1b9ca7a1bd682a3bbda --- /dev/null +++ b/results/metrics/r1-qwen-7b_bnb_nf4_w4_restored_math500_run0_metrics.json @@ -0,0 +1,47 @@ +{ + "model": "r1-qwen-7b", + "quantization": "bnb_nf4_w4_restored", + "accuracy": 0.732, + "accuracy_delta": 0.0, + "avg_ffs": 0.6865671641791045, + "median_ffs": 0.0, + "ecr": 0.937771346435757, + "ssr_curve": [ + 0.864, + 0.7596741344195519, + 0.7573221757322176, + 0.7532467532467533, + 0.7443946188340808, + 0.7390300230946882, + 0.7305825242718447, + 0.7197943444730077, + 0.7008310249307479, + 0.6900584795321637, + 0.671875, + 0.6521739130434783, + 0.6330935251798561, + 0.6159695817490495, + 0.592741935483871, + 0.5726495726495726, + 0.5380952380952381, + 0.5303030303030303, + 0.5161290322580645, + 0.5054945054945055, + 0.46706586826347307, + 0.4591194968553459, + 0.4383561643835616, + 0.43884892086330934, + 0.4263565891472868, + 0.4083333333333333, + 0.43119266055045874, + 0.44, + 0.44086021505376344, + 0.4318181818181818 + ], + "error_type_dist": { + "conceptual": 0.08551593562638056, + "methodological": 0.5979804354686021, + "executional": 0.08961817608078258, + "logical": 0.22688545282423478 + } +} \ No newline at end of file diff --git a/results/metrics/r1-qwen-7b_gptq_w4_gpqa_run0_sig.json b/results/metrics/r1-qwen-7b_gptq_w4_gpqa_run0_sig.json new file mode 100644 index 0000000000000000000000000000000000000000..8da9faf159e7dc120b3f64eac4bac80545c99970 --- /dev/null +++ b/results/metrics/r1-qwen-7b_gptq_w4_gpqa_run0_sig.json @@ -0,0 +1,7 @@ +{ + "n_pairs": 198, + "delta_acc_observed": 0.005050505050505055, + "delta_acc_ci_lo": -0.06060606060606061, + "delta_acc_ci_hi": 0.07070707070707072, + "p_value": 0.9324 +} \ No newline at end of file diff --git a/results/metrics/r1-qwen-7b_gptq_w4_gsm8k_run0_metrics.json b/results/metrics/r1-qwen-7b_gptq_w4_gsm8k_run0_metrics.json new file mode 100644 index 0000000000000000000000000000000000000000..0281005e435ee8ce45ce7c510806baf89c13bc32 --- /dev/null +++ b/results/metrics/r1-qwen-7b_gptq_w4_gsm8k_run0_metrics.json @@ -0,0 +1,47 @@ +{ + "model": "r1-qwen-7b", + "quantization": "gptq_w4", + "accuracy": 0.7778620166793025, + "accuracy_delta": 0.0, + "avg_ffs": 0.8361774744027304, + "median_ffs": 0.0, + "ecr": 0.7826780353954256, + "ssr_curve": [ + 0.8749052312357847, + 0.8146453089244852, + 0.801452784503632, + 0.789103690685413, + 0.787847579814624, + 0.7726161369193154, + 0.7604690117252931, + 0.7142857142857143, + 0.6920289855072463, + 0.6538461538461539, + 0.5955882352941176, + 0.5691056910569106, + 0.5607476635514018, + 0.5612244897959183, + 0.5287356321839081, + 0.4935064935064935, + 0.47761194029850745, + 0.4696969696969697, + 0.4426229508196721, + 0.4406779661016949, + 0.42592592592592593, + 0.40384615384615385, + 0.39215686274509803, + 0.3541666666666667, + 0.3, + 0.2631578947368421, + 0.23529411764705882, + 0.21212121212121213, + 0.1111111111111111, + 0.1111111111111111 + ], + "error_type_dist": { + "conceptual": 0.7540856031128405, + "executional": 0.026070038910505838, + "logical": 0.21945525291828794, + "methodological": 0.0003891050583657588 + } +} \ No newline at end of file diff --git a/results/metrics/r1-qwen-7b_gptq_w4_gsm8k_run0_sig.json b/results/metrics/r1-qwen-7b_gptq_w4_gsm8k_run0_sig.json new file mode 100644 index 0000000000000000000000000000000000000000..ece35ab2407a9367df1b1dc1c052e50292b1ec47 --- /dev/null +++ b/results/metrics/r1-qwen-7b_gptq_w4_gsm8k_run0_sig.json @@ -0,0 +1,7 @@ +{ + "n_pairs": 1319, + "delta_acc_observed": -0.04245640636846104, + "delta_acc_ci_lo": -0.06671721000758146, + "delta_acc_ci_hi": -0.0181956027293404, + "p_value": 0.0004 +} \ No newline at end of file diff --git a/results/metrics/r1-qwen-7b_gptq_w4_math500_run0_ci.json b/results/metrics/r1-qwen-7b_gptq_w4_math500_run0_ci.json new file mode 100644 index 0000000000000000000000000000000000000000..92877580649722d13eb69d84ef113a7c7f7878a2 --- /dev/null +++ b/results/metrics/r1-qwen-7b_gptq_w4_math500_run0_ci.json @@ -0,0 +1,116 @@ +{ + "n": 500, + "accuracy": { + "mean": 0.7079028, + "ci_lo": 0.666, + "ci_hi": 0.748 + }, + "avg_ffs": { + "mean": 0.6578699173500949, + "ci_lo": 0.525304111679684, + "ci_hi": 0.7985074626865671 + }, + "ecr": { + "mean": 0.9670660112144062, + "ci_lo": 0.9563754956312882, + "ci_hi": 0.9763004323932382 + }, + "ssr_curve_ci": { + "mean": [ + 0.8559140000000026, + 0.7359647999999993, + 0.7196828180469941, + 0.7104443809411035, + 0.7029715086286611, + 0.6968827139188897, + 0.6892034877413867, + 0.673461840320216, + 0.6624581076850949, + 0.6513704275750692, + 0.6410072848870259, + 0.6290080981787373, + 0.6206027954728983, + 0.6030029531007869, + 0.5890062236349357, + 0.5711888855261541, + 0.5519206259297451, + 0.5298419509418649, + 0.5076336708142006, + 0.5017249096749363, + 0.4903201984532143, + 0.47058114703043263, + 0.44826534867823975, + 0.4219284623717116, + 0.4183482757773561, + 0.39919298053467706, + 0.3978373463125384, + 0.3898850401278166, + 0.37048571141165043, + 0.350463944852927 + ], + "ci_lo": [ + 0.824, + 0.696, + 0.678714859437751, + 0.6686991869918699, + 0.6611226611226612, + 0.6539278131634819, + 0.64537090850484, + 0.6276346604215457, + 0.6148325358851675, + 0.6020408163265306, + 0.5906732192677788, + 0.5776537615293826, + 0.5681801712779975, + 0.5484824046920821, + 0.5327042116761742, + 0.5136008765922476, + 0.49285455486542445, + 0.46969289670902575, + 0.44485635825940006, + 0.43643699692428506, + 0.42210824766031996, + 0.400990099009901, + 0.37627370849233926, + 0.350253807106599, + 0.3442622950819672, + 0.32386363636363635, + 0.3220338983050847, + 0.31055365174555577, + 0.29052893542397357, + 0.26811594202898553 + ], + "ci_hi": [ + 0.886, + 0.774, + 0.7585513078470825, + 0.75, + 0.7433264887063655, + 0.7377424640614505, + 0.7304367192795584, + 0.7173396674584323, + 0.7067307692307693, + 0.6976786013050025, + 0.689078469225528, + 0.6781025444899792, + 0.6696998284734134, + 0.6548698442270868, + 0.641976221783914, + 0.6258741258741258, + 0.6084212747994934, + 0.5904059040590406, + 0.5702127659574469, + 0.566674107142857, + 0.5576962809917355, + 0.5402843601895735, + 0.519612696563774, + 0.4944459706959707, + 0.4906832298136646, + 0.47398843930635837, + 0.47368421052631576, + 0.4675428737928736, + 0.45, + 0.4330708661417323 + ] + } +} \ No newline at end of file diff --git a/results/metrics/r1-qwen-7b_gptq_w4_math500_run0_sig.json b/results/metrics/r1-qwen-7b_gptq_w4_math500_run0_sig.json new file mode 100644 index 0000000000000000000000000000000000000000..39afb0277073aa1323d6c62d745a6d7b65866b64 --- /dev/null +++ b/results/metrics/r1-qwen-7b_gptq_w4_math500_run0_sig.json @@ -0,0 +1,7 @@ +{ + "n_pairs": 500, + "delta_acc_observed": 0.010000000000000009, + "delta_acc_ci_lo": -0.027999999999999917, + "delta_acc_ci_hi": 0.04800000000000004, + "p_value": 0.6272 +} \ No newline at end of file diff --git a/results/metrics/r1-qwen-7b_gptq_w4_restored_gpqa_run0_ci.json b/results/metrics/r1-qwen-7b_gptq_w4_restored_gpqa_run0_ci.json new file mode 100644 index 0000000000000000000000000000000000000000..4a0e2e3db406b126bf5109cdc7ce0ed0880ca06a --- /dev/null +++ b/results/metrics/r1-qwen-7b_gptq_w4_restored_gpqa_run0_ci.json @@ -0,0 +1,116 @@ +{ + "n": 198, + "accuracy": { + "mean": 0.2424969696969697, + "ci_lo": 0.18181818181818182, + "ci_hi": 0.30303030303030304 + }, + "avg_ffs": { + "mean": 0.5126595829348064, + "ci_lo": 0.42068196793323087, + "ci_hi": 0.6038993724396409 + }, + "ecr": { + "mean": 0.9860350245959504, + "ci_lo": 0.9783709172691426, + "ci_hi": 0.9922428591190284 + }, + "ssr_curve_ci": { + "mean": [ + 0.6007757575757552, + 0.2725252525252526, + 0.2398967773571047, + 0.23823852137562612, + 0.23507314721558456, + 0.23339110828944962, + 0.23600453331700205, + 0.22668237237037528, + 0.2241887746693302, + 0.21747250820568098, + 0.21885429268131085, + 0.20540262706023915, + 0.20806780141130415, + 0.2054958300213581, + 0.20967845846918703, + 0.21111711162633573, + 0.207097119039766, + 0.20010668722980043, + 0.20592321066138972, + 0.20622589452979237, + 0.20622589452979237, + 0.19825100429982703, + 0.20313645286316542, + 0.191474260213221, + 0.18532309369698993, + 0.1924951245607718, + 0.17881990396805864, + 0.17586970778787886, + 0.17090354272737313, + 0.17958689817511037 + ], + "ci_lo": [ + 0.5303030303030303, + 0.20707070707070707, + 0.17857142857142858, + 0.17799569329505152, + 0.17318435754189945, + 0.17222222222222222, + 0.1748522683949676, + 0.16470588235294117, + 0.16147127329192548, + 0.1562398538961039, + 0.1568627450980392, + 0.1437908496732026, + 0.14598159975669098, + 0.14285714285714285, + 0.1456953642384106, + 0.14705368161250515, + 0.14285714285714285, + 0.1357142857142857, + 0.1397016073871409, + 0.13934426229508196, + 0.13934426229508196, + 0.13114754098360656, + 0.13385826771653545, + 0.12068167305236271, + 0.11428571428571428, + 0.1188118811881188, + 0.10416666666666667, + 0.0989010989010989, + 0.09210526315789473, + 0.0978194591728526 + ], + "ci_hi": [ + 0.6717171717171717, + 0.3383838383838384, + 0.30256410256410254, + 0.3010752688172043, + 0.2973086850895838, + 0.2956989247311828, + 0.2988505747126437, + 0.28994082840236685, + 0.2883435582822086, + 0.2822085889570552, + 0.28378596338273754, + 0.26973684210526316, + 0.2733333333333333, + 0.2709792626728109, + 0.27631578947368424, + 0.2777777777777778, + 0.27388753162900265, + 0.26811594202898553, + 0.27481393129770987, + 0.27731245782984726, + 0.27731245782984726, + 0.2698412698412698, + 0.27643183797909393, + 0.26666666666666666, + 0.2616935849741401, + 0.2719298245614035, + 0.25742574257425743, + 0.25757966260543574, + 0.2564102564102564, + 0.2702782193958663 + ] + } +} \ No newline at end of file diff --git a/results/metrics/r1-qwen-7b_gptq_w4_restored_gsm8k_run0_ci.json b/results/metrics/r1-qwen-7b_gptq_w4_restored_gsm8k_run0_ci.json new file mode 100644 index 0000000000000000000000000000000000000000..41b29669efa3b252d08ce1e028f4e02ccbee0520 --- /dev/null +++ b/results/metrics/r1-qwen-7b_gptq_w4_restored_gsm8k_run0_ci.json @@ -0,0 +1,116 @@ +{ + "n": 1319, + "accuracy": { + "mean": 0.735281880212282, + "ci_lo": 0.711144806671721, + "ci_hi": 0.7589082638362395 + }, + "avg_ffs": { + "mean": 0.3605896705419057, + "ci_lo": 0.28362573099415206, + "ci_hi": 0.44507171075231877 + }, + "ecr": { + "mean": 0.8362798170191948, + "ci_lo": 0.8047071829101601, + "ci_hi": 0.8672415708790288 + }, + "ssr_curve_ci": { + "mean": [ + 0.801242608036393, + 0.7635211662560719, + 0.7653905253747193, + 0.8079583003647913, + 0.820835687955938, + 0.8148859909445869, + 0.8072274496856483, + 0.7956425929619408, + 0.7703819646432624, + 0.7123517689080175, + 0.702720625462509, + 0.6352320582023624, + 0.7219945255339464, + 0.7219945255339464, + 0.7136042002061297, + 0.7687567973736393, + 0.7279798243115667, + 0.7009560491881683, + 0.7009560491881683, + 0.7009560491881683, + 0.668395680607661, + 0.668395680607661, + 0.7516293152757753, + 0.7516293152757753, + 0.7516293152757753, + 0.7516293152757753, + 0.7516293152757753, + 0.71676658011091, + 0.6700883716292606, + 0.6700883716292606 + ], + "ci_lo": [ + 0.7793783169067475, + 0.7389541499330655, + 0.7401991175019522, + 0.7817366093795789, + 0.7920107438016528, + 0.7826824433791626, + 0.7662328516553869, + 0.7432369285173023, + 0.6935483870967742, + 0.6031746031746031, + 0.5483870967741935, + 0.42857142857142855, + 0.5, + 0.5, + 0.45, + 0.5, + 0.42857142857142855, + 0.375, + 0.375, + 0.375, + 0.3333333333333333, + 0.3333333333333333, + 0.4, + 0.4, + 0.4, + 0.4, + 0.4, + 0.3333333333333333, + 0.24166666666666675, + 0.24166666666666675 + ], + "ci_hi": [ + 0.8225928733889311, + 0.7868188697693264, + 0.7897165102915477, + 0.8322301181158798, + 0.847457627118644, + 0.8451178451178452, + 0.8465230785668495, + 0.8458154368575624, + 0.8425215077282006, + 0.8181818181818182, + 0.84375, + 0.8333333333333334, + 0.9285714285714286, + 0.9285714285714286, + 0.9375, + 1.0, + 1.0, + 1.0, + 1.0, + 1.0, + 1.0, + 1.0, + 1.0, + 1.0, + 1.0, + 1.0, + 1.0, + 1.0, + 1.0, + 1.0 + ] + } +} \ No newline at end of file diff --git a/results/metrics/r1-qwen-7b_gptq_w4_restored_gsm8k_run0_metrics.json b/results/metrics/r1-qwen-7b_gptq_w4_restored_gsm8k_run0_metrics.json new file mode 100644 index 0000000000000000000000000000000000000000..6b21ce60c469f2aed5b96489fc057c9ce54819b9 --- /dev/null +++ b/results/metrics/r1-qwen-7b_gptq_w4_restored_gsm8k_run0_metrics.json @@ -0,0 +1,47 @@ +{ + "model": "r1-qwen-7b", + "quantization": "gptq_w4_restored", + "accuracy": 0.7354056103108415, + "accuracy_delta": 0.0, + "avg_ffs": 0.36103151862464183, + "median_ffs": 0.0, + "ecr": 0.8360355739951979, + "ssr_curve": [ + 0.8013646702047005, + 0.7637540453074434, + 0.7657657657657657, + 0.8081140350877193, + 0.8209366391184573, + 0.815, + 0.8072916666666666, + 0.7956521739130434, + 0.7704918032786885, + 0.7121212121212122, + 0.7027027027027027, + 0.6363636363636364, + 0.7222222222222222, + 0.7222222222222222, + 0.7142857142857143, + 0.7692307692307693, + 0.7272727272727273, + 0.7, + 0.7, + 0.7, + 0.6666666666666666, + 0.6666666666666666, + 0.75, + 0.75, + 0.75, + 0.75, + 0.75, + 0.7142857142857143, + 0.6666666666666666, + 0.6666666666666666 + ], + "error_type_dist": { + "conceptual": 0.7509238728750924, + "executional": 0.028824833702882482, + "logical": 0.21951219512195122, + "methodological": 0.0007390983000739098 + } +} \ No newline at end of file diff --git a/results/metrics/r1-qwen-7b_gptq_w4_restored_math500_run0_ci.json b/results/metrics/r1-qwen-7b_gptq_w4_restored_math500_run0_ci.json new file mode 100644 index 0000000000000000000000000000000000000000..2ccfb6ec69afc02e286035dc09722242573458d9 --- /dev/null +++ b/results/metrics/r1-qwen-7b_gptq_w4_restored_math500_run0_ci.json @@ -0,0 +1,116 @@ +{ + "n": 500, + "accuracy": { + "mean": 0.7175691999999999, + "ci_lo": 0.678, + "ci_hi": 0.756 + }, + "avg_ffs": { + "mean": 0.6449529254868427, + "ci_lo": 0.49659746657283604, + "ci_hi": 0.8187931621437603 + }, + "ecr": { + "mean": 0.9587974000235455, + "ci_lo": 0.9450916180247755, + "ci_hi": 0.9710610978697083 + }, + "ssr_curve_ci": { + "mean": [ + 0.8538436000000013, + 0.73989510868186, + 0.7292787336620187, + 0.7241725730250751, + 0.714264446668119, + 0.7075252422883055, + 0.7019986294297477, + 0.688062005546298, + 0.6709138243211936, + 0.6545866193203543, + 0.6307664978467706, + 0.6095833290919246, + 0.5926435920942383, + 0.5748647084053081, + 0.552178394864246, + 0.542072889321191, + 0.5197077157050737, + 0.5021866179737599, + 0.4923274202288306, + 0.47320031402458135, + 0.44902443708114187, + 0.4203544990740653, + 0.3998120028087843, + 0.3685070948772266, + 0.3719396552430148, + 0.3653497632044509, + 0.34453166582907757, + 0.3361715717346733, + 0.34015783300473434, + 0.3215607069951452 + ], + "ci_lo": [ + 0.822, + 0.7003947368421052, + 0.688793258330838, + 0.6825053995680346, + 0.6725265127885216, + 0.6635933933416599, + 0.6570726027873354, + 0.6406228885135136, + 0.6225288962033068, + 0.6042250231388505, + 0.5768084822619365, + 0.554379818974187, + 0.5357142857142857, + 0.5158730158730159, + 0.489957806936247, + 0.4798085831820035, + 0.45414847161572053, + 0.4360129788251663, + 0.42452485043718363, + 0.4021150384552446, + 0.37432998717828436, + 0.34545454545454546, + 0.3225806451612903, + 0.2890542473591549, + 0.2907801418439716, + 0.2802954545454545, + 0.25892857142857145, + 0.24561004784688995, + 0.24444444444444444, + 0.2222222222222222 + ], + "ci_hi": [ + 0.884, + 0.7784552845528455, + 0.76875, + 0.7639515170094568, + 0.7562076749435666, + 0.75, + 0.7457627118644068, + 0.734249785958904, + 0.7187542229729729, + 0.7040291211101765, + 0.6830121178906344, + 0.6645579448069816, + 0.6501810027428401, + 0.6342412451361867, + 0.6147929114501082, + 0.6060606060606061, + 0.5848322292626726, + 0.5700494378978734, + 0.5625, + 0.5463427206437013, + 0.5235294117647059, + 0.4968553459119497, + 0.48026315789473684, + 0.44966442953020136, + 0.4539168278529978, + 0.45185185185185184, + 0.4329896907216495, + 0.4260905203136136, + 0.4375, + 0.42168674698795183 + ] + } +} \ No newline at end of file diff --git a/results/metrics/r1-qwen-7b_gptq_w4_restored_math500_run0_metrics.json b/results/metrics/r1-qwen-7b_gptq_w4_restored_math500_run0_metrics.json new file mode 100644 index 0000000000000000000000000000000000000000..fe7ca9270bd1f2245b87359ca2d374561aceee48 --- /dev/null +++ b/results/metrics/r1-qwen-7b_gptq_w4_restored_math500_run0_metrics.json @@ -0,0 +1,47 @@ +{ + "model": "r1-qwen-7b", + "quantization": "gptq_w4_restored", + "accuracy": 0.718, + "accuracy_delta": 0.0, + "avg_ffs": 0.6453900709219859, + "median_ffs": 0.0, + "ecr": 0.9587642209632271, + "ssr_curve": [ + 0.854, + 0.7403651115618661, + 0.7297297297297297, + 0.7245119305856833, + 0.7146067415730337, + 0.7079439252336449, + 0.7024390243902439, + 0.6884816753926701, + 0.6713091922005571, + 0.6549707602339181, + 0.63125, + 0.61, + 0.5928571428571429, + 0.575187969924812, + 0.5524193548387096, + 0.5423728813559322, + 0.52, + 0.5023474178403756, + 0.49261083743842365, + 0.4734042553191489, + 0.449438202247191, + 0.42073170731707316, + 0.4, + 0.36879432624113473, + 0.3722627737226277, + 0.36585365853658536, + 0.34513274336283184, + 0.33653846153846156, + 0.3404255319148936, + 0.3218390804597701 + ], + "error_type_dist": { + "conceptual": 0.0854205114684946, + "methodological": 0.6029528078038492, + "executional": 0.07434748220406011, + "logical": 0.2372791985235961 + } +} \ No newline at end of file diff --git a/sparse_logs/15-04-2026_15.43.34.log b/sparse_logs/15-04-2026_15.43.34.log new file mode 100644 index 0000000000000000000000000000000000000000..aa8d388564a8144788cebc6ac6cf76cda0d96819 --- /dev/null +++ b/sparse_logs/15-04-2026_15.43.34.log @@ -0,0 +1,2521 @@ +2026-04-15 15:43:34.167 | INFO | llmcompressor.metrics.logger:_create_default_logger:357 - Logging all LLM Compressor modifier-level logs to sparse_logs/15-04-2026_15.43.34.log +2026-04-15 15:43:34.168 | DEBUG | llmcompressor.core.lifecycle:initialize:92 - Initializing compression lifecycle +2026-04-15 15:43:34.168 | INFO | llmcompressor.recipe.recipe:from_modifiers:68 - Creating recipe from modifiers +2026-04-15 15:43:34.485 | DEBUG | llmcompressor.core.lifecycle:initialize:105 - Initialized modifier: config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=False ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False +2026-04-15 15:43:34.485 | INFO | llmcompressor.core.lifecycle:initialize:110 - Compression lifecycle initialized for 1 modifiers +2026-04-15 15:43:34.486 | INFO | llmcompressor.pipelines.independent.pipeline:IndependentPipeline:43 - Inferred `SequentialPipeline` for `GPTQModifier` +2026-04-15 15:43:53.651 | DEBUG | llmcompressor.pipelines.sequential.ast_utils.auto_wrapper:_wrap_stmt:267 - ---- Autowrapper ---- +2026-04-15 15:43:53.652 | DEBUG | llmcompressor.pipelines.sequential.ast_utils.auto_wrapper:_wrap_stmt:268 - @torch.fx.wrap +def wrapped_0(input_ids, inputs_embeds): + if (input_ids is None) ^ (inputs_embeds is not None): + raise ValueError('You must specify exactly one of input_ids or inputs_embeds') + return () +2026-04-15 15:43:53.652 | DEBUG | llmcompressor.pipelines.sequential.ast_utils.auto_wrapper:_wrap_stmt:269 - --------------------- +2026-04-15 15:43:53.652 | DEBUG | llmcompressor.pipelines.sequential.ast_utils.auto_wrapper:_wrap_stmt:267 - ---- Autowrapper ---- +2026-04-15 15:43:53.653 | DEBUG | llmcompressor.pipelines.sequential.ast_utils.auto_wrapper:_wrap_stmt:268 - @torch.fx.wrap +def wrapped_1(input_ids, inputs_embeds): + if inputs_embeds is None: + inputs_embeds = self.embed_tokens(input_ids) + return (inputs_embeds,) +2026-04-15 15:43:53.653 | DEBUG | llmcompressor.pipelines.sequential.ast_utils.auto_wrapper:_wrap_stmt:269 - --------------------- +2026-04-15 15:43:53.653 | DEBUG | llmcompressor.pipelines.sequential.ast_utils.auto_wrapper:_wrap_stmt:267 - ---- Autowrapper ---- +2026-04-15 15:43:53.654 | DEBUG | llmcompressor.pipelines.sequential.ast_utils.auto_wrapper:_wrap_stmt:268 - @torch.fx.wrap +def wrapped_2(past_key_values, use_cache): + if use_cache and past_key_values is None: + past_key_values = DynamicCache(config=self.config) + return (past_key_values,) +2026-04-15 15:43:53.654 | DEBUG | llmcompressor.pipelines.sequential.ast_utils.auto_wrapper:_wrap_stmt:269 - --------------------- +2026-04-15 15:43:53.654 | DEBUG | llmcompressor.pipelines.sequential.ast_utils.auto_wrapper:_wrap_stmt:267 - ---- Autowrapper ---- +2026-04-15 15:43:53.655 | DEBUG | llmcompressor.pipelines.sequential.ast_utils.auto_wrapper:_wrap_stmt:268 - @torch.fx.wrap +def wrapped_3(cache_position, inputs_embeds, past_key_values, *, past_seen_tokens=None): + if cache_position is None: + past_seen_tokens = past_key_values.get_seq_length() if past_key_values is not None else 0 + cache_position = torch.arange(past_seen_tokens, past_seen_tokens + inputs_embeds.shape[1], device=inputs_embeds.device) + return (cache_position, past_seen_tokens) +2026-04-15 15:43:53.655 | DEBUG | llmcompressor.pipelines.sequential.ast_utils.auto_wrapper:_wrap_stmt:269 - --------------------- +2026-04-15 15:43:53.655 | DEBUG | llmcompressor.pipelines.sequential.ast_utils.auto_wrapper:_wrap_stmt:267 - ---- Autowrapper ---- +2026-04-15 15:43:53.656 | DEBUG | llmcompressor.pipelines.sequential.ast_utils.auto_wrapper:_wrap_stmt:268 - @torch.fx.wrap +def wrapped_4(cache_position, position_ids): + if position_ids is None: + position_ids = cache_position.unsqueeze(0) + return (position_ids,) +2026-04-15 15:43:53.656 | DEBUG | llmcompressor.pipelines.sequential.ast_utils.auto_wrapper:_wrap_stmt:269 - --------------------- +2026-04-15 15:43:53.657 | DEBUG | llmcompressor.pipelines.sequential.ast_utils.auto_wrapper:_wrap_stmt:267 - ---- Autowrapper ---- +2026-04-15 15:43:53.657 | DEBUG | llmcompressor.pipelines.sequential.ast_utils.auto_wrapper:_wrap_stmt:268 - @torch.fx.wrap +def wrapped_5(attention_mask, cache_position, inputs_embeds, past_key_values, position_ids, *, causal_mask_mapping=None, mask_kwargs=None): + if not isinstance((causal_mask_mapping := attention_mask), dict): + mask_kwargs = {'config': self.config, 'input_embeds': inputs_embeds, 'attention_mask': attention_mask, 'cache_position': cache_position, 'past_key_values': past_key_values, 'position_ids': position_ids} + causal_mask_mapping = {'full_attention': create_causal_mask(**mask_kwargs)} + if self.has_sliding_layers: + causal_mask_mapping['sliding_attention'] = create_sliding_window_causal_mask(**mask_kwargs) + return (causal_mask_mapping, mask_kwargs) +2026-04-15 15:43:53.657 | DEBUG | llmcompressor.pipelines.sequential.ast_utils.auto_wrapper:_wrap_stmt:269 - --------------------- +2026-04-15 15:43:53.664 | DEBUG | llmcompressor.pipelines.sequential.ast_utils.auto_wrapper:_wrap_stmt:267 - ---- Autowrapper ---- +2026-04-15 15:43:53.664 | DEBUG | llmcompressor.pipelines.sequential.ast_utils.auto_wrapper:_wrap_stmt:268 - @torch.fx.wrap +def wrapped_0(kwargs, labels, logits, loss): + if labels is not None: + loss = self.loss_function(logits=logits, labels=labels, vocab_size=self.config.vocab_size, **kwargs) + return (loss,) +2026-04-15 15:43:53.665 | DEBUG | llmcompressor.pipelines.sequential.ast_utils.auto_wrapper:_wrap_stmt:269 - --------------------- +2026-04-15 15:43:54.048 | DEBUG | llmcompressor.core.lifecycle:event:195 - Handling event: EventType.CALIBRATION_EPOCH_START +2026-04-15 15:43:54.071 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.071 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.071 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.072 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.072 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.072 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.072 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.072 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.072 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.073 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.073 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.073 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.073 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.073 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.073 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.073 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.074 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.074 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.074 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.074 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.074 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.074 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.074 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.075 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.075 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.075 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.075 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.075 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.075 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.075 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.076 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.076 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.076 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.076 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.076 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.076 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.076 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.077 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.077 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.077 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.077 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.077 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.077 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.077 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.078 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.078 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.078 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.078 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.078 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.078 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.079 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.079 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.079 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.079 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.079 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.079 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.080 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.080 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.080 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.080 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.080 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.080 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.080 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.081 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.081 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.081 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.081 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.081 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.081 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.081 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.082 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.082 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.082 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.082 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.082 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.082 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.083 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.083 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.083 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.083 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.083 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.083 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.083 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.084 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.084 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.084 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.084 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.084 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.084 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.084 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.085 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.085 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.085 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.085 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.085 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.085 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.086 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.086 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.086 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.086 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.086 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.086 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.086 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.087 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.087 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.087 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.087 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.087 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.087 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.087 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.088 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.088 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.088 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.088 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.088 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.088 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.089 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.089 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.089 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.089 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.089 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.089 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.090 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.090 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.090 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.090 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.090 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.090 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.091 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.091 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.091 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.091 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.091 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.091 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.091 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.092 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.092 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.092 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.092 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.092 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.092 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.093 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.093 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.093 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.093 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.093 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.093 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.094 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.094 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.094 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.094 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.094 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.094 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.094 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.095 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.095 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.095 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.095 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.095 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.095 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.095 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.096 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.096 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.096 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.096 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.096 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.096 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.097 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.097 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.097 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.097 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.097 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.097 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.098 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.098 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.098 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.098 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.098 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.098 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.099 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.099 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.099 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.099 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.099 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.099 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.100 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.100 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.100 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.100 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.100 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.100 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.101 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.101 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.101 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.101 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.101 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.101 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.102 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.102 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.102 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.102 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.102 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.102 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.103 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.103 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.103 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.103 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.103 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.103 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.104 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.104 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.104 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.104 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.104 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.104 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.104 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.105 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.105 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.105 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.105 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.105 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.105 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.105 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.105 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.106 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.106 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.106 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.106 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.106 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.106 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.107 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.107 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.107 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.107 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.107 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.107 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.108 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.108 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.108 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.108 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.108 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.108 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.108 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.109 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.109 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.109 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.109 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.109 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.109 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.109 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.110 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.110 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.110 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.110 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.110 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.110 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.110 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.111 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.111 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.112 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.112 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.112 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.112 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.112 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.112 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.112 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.113 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.113 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.113 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.113 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.113 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.113 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.114 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.114 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.114 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.114 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.114 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.114 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.114 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.115 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.115 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.115 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.115 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.115 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.115 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.115 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.116 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.116 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.116 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.116 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.116 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.116 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.116 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.117 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.117 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.117 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.117 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.117 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.117 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.118 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.118 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.118 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.118 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.118 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.118 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.118 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.119 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.119 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.119 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.119 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.119 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.119 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.119 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.120 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.120 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.120 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.120 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.120 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.120 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.121 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.121 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.121 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.121 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.121 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.121 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.122 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.122 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.122 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.122 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.122 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.122 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.122 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.123 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.123 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.123 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.123 | DEBUG | llmcompressor.modifiers.utils.hooks:register_hook:98 - config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False added +2026-04-15 15:43:54.123 | DEBUG | llmcompressor.core.lifecycle:event:205 - Updated event with modifier: config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False +2026-04-15 15:44:00.647 | DEBUG | llmcompressor.core.lifecycle:event:195 - Handling event: EventType.SEQUENTIAL_EPOCH_END +2026-04-15 15:44:00.647 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.0.self_attn.q_proj using 128 samples +2026-04-15 15:44:04.137 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.49s +2026-04-15 15:44:04.137 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 118.68 +2026-04-15 15:44:04.138 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:44:04.138 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 21.49% | total memory: 25 GB +2026-04-15 15:44:04.138 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 53.05344 MB +2026-04-15 15:44:04.138 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.0.self_attn.k_proj using 128 samples +2026-04-15 15:44:07.225 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.09s +2026-04-15 15:44:07.225 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 42.77 +2026-04-15 15:44:07.225 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:44:07.225 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 21.49% | total memory: 25 GB +2026-04-15 15:44:07.225 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 10.610688 MB +2026-04-15 15:44:07.225 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.0.self_attn.v_proj using 128 samples +2026-04-15 15:44:10.299 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.07s +2026-04-15 15:44:10.299 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 8.49 +2026-04-15 15:44:10.300 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:44:10.300 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 21.49% | total memory: 25 GB +2026-04-15 15:44:10.300 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 10.610688 MB +2026-04-15 15:44:10.300 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.0.self_attn.o_proj using 128 samples +2026-04-15 15:44:13.520 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.22s +2026-04-15 15:44:13.521 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 115.30 +2026-04-15 15:44:13.521 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:44:13.521 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 21.49% | total memory: 25 GB +2026-04-15 15:44:13.522 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 53.0432 MB +2026-04-15 15:44:13.522 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.0.mlp.gate_proj using 128 samples +2026-04-15 15:44:17.007 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.48s +2026-04-15 15:44:17.007 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 64.79 +2026-04-15 15:44:17.007 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:44:17.007 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 21.49% | total memory: 25 GB +2026-04-15 15:44:17.007 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 15:44:17.008 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.0.mlp.up_proj using 128 samples +2026-04-15 15:44:20.493 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.49s +2026-04-15 15:44:20.494 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 55.77 +2026-04-15 15:44:20.494 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:44:20.494 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 22.54% | total memory: 25 GB +2026-04-15 15:44:20.495 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 15:44:20.495 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.0.mlp.down_proj using 128 samples +2026-04-15 15:44:29.432 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 8.94s +2026-04-15 15:44:29.432 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 50.85 +2026-04-15 15:44:29.433 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:44:29.433 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 28.48% | total memory: 25 GB +2026-04-15 15:44:29.433 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 15:44:29.433 | DEBUG | llmcompressor.core.lifecycle:event:205 - Updated event with modifier: config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False +2026-04-15 15:44:33.860 | DEBUG | llmcompressor.core.lifecycle:event:195 - Handling event: EventType.SEQUENTIAL_EPOCH_END +2026-04-15 15:44:33.860 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.1.self_attn.q_proj using 128 samples +2026-04-15 15:44:37.516 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.66s +2026-04-15 15:44:37.517 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 3.17 +2026-04-15 15:44:37.517 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:44:37.518 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 15:44:37.518 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 53.05344 MB +2026-04-15 15:44:37.518 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.1.self_attn.k_proj using 128 samples +2026-04-15 15:44:40.799 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.28s +2026-04-15 15:44:40.800 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 1.04 +2026-04-15 15:44:40.800 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:44:40.800 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 15:44:40.800 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 10.610688 MB +2026-04-15 15:44:40.800 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.1.self_attn.v_proj using 128 samples +2026-04-15 15:44:44.062 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.26s +2026-04-15 15:44:44.063 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 0.47 +2026-04-15 15:44:44.063 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:44:44.063 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 15:44:44.063 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 10.610688 MB +2026-04-15 15:44:44.063 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.1.self_attn.o_proj using 128 samples +2026-04-15 15:44:47.478 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.41s +2026-04-15 15:44:47.478 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 1.29 +2026-04-15 15:44:47.478 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:44:47.478 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 15:44:47.478 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 53.0432 MB +2026-04-15 15:44:47.479 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.1.mlp.gate_proj using 128 samples +2026-04-15 15:44:51.148 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.67s +2026-04-15 15:44:51.148 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 487.44 +2026-04-15 15:44:51.149 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:44:51.149 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 15:44:51.149 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 15:44:51.149 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.1.mlp.up_proj using 128 samples +2026-04-15 15:44:54.821 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.67s +2026-04-15 15:44:54.821 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 157.32 +2026-04-15 15:44:54.822 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:44:54.822 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 16.49% | total memory: 25 GB +2026-04-15 15:44:54.822 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 15:44:54.822 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.1.mlp.down_proj using 128 samples +2026-04-15 15:45:04.352 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 9.53s +2026-04-15 15:45:04.353 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 6.11 +2026-04-15 15:45:04.354 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:45:04.354 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 22.43% | total memory: 25 GB +2026-04-15 15:45:04.355 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 15:45:04.356 | DEBUG | llmcompressor.core.lifecycle:event:205 - Updated event with modifier: config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False +2026-04-15 15:45:07.603 | DEBUG | llmcompressor.core.lifecycle:event:195 - Handling event: EventType.SEQUENTIAL_EPOCH_END +2026-04-15 15:45:07.604 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.2.self_attn.q_proj using 128 samples +2026-04-15 15:45:11.046 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.44s +2026-04-15 15:45:11.047 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 11.68 +2026-04-15 15:45:11.047 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:45:11.047 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 15:45:11.047 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 53.05344 MB +2026-04-15 15:45:11.048 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.2.self_attn.k_proj using 128 samples +2026-04-15 15:45:14.367 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.32s +2026-04-15 15:45:14.367 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 4.38 +2026-04-15 15:45:14.367 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:45:14.367 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 15:45:14.367 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 10.610688 MB +2026-04-15 15:45:14.368 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.2.self_attn.v_proj using 128 samples +2026-04-15 15:45:17.711 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.34s +2026-04-15 15:45:17.712 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 1.68 +2026-04-15 15:45:17.712 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:45:17.712 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 15:45:17.712 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 10.610688 MB +2026-04-15 15:45:17.712 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.2.self_attn.o_proj using 128 samples +2026-04-15 15:45:21.177 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.46s +2026-04-15 15:45:21.177 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 10.09 +2026-04-15 15:45:21.177 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:45:21.177 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 15:45:21.178 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 53.0432 MB +2026-04-15 15:45:21.178 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.2.mlp.gate_proj using 128 samples +2026-04-15 15:45:24.888 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.71s +2026-04-15 15:45:24.888 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 844.41 +2026-04-15 15:45:24.889 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:45:24.889 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 15:45:24.889 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 15:45:24.889 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.2.mlp.up_proj using 128 samples +2026-04-15 15:45:28.598 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.71s +2026-04-15 15:45:28.598 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 445.21 +2026-04-15 15:45:28.599 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:45:28.599 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 15:45:28.599 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 15:45:28.599 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.2.mlp.down_proj using 128 samples +2026-04-15 15:45:38.204 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 9.60s +2026-04-15 15:45:38.204 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 30.15 +2026-04-15 15:45:38.205 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:45:38.205 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 21.38% | total memory: 25 GB +2026-04-15 15:45:38.205 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 15:45:38.206 | DEBUG | llmcompressor.core.lifecycle:event:205 - Updated event with modifier: config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False +2026-04-15 15:45:41.528 | DEBUG | llmcompressor.core.lifecycle:event:195 - Handling event: EventType.SEQUENTIAL_EPOCH_END +2026-04-15 15:45:41.528 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.3.self_attn.q_proj using 128 samples +2026-04-15 15:45:44.933 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.40s +2026-04-15 15:45:44.933 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 70.93 +2026-04-15 15:45:44.933 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:45:44.933 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 15:45:44.933 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 53.05344 MB +2026-04-15 15:45:44.934 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.3.self_attn.k_proj using 128 samples +2026-04-15 15:45:48.238 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.30s +2026-04-15 15:45:48.239 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 26.31 +2026-04-15 15:45:48.239 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:45:48.239 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 15:45:48.240 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 10.610688 MB +2026-04-15 15:45:48.240 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.3.self_attn.v_proj using 128 samples +2026-04-15 15:45:51.528 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.29s +2026-04-15 15:45:51.529 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 9.85 +2026-04-15 15:45:51.529 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:45:51.529 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 15:45:51.529 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 10.610688 MB +2026-04-15 15:45:51.529 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.3.self_attn.o_proj using 128 samples +2026-04-15 15:45:54.979 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.45s +2026-04-15 15:45:54.979 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 28.80 +2026-04-15 15:45:54.979 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:45:54.979 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 15:45:54.979 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 53.0432 MB +2026-04-15 15:45:54.980 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.3.mlp.gate_proj using 128 samples +2026-04-15 15:45:58.711 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.73s +2026-04-15 15:45:58.712 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 1213.98 +2026-04-15 15:45:58.712 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:45:58.713 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 15:45:58.713 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 15:45:58.713 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.3.mlp.up_proj using 128 samples +2026-04-15 15:46:02.413 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.70s +2026-04-15 15:46:02.413 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 571.55 +2026-04-15 15:46:02.414 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:46:02.414 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 16.49% | total memory: 25 GB +2026-04-15 15:46:02.414 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 15:46:02.414 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.3.mlp.down_proj using 128 samples +2026-04-15 15:46:12.050 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 9.64s +2026-04-15 15:46:12.051 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 62.24 +2026-04-15 15:46:12.051 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:46:12.051 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 22.43% | total memory: 25 GB +2026-04-15 15:46:12.052 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 15:46:12.052 | DEBUG | llmcompressor.core.lifecycle:event:205 - Updated event with modifier: config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False +2026-04-15 15:46:15.210 | DEBUG | llmcompressor.core.lifecycle:event:195 - Handling event: EventType.SEQUENTIAL_EPOCH_END +2026-04-15 15:46:15.210 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.4.self_attn.q_proj using 128 samples +2026-04-15 15:46:18.641 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.43s +2026-04-15 15:46:18.641 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 94.58 +2026-04-15 15:46:18.642 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:46:18.642 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 15:46:18.642 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 53.05344 MB +2026-04-15 15:46:18.642 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.4.self_attn.k_proj using 128 samples +2026-04-15 15:46:21.918 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.28s +2026-04-15 15:46:21.918 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 35.15 +2026-04-15 15:46:21.919 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:46:21.919 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 15:46:21.919 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 10.610688 MB +2026-04-15 15:46:21.919 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.4.self_attn.v_proj using 128 samples +2026-04-15 15:46:25.229 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.31s +2026-04-15 15:46:25.230 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 13.41 +2026-04-15 15:46:25.230 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:46:25.230 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 15:46:25.230 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 10.610688 MB +2026-04-15 15:46:25.231 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.4.self_attn.o_proj using 128 samples +2026-04-15 15:46:28.656 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.43s +2026-04-15 15:46:28.656 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 52.33 +2026-04-15 15:46:28.657 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:46:28.657 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 15:46:28.657 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 53.0432 MB +2026-04-15 15:46:28.657 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.4.mlp.gate_proj using 128 samples +2026-04-15 15:46:32.400 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.74s +2026-04-15 15:46:32.401 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 2206.95 +2026-04-15 15:46:32.401 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:46:32.401 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 15:46:32.401 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 15:46:32.401 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.4.mlp.up_proj using 128 samples +2026-04-15 15:46:36.090 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.69s +2026-04-15 15:46:36.090 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 1231.14 +2026-04-15 15:46:36.090 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:46:36.090 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 15:46:36.090 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 15:46:36.091 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.4.mlp.down_proj using 128 samples +2026-04-15 15:46:45.646 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 9.55s +2026-04-15 15:46:45.646 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 195405.66 +2026-04-15 15:46:45.646 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:46:45.647 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 21.38% | total memory: 25 GB +2026-04-15 15:46:45.647 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 15:46:45.647 | DEBUG | llmcompressor.core.lifecycle:event:205 - Updated event with modifier: config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False +2026-04-15 15:46:48.811 | DEBUG | llmcompressor.core.lifecycle:event:195 - Handling event: EventType.SEQUENTIAL_EPOCH_END +2026-04-15 15:46:48.811 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.5.self_attn.q_proj using 128 samples +2026-04-15 15:46:52.241 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.43s +2026-04-15 15:46:52.242 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 448.94 +2026-04-15 15:46:52.242 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:46:52.242 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 15:46:52.242 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 53.05344 MB +2026-04-15 15:46:52.242 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.5.self_attn.k_proj using 128 samples +2026-04-15 15:46:55.569 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.33s +2026-04-15 15:46:55.570 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 139.85 +2026-04-15 15:46:55.570 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:46:55.570 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 15:46:55.571 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 10.610688 MB +2026-04-15 15:46:55.571 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.5.self_attn.v_proj using 128 samples +2026-04-15 15:46:58.871 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.30s +2026-04-15 15:46:58.871 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 67.12 +2026-04-15 15:46:58.872 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:46:58.872 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 15:46:58.872 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 10.610688 MB +2026-04-15 15:46:58.872 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.5.self_attn.o_proj using 128 samples +2026-04-15 15:47:02.363 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.49s +2026-04-15 15:47:02.363 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 117.14 +2026-04-15 15:47:02.363 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:47:02.364 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 15:47:02.364 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 53.0432 MB +2026-04-15 15:47:02.364 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.5.mlp.gate_proj using 128 samples +2026-04-15 15:47:06.109 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.75s +2026-04-15 15:47:06.110 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 4313.08 +2026-04-15 15:47:06.110 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:47:06.110 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 15:47:06.111 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 15:47:06.111 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.5.mlp.up_proj using 128 samples +2026-04-15 15:47:09.811 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.70s +2026-04-15 15:47:09.812 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 2242.96 +2026-04-15 15:47:09.812 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:47:09.812 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 17.75% | total memory: 25 GB +2026-04-15 15:47:09.812 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 15:47:09.813 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.5.mlp.down_proj using 128 samples +2026-04-15 15:47:21.016 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 11.20s +2026-04-15 15:47:21.017 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 1059.16 +2026-04-15 15:47:21.017 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:47:21.017 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 76.80% | total memory: 25 GB +2026-04-15 15:47:21.018 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 15:47:21.018 | DEBUG | llmcompressor.core.lifecycle:event:205 - Updated event with modifier: config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False +2026-04-15 15:47:25.193 | DEBUG | llmcompressor.core.lifecycle:event:195 - Handling event: EventType.SEQUENTIAL_EPOCH_END +2026-04-15 15:47:25.193 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.6.self_attn.q_proj using 128 samples +2026-04-15 15:47:28.636 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.44s +2026-04-15 15:47:28.637 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 570.37 +2026-04-15 15:47:28.637 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:47:28.637 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 69.86% | total memory: 25 GB +2026-04-15 15:47:28.637 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 53.05344 MB +2026-04-15 15:47:28.638 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.6.self_attn.k_proj using 128 samples +2026-04-15 15:47:31.899 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.26s +2026-04-15 15:47:31.899 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 165.26 +2026-04-15 15:47:31.899 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:47:31.899 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 69.86% | total memory: 25 GB +2026-04-15 15:47:31.899 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 10.610688 MB +2026-04-15 15:47:31.900 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.6.self_attn.v_proj using 128 samples +2026-04-15 15:47:35.175 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.27s +2026-04-15 15:47:35.175 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 98.26 +2026-04-15 15:47:35.175 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:47:35.175 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 69.86% | total memory: 25 GB +2026-04-15 15:47:35.176 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 10.610688 MB +2026-04-15 15:47:35.176 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.6.self_attn.o_proj using 128 samples +2026-04-15 15:47:38.563 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.39s +2026-04-15 15:47:38.564 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 87.50 +2026-04-15 15:47:38.564 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:47:38.564 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 69.86% | total memory: 25 GB +2026-04-15 15:47:38.564 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 53.0432 MB +2026-04-15 15:47:38.564 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.6.mlp.gate_proj using 128 samples +2026-04-15 15:47:42.272 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.71s +2026-04-15 15:47:42.272 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 6580.07 +2026-04-15 15:47:42.273 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:47:42.273 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 69.86% | total memory: 25 GB +2026-04-15 15:47:42.274 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 15:47:42.274 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.6.mlp.up_proj using 128 samples +2026-04-15 15:47:45.983 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.71s +2026-04-15 15:47:45.984 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 3631.28 +2026-04-15 15:47:45.984 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:47:45.984 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 69.86% | total memory: 25 GB +2026-04-15 15:47:45.984 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 15:47:45.985 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.6.mlp.down_proj using 128 samples +2026-04-15 15:47:55.474 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 9.49s +2026-04-15 15:47:55.475 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 689.57 +2026-04-15 15:47:55.475 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:47:55.475 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 75.80% | total memory: 25 GB +2026-04-15 15:47:55.475 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 15:47:55.475 | DEBUG | llmcompressor.core.lifecycle:event:205 - Updated event with modifier: config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False +2026-04-15 15:47:58.843 | DEBUG | llmcompressor.core.lifecycle:event:195 - Handling event: EventType.SEQUENTIAL_EPOCH_END +2026-04-15 15:47:58.843 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.7.self_attn.q_proj using 128 samples +2026-04-15 15:48:02.281 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.44s +2026-04-15 15:48:02.281 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 893.26 +2026-04-15 15:48:02.281 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:48:02.281 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 69.87% | total memory: 25 GB +2026-04-15 15:48:02.282 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 53.05344 MB +2026-04-15 15:48:02.282 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.7.self_attn.k_proj using 128 samples +2026-04-15 15:48:05.540 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.26s +2026-04-15 15:48:05.540 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 255.64 +2026-04-15 15:48:05.541 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:48:05.541 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 69.87% | total memory: 25 GB +2026-04-15 15:48:05.541 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 10.610688 MB +2026-04-15 15:48:05.541 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.7.self_attn.v_proj using 128 samples +2026-04-15 15:48:08.827 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.29s +2026-04-15 15:48:08.828 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 155.41 +2026-04-15 15:48:08.828 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:48:08.828 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 69.87% | total memory: 25 GB +2026-04-15 15:48:08.828 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 10.610688 MB +2026-04-15 15:48:08.829 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.7.self_attn.o_proj using 128 samples +2026-04-15 15:48:12.273 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.44s +2026-04-15 15:48:12.274 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 86.14 +2026-04-15 15:48:12.274 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:48:12.274 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 69.87% | total memory: 25 GB +2026-04-15 15:48:12.274 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 53.0432 MB +2026-04-15 15:48:12.274 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.7.mlp.gate_proj using 128 samples +2026-04-15 15:48:15.997 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.72s +2026-04-15 15:48:15.997 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 8130.64 +2026-04-15 15:48:15.998 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:48:15.998 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 69.87% | total memory: 25 GB +2026-04-15 15:48:15.998 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 15:48:15.998 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.7.mlp.up_proj using 128 samples +2026-04-15 15:48:19.674 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.68s +2026-04-15 15:48:19.674 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 4653.96 +2026-04-15 15:48:19.674 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:48:19.674 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 70.92% | total memory: 25 GB +2026-04-15 15:48:19.674 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 15:48:19.675 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.7.mlp.down_proj using 128 samples +2026-04-15 15:48:29.232 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 9.56s +2026-04-15 15:48:29.232 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 196.37 +2026-04-15 15:48:29.232 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:48:29.233 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 76.86% | total memory: 25 GB +2026-04-15 15:48:29.233 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 15:48:29.233 | DEBUG | llmcompressor.core.lifecycle:event:205 - Updated event with modifier: config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False +2026-04-15 15:48:32.406 | DEBUG | llmcompressor.core.lifecycle:event:195 - Handling event: EventType.SEQUENTIAL_EPOCH_END +2026-04-15 15:48:32.406 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.8.self_attn.q_proj using 128 samples +2026-04-15 15:48:35.824 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.42s +2026-04-15 15:48:35.825 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 820.72 +2026-04-15 15:48:35.825 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:48:35.825 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 69.86% | total memory: 25 GB +2026-04-15 15:48:35.825 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 53.05344 MB +2026-04-15 15:48:35.826 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.8.self_attn.k_proj using 128 samples +2026-04-15 15:48:39.097 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.27s +2026-04-15 15:48:39.097 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 247.05 +2026-04-15 15:48:39.098 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:48:39.098 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 69.86% | total memory: 25 GB +2026-04-15 15:48:39.098 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 10.610688 MB +2026-04-15 15:48:39.098 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.8.self_attn.v_proj using 128 samples +2026-04-15 15:48:42.403 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.30s +2026-04-15 15:48:42.403 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 141.44 +2026-04-15 15:48:42.404 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:48:42.404 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 69.86% | total memory: 25 GB +2026-04-15 15:48:42.404 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 10.610688 MB +2026-04-15 15:48:42.404 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.8.self_attn.o_proj using 128 samples +2026-04-15 15:48:45.818 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.41s +2026-04-15 15:48:45.818 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 70.22 +2026-04-15 15:48:45.819 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:48:45.819 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 69.86% | total memory: 25 GB +2026-04-15 15:48:45.819 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 53.0432 MB +2026-04-15 15:48:45.819 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.8.mlp.gate_proj using 128 samples +2026-04-15 15:48:49.533 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.71s +2026-04-15 15:48:49.533 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 5702.51 +2026-04-15 15:48:49.534 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:48:49.534 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 69.86% | total memory: 25 GB +2026-04-15 15:48:49.534 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 15:48:49.534 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.8.mlp.up_proj using 128 samples +2026-04-15 15:48:53.229 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.69s +2026-04-15 15:48:53.229 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 3339.95 +2026-04-15 15:48:53.230 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:48:53.230 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 69.86% | total memory: 25 GB +2026-04-15 15:48:53.230 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 15:48:53.230 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.8.mlp.down_proj using 128 samples +2026-04-15 15:49:02.798 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 9.57s +2026-04-15 15:49:02.798 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 238.11 +2026-04-15 15:49:02.799 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:49:02.799 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 75.80% | total memory: 25 GB +2026-04-15 15:49:02.799 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 15:49:02.799 | DEBUG | llmcompressor.core.lifecycle:event:205 - Updated event with modifier: config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False +2026-04-15 15:49:06.021 | DEBUG | llmcompressor.core.lifecycle:event:195 - Handling event: EventType.SEQUENTIAL_EPOCH_END +2026-04-15 15:49:06.021 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.9.self_attn.q_proj using 128 samples +2026-04-15 15:49:09.407 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.39s +2026-04-15 15:49:09.408 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 857.54 +2026-04-15 15:49:09.408 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:49:09.408 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 69.86% | total memory: 25 GB +2026-04-15 15:49:09.408 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 53.05344 MB +2026-04-15 15:49:09.409 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.9.self_attn.k_proj using 128 samples +2026-04-15 15:49:12.672 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.26s +2026-04-15 15:49:12.673 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 247.76 +2026-04-15 15:49:12.673 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:49:12.673 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 69.86% | total memory: 25 GB +2026-04-15 15:49:12.673 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 10.610688 MB +2026-04-15 15:49:12.673 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.9.self_attn.v_proj using 128 samples +2026-04-15 15:49:15.939 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.27s +2026-04-15 15:49:15.940 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 157.81 +2026-04-15 15:49:15.940 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:49:15.940 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 69.86% | total memory: 25 GB +2026-04-15 15:49:15.940 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 10.610688 MB +2026-04-15 15:49:15.941 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.9.self_attn.o_proj using 128 samples +2026-04-15 15:49:19.323 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.38s +2026-04-15 15:49:19.323 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 134.13 +2026-04-15 15:49:19.324 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:49:19.324 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 69.86% | total memory: 25 GB +2026-04-15 15:49:19.324 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 53.0432 MB +2026-04-15 15:49:19.324 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.9.mlp.gate_proj using 128 samples +2026-04-15 15:49:22.975 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.65s +2026-04-15 15:49:22.975 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 2739.85 +2026-04-15 15:49:22.975 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:49:22.975 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 69.86% | total memory: 25 GB +2026-04-15 15:49:22.976 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 15:49:22.976 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.9.mlp.up_proj using 128 samples +2026-04-15 15:49:26.597 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.62s +2026-04-15 15:49:26.597 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 2445.99 +2026-04-15 15:49:26.597 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:49:26.597 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 70.91% | total memory: 25 GB +2026-04-15 15:49:26.597 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 15:49:26.598 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.9.mlp.down_proj using 128 samples +2026-04-15 15:49:36.092 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 9.49s +2026-04-15 15:49:36.093 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 366.06 +2026-04-15 15:49:36.093 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:49:36.093 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 76.85% | total memory: 25 GB +2026-04-15 15:49:36.093 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 15:49:36.094 | DEBUG | llmcompressor.core.lifecycle:event:205 - Updated event with modifier: config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False +2026-04-15 15:49:39.460 | DEBUG | llmcompressor.core.lifecycle:event:195 - Handling event: EventType.SEQUENTIAL_EPOCH_END +2026-04-15 15:49:39.461 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.10.self_attn.q_proj using 128 samples +2026-04-15 15:49:42.883 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.42s +2026-04-15 15:49:42.884 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 1392.99 +2026-04-15 15:49:42.884 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:49:42.884 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 69.86% | total memory: 25 GB +2026-04-15 15:49:42.884 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 53.05344 MB +2026-04-15 15:49:42.885 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.10.self_attn.k_proj using 128 samples +2026-04-15 15:49:46.124 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.24s +2026-04-15 15:49:46.124 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 392.46 +2026-04-15 15:49:46.124 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:49:46.125 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 69.86% | total memory: 25 GB +2026-04-15 15:49:46.125 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 10.610688 MB +2026-04-15 15:49:46.125 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.10.self_attn.v_proj using 128 samples +2026-04-15 15:49:49.385 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.26s +2026-04-15 15:49:49.385 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 229.66 +2026-04-15 15:49:49.386 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:49:49.386 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 69.86% | total memory: 25 GB +2026-04-15 15:49:49.386 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 10.610688 MB +2026-04-15 15:49:49.386 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.10.self_attn.o_proj using 128 samples +2026-04-15 15:49:52.785 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.40s +2026-04-15 15:49:52.785 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 98.31 +2026-04-15 15:49:52.786 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:49:52.786 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 69.86% | total memory: 25 GB +2026-04-15 15:49:52.786 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 53.0432 MB +2026-04-15 15:49:52.786 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.10.mlp.gate_proj using 128 samples +2026-04-15 15:49:56.526 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.74s +2026-04-15 15:49:56.526 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 3103.97 +2026-04-15 15:49:56.527 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:49:56.527 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 69.86% | total memory: 25 GB +2026-04-15 15:49:56.527 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 15:49:56.527 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.10.mlp.up_proj using 128 samples +2026-04-15 15:50:00.256 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.73s +2026-04-15 15:50:00.256 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 2775.18 +2026-04-15 15:50:00.256 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:50:00.257 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 69.86% | total memory: 25 GB +2026-04-15 15:50:00.257 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 15:50:00.257 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.10.mlp.down_proj using 128 samples +2026-04-15 15:50:09.747 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 9.49s +2026-04-15 15:50:09.748 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 438.09 +2026-04-15 15:50:09.748 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:50:09.749 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 75.80% | total memory: 25 GB +2026-04-15 15:50:09.749 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 15:50:09.749 | DEBUG | llmcompressor.core.lifecycle:event:205 - Updated event with modifier: config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False +2026-04-15 15:50:13.015 | DEBUG | llmcompressor.core.lifecycle:event:195 - Handling event: EventType.SEQUENTIAL_EPOCH_END +2026-04-15 15:50:13.015 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.11.self_attn.q_proj using 128 samples +2026-04-15 15:50:16.396 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.38s +2026-04-15 15:50:16.397 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 932.91 +2026-04-15 15:50:16.397 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:50:16.397 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 69.86% | total memory: 25 GB +2026-04-15 15:50:16.397 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 53.05344 MB +2026-04-15 15:50:16.398 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.11.self_attn.k_proj using 128 samples +2026-04-15 15:50:19.618 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.22s +2026-04-15 15:50:19.619 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 260.85 +2026-04-15 15:50:19.619 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:50:19.619 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 69.86% | total memory: 25 GB +2026-04-15 15:50:19.619 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 10.610688 MB +2026-04-15 15:50:19.619 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.11.self_attn.v_proj using 128 samples +2026-04-15 15:50:22.867 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.25s +2026-04-15 15:50:22.867 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 150.46 +2026-04-15 15:50:22.867 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:50:22.867 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 69.86% | total memory: 25 GB +2026-04-15 15:50:22.867 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 10.610688 MB +2026-04-15 15:50:22.868 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.11.self_attn.o_proj using 128 samples +2026-04-15 15:50:26.232 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.36s +2026-04-15 15:50:26.233 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 186.73 +2026-04-15 15:50:26.233 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:50:26.233 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 69.86% | total memory: 25 GB +2026-04-15 15:50:26.233 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 53.0432 MB +2026-04-15 15:50:26.234 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.11.mlp.gate_proj using 128 samples +2026-04-15 15:50:29.913 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.68s +2026-04-15 15:50:29.913 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 3998.73 +2026-04-15 15:50:29.914 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:50:29.914 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 69.86% | total memory: 25 GB +2026-04-15 15:50:29.914 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 15:50:29.914 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.11.mlp.up_proj using 128 samples +2026-04-15 15:50:33.624 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.71s +2026-04-15 15:50:33.624 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 3178.06 +2026-04-15 15:50:33.625 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:50:33.625 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 70.91% | total memory: 25 GB +2026-04-15 15:50:33.625 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 15:50:33.625 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.11.mlp.down_proj using 128 samples +2026-04-15 15:50:43.307 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 9.68s +2026-04-15 15:50:43.307 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 413.78 +2026-04-15 15:50:43.308 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:50:43.308 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 76.85% | total memory: 25 GB +2026-04-15 15:50:43.308 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 15:50:43.308 | DEBUG | llmcompressor.core.lifecycle:event:205 - Updated event with modifier: config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False +2026-04-15 15:50:46.558 | DEBUG | llmcompressor.core.lifecycle:event:195 - Handling event: EventType.SEQUENTIAL_EPOCH_END +2026-04-15 15:50:46.558 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.12.self_attn.q_proj using 128 samples +2026-04-15 15:50:49.945 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.39s +2026-04-15 15:50:49.946 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 1199.40 +2026-04-15 15:50:49.946 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:50:49.946 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 69.86% | total memory: 25 GB +2026-04-15 15:50:49.947 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 53.05344 MB +2026-04-15 15:50:49.947 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.12.self_attn.k_proj using 128 samples +2026-04-15 15:50:53.182 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.23s +2026-04-15 15:50:53.182 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 335.82 +2026-04-15 15:50:53.182 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:50:53.182 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 69.86% | total memory: 25 GB +2026-04-15 15:50:53.183 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 10.610688 MB +2026-04-15 15:50:53.183 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.12.self_attn.v_proj using 128 samples +2026-04-15 15:50:56.402 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.22s +2026-04-15 15:50:56.402 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 192.34 +2026-04-15 15:50:56.402 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:50:56.402 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 69.86% | total memory: 25 GB +2026-04-15 15:50:56.402 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 10.610688 MB +2026-04-15 15:50:56.402 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.12.self_attn.o_proj using 128 samples +2026-04-15 15:50:59.795 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.39s +2026-04-15 15:50:59.796 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 206.51 +2026-04-15 15:50:59.796 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:50:59.796 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 69.86% | total memory: 25 GB +2026-04-15 15:50:59.796 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 53.0432 MB +2026-04-15 15:50:59.796 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.12.mlp.gate_proj using 128 samples +2026-04-15 15:51:03.463 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.67s +2026-04-15 15:51:03.463 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 3764.56 +2026-04-15 15:51:03.463 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:51:03.464 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 69.86% | total memory: 25 GB +2026-04-15 15:51:03.464 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 15:51:03.464 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.12.mlp.up_proj using 128 samples +2026-04-15 15:51:07.075 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.61s +2026-04-15 15:51:07.076 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 3526.55 +2026-04-15 15:51:07.076 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:51:07.076 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 69.86% | total memory: 25 GB +2026-04-15 15:51:07.076 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 15:51:07.077 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.12.mlp.down_proj using 128 samples +2026-04-15 15:51:16.474 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 9.40s +2026-04-15 15:51:16.475 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 530.73 +2026-04-15 15:51:16.475 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:51:16.475 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 75.80% | total memory: 25 GB +2026-04-15 15:51:16.475 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 15:51:16.476 | DEBUG | llmcompressor.core.lifecycle:event:205 - Updated event with modifier: config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False +2026-04-15 15:51:19.740 | DEBUG | llmcompressor.core.lifecycle:event:195 - Handling event: EventType.SEQUENTIAL_EPOCH_END +2026-04-15 15:51:19.741 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.13.self_attn.q_proj using 128 samples +2026-04-15 15:51:23.135 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.39s +2026-04-15 15:51:23.135 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 1284.99 +2026-04-15 15:51:23.135 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:51:23.136 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 69.86% | total memory: 25 GB +2026-04-15 15:51:23.136 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 53.05344 MB +2026-04-15 15:51:23.136 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.13.self_attn.k_proj using 128 samples +2026-04-15 15:51:26.382 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.25s +2026-04-15 15:51:26.382 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 364.14 +2026-04-15 15:51:26.383 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:51:26.383 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 69.86% | total memory: 25 GB +2026-04-15 15:51:26.383 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 10.610688 MB +2026-04-15 15:51:26.383 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.13.self_attn.v_proj using 128 samples +2026-04-15 15:51:29.623 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.24s +2026-04-15 15:51:29.624 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 233.06 +2026-04-15 15:51:29.624 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:51:29.624 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 69.86% | total memory: 25 GB +2026-04-15 15:51:29.624 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 10.610688 MB +2026-04-15 15:51:29.624 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.13.self_attn.o_proj using 128 samples +2026-04-15 15:51:33.045 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.42s +2026-04-15 15:51:33.045 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 212.39 +2026-04-15 15:51:33.046 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:51:33.046 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 69.86% | total memory: 25 GB +2026-04-15 15:51:33.046 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 53.0432 MB +2026-04-15 15:51:33.046 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.13.mlp.gate_proj using 128 samples +2026-04-15 15:51:36.721 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.67s +2026-04-15 15:51:36.721 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 4320.32 +2026-04-15 15:51:36.722 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:51:36.722 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 69.86% | total memory: 25 GB +2026-04-15 15:51:36.722 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 15:51:36.722 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.13.mlp.up_proj using 128 samples +2026-04-15 15:51:40.406 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.68s +2026-04-15 15:51:40.406 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 4043.76 +2026-04-15 15:51:40.406 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:51:40.407 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 70.91% | total memory: 25 GB +2026-04-15 15:51:40.407 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 15:51:40.407 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.13.mlp.down_proj using 128 samples +2026-04-15 15:51:50.021 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 9.61s +2026-04-15 15:51:50.022 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 609.84 +2026-04-15 15:51:50.022 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:51:50.022 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 76.85% | total memory: 25 GB +2026-04-15 15:51:50.022 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 15:51:50.023 | DEBUG | llmcompressor.core.lifecycle:event:205 - Updated event with modifier: config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False +2026-04-15 15:51:53.317 | DEBUG | llmcompressor.core.lifecycle:event:195 - Handling event: EventType.SEQUENTIAL_EPOCH_END +2026-04-15 15:51:53.318 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.14.self_attn.q_proj using 128 samples +2026-04-15 15:51:56.726 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.41s +2026-04-15 15:51:56.726 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 1613.83 +2026-04-15 15:51:56.727 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:51:56.727 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 69.86% | total memory: 25 GB +2026-04-15 15:51:56.727 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 53.05344 MB +2026-04-15 15:51:56.727 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.14.self_attn.k_proj using 128 samples +2026-04-15 15:52:00.004 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.28s +2026-04-15 15:52:00.004 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 467.33 +2026-04-15 15:52:00.004 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:52:00.005 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 69.86% | total memory: 25 GB +2026-04-15 15:52:00.005 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 10.610688 MB +2026-04-15 15:52:00.005 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.14.self_attn.v_proj using 128 samples +2026-04-15 15:52:03.299 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.29s +2026-04-15 15:52:03.300 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 247.08 +2026-04-15 15:52:03.300 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:52:03.300 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 69.86% | total memory: 25 GB +2026-04-15 15:52:03.300 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 10.610688 MB +2026-04-15 15:52:03.301 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.14.self_attn.o_proj using 128 samples +2026-04-15 15:52:06.721 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.42s +2026-04-15 15:52:06.721 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 213.98 +2026-04-15 15:52:06.721 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:52:06.721 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 69.86% | total memory: 25 GB +2026-04-15 15:52:06.722 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 53.0432 MB +2026-04-15 15:52:06.722 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.14.mlp.gate_proj using 128 samples +2026-04-15 15:52:10.523 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.80s +2026-04-15 15:52:10.524 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 4363.41 +2026-04-15 15:52:10.524 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:52:10.524 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 69.86% | total memory: 25 GB +2026-04-15 15:52:10.524 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 15:52:10.524 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.14.mlp.up_proj using 128 samples +2026-04-15 15:52:14.442 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.92s +2026-04-15 15:52:14.442 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 4105.87 +2026-04-15 15:52:14.443 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:52:14.443 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 69.86% | total memory: 25 GB +2026-04-15 15:52:14.443 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 15:52:14.443 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.14.mlp.down_proj using 128 samples +2026-04-15 15:52:24.529 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 10.09s +2026-04-15 15:52:24.530 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 635.41 +2026-04-15 15:52:24.530 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:52:24.530 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 21.38% | total memory: 25 GB +2026-04-15 15:52:24.530 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 15:52:24.531 | DEBUG | llmcompressor.core.lifecycle:event:205 - Updated event with modifier: config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False +2026-04-15 15:52:27.752 | DEBUG | llmcompressor.core.lifecycle:event:195 - Handling event: EventType.SEQUENTIAL_EPOCH_END +2026-04-15 15:52:27.753 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.15.self_attn.q_proj using 128 samples +2026-04-15 15:52:31.269 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.52s +2026-04-15 15:52:31.270 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 1428.68 +2026-04-15 15:52:31.270 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:52:31.270 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 15:52:31.270 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 53.05344 MB +2026-04-15 15:52:31.271 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.15.self_attn.k_proj using 128 samples +2026-04-15 15:52:34.632 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.36s +2026-04-15 15:52:34.632 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 395.67 +2026-04-15 15:52:34.632 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:52:34.632 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 15:52:34.632 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 10.610688 MB +2026-04-15 15:52:34.633 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.15.self_attn.v_proj using 128 samples +2026-04-15 15:52:37.981 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.35s +2026-04-15 15:52:37.981 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 273.60 +2026-04-15 15:52:37.982 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:52:37.982 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 15:52:37.982 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 10.610688 MB +2026-04-15 15:52:37.982 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.15.self_attn.o_proj using 128 samples +2026-04-15 15:52:41.492 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.51s +2026-04-15 15:52:41.492 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 257.63 +2026-04-15 15:52:41.492 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:52:41.493 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 15:52:41.493 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 53.0432 MB +2026-04-15 15:52:41.493 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.15.mlp.gate_proj using 128 samples +2026-04-15 15:52:45.297 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.80s +2026-04-15 15:52:45.298 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 4574.48 +2026-04-15 15:52:45.298 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:52:45.298 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 15:52:45.298 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 15:52:45.298 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.15.mlp.up_proj using 128 samples +2026-04-15 15:52:49.109 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.81s +2026-04-15 15:52:49.110 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 4149.50 +2026-04-15 15:52:49.110 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:52:49.110 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 16.49% | total memory: 25 GB +2026-04-15 15:52:49.110 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 15:52:49.110 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.15.mlp.down_proj using 128 samples +2026-04-15 15:52:59.019 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 9.91s +2026-04-15 15:52:59.019 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 640.84 +2026-04-15 15:52:59.020 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:52:59.020 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 22.43% | total memory: 25 GB +2026-04-15 15:52:59.020 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 15:52:59.020 | DEBUG | llmcompressor.core.lifecycle:event:205 - Updated event with modifier: config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False +2026-04-15 15:53:02.434 | DEBUG | llmcompressor.core.lifecycle:event:195 - Handling event: EventType.SEQUENTIAL_EPOCH_END +2026-04-15 15:53:02.434 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.16.self_attn.q_proj using 128 samples +2026-04-15 15:53:05.940 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.51s +2026-04-15 15:53:05.940 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 1233.65 +2026-04-15 15:53:05.940 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:53:05.940 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 15:53:05.940 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 53.05344 MB +2026-04-15 15:53:05.941 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.16.self_attn.k_proj using 128 samples +2026-04-15 15:53:09.319 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.38s +2026-04-15 15:53:09.320 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 386.13 +2026-04-15 15:53:09.320 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:53:09.320 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 15:53:09.320 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 10.610688 MB +2026-04-15 15:53:09.321 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.16.self_attn.v_proj using 128 samples +2026-04-15 15:53:12.707 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.39s +2026-04-15 15:53:12.707 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 184.10 +2026-04-15 15:53:12.707 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:53:12.707 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 15:53:12.707 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 10.610688 MB +2026-04-15 15:53:12.708 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.16.self_attn.o_proj using 128 samples +2026-04-15 15:53:16.248 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.54s +2026-04-15 15:53:16.249 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 211.13 +2026-04-15 15:53:16.249 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:53:16.249 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 15:53:16.250 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 53.0432 MB +2026-04-15 15:53:16.250 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.16.mlp.gate_proj using 128 samples +2026-04-15 15:53:20.126 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.88s +2026-04-15 15:53:20.126 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 3809.93 +2026-04-15 15:53:20.127 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:53:20.127 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 15:53:20.127 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 15:53:20.127 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.16.mlp.up_proj using 128 samples +2026-04-15 15:53:23.944 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.82s +2026-04-15 15:53:23.945 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 3741.67 +2026-04-15 15:53:23.945 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:53:23.945 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 15:53:23.945 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 15:53:23.946 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.16.mlp.down_proj using 128 samples +2026-04-15 15:53:33.872 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 9.93s +2026-04-15 15:53:33.872 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 622.10 +2026-04-15 15:53:33.873 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:53:33.873 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 21.38% | total memory: 25 GB +2026-04-15 15:53:33.873 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 15:53:33.873 | DEBUG | llmcompressor.core.lifecycle:event:205 - Updated event with modifier: config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False +2026-04-15 15:53:37.203 | DEBUG | llmcompressor.core.lifecycle:event:195 - Handling event: EventType.SEQUENTIAL_EPOCH_END +2026-04-15 15:53:37.204 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.17.self_attn.q_proj using 128 samples +2026-04-15 15:53:40.735 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.53s +2026-04-15 15:53:40.735 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 1419.22 +2026-04-15 15:53:40.736 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:53:40.736 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 15:53:40.736 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 53.05344 MB +2026-04-15 15:53:40.737 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.17.self_attn.k_proj using 128 samples +2026-04-15 15:53:44.106 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.37s +2026-04-15 15:53:44.106 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 411.44 +2026-04-15 15:53:44.106 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:53:44.106 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 15:53:44.107 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 10.610688 MB +2026-04-15 15:53:44.107 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.17.self_attn.v_proj using 128 samples +2026-04-15 15:53:47.482 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.37s +2026-04-15 15:53:47.483 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 197.82 +2026-04-15 15:53:47.483 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:53:47.484 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 15:53:47.484 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 10.610688 MB +2026-04-15 15:53:47.484 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.17.self_attn.o_proj using 128 samples +2026-04-15 15:53:50.920 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.44s +2026-04-15 15:53:50.921 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 226.67 +2026-04-15 15:53:50.921 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:53:50.921 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 15:53:50.921 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 53.0432 MB +2026-04-15 15:53:50.921 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.17.mlp.gate_proj using 128 samples +2026-04-15 15:53:54.551 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.63s +2026-04-15 15:53:54.551 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 3560.58 +2026-04-15 15:53:54.552 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:53:54.552 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 15:53:54.552 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 15:53:54.552 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.17.mlp.up_proj using 128 samples +2026-04-15 15:53:58.343 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.79s +2026-04-15 15:53:58.343 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 3611.33 +2026-04-15 15:53:58.343 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:53:58.343 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 16.49% | total memory: 25 GB +2026-04-15 15:53:58.344 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 15:53:58.344 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.17.mlp.down_proj using 128 samples +2026-04-15 15:54:08.103 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 9.76s +2026-04-15 15:54:08.104 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 597.52 +2026-04-15 15:54:08.104 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:54:08.104 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 22.43% | total memory: 25 GB +2026-04-15 15:54:08.104 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 15:54:08.105 | DEBUG | llmcompressor.core.lifecycle:event:205 - Updated event with modifier: config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False +2026-04-15 15:54:11.322 | DEBUG | llmcompressor.core.lifecycle:event:195 - Handling event: EventType.SEQUENTIAL_EPOCH_END +2026-04-15 15:54:11.323 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.18.self_attn.q_proj using 128 samples +2026-04-15 15:54:14.801 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.48s +2026-04-15 15:54:14.802 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 1356.98 +2026-04-15 15:54:14.802 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:54:14.802 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 15:54:14.803 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 53.05344 MB +2026-04-15 15:54:14.803 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.18.self_attn.k_proj using 128 samples +2026-04-15 15:54:18.128 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.32s +2026-04-15 15:54:18.128 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 389.99 +2026-04-15 15:54:18.128 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:54:18.128 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 15:54:18.129 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 10.610688 MB +2026-04-15 15:54:18.129 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.18.self_attn.v_proj using 128 samples +2026-04-15 15:54:21.439 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.31s +2026-04-15 15:54:21.439 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 220.06 +2026-04-15 15:54:21.440 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:54:21.440 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 15:54:21.440 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 10.610688 MB +2026-04-15 15:54:21.441 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.18.self_attn.o_proj using 128 samples +2026-04-15 15:54:24.938 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.50s +2026-04-15 15:54:24.938 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 253.21 +2026-04-15 15:54:24.939 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:54:24.939 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 15:54:24.939 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 53.0432 MB +2026-04-15 15:54:24.940 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.18.mlp.gate_proj using 128 samples +2026-04-15 15:54:28.731 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.79s +2026-04-15 15:54:28.731 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 3409.22 +2026-04-15 15:54:28.731 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:54:28.731 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 15:54:28.731 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 15:54:28.732 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.18.mlp.up_proj using 128 samples +2026-04-15 15:54:32.544 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.81s +2026-04-15 15:54:32.545 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 3566.92 +2026-04-15 15:54:32.545 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:54:32.545 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 15:54:32.545 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 15:54:32.545 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.18.mlp.down_proj using 128 samples +2026-04-15 15:54:42.172 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 9.63s +2026-04-15 15:54:42.173 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 603.96 +2026-04-15 15:54:42.173 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:54:42.173 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 21.38% | total memory: 25 GB +2026-04-15 15:54:42.174 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 15:54:42.174 | DEBUG | llmcompressor.core.lifecycle:event:205 - Updated event with modifier: config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False +2026-04-15 15:54:45.411 | DEBUG | llmcompressor.core.lifecycle:event:195 - Handling event: EventType.SEQUENTIAL_EPOCH_END +2026-04-15 15:54:45.411 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.19.self_attn.q_proj using 128 samples +2026-04-15 15:54:48.893 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.48s +2026-04-15 15:54:48.893 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 1598.55 +2026-04-15 15:54:48.894 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:54:48.894 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 15:54:48.894 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 53.05344 MB +2026-04-15 15:54:48.895 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.19.self_attn.k_proj using 128 samples +2026-04-15 15:54:52.194 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.30s +2026-04-15 15:54:52.195 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 428.54 +2026-04-15 15:54:52.195 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:54:52.195 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 15:54:52.195 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 10.610688 MB +2026-04-15 15:54:52.195 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.19.self_attn.v_proj using 128 samples +2026-04-15 15:54:55.499 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.30s +2026-04-15 15:54:55.500 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 234.31 +2026-04-15 15:54:55.500 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:54:55.500 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 15:54:55.500 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 10.610688 MB +2026-04-15 15:54:55.500 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.19.self_attn.o_proj using 128 samples +2026-04-15 15:54:58.999 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.50s +2026-04-15 15:54:59.000 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 193.04 +2026-04-15 15:54:59.000 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:54:59.000 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 15:54:59.001 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 53.0432 MB +2026-04-15 15:54:59.001 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.19.mlp.gate_proj using 128 samples +2026-04-15 15:55:02.867 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.87s +2026-04-15 15:55:02.867 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 3512.02 +2026-04-15 15:55:02.868 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:55:02.868 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 15:55:02.868 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 15:55:02.868 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.19.mlp.up_proj using 128 samples +2026-04-15 15:55:06.731 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.86s +2026-04-15 15:55:06.732 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 3678.19 +2026-04-15 15:55:06.732 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:55:06.732 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 16.49% | total memory: 25 GB +2026-04-15 15:55:06.732 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 15:55:06.732 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.19.mlp.down_proj using 128 samples +2026-04-15 15:55:16.526 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 9.79s +2026-04-15 15:55:16.526 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 635.19 +2026-04-15 15:55:16.527 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:55:16.527 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 22.43% | total memory: 25 GB +2026-04-15 15:55:16.527 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 15:55:16.528 | DEBUG | llmcompressor.core.lifecycle:event:205 - Updated event with modifier: config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False +2026-04-15 15:55:19.728 | DEBUG | llmcompressor.core.lifecycle:event:195 - Handling event: EventType.SEQUENTIAL_EPOCH_END +2026-04-15 15:55:19.728 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.20.self_attn.q_proj using 128 samples +2026-04-15 15:55:23.195 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.47s +2026-04-15 15:55:23.196 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 1529.32 +2026-04-15 15:55:23.197 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:55:23.197 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 15:55:23.197 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 53.05344 MB +2026-04-15 15:55:23.197 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.20.self_attn.k_proj using 128 samples +2026-04-15 15:55:26.519 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.32s +2026-04-15 15:55:26.519 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 441.86 +2026-04-15 15:55:26.519 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:55:26.520 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 15:55:26.520 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 10.610688 MB +2026-04-15 15:55:26.520 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.20.self_attn.v_proj using 128 samples +2026-04-15 15:55:29.828 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.31s +2026-04-15 15:55:29.829 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 216.74 +2026-04-15 15:55:29.829 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:55:29.829 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 15:55:29.829 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 10.610688 MB +2026-04-15 15:55:29.830 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.20.self_attn.o_proj using 128 samples +2026-04-15 15:55:33.292 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.46s +2026-04-15 15:55:33.293 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 258.15 +2026-04-15 15:55:33.294 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:55:33.294 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 15:55:33.295 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 53.0432 MB +2026-04-15 15:55:33.295 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.20.mlp.gate_proj using 128 samples +2026-04-15 15:55:37.074 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.78s +2026-04-15 15:55:37.074 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 3218.17 +2026-04-15 15:55:37.074 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:55:37.074 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 15:55:37.074 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 15:55:37.075 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.20.mlp.up_proj using 128 samples +2026-04-15 15:55:40.906 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.83s +2026-04-15 15:55:40.906 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 3557.13 +2026-04-15 15:55:40.907 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:55:40.907 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 15:55:40.907 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 15:55:40.908 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.20.mlp.down_proj using 128 samples +2026-04-15 15:55:50.637 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 9.73s +2026-04-15 15:55:50.638 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 656.90 +2026-04-15 15:55:50.638 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:55:50.639 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 21.38% | total memory: 25 GB +2026-04-15 15:55:50.639 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 15:55:50.639 | DEBUG | llmcompressor.core.lifecycle:event:205 - Updated event with modifier: config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False +2026-04-15 15:55:53.860 | DEBUG | llmcompressor.core.lifecycle:event:195 - Handling event: EventType.SEQUENTIAL_EPOCH_END +2026-04-15 15:55:53.860 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.21.self_attn.q_proj using 128 samples +2026-04-15 15:55:57.376 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.52s +2026-04-15 15:55:57.377 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 1420.07 +2026-04-15 15:55:57.377 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:55:57.377 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 15:55:57.378 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 53.05344 MB +2026-04-15 15:55:57.378 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.21.self_attn.k_proj using 128 samples +2026-04-15 15:56:00.698 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.32s +2026-04-15 15:56:00.699 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 410.86 +2026-04-15 15:56:00.699 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:56:00.699 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 15:56:00.699 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 10.610688 MB +2026-04-15 15:56:00.700 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.21.self_attn.v_proj using 128 samples +2026-04-15 15:56:04.037 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.34s +2026-04-15 15:56:04.037 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 196.97 +2026-04-15 15:56:04.037 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:56:04.037 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 15:56:04.038 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 10.610688 MB +2026-04-15 15:56:04.038 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.21.self_attn.o_proj using 128 samples +2026-04-15 15:56:07.521 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.48s +2026-04-15 15:56:07.521 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 271.23 +2026-04-15 15:56:07.522 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:56:07.522 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 15:56:07.522 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 53.0432 MB +2026-04-15 15:56:07.523 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.21.mlp.gate_proj using 128 samples +2026-04-15 15:56:11.342 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.82s +2026-04-15 15:56:11.342 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 3193.73 +2026-04-15 15:56:11.342 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:56:11.342 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 15:56:11.342 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 15:56:11.343 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.21.mlp.up_proj using 128 samples +2026-04-15 15:56:15.197 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.85s +2026-04-15 15:56:15.197 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 3543.95 +2026-04-15 15:56:15.197 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:56:15.197 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 16.49% | total memory: 25 GB +2026-04-15 15:56:15.197 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 15:56:15.198 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.21.mlp.down_proj using 128 samples +2026-04-15 15:56:25.030 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 9.83s +2026-04-15 15:56:25.031 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 635.42 +2026-04-15 15:56:25.031 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:56:25.032 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 22.43% | total memory: 25 GB +2026-04-15 15:56:25.032 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 15:56:25.032 | DEBUG | llmcompressor.core.lifecycle:event:205 - Updated event with modifier: config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False +2026-04-15 15:56:28.263 | DEBUG | llmcompressor.core.lifecycle:event:195 - Handling event: EventType.SEQUENTIAL_EPOCH_END +2026-04-15 15:56:28.263 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.22.self_attn.q_proj using 128 samples +2026-04-15 15:56:31.766 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.50s +2026-04-15 15:56:31.767 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 1580.42 +2026-04-15 15:56:31.767 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:56:31.767 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 15:56:31.768 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 53.05344 MB +2026-04-15 15:56:31.768 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.22.self_attn.k_proj using 128 samples +2026-04-15 15:56:35.089 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.32s +2026-04-15 15:56:35.089 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 449.72 +2026-04-15 15:56:35.089 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:56:35.089 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 15:56:35.090 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 10.610688 MB +2026-04-15 15:56:35.090 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.22.self_attn.v_proj using 128 samples +2026-04-15 15:56:38.413 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.32s +2026-04-15 15:56:38.413 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 299.31 +2026-04-15 15:56:38.413 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:56:38.413 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 15:56:38.414 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 10.610688 MB +2026-04-15 15:56:38.414 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.22.self_attn.o_proj using 128 samples +2026-04-15 15:56:41.911 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.50s +2026-04-15 15:56:41.912 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 298.43 +2026-04-15 15:56:41.912 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:56:41.912 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 15:56:41.912 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 53.0432 MB +2026-04-15 15:56:41.913 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.22.mlp.gate_proj using 128 samples +2026-04-15 15:56:45.766 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.85s +2026-04-15 15:56:45.766 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 3244.10 +2026-04-15 15:56:45.767 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:56:45.767 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 15:56:45.767 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 15:56:45.767 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.22.mlp.up_proj using 128 samples +2026-04-15 15:56:49.592 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.82s +2026-04-15 15:56:49.592 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 3608.44 +2026-04-15 15:56:49.592 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:56:49.593 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 15:56:49.593 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 15:56:49.593 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.22.mlp.down_proj using 128 samples +2026-04-15 15:56:59.287 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 9.69s +2026-04-15 15:56:59.288 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 678.47 +2026-04-15 15:56:59.288 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:56:59.288 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 21.38% | total memory: 25 GB +2026-04-15 15:56:59.289 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 15:56:59.289 | DEBUG | llmcompressor.core.lifecycle:event:205 - Updated event with modifier: config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False +2026-04-15 15:57:02.497 | DEBUG | llmcompressor.core.lifecycle:event:195 - Handling event: EventType.SEQUENTIAL_EPOCH_END +2026-04-15 15:57:02.497 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.23.self_attn.q_proj using 128 samples +2026-04-15 15:57:05.942 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.44s +2026-04-15 15:57:05.943 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 1555.15 +2026-04-15 15:57:05.943 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:57:05.943 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 15:57:05.943 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 53.05344 MB +2026-04-15 15:57:05.944 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.23.self_attn.k_proj using 128 samples +2026-04-15 15:57:09.212 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.27s +2026-04-15 15:57:09.212 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 416.15 +2026-04-15 15:57:09.212 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:57:09.213 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 15:57:09.213 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 10.610688 MB +2026-04-15 15:57:09.213 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.23.self_attn.v_proj using 128 samples +2026-04-15 15:57:12.527 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.31s +2026-04-15 15:57:12.528 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 292.92 +2026-04-15 15:57:12.528 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:57:12.528 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 15:57:12.528 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 10.610688 MB +2026-04-15 15:57:12.529 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.23.self_attn.o_proj using 128 samples +2026-04-15 15:57:15.985 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.46s +2026-04-15 15:57:15.986 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 313.17 +2026-04-15 15:57:15.987 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:57:15.987 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 15:57:15.987 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 53.0432 MB +2026-04-15 15:57:15.987 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.23.mlp.gate_proj using 128 samples +2026-04-15 15:57:19.770 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.78s +2026-04-15 15:57:19.770 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 3443.85 +2026-04-15 15:57:19.771 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:57:19.771 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 15:57:19.771 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 15:57:19.771 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.23.mlp.up_proj using 128 samples +2026-04-15 15:57:23.585 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.81s +2026-04-15 15:57:23.585 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 3637.27 +2026-04-15 15:57:23.587 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:57:23.587 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 16.49% | total memory: 25 GB +2026-04-15 15:57:23.587 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 15:57:23.588 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.23.mlp.down_proj using 128 samples +2026-04-15 15:57:33.222 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 9.63s +2026-04-15 15:57:33.223 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 661.69 +2026-04-15 15:57:33.223 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:57:33.223 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 22.43% | total memory: 25 GB +2026-04-15 15:57:33.223 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 15:57:33.223 | DEBUG | llmcompressor.core.lifecycle:event:205 - Updated event with modifier: config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False +2026-04-15 15:57:36.454 | DEBUG | llmcompressor.core.lifecycle:event:195 - Handling event: EventType.SEQUENTIAL_EPOCH_END +2026-04-15 15:57:36.454 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.24.self_attn.q_proj using 128 samples +2026-04-15 15:57:39.807 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.35s +2026-04-15 15:57:39.808 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 1722.65 +2026-04-15 15:57:39.808 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:57:39.809 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 15:57:39.809 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 53.05344 MB +2026-04-15 15:57:39.809 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.24.self_attn.k_proj using 128 samples +2026-04-15 15:57:42.967 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.16s +2026-04-15 15:57:42.967 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 542.39 +2026-04-15 15:57:42.968 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:57:42.968 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 15:57:42.968 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 10.610688 MB +2026-04-15 15:57:42.968 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.24.self_attn.v_proj using 128 samples +2026-04-15 15:57:46.167 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.20s +2026-04-15 15:57:46.167 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 247.72 +2026-04-15 15:57:46.168 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:57:46.168 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 15:57:46.168 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 10.610688 MB +2026-04-15 15:57:46.168 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.24.self_attn.o_proj using 128 samples +2026-04-15 15:57:49.468 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.30s +2026-04-15 15:57:49.469 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 333.69 +2026-04-15 15:57:49.469 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:57:49.469 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 15:57:49.469 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 53.0432 MB +2026-04-15 15:57:49.470 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.24.mlp.gate_proj using 128 samples +2026-04-15 15:57:53.109 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.64s +2026-04-15 15:57:53.110 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 3220.40 +2026-04-15 15:57:53.110 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:57:53.110 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 15:57:53.111 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 15:57:53.111 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.24.mlp.up_proj using 128 samples +2026-04-15 15:57:56.739 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.63s +2026-04-15 15:57:56.739 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 3445.17 +2026-04-15 15:57:56.740 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:57:56.740 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 15:57:56.740 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 15:57:56.740 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.24.mlp.down_proj using 128 samples +2026-04-15 15:58:06.050 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 9.31s +2026-04-15 15:58:06.050 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 592.58 +2026-04-15 15:58:06.051 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:58:06.051 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 21.38% | total memory: 25 GB +2026-04-15 15:58:06.051 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 15:58:06.052 | DEBUG | llmcompressor.core.lifecycle:event:205 - Updated event with modifier: config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False +2026-04-15 15:58:09.215 | DEBUG | llmcompressor.core.lifecycle:event:195 - Handling event: EventType.SEQUENTIAL_EPOCH_END +2026-04-15 15:58:09.215 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.25.self_attn.q_proj using 128 samples +2026-04-15 15:58:12.491 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.28s +2026-04-15 15:58:12.491 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 1820.92 +2026-04-15 15:58:12.492 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:58:12.492 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 15:58:12.492 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 53.05344 MB +2026-04-15 15:58:12.492 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.25.self_attn.k_proj using 128 samples +2026-04-15 15:58:15.613 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.12s +2026-04-15 15:58:15.613 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 513.15 +2026-04-15 15:58:15.613 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:58:15.613 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 15:58:15.614 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 10.610688 MB +2026-04-15 15:58:15.614 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.25.self_attn.v_proj using 128 samples +2026-04-15 15:58:18.756 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.14s +2026-04-15 15:58:18.757 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 311.12 +2026-04-15 15:58:18.757 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:58:18.757 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 15:58:18.758 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 10.610688 MB +2026-04-15 15:58:18.758 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.25.self_attn.o_proj using 128 samples +2026-04-15 15:58:22.043 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.29s +2026-04-15 15:58:22.044 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 195.23 +2026-04-15 15:58:22.044 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:58:22.044 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 15:58:22.044 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 53.0432 MB +2026-04-15 15:58:22.044 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.25.mlp.gate_proj using 128 samples +2026-04-15 15:58:25.626 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.58s +2026-04-15 15:58:25.627 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 3175.35 +2026-04-15 15:58:25.628 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:58:25.628 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 15:58:25.628 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 15:58:25.629 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.25.mlp.up_proj using 128 samples +2026-04-15 15:58:29.226 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.60s +2026-04-15 15:58:29.226 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 3520.01 +2026-04-15 15:58:29.227 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:58:29.227 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 16.49% | total memory: 25 GB +2026-04-15 15:58:29.227 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 15:58:29.227 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.25.mlp.down_proj using 128 samples +2026-04-15 15:58:38.496 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 9.27s +2026-04-15 15:58:38.497 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 646.86 +2026-04-15 15:58:38.497 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:58:38.497 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 22.43% | total memory: 25 GB +2026-04-15 15:58:38.497 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 15:58:38.498 | DEBUG | llmcompressor.core.lifecycle:event:205 - Updated event with modifier: config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False +2026-04-15 15:58:41.672 | DEBUG | llmcompressor.core.lifecycle:event:195 - Handling event: EventType.SEQUENTIAL_EPOCH_END +2026-04-15 15:58:41.672 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.26.self_attn.q_proj using 128 samples +2026-04-15 15:58:45.047 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.37s +2026-04-15 15:58:45.048 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 1714.87 +2026-04-15 15:58:45.048 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:58:45.048 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 15:58:45.048 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 53.05344 MB +2026-04-15 15:58:45.049 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.26.self_attn.k_proj using 128 samples +2026-04-15 15:58:48.270 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.22s +2026-04-15 15:58:48.270 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 485.60 +2026-04-15 15:58:48.270 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:58:48.270 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 15:58:48.271 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 10.610688 MB +2026-04-15 15:58:48.271 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.26.self_attn.v_proj using 128 samples +2026-04-15 15:58:51.524 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.25s +2026-04-15 15:58:51.525 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 214.30 +2026-04-15 15:58:51.525 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:58:51.525 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 15:58:51.526 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 10.610688 MB +2026-04-15 15:58:51.526 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.26.self_attn.o_proj using 128 samples +2026-04-15 15:58:54.883 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.36s +2026-04-15 15:58:54.883 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 245.05 +2026-04-15 15:58:54.883 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:58:54.883 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 15:58:54.883 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 53.0432 MB +2026-04-15 15:58:54.884 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.26.mlp.gate_proj using 128 samples +2026-04-15 15:58:58.562 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.68s +2026-04-15 15:58:58.563 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 3363.41 +2026-04-15 15:58:58.563 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:58:58.563 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 15:58:58.564 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 15:58:58.564 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.26.mlp.up_proj using 128 samples +2026-04-15 15:59:02.207 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.64s +2026-04-15 15:59:02.208 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 3875.71 +2026-04-15 15:59:02.208 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:59:02.208 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 15:59:02.208 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 15:59:02.208 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.26.mlp.down_proj using 128 samples +2026-04-15 15:59:11.816 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 9.61s +2026-04-15 15:59:11.817 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 779.04 +2026-04-15 15:59:11.817 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:59:11.817 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 21.38% | total memory: 25 GB +2026-04-15 15:59:11.817 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 15:59:11.817 | DEBUG | llmcompressor.core.lifecycle:event:205 - Updated event with modifier: config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False +2026-04-15 15:59:15.027 | DEBUG | llmcompressor.core.lifecycle:event:195 - Handling event: EventType.SEQUENTIAL_EPOCH_END +2026-04-15 15:59:15.027 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.27.self_attn.q_proj using 128 samples +2026-04-15 15:59:18.411 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.38s +2026-04-15 15:59:18.411 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 1708.06 +2026-04-15 15:59:18.412 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:59:18.412 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 15:59:18.412 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 53.05344 MB +2026-04-15 15:59:18.413 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.27.self_attn.k_proj using 128 samples +2026-04-15 15:59:21.637 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.22s +2026-04-15 15:59:21.638 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 509.08 +2026-04-15 15:59:21.638 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:59:21.638 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 15:59:21.638 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 10.610688 MB +2026-04-15 15:59:21.639 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.27.self_attn.v_proj using 128 samples +2026-04-15 15:59:24.862 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.22s +2026-04-15 15:59:24.863 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 270.57 +2026-04-15 15:59:24.863 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:59:24.863 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 15:59:24.863 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 10.610688 MB +2026-04-15 15:59:24.863 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.27.self_attn.o_proj using 128 samples +2026-04-15 15:59:28.255 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.39s +2026-04-15 15:59:28.255 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 391.07 +2026-04-15 15:59:28.255 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:59:28.256 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 15:59:28.256 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 53.0432 MB +2026-04-15 15:59:28.256 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.27.mlp.gate_proj using 128 samples +2026-04-15 15:59:31.986 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.73s +2026-04-15 15:59:31.987 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 3563.34 +2026-04-15 15:59:31.987 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:59:31.987 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 15:59:31.987 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 15:59:31.988 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.27.mlp.up_proj using 128 samples +2026-04-15 15:59:35.689 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.70s +2026-04-15 15:59:35.689 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 4103.65 +2026-04-15 15:59:35.690 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:59:35.690 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 16.49% | total memory: 25 GB +2026-04-15 15:59:35.690 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 15:59:35.690 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.27.mlp.down_proj using 128 samples +2026-04-15 15:59:45.295 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 9.60s +2026-04-15 15:59:45.295 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 930.76 +2026-04-15 15:59:45.296 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:59:45.296 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 22.43% | total memory: 25 GB +2026-04-15 15:59:45.296 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 15:59:45.297 | DEBUG | llmcompressor.core.lifecycle:event:205 - Updated event with modifier: config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False +2026-04-15 15:59:48.507 | DEBUG | llmcompressor.core.lifecycle:event:195 - Handling event: EventType.SEQUENTIAL_EPOCH_END +2026-04-15 15:59:48.507 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.28.self_attn.q_proj using 128 samples +2026-04-15 15:59:51.919 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.41s +2026-04-15 15:59:51.920 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 1673.87 +2026-04-15 15:59:51.920 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:59:51.921 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 15:59:51.921 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 53.05344 MB +2026-04-15 15:59:51.921 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.28.self_attn.k_proj using 128 samples +2026-04-15 15:59:55.187 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.27s +2026-04-15 15:59:55.187 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 412.82 +2026-04-15 15:59:55.187 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:59:55.187 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 15:59:55.188 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 10.610688 MB +2026-04-15 15:59:55.188 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.28.self_attn.v_proj using 128 samples +2026-04-15 15:59:58.468 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.28s +2026-04-15 15:59:58.469 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 386.43 +2026-04-15 15:59:58.469 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 15:59:58.470 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 15:59:58.470 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 10.610688 MB +2026-04-15 15:59:58.470 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.28.self_attn.o_proj using 128 samples +2026-04-15 16:00:01.922 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.45s +2026-04-15 16:00:01.923 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 386.95 +2026-04-15 16:00:01.924 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:00:01.924 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:00:01.924 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 53.0432 MB +2026-04-15 16:00:01.924 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.28.mlp.gate_proj using 128 samples +2026-04-15 16:00:05.663 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.74s +2026-04-15 16:00:05.666 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 3527.68 +2026-04-15 16:00:05.666 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:00:05.666 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:00:05.666 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 16:00:05.667 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.28.mlp.up_proj using 128 samples +2026-04-15 16:00:09.428 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.76s +2026-04-15 16:00:09.429 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 4074.65 +2026-04-15 16:00:09.429 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:00:09.429 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:00:09.430 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 16:00:09.430 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.28.mlp.down_proj using 128 samples +2026-04-15 16:00:19.023 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 9.59s +2026-04-15 16:00:19.024 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 922.87 +2026-04-15 16:00:19.024 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:00:19.024 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 21.38% | total memory: 25 GB +2026-04-15 16:00:19.024 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 16:00:19.025 | DEBUG | llmcompressor.core.lifecycle:event:205 - Updated event with modifier: config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False +2026-04-15 16:00:22.225 | DEBUG | llmcompressor.core.lifecycle:event:195 - Handling event: EventType.SEQUENTIAL_EPOCH_END +2026-04-15 16:00:22.225 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.29.self_attn.q_proj using 128 samples +2026-04-15 16:00:25.660 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.43s +2026-04-15 16:00:25.661 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 1857.16 +2026-04-15 16:00:25.662 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:00:25.662 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:00:25.662 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 53.05344 MB +2026-04-15 16:00:25.662 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.29.self_attn.k_proj using 128 samples +2026-04-15 16:00:28.938 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.28s +2026-04-15 16:00:28.939 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 511.46 +2026-04-15 16:00:28.939 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:00:28.939 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:00:28.939 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 10.610688 MB +2026-04-15 16:00:28.939 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.29.self_attn.v_proj using 128 samples +2026-04-15 16:00:32.207 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.27s +2026-04-15 16:00:32.207 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 371.78 +2026-04-15 16:00:32.207 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:00:32.207 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:00:32.208 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 10.610688 MB +2026-04-15 16:00:32.208 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.29.self_attn.o_proj using 128 samples +2026-04-15 16:00:35.614 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.41s +2026-04-15 16:00:35.615 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 325.11 +2026-04-15 16:00:35.615 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:00:35.615 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:00:35.615 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 53.0432 MB +2026-04-15 16:00:35.616 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.29.mlp.gate_proj using 128 samples +2026-04-15 16:00:39.396 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.78s +2026-04-15 16:00:39.397 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 3521.71 +2026-04-15 16:00:39.397 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:00:39.397 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:00:39.397 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 16:00:39.397 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.29.mlp.up_proj using 128 samples +2026-04-15 16:00:43.168 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.77s +2026-04-15 16:00:43.169 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 4048.29 +2026-04-15 16:00:43.169 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:00:43.169 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 16.49% | total memory: 25 GB +2026-04-15 16:00:43.170 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 16:00:43.170 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.29.mlp.down_proj using 128 samples +2026-04-15 16:00:52.872 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 9.70s +2026-04-15 16:00:52.873 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 950.32 +2026-04-15 16:00:52.874 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:00:52.874 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 22.43% | total memory: 25 GB +2026-04-15 16:00:52.874 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 16:00:52.874 | DEBUG | llmcompressor.core.lifecycle:event:205 - Updated event with modifier: config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False +2026-04-15 16:00:56.096 | DEBUG | llmcompressor.core.lifecycle:event:195 - Handling event: EventType.SEQUENTIAL_EPOCH_END +2026-04-15 16:00:56.097 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.30.self_attn.q_proj using 128 samples +2026-04-15 16:00:59.536 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.44s +2026-04-15 16:00:59.537 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 1735.50 +2026-04-15 16:00:59.537 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:00:59.537 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:00:59.537 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 53.05344 MB +2026-04-15 16:00:59.537 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.30.self_attn.k_proj using 128 samples +2026-04-15 16:01:02.799 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.26s +2026-04-15 16:01:02.800 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 449.43 +2026-04-15 16:01:02.800 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:01:02.800 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:01:02.800 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 10.610688 MB +2026-04-15 16:01:02.801 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.30.self_attn.v_proj using 128 samples +2026-04-15 16:01:06.054 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.25s +2026-04-15 16:01:06.054 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 416.07 +2026-04-15 16:01:06.054 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:01:06.054 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:01:06.054 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 10.610688 MB +2026-04-15 16:01:06.055 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.30.self_attn.o_proj using 128 samples +2026-04-15 16:01:09.487 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.43s +2026-04-15 16:01:09.488 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 410.76 +2026-04-15 16:01:09.488 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:01:09.488 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:01:09.488 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 53.0432 MB +2026-04-15 16:01:09.489 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.30.mlp.gate_proj using 128 samples +2026-04-15 16:01:13.268 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.78s +2026-04-15 16:01:13.269 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 3684.13 +2026-04-15 16:01:13.269 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:01:13.269 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:01:13.270 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 16:01:13.270 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.30.mlp.up_proj using 128 samples +2026-04-15 16:01:17.005 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.74s +2026-04-15 16:01:17.006 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 4178.94 +2026-04-15 16:01:17.006 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:01:17.006 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:01:17.006 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 16:01:17.007 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.30.mlp.down_proj using 128 samples +2026-04-15 16:01:26.710 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 9.70s +2026-04-15 16:01:26.710 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 1028.78 +2026-04-15 16:01:26.711 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:01:26.711 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 21.38% | total memory: 25 GB +2026-04-15 16:01:26.711 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 16:01:26.711 | DEBUG | llmcompressor.core.lifecycle:event:205 - Updated event with modifier: config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False +2026-04-15 16:01:29.928 | DEBUG | llmcompressor.core.lifecycle:event:195 - Handling event: EventType.SEQUENTIAL_EPOCH_END +2026-04-15 16:01:29.928 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.31.self_attn.q_proj using 128 samples +2026-04-15 16:01:33.540 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.61s +2026-04-15 16:01:33.541 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 1779.48 +2026-04-15 16:01:33.542 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:01:33.542 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:01:33.542 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 53.05344 MB +2026-04-15 16:01:33.543 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.31.self_attn.k_proj using 128 samples +2026-04-15 16:01:36.814 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.27s +2026-04-15 16:01:36.814 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 454.29 +2026-04-15 16:01:36.814 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:01:36.814 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:01:36.814 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 10.610688 MB +2026-04-15 16:01:36.815 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.31.self_attn.v_proj using 128 samples +2026-04-15 16:01:40.096 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.28s +2026-04-15 16:01:40.096 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 395.05 +2026-04-15 16:01:40.096 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:01:40.097 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:01:40.097 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 10.610688 MB +2026-04-15 16:01:40.097 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.31.self_attn.o_proj using 128 samples +2026-04-15 16:01:43.534 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.44s +2026-04-15 16:01:43.535 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 395.06 +2026-04-15 16:01:43.535 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:01:43.536 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:01:43.536 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 53.0432 MB +2026-04-15 16:01:43.536 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.31.mlp.gate_proj using 128 samples +2026-04-15 16:01:47.360 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.82s +2026-04-15 16:01:47.360 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 4132.20 +2026-04-15 16:01:47.361 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:01:47.361 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:01:47.361 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 16:01:47.361 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.31.mlp.up_proj using 128 samples +2026-04-15 16:01:51.183 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.82s +2026-04-15 16:01:51.183 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 4571.82 +2026-04-15 16:01:51.184 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:01:51.184 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 16.49% | total memory: 25 GB +2026-04-15 16:01:51.184 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 16:01:51.185 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.31.mlp.down_proj using 128 samples +2026-04-15 16:02:00.847 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 9.66s +2026-04-15 16:02:00.848 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 1219.91 +2026-04-15 16:02:00.848 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:02:00.849 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 22.43% | total memory: 25 GB +2026-04-15 16:02:00.849 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 16:02:00.849 | DEBUG | llmcompressor.core.lifecycle:event:205 - Updated event with modifier: config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False +2026-04-15 16:02:04.060 | DEBUG | llmcompressor.core.lifecycle:event:195 - Handling event: EventType.SEQUENTIAL_EPOCH_END +2026-04-15 16:02:04.060 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.32.self_attn.q_proj using 128 samples +2026-04-15 16:02:07.480 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.42s +2026-04-15 16:02:07.481 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 1970.36 +2026-04-15 16:02:07.482 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:02:07.482 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:02:07.482 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 53.05344 MB +2026-04-15 16:02:07.482 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.32.self_attn.k_proj using 128 samples +2026-04-15 16:02:10.751 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.27s +2026-04-15 16:02:10.752 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 473.54 +2026-04-15 16:02:10.752 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:02:10.752 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:02:10.753 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 10.610688 MB +2026-04-15 16:02:10.753 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.32.self_attn.v_proj using 128 samples +2026-04-15 16:02:14.004 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.25s +2026-04-15 16:02:14.005 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 530.00 +2026-04-15 16:02:14.005 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:02:14.005 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:02:14.005 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 10.610688 MB +2026-04-15 16:02:14.006 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.32.self_attn.o_proj using 128 samples +2026-04-15 16:02:17.378 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.37s +2026-04-15 16:02:17.379 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 391.34 +2026-04-15 16:02:17.379 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:02:17.379 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:02:17.380 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 53.0432 MB +2026-04-15 16:02:17.380 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.32.mlp.gate_proj using 128 samples +2026-04-15 16:02:21.084 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.70s +2026-04-15 16:02:21.084 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 4389.55 +2026-04-15 16:02:21.084 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:02:21.084 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:02:21.084 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 16:02:21.085 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.32.mlp.up_proj using 128 samples +2026-04-15 16:02:24.801 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.72s +2026-04-15 16:02:24.802 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 4755.79 +2026-04-15 16:02:24.802 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:02:24.802 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:02:24.802 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 16:02:24.802 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.32.mlp.down_proj using 128 samples +2026-04-15 16:02:34.358 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 9.55s +2026-04-15 16:02:34.358 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 1371.41 +2026-04-15 16:02:34.359 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:02:34.359 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 21.38% | total memory: 25 GB +2026-04-15 16:02:34.359 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 16:02:34.359 | DEBUG | llmcompressor.core.lifecycle:event:205 - Updated event with modifier: config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False +2026-04-15 16:02:37.556 | DEBUG | llmcompressor.core.lifecycle:event:195 - Handling event: EventType.SEQUENTIAL_EPOCH_END +2026-04-15 16:02:37.557 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.33.self_attn.q_proj using 128 samples +2026-04-15 16:02:40.954 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.40s +2026-04-15 16:02:40.954 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 2048.45 +2026-04-15 16:02:40.955 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:02:40.955 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:02:40.955 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 53.05344 MB +2026-04-15 16:02:40.956 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.33.self_attn.k_proj using 128 samples +2026-04-15 16:02:44.164 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.21s +2026-04-15 16:02:44.164 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 487.90 +2026-04-15 16:02:44.164 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:02:44.164 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:02:44.165 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 10.610688 MB +2026-04-15 16:02:44.165 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.33.self_attn.v_proj using 128 samples +2026-04-15 16:02:47.388 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.22s +2026-04-15 16:02:47.389 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 500.24 +2026-04-15 16:02:47.389 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:02:47.389 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:02:47.390 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 10.610688 MB +2026-04-15 16:02:47.390 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.33.self_attn.o_proj using 128 samples +2026-04-15 16:02:50.755 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.36s +2026-04-15 16:02:50.755 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 484.31 +2026-04-15 16:02:50.756 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:02:50.756 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:02:50.756 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 53.0432 MB +2026-04-15 16:02:50.756 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.33.mlp.gate_proj using 128 samples +2026-04-15 16:02:54.440 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.68s +2026-04-15 16:02:54.440 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 5296.09 +2026-04-15 16:02:54.441 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:02:54.441 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:02:54.441 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 16:02:54.441 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.33.mlp.up_proj using 128 samples +2026-04-15 16:02:58.129 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.69s +2026-04-15 16:02:58.130 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 5635.88 +2026-04-15 16:02:58.130 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:02:58.130 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 16.49% | total memory: 25 GB +2026-04-15 16:02:58.131 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 16:02:58.131 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.33.mlp.down_proj using 128 samples +2026-04-15 16:03:07.663 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 9.53s +2026-04-15 16:03:07.664 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 1989.15 +2026-04-15 16:03:07.664 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:03:07.664 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 22.43% | total memory: 25 GB +2026-04-15 16:03:07.665 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 16:03:07.665 | DEBUG | llmcompressor.core.lifecycle:event:205 - Updated event with modifier: config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False +2026-04-15 16:03:10.894 | DEBUG | llmcompressor.core.lifecycle:event:195 - Handling event: EventType.SEQUENTIAL_EPOCH_END +2026-04-15 16:03:10.895 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.34.self_attn.q_proj using 128 samples +2026-04-15 16:03:14.269 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.37s +2026-04-15 16:03:14.270 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 2260.54 +2026-04-15 16:03:14.270 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:03:14.270 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:03:14.270 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 53.05344 MB +2026-04-15 16:03:14.270 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.34.self_attn.k_proj using 128 samples +2026-04-15 16:03:17.473 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.20s +2026-04-15 16:03:17.473 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 487.08 +2026-04-15 16:03:17.474 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:03:17.474 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:03:17.474 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 10.610688 MB +2026-04-15 16:03:17.475 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.34.self_attn.v_proj using 128 samples +2026-04-15 16:03:20.654 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.18s +2026-04-15 16:03:20.655 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 600.37 +2026-04-15 16:03:20.655 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:03:20.655 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:03:20.655 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 10.610688 MB +2026-04-15 16:03:20.655 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.34.self_attn.o_proj using 128 samples +2026-04-15 16:03:23.974 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.32s +2026-04-15 16:03:23.974 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 413.67 +2026-04-15 16:03:23.975 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:03:23.975 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:03:23.975 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 53.0432 MB +2026-04-15 16:03:23.975 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.34.mlp.gate_proj using 128 samples +2026-04-15 16:03:27.909 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.93s +2026-04-15 16:03:27.910 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 6272.23 +2026-04-15 16:03:27.910 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:03:27.910 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:03:27.911 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 16:03:27.911 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.34.mlp.up_proj using 128 samples +2026-04-15 16:03:31.572 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.66s +2026-04-15 16:03:31.572 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 6507.12 +2026-04-15 16:03:31.573 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:03:31.573 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:03:31.573 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 16:03:31.573 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.34.mlp.down_proj using 128 samples +2026-04-15 16:03:41.048 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 9.47s +2026-04-15 16:03:41.048 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 2401.74 +2026-04-15 16:03:41.049 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:03:41.049 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 21.38% | total memory: 25 GB +2026-04-15 16:03:41.049 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 16:03:41.049 | DEBUG | llmcompressor.core.lifecycle:event:205 - Updated event with modifier: config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False +2026-04-15 16:03:44.548 | DEBUG | llmcompressor.core.lifecycle:event:195 - Handling event: EventType.SEQUENTIAL_EPOCH_END +2026-04-15 16:03:44.548 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.35.self_attn.q_proj using 128 samples +2026-04-15 16:03:47.900 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.35s +2026-04-15 16:03:47.900 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 2067.54 +2026-04-15 16:03:47.901 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:03:47.901 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:03:47.901 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 53.05344 MB +2026-04-15 16:03:47.901 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.35.self_attn.k_proj using 128 samples +2026-04-15 16:03:51.102 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.20s +2026-04-15 16:03:51.103 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 497.91 +2026-04-15 16:03:51.103 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:03:51.103 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:03:51.103 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 10.610688 MB +2026-04-15 16:03:51.103 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.35.self_attn.v_proj using 128 samples +2026-04-15 16:03:54.282 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.18s +2026-04-15 16:03:54.283 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 539.01 +2026-04-15 16:03:54.283 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:03:54.283 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:03:54.283 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 10.610688 MB +2026-04-15 16:03:54.283 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.35.self_attn.o_proj using 128 samples +2026-04-15 16:03:57.617 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.33s +2026-04-15 16:03:57.618 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 881.93 +2026-04-15 16:03:57.618 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:03:57.619 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:03:57.619 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 53.0432 MB +2026-04-15 16:03:57.619 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.35.mlp.gate_proj using 128 samples +2026-04-15 16:04:01.279 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.66s +2026-04-15 16:04:01.279 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 7204.15 +2026-04-15 16:04:01.280 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:04:01.280 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:04:01.280 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 16:04:01.280 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.35.mlp.up_proj using 128 samples +2026-04-15 16:04:04.947 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.67s +2026-04-15 16:04:04.948 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 7342.12 +2026-04-15 16:04:04.948 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:04:04.948 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 16.49% | total memory: 25 GB +2026-04-15 16:04:04.948 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 16:04:04.948 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.35.mlp.down_proj using 128 samples +2026-04-15 16:04:14.461 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 9.51s +2026-04-15 16:04:14.462 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 3014.80 +2026-04-15 16:04:14.462 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:04:14.462 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 22.43% | total memory: 25 GB +2026-04-15 16:04:14.462 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 16:04:14.463 | DEBUG | llmcompressor.core.lifecycle:event:205 - Updated event with modifier: config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False +2026-04-15 16:04:17.934 | DEBUG | llmcompressor.core.lifecycle:event:195 - Handling event: EventType.SEQUENTIAL_EPOCH_END +2026-04-15 16:04:17.934 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.36.self_attn.q_proj using 128 samples +2026-04-15 16:04:21.324 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.39s +2026-04-15 16:04:21.325 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 2544.02 +2026-04-15 16:04:21.325 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:04:21.325 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:04:21.326 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 53.05344 MB +2026-04-15 16:04:21.326 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.36.self_attn.k_proj using 128 samples +2026-04-15 16:04:24.526 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.20s +2026-04-15 16:04:24.527 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 523.54 +2026-04-15 16:04:24.527 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:04:24.527 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:04:24.527 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 10.610688 MB +2026-04-15 16:04:24.527 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.36.self_attn.v_proj using 128 samples +2026-04-15 16:04:27.732 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.20s +2026-04-15 16:04:27.732 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 971.65 +2026-04-15 16:04:27.732 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:04:27.732 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:04:27.733 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 10.610688 MB +2026-04-15 16:04:27.733 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.36.self_attn.o_proj using 128 samples +2026-04-15 16:04:31.083 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.35s +2026-04-15 16:04:31.084 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 607.68 +2026-04-15 16:04:31.084 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:04:31.084 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:04:31.084 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 53.0432 MB +2026-04-15 16:04:31.085 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.36.mlp.gate_proj using 128 samples +2026-04-15 16:04:34.764 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.68s +2026-04-15 16:04:34.764 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 7853.79 +2026-04-15 16:04:34.764 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:04:34.764 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:04:34.765 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 16:04:34.765 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.36.mlp.up_proj using 128 samples +2026-04-15 16:04:38.387 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.62s +2026-04-15 16:04:38.388 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 8074.69 +2026-04-15 16:04:38.388 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:04:38.388 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:04:38.388 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 16:04:38.389 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.36.mlp.down_proj using 128 samples +2026-04-15 16:04:47.755 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 9.37s +2026-04-15 16:04:47.755 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 3707.10 +2026-04-15 16:04:47.756 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:04:47.756 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 21.38% | total memory: 25 GB +2026-04-15 16:04:47.756 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 16:04:47.756 | DEBUG | llmcompressor.core.lifecycle:event:205 - Updated event with modifier: config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False +2026-04-15 16:04:51.193 | DEBUG | llmcompressor.core.lifecycle:event:195 - Handling event: EventType.SEQUENTIAL_EPOCH_END +2026-04-15 16:04:51.193 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.37.self_attn.q_proj using 128 samples +2026-04-15 16:04:54.556 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.36s +2026-04-15 16:04:54.556 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 2836.41 +2026-04-15 16:04:54.556 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:04:54.557 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:04:54.557 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 53.05344 MB +2026-04-15 16:04:54.557 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.37.self_attn.k_proj using 128 samples +2026-04-15 16:04:57.763 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.21s +2026-04-15 16:04:57.764 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 615.35 +2026-04-15 16:04:57.764 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:04:57.764 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:04:57.764 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 10.610688 MB +2026-04-15 16:04:57.764 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.37.self_attn.v_proj using 128 samples +2026-04-15 16:05:00.971 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.21s +2026-04-15 16:05:00.971 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 912.78 +2026-04-15 16:05:00.971 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:05:00.972 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:05:00.972 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 10.610688 MB +2026-04-15 16:05:00.972 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.37.self_attn.o_proj using 128 samples +2026-04-15 16:05:04.344 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.37s +2026-04-15 16:05:04.345 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 855.88 +2026-04-15 16:05:04.345 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:05:04.345 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:05:04.345 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 53.0432 MB +2026-04-15 16:05:04.345 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.37.mlp.gate_proj using 128 samples +2026-04-15 16:05:08.012 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.67s +2026-04-15 16:05:08.012 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 8925.87 +2026-04-15 16:05:08.012 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:05:08.012 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:05:08.013 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 16:05:08.013 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.37.mlp.up_proj using 128 samples +2026-04-15 16:05:11.690 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.68s +2026-04-15 16:05:11.690 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 9217.24 +2026-04-15 16:05:11.691 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:05:11.691 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 16.49% | total memory: 25 GB +2026-04-15 16:05:11.691 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 16:05:11.691 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.37.mlp.down_proj using 128 samples +2026-04-15 16:05:21.194 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 9.50s +2026-04-15 16:05:21.195 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 4388.04 +2026-04-15 16:05:21.195 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:05:21.195 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 22.43% | total memory: 25 GB +2026-04-15 16:05:21.196 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 16:05:21.196 | DEBUG | llmcompressor.core.lifecycle:event:205 - Updated event with modifier: config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False +2026-04-15 16:05:24.659 | DEBUG | llmcompressor.core.lifecycle:event:195 - Handling event: EventType.SEQUENTIAL_EPOCH_END +2026-04-15 16:05:24.660 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.38.self_attn.q_proj using 128 samples +2026-04-15 16:05:28.073 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.41s +2026-04-15 16:05:28.074 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 2892.21 +2026-04-15 16:05:28.074 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:05:28.074 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:05:28.074 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 53.05344 MB +2026-04-15 16:05:28.074 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.38.self_attn.k_proj using 128 samples +2026-04-15 16:05:31.319 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.24s +2026-04-15 16:05:31.319 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 626.47 +2026-04-15 16:05:31.319 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:05:31.320 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:05:31.320 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 10.610688 MB +2026-04-15 16:05:31.320 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.38.self_attn.v_proj using 128 samples +2026-04-15 16:05:34.585 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.26s +2026-04-15 16:05:34.585 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 955.41 +2026-04-15 16:05:34.585 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:05:34.586 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:05:34.586 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 10.610688 MB +2026-04-15 16:05:34.586 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.38.self_attn.o_proj using 128 samples +2026-04-15 16:05:37.996 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.41s +2026-04-15 16:05:37.997 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 818.42 +2026-04-15 16:05:37.997 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:05:37.997 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:05:37.997 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 53.0432 MB +2026-04-15 16:05:37.998 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.38.mlp.gate_proj using 128 samples +2026-04-15 16:05:41.800 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.80s +2026-04-15 16:05:41.800 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 10109.41 +2026-04-15 16:05:41.801 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:05:41.801 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:05:41.801 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 16:05:41.801 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.38.mlp.up_proj using 128 samples +2026-04-15 16:05:45.465 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.66s +2026-04-15 16:05:45.466 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 10557.55 +2026-04-15 16:05:45.466 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:05:45.467 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:05:45.467 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 16:05:45.467 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.38.mlp.down_proj using 128 samples +2026-04-15 16:05:54.892 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 9.42s +2026-04-15 16:05:54.892 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 5141.34 +2026-04-15 16:05:54.892 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:05:54.892 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 21.38% | total memory: 25 GB +2026-04-15 16:05:54.892 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 16:05:54.893 | DEBUG | llmcompressor.core.lifecycle:event:205 - Updated event with modifier: config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False +2026-04-15 16:05:58.327 | DEBUG | llmcompressor.core.lifecycle:event:195 - Handling event: EventType.SEQUENTIAL_EPOCH_END +2026-04-15 16:05:58.327 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.39.self_attn.q_proj using 128 samples +2026-04-15 16:06:01.688 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.36s +2026-04-15 16:06:01.688 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 3006.95 +2026-04-15 16:06:01.688 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:06:01.689 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:06:01.689 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 53.05344 MB +2026-04-15 16:06:01.689 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.39.self_attn.k_proj using 128 samples +2026-04-15 16:06:04.880 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.19s +2026-04-15 16:06:04.881 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 637.68 +2026-04-15 16:06:04.881 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:06:04.881 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:06:04.881 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 10.610688 MB +2026-04-15 16:06:04.881 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.39.self_attn.v_proj using 128 samples +2026-04-15 16:06:08.076 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.19s +2026-04-15 16:06:08.077 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 1141.71 +2026-04-15 16:06:08.077 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:06:08.077 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:06:08.077 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 10.610688 MB +2026-04-15 16:06:08.077 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.39.self_attn.o_proj using 128 samples +2026-04-15 16:06:11.420 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.34s +2026-04-15 16:06:11.421 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 1292.78 +2026-04-15 16:06:11.421 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:06:11.421 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:06:11.421 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 53.0432 MB +2026-04-15 16:06:11.421 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.39.mlp.gate_proj using 128 samples +2026-04-15 16:06:15.094 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.67s +2026-04-15 16:06:15.095 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 11360.30 +2026-04-15 16:06:15.095 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:06:15.095 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:06:15.096 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 16:06:15.096 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.39.mlp.up_proj using 128 samples +2026-04-15 16:06:18.723 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.63s +2026-04-15 16:06:18.723 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 11987.16 +2026-04-15 16:06:18.724 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:06:18.724 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 16.49% | total memory: 25 GB +2026-04-15 16:06:18.724 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 16:06:18.724 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.39.mlp.down_proj using 128 samples +2026-04-15 16:06:28.161 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 9.44s +2026-04-15 16:06:28.161 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 6399.00 +2026-04-15 16:06:28.161 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:06:28.161 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 22.43% | total memory: 25 GB +2026-04-15 16:06:28.162 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 16:06:28.162 | DEBUG | llmcompressor.core.lifecycle:event:205 - Updated event with modifier: config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False +2026-04-15 16:06:31.594 | DEBUG | llmcompressor.core.lifecycle:event:195 - Handling event: EventType.SEQUENTIAL_EPOCH_END +2026-04-15 16:06:31.594 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.40.self_attn.q_proj using 128 samples +2026-04-15 16:06:34.978 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.38s +2026-04-15 16:06:34.978 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 3366.48 +2026-04-15 16:06:34.979 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:06:34.979 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:06:34.979 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 53.05344 MB +2026-04-15 16:06:34.979 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.40.self_attn.k_proj using 128 samples +2026-04-15 16:06:38.185 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.21s +2026-04-15 16:06:38.185 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 692.37 +2026-04-15 16:06:38.185 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:06:38.186 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:06:38.186 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 10.610688 MB +2026-04-15 16:06:38.186 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.40.self_attn.v_proj using 128 samples +2026-04-15 16:06:41.410 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.22s +2026-04-15 16:06:41.410 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 1453.99 +2026-04-15 16:06:41.411 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:06:41.411 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:06:41.411 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 10.610688 MB +2026-04-15 16:06:41.411 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.40.self_attn.o_proj using 128 samples +2026-04-15 16:06:44.787 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.38s +2026-04-15 16:06:44.788 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 1201.14 +2026-04-15 16:06:44.788 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:06:44.788 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:06:44.788 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 53.0432 MB +2026-04-15 16:06:44.788 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.40.mlp.gate_proj using 128 samples +2026-04-15 16:06:48.433 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.64s +2026-04-15 16:06:48.433 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 12449.85 +2026-04-15 16:06:48.433 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:06:48.433 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:06:48.433 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 16:06:48.434 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.40.mlp.up_proj using 128 samples +2026-04-15 16:06:52.047 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.61s +2026-04-15 16:06:52.047 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 13229.50 +2026-04-15 16:06:52.048 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:06:52.048 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:06:52.048 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 16:06:52.048 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.40.mlp.down_proj using 128 samples +2026-04-15 16:07:01.372 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 9.32s +2026-04-15 16:07:01.373 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 7477.60 +2026-04-15 16:07:01.373 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:07:01.373 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 21.38% | total memory: 25 GB +2026-04-15 16:07:01.373 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 16:07:01.373 | DEBUG | llmcompressor.core.lifecycle:event:205 - Updated event with modifier: config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False +2026-04-15 16:07:04.787 | DEBUG | llmcompressor.core.lifecycle:event:195 - Handling event: EventType.SEQUENTIAL_EPOCH_END +2026-04-15 16:07:04.787 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.41.self_attn.q_proj using 128 samples +2026-04-15 16:07:08.152 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.36s +2026-04-15 16:07:08.152 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 3593.10 +2026-04-15 16:07:08.153 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:07:08.153 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:07:08.153 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 53.05344 MB +2026-04-15 16:07:08.153 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.41.self_attn.k_proj using 128 samples +2026-04-15 16:07:11.382 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.23s +2026-04-15 16:07:11.383 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 718.40 +2026-04-15 16:07:11.383 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:07:11.384 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:07:11.384 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 10.610688 MB +2026-04-15 16:07:11.384 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.41.self_attn.v_proj using 128 samples +2026-04-15 16:07:14.616 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.23s +2026-04-15 16:07:14.617 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 1542.13 +2026-04-15 16:07:14.617 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:07:14.617 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:07:14.617 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 10.610688 MB +2026-04-15 16:07:14.617 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.41.self_attn.o_proj using 128 samples +2026-04-15 16:07:17.978 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.36s +2026-04-15 16:07:17.979 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 1276.55 +2026-04-15 16:07:17.979 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:07:17.979 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:07:17.980 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 53.0432 MB +2026-04-15 16:07:17.980 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.41.mlp.gate_proj using 128 samples +2026-04-15 16:07:21.499 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.52s +2026-04-15 16:07:21.500 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 13610.60 +2026-04-15 16:07:21.500 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:07:21.500 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:07:21.501 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 16:07:21.501 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.41.mlp.up_proj using 128 samples +2026-04-15 16:07:25.006 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.50s +2026-04-15 16:07:25.006 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 14692.79 +2026-04-15 16:07:25.006 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:07:25.007 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 16.49% | total memory: 25 GB +2026-04-15 16:07:25.007 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 16:07:25.007 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.41.mlp.down_proj using 128 samples +2026-04-15 16:07:34.134 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 9.13s +2026-04-15 16:07:34.134 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 8930.47 +2026-04-15 16:07:34.135 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:07:34.135 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 22.43% | total memory: 25 GB +2026-04-15 16:07:34.135 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 16:07:34.135 | DEBUG | llmcompressor.core.lifecycle:event:205 - Updated event with modifier: config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False +2026-04-15 16:07:37.496 | DEBUG | llmcompressor.core.lifecycle:event:195 - Handling event: EventType.SEQUENTIAL_EPOCH_END +2026-04-15 16:07:37.497 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.42.self_attn.q_proj using 128 samples +2026-04-15 16:07:40.721 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.22s +2026-04-15 16:07:40.722 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 3612.23 +2026-04-15 16:07:40.722 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:07:40.722 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:07:40.722 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 53.05344 MB +2026-04-15 16:07:40.723 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.42.self_attn.k_proj using 128 samples +2026-04-15 16:07:43.829 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.11s +2026-04-15 16:07:43.830 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 689.86 +2026-04-15 16:07:43.830 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:07:43.830 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:07:43.830 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 10.610688 MB +2026-04-15 16:07:43.831 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.42.self_attn.v_proj using 128 samples +2026-04-15 16:07:46.920 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.09s +2026-04-15 16:07:46.921 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 1527.44 +2026-04-15 16:07:46.922 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:07:46.922 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:07:46.922 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 10.610688 MB +2026-04-15 16:07:46.922 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.42.self_attn.o_proj using 128 samples +2026-04-15 16:07:50.156 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.23s +2026-04-15 16:07:50.157 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 1164.77 +2026-04-15 16:07:50.157 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:07:50.157 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:07:50.157 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 53.0432 MB +2026-04-15 16:07:50.157 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.42.mlp.gate_proj using 128 samples +2026-04-15 16:07:53.697 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.54s +2026-04-15 16:07:53.698 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 14922.95 +2026-04-15 16:07:53.698 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:07:53.698 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:07:53.698 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 16:07:53.698 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.42.mlp.up_proj using 128 samples +2026-04-15 16:07:57.183 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.48s +2026-04-15 16:07:57.184 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 16300.43 +2026-04-15 16:07:57.184 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:07:57.184 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:07:57.184 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 16:07:57.184 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.42.mlp.down_proj using 128 samples +2026-04-15 16:08:06.270 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 9.08s +2026-04-15 16:08:06.271 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 10789.12 +2026-04-15 16:08:06.271 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:08:06.271 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 21.38% | total memory: 25 GB +2026-04-15 16:08:06.271 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 16:08:06.272 | DEBUG | llmcompressor.core.lifecycle:event:205 - Updated event with modifier: config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False +2026-04-15 16:08:09.654 | DEBUG | llmcompressor.core.lifecycle:event:195 - Handling event: EventType.SEQUENTIAL_EPOCH_END +2026-04-15 16:08:09.654 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.43.self_attn.q_proj using 128 samples +2026-04-15 16:08:12.919 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.26s +2026-04-15 16:08:12.920 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 3828.64 +2026-04-15 16:08:12.920 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:08:12.920 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:08:12.920 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 53.05344 MB +2026-04-15 16:08:12.921 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.43.self_attn.k_proj using 128 samples +2026-04-15 16:08:16.012 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.09s +2026-04-15 16:08:16.013 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 715.10 +2026-04-15 16:08:16.013 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:08:16.013 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:08:16.013 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 10.610688 MB +2026-04-15 16:08:16.014 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.43.self_attn.v_proj using 128 samples +2026-04-15 16:08:19.141 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.13s +2026-04-15 16:08:19.142 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 2478.94 +2026-04-15 16:08:19.142 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:08:19.142 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:08:19.142 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 10.610688 MB +2026-04-15 16:08:19.143 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.43.self_attn.o_proj using 128 samples +2026-04-15 16:08:22.403 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.26s +2026-04-15 16:08:22.403 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 2146.43 +2026-04-15 16:08:22.403 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:08:22.404 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:08:22.404 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 53.0432 MB +2026-04-15 16:08:22.404 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.43.mlp.gate_proj using 128 samples +2026-04-15 16:08:25.939 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.53s +2026-04-15 16:08:25.939 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 16573.05 +2026-04-15 16:08:25.940 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:08:25.940 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:08:25.940 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 16:08:25.940 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.43.mlp.up_proj using 128 samples +2026-04-15 16:08:29.448 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.51s +2026-04-15 16:08:29.449 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 18439.83 +2026-04-15 16:08:29.449 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:08:29.449 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 16.49% | total memory: 25 GB +2026-04-15 16:08:29.449 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 16:08:29.449 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.43.mlp.down_proj using 128 samples +2026-04-15 16:08:38.552 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 9.10s +2026-04-15 16:08:38.552 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 13438.67 +2026-04-15 16:08:38.553 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:08:38.553 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 22.43% | total memory: 25 GB +2026-04-15 16:08:38.553 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 16:08:38.553 | DEBUG | llmcompressor.core.lifecycle:event:205 - Updated event with modifier: config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False +2026-04-15 16:08:41.952 | DEBUG | llmcompressor.core.lifecycle:event:195 - Handling event: EventType.SEQUENTIAL_EPOCH_END +2026-04-15 16:08:41.953 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.44.self_attn.q_proj using 128 samples +2026-04-15 16:08:45.195 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.24s +2026-04-15 16:08:45.195 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 3732.74 +2026-04-15 16:08:45.195 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:08:45.195 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:08:45.195 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 53.05344 MB +2026-04-15 16:08:45.196 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.44.self_attn.k_proj using 128 samples +2026-04-15 16:08:48.471 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.28s +2026-04-15 16:08:48.472 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 649.77 +2026-04-15 16:08:48.472 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:08:48.472 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:08:48.472 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 10.610688 MB +2026-04-15 16:08:48.473 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.44.self_attn.v_proj using 128 samples +2026-04-15 16:08:51.663 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.19s +2026-04-15 16:08:51.663 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 2740.42 +2026-04-15 16:08:51.663 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:08:51.663 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:08:51.663 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 10.610688 MB +2026-04-15 16:08:51.664 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.44.self_attn.o_proj using 128 samples +2026-04-15 16:08:54.994 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.33s +2026-04-15 16:08:54.994 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 3431.09 +2026-04-15 16:08:54.995 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:08:54.995 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:08:54.995 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 53.0432 MB +2026-04-15 16:08:54.995 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.44.mlp.gate_proj using 128 samples +2026-04-15 16:08:58.648 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.65s +2026-04-15 16:08:58.649 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 17453.86 +2026-04-15 16:08:58.649 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:08:58.649 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:08:58.649 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 16:08:58.650 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.44.mlp.up_proj using 128 samples +2026-04-15 16:09:02.280 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.63s +2026-04-15 16:09:02.280 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 19786.39 +2026-04-15 16:09:02.280 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:09:02.280 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:09:02.280 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 16:09:02.281 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.44.mlp.down_proj using 128 samples +2026-04-15 16:09:11.693 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 9.41s +2026-04-15 16:09:11.693 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 28872.82 +2026-04-15 16:09:11.693 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:09:11.693 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 21.38% | total memory: 25 GB +2026-04-15 16:09:11.693 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 16:09:11.694 | DEBUG | llmcompressor.core.lifecycle:event:205 - Updated event with modifier: config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False +2026-04-15 16:09:15.119 | DEBUG | llmcompressor.core.lifecycle:event:195 - Handling event: EventType.SEQUENTIAL_EPOCH_END +2026-04-15 16:09:15.119 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.45.self_attn.q_proj using 128 samples +2026-04-15 16:09:18.329 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.21s +2026-04-15 16:09:18.329 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 3900.74 +2026-04-15 16:09:18.329 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:09:18.329 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:09:18.329 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 53.05344 MB +2026-04-15 16:09:18.330 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.45.self_attn.k_proj using 128 samples +2026-04-15 16:09:21.398 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.07s +2026-04-15 16:09:21.398 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 657.64 +2026-04-15 16:09:21.398 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:09:21.398 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:09:21.398 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 10.610688 MB +2026-04-15 16:09:21.399 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.45.self_attn.v_proj using 128 samples +2026-04-15 16:09:24.461 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.06s +2026-04-15 16:09:24.462 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 3286.53 +2026-04-15 16:09:24.462 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:09:24.462 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:09:24.462 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 10.610688 MB +2026-04-15 16:09:24.462 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.45.self_attn.o_proj using 128 samples +2026-04-15 16:09:27.668 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.21s +2026-04-15 16:09:27.668 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 3629.87 +2026-04-15 16:09:27.669 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:09:27.669 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:09:27.669 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 53.0432 MB +2026-04-15 16:09:27.669 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.45.mlp.gate_proj using 128 samples +2026-04-15 16:09:31.145 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.48s +2026-04-15 16:09:31.146 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 18677.84 +2026-04-15 16:09:31.146 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:09:31.146 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:09:31.146 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 16:09:31.146 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.45.mlp.up_proj using 128 samples +2026-04-15 16:09:34.615 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.47s +2026-04-15 16:09:34.616 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 20841.60 +2026-04-15 16:09:34.616 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:09:34.616 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 16.49% | total memory: 25 GB +2026-04-15 16:09:34.616 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 16:09:34.616 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.45.mlp.down_proj using 128 samples +2026-04-15 16:09:43.582 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 8.97s +2026-04-15 16:09:43.582 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 23877.44 +2026-04-15 16:09:43.583 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:09:43.583 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 22.43% | total memory: 25 GB +2026-04-15 16:09:43.583 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 16:09:43.583 | DEBUG | llmcompressor.core.lifecycle:event:205 - Updated event with modifier: config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False +2026-04-15 16:09:46.855 | DEBUG | llmcompressor.core.lifecycle:event:195 - Handling event: EventType.SEQUENTIAL_EPOCH_END +2026-04-15 16:09:46.855 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.46.self_attn.q_proj using 128 samples +2026-04-15 16:09:50.029 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.17s +2026-04-15 16:09:50.029 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 3601.02 +2026-04-15 16:09:50.030 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:09:50.030 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:09:50.030 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 53.05344 MB +2026-04-15 16:09:50.030 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.46.self_attn.k_proj using 128 samples +2026-04-15 16:09:53.064 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.03s +2026-04-15 16:09:53.064 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 609.48 +2026-04-15 16:09:53.064 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:09:53.064 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:09:53.064 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 10.610688 MB +2026-04-15 16:09:53.065 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.46.self_attn.v_proj using 128 samples +2026-04-15 16:09:56.105 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.04s +2026-04-15 16:09:56.106 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 3442.43 +2026-04-15 16:09:56.106 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:09:56.106 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:09:56.106 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 10.610688 MB +2026-04-15 16:09:56.106 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.46.self_attn.o_proj using 128 samples +2026-04-15 16:09:59.296 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.19s +2026-04-15 16:09:59.296 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 6393.74 +2026-04-15 16:09:59.297 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:09:59.297 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:09:59.297 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 53.0432 MB +2026-04-15 16:09:59.297 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.46.mlp.gate_proj using 128 samples +2026-04-15 16:10:02.800 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.50s +2026-04-15 16:10:02.801 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 19993.11 +2026-04-15 16:10:02.801 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:10:02.801 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:10:02.801 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 16:10:02.802 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.46.mlp.up_proj using 128 samples +2026-04-15 16:10:06.268 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.47s +2026-04-15 16:10:06.268 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 21493.32 +2026-04-15 16:10:06.269 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:10:06.269 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:10:06.269 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 16:10:06.269 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.46.mlp.down_proj using 128 samples +2026-04-15 16:10:15.253 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 8.98s +2026-04-15 16:10:15.254 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 26948.23 +2026-04-15 16:10:15.254 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:10:15.254 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 21.38% | total memory: 25 GB +2026-04-15 16:10:15.254 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 16:10:15.255 | DEBUG | llmcompressor.core.lifecycle:event:205 - Updated event with modifier: config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False +2026-04-15 16:10:18.629 | DEBUG | llmcompressor.core.lifecycle:event:195 - Handling event: EventType.SEQUENTIAL_EPOCH_END +2026-04-15 16:10:18.629 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.47.self_attn.q_proj using 128 samples +2026-04-15 16:10:21.857 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.23s +2026-04-15 16:10:21.858 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 3060.20 +2026-04-15 16:10:21.858 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:10:21.858 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:10:21.858 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 53.05344 MB +2026-04-15 16:10:21.858 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.47.self_attn.k_proj using 128 samples +2026-04-15 16:10:24.927 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.07s +2026-04-15 16:10:24.927 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 506.19 +2026-04-15 16:10:24.927 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:10:24.927 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:10:24.927 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 10.610688 MB +2026-04-15 16:10:24.928 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.47.self_attn.v_proj using 128 samples +2026-04-15 16:10:27.985 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.06s +2026-04-15 16:10:27.986 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 1778.24 +2026-04-15 16:10:27.986 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:10:27.986 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:10:27.986 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 10.610688 MB +2026-04-15 16:10:27.986 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.47.self_attn.o_proj using 128 samples +2026-04-15 16:10:31.162 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.18s +2026-04-15 16:10:31.162 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 2978.61 +2026-04-15 16:10:31.162 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:10:31.162 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:10:31.162 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 53.0432 MB +2026-04-15 16:10:31.163 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.47.mlp.gate_proj using 128 samples +2026-04-15 16:10:34.666 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.50s +2026-04-15 16:10:34.666 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 21124.47 +2026-04-15 16:10:34.667 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:10:34.667 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 15.44% | total memory: 25 GB +2026-04-15 16:10:34.667 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 16:10:34.667 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.47.mlp.up_proj using 128 samples +2026-04-15 16:10:38.141 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 3.47s +2026-04-15 16:10:38.141 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 22059.31 +2026-04-15 16:10:38.142 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:10:38.142 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 16.49% | total memory: 25 GB +2026-04-15 16:10:38.142 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 16:10:38.142 | INFO | llmcompressor.modifiers.quantization.gptq.base:compress_modules:251 - Quantizing model.layers.47.mlp.down_proj using 128 samples +2026-04-15 16:10:47.137 | METRIC | llmcompressor.utils.metric_logging:compress:136 - time 8.99s +2026-04-15 16:10:47.137 | METRIC | llmcompressor.utils.metric_logging:compress:138 - error 35767.55 +2026-04-15 16:10:47.137 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 0 | usage: 68.38% | total memory: 25 GB +2026-04-15 16:10:47.137 | METRIC | llmcompressor.utils.metric_logging:compress:145 - GPU 1 | usage: 22.43% | total memory: 25 GB +2026-04-15 16:10:47.137 | METRIC | llmcompressor.utils.metric_logging:compress:154 - Compressed module size: 143.21664 MB +2026-04-15 16:10:47.138 | DEBUG | llmcompressor.core.lifecycle:event:205 - Updated event with modifier: config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False +2026-04-15 16:10:49.203 | DEBUG | llmcompressor.core.lifecycle:event:195 - Handling event: EventType.SEQUENTIAL_EPOCH_END +2026-04-15 16:10:49.204 | DEBUG | llmcompressor.core.lifecycle:event:205 - Updated event with modifier: config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=False sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False +2026-04-15 16:10:51.041 | DEBUG | llmcompressor.core.lifecycle:event:195 - Handling event: EventType.CALIBRATION_EPOCH_END +2026-04-15 16:10:51.051 | DEBUG | llmcompressor.core.lifecycle:event:205 - Updated event with modifier: config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=False started_=True ended_=True sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False +2026-04-15 16:10:51.055 | DEBUG | llmcompressor.core.lifecycle:finalize:134 - Finalizing compression lifecycle +2026-04-15 16:10:51.055 | DEBUG | llmcompressor.core.lifecycle:finalize:138 - Finalized modifier: config_groups={'group_0': QuantizationScheme(targets=['Linear'], weights=QuantizationArgs(num_bits=4, type='int', symmetric=True, group_size=128, strategy='group', block_structure=None, dynamic=False, actorder=, observer='minmax', observer_kwargs={}), input_activations=None, output_activations=None, format=None)} targets=['Linear'] ignore=['lm_head'] scheme=None kv_cache_scheme=None index=None group=None start=None end=None update=None initialized_=True finalized_=True started_=True ended_=True sequential_targets=None block_size=128 dampening_frac=0.01 actorder=static offload_hessians=False +2026-04-15 16:10:51.055 | INFO | llmcompressor.core.lifecycle:finalize:144 - Compression lifecycle finalized for 1 modifiers +2026-04-15 16:10:51.142 | INFO | llmcompressor.transformers.compression.compressed_tensors_utils:get_model_compressor:193 - skip_sparsity_compression_stats set to True. Skipping sparsity compression statistic calculations. No sparsity compressor will be applied. +2026-04-15 16:12:01.122 | DEBUG | llmcompressor.transformers.utils.helpers:recipe_from_huggingface_model_id:146 - Unable to find recipe recipe.yaml for model ID: deepseek-ai/DeepSeek-R1-Distill-Qwen-14B: 404 Client Error. (Request ID: Root=1-69df5660-695f2f853d22f0857d5cb107;aff73b7e-166e-40fb-a890-6dc38cc9283a) + +Entry Not Found for url: https://huggingface.co/deepseek-ai/DeepSeek-R1-Distill-Qwen-14B/resolve/main/recipe.yaml..Skipping recipe resolution. +2026-04-15 16:12:01.127 | DEBUG | llmcompressor.transformers.utils.helpers:infer_recipe_from_model_path:112 - Failed to infer the recipe from the model_path