phi3-lora-openbookqa3 / trainer_state.json
mikasenghaas's picture
Upload folder using huggingface_hub
801acfa verified
{
"best_metric": 1.6195639371871948,
"best_model_checkpoint": "./output/checkpoints/2024-06-11_15-20-56/checkpoint-10",
"epoch": 3.0,
"eval_steps": 1,
"global_step": 111,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.02702702702702703,
"grad_norm": 3.045227289199829,
"learning_rate": 3.3333333333333335e-05,
"loss": 5.5922,
"step": 1
},
{
"epoch": 0.02702702702702703,
"eval_loss": 5.6431684494018555,
"eval_runtime": 11.0821,
"eval_samples_per_second": 11.189,
"eval_steps_per_second": 0.722,
"step": 1
},
{
"epoch": 0.05405405405405406,
"grad_norm": 3.084507942199707,
"learning_rate": 6.666666666666667e-05,
"loss": 5.6812,
"step": 2
},
{
"epoch": 0.05405405405405406,
"eval_loss": 5.540425777435303,
"eval_runtime": 11.083,
"eval_samples_per_second": 11.188,
"eval_steps_per_second": 0.722,
"step": 2
},
{
"epoch": 0.08108108108108109,
"grad_norm": 3.1143131256103516,
"learning_rate": 0.0001,
"loss": 5.5805,
"step": 3
},
{
"epoch": 0.08108108108108109,
"eval_loss": 5.062227725982666,
"eval_runtime": 11.1285,
"eval_samples_per_second": 11.143,
"eval_steps_per_second": 0.719,
"step": 3
},
{
"epoch": 0.10810810810810811,
"grad_norm": 3.3044817447662354,
"learning_rate": 0.00013333333333333334,
"loss": 5.0744,
"step": 4
},
{
"epoch": 0.10810810810810811,
"eval_loss": 4.176051139831543,
"eval_runtime": 11.2317,
"eval_samples_per_second": 11.04,
"eval_steps_per_second": 0.712,
"step": 4
},
{
"epoch": 0.13513513513513514,
"grad_norm": 3.3503236770629883,
"learning_rate": 0.0001666666666666667,
"loss": 4.1348,
"step": 5
},
{
"epoch": 0.13513513513513514,
"eval_loss": 3.137871265411377,
"eval_runtime": 11.1384,
"eval_samples_per_second": 11.133,
"eval_steps_per_second": 0.718,
"step": 5
},
{
"epoch": 0.16216216216216217,
"grad_norm": 2.9682626724243164,
"learning_rate": 0.0002,
"loss": 3.0425,
"step": 6
},
{
"epoch": 0.16216216216216217,
"eval_loss": 2.459784746170044,
"eval_runtime": 11.1086,
"eval_samples_per_second": 11.163,
"eval_steps_per_second": 0.72,
"step": 6
},
{
"epoch": 0.1891891891891892,
"grad_norm": 1.8949077129364014,
"learning_rate": 0.00023333333333333336,
"loss": 2.3291,
"step": 7
},
{
"epoch": 0.1891891891891892,
"eval_loss": 2.064051866531372,
"eval_runtime": 11.1293,
"eval_samples_per_second": 11.142,
"eval_steps_per_second": 0.719,
"step": 7
},
{
"epoch": 0.21621621621621623,
"grad_norm": 1.2975975275039673,
"learning_rate": 0.0002666666666666667,
"loss": 1.9481,
"step": 8
},
{
"epoch": 0.21621621621621623,
"eval_loss": 1.826060175895691,
"eval_runtime": 11.1457,
"eval_samples_per_second": 11.125,
"eval_steps_per_second": 0.718,
"step": 8
},
{
"epoch": 0.24324324324324326,
"grad_norm": 0.9844012260437012,
"learning_rate": 0.00030000000000000003,
"loss": 1.6863,
"step": 9
},
{
"epoch": 0.24324324324324326,
"eval_loss": 1.7158275842666626,
"eval_runtime": 11.1384,
"eval_samples_per_second": 11.133,
"eval_steps_per_second": 0.718,
"step": 9
},
{
"epoch": 0.2702702702702703,
"grad_norm": 0.9899163246154785,
"learning_rate": 0.0003333333333333334,
"loss": 1.6293,
"step": 10
},
{
"epoch": 0.2702702702702703,
"eval_loss": 1.6195639371871948,
"eval_runtime": 11.1633,
"eval_samples_per_second": 11.108,
"eval_steps_per_second": 0.717,
"step": 10
},
{
"epoch": 0.2972972972972973,
"grad_norm": 0.2760697901248932,
"learning_rate": 0.00036666666666666667,
"loss": 1.5182,
"step": 11
},
{
"epoch": 0.2972972972972973,
"eval_loss": 1.5875523090362549,
"eval_runtime": 11.1331,
"eval_samples_per_second": 11.138,
"eval_steps_per_second": 0.719,
"step": 11
},
{
"epoch": 0.32432432432432434,
"grad_norm": 0.4246121644973755,
"learning_rate": 0.0004,
"loss": 1.432,
"step": 12
},
{
"epoch": 0.32432432432432434,
"eval_loss": 1.574428677558899,
"eval_runtime": 11.1913,
"eval_samples_per_second": 11.08,
"eval_steps_per_second": 0.715,
"step": 12
},
{
"epoch": 0.35135135135135137,
"grad_norm": 0.4968096911907196,
"learning_rate": 0.000395959595959596,
"loss": 1.3155,
"step": 13
},
{
"epoch": 0.35135135135135137,
"eval_loss": 1.6574310064315796,
"eval_runtime": 11.1636,
"eval_samples_per_second": 11.108,
"eval_steps_per_second": 0.717,
"step": 13
},
{
"epoch": 0.3783783783783784,
"grad_norm": 0.2221263200044632,
"learning_rate": 0.0003919191919191919,
"loss": 1.2013,
"step": 14
},
{
"epoch": 0.3783783783783784,
"eval_loss": 1.7277523279190063,
"eval_runtime": 11.2226,
"eval_samples_per_second": 11.049,
"eval_steps_per_second": 0.713,
"step": 14
},
{
"epoch": 0.40540540540540543,
"grad_norm": 0.24387121200561523,
"learning_rate": 0.0003878787878787879,
"loss": 1.1303,
"step": 15
},
{
"epoch": 0.40540540540540543,
"eval_loss": 1.7132700681686401,
"eval_runtime": 11.1622,
"eval_samples_per_second": 11.109,
"eval_steps_per_second": 0.717,
"step": 15
},
{
"epoch": 0.43243243243243246,
"grad_norm": 0.2101132869720459,
"learning_rate": 0.00038383838383838383,
"loss": 1.1181,
"step": 16
},
{
"epoch": 0.43243243243243246,
"eval_loss": 1.6620469093322754,
"eval_runtime": 11.1581,
"eval_samples_per_second": 11.113,
"eval_steps_per_second": 0.717,
"step": 16
},
{
"epoch": 0.4594594594594595,
"grad_norm": 0.1956334114074707,
"learning_rate": 0.0003797979797979798,
"loss": 1.0897,
"step": 17
},
{
"epoch": 0.4594594594594595,
"eval_loss": 1.6307740211486816,
"eval_runtime": 11.1407,
"eval_samples_per_second": 11.13,
"eval_steps_per_second": 0.718,
"step": 17
},
{
"epoch": 0.4864864864864865,
"grad_norm": 0.1778702437877655,
"learning_rate": 0.0003757575757575758,
"loss": 1.0865,
"step": 18
},
{
"epoch": 0.4864864864864865,
"eval_loss": 1.6229671239852905,
"eval_runtime": 11.2453,
"eval_samples_per_second": 11.027,
"eval_steps_per_second": 0.711,
"step": 18
},
{
"epoch": 0.5135135135135135,
"grad_norm": 0.15137992799282074,
"learning_rate": 0.0003717171717171717,
"loss": 1.0591,
"step": 19
},
{
"epoch": 0.5135135135135135,
"eval_loss": 1.6271958351135254,
"eval_runtime": 11.1705,
"eval_samples_per_second": 11.101,
"eval_steps_per_second": 0.716,
"step": 19
},
{
"epoch": 0.5405405405405406,
"grad_norm": 0.11392053216695786,
"learning_rate": 0.0003676767676767677,
"loss": 1.0392,
"step": 20
},
{
"epoch": 0.5405405405405406,
"eval_loss": 1.6411267518997192,
"eval_runtime": 11.1409,
"eval_samples_per_second": 11.13,
"eval_steps_per_second": 0.718,
"step": 20
},
{
"epoch": 0.5675675675675675,
"grad_norm": 0.0866568386554718,
"learning_rate": 0.00036363636363636367,
"loss": 1.0073,
"step": 21
},
{
"epoch": 0.5675675675675675,
"eval_loss": 1.6590908765792847,
"eval_runtime": 11.1242,
"eval_samples_per_second": 11.147,
"eval_steps_per_second": 0.719,
"step": 21
},
{
"epoch": 0.5945945945945946,
"grad_norm": 0.10577922314405441,
"learning_rate": 0.0003595959595959596,
"loss": 0.9898,
"step": 22
},
{
"epoch": 0.5945945945945946,
"eval_loss": 1.6635982990264893,
"eval_runtime": 11.2221,
"eval_samples_per_second": 11.05,
"eval_steps_per_second": 0.713,
"step": 22
},
{
"epoch": 0.6216216216216216,
"grad_norm": 0.1029527559876442,
"learning_rate": 0.00035555555555555557,
"loss": 1.0017,
"step": 23
},
{
"epoch": 0.6216216216216216,
"eval_loss": 1.6551401615142822,
"eval_runtime": 11.1896,
"eval_samples_per_second": 11.082,
"eval_steps_per_second": 0.715,
"step": 23
},
{
"epoch": 0.6486486486486487,
"grad_norm": 0.09740535914897919,
"learning_rate": 0.00035151515151515155,
"loss": 1.0052,
"step": 24
},
{
"epoch": 0.6486486486486487,
"eval_loss": 1.642539620399475,
"eval_runtime": 11.1624,
"eval_samples_per_second": 11.109,
"eval_steps_per_second": 0.717,
"step": 24
},
{
"epoch": 0.6756756756756757,
"grad_norm": 0.09230296313762665,
"learning_rate": 0.00034747474747474753,
"loss": 0.9873,
"step": 25
},
{
"epoch": 0.6756756756756757,
"eval_loss": 1.6282507181167603,
"eval_runtime": 11.0959,
"eval_samples_per_second": 11.175,
"eval_steps_per_second": 0.721,
"step": 25
},
{
"epoch": 0.7027027027027027,
"grad_norm": 0.08403297513723373,
"learning_rate": 0.00034343434343434346,
"loss": 0.9743,
"step": 26
},
{
"epoch": 0.7027027027027027,
"eval_loss": 1.6223111152648926,
"eval_runtime": 11.143,
"eval_samples_per_second": 11.128,
"eval_steps_per_second": 0.718,
"step": 26
},
{
"epoch": 0.7297297297297297,
"grad_norm": 0.0808207094669342,
"learning_rate": 0.00033939393939393943,
"loss": 0.9752,
"step": 27
},
{
"epoch": 0.7297297297297297,
"eval_loss": 1.6211432218551636,
"eval_runtime": 11.2057,
"eval_samples_per_second": 11.066,
"eval_steps_per_second": 0.714,
"step": 27
},
{
"epoch": 0.7567567567567568,
"grad_norm": 0.08751657605171204,
"learning_rate": 0.00033535353535353536,
"loss": 0.984,
"step": 28
},
{
"epoch": 0.7567567567567568,
"eval_loss": 1.6278128623962402,
"eval_runtime": 11.2228,
"eval_samples_per_second": 11.049,
"eval_steps_per_second": 0.713,
"step": 28
},
{
"epoch": 0.7837837837837838,
"grad_norm": 0.08614286035299301,
"learning_rate": 0.00033131313131313134,
"loss": 0.9609,
"step": 29
},
{
"epoch": 0.7837837837837838,
"eval_loss": 1.640842318534851,
"eval_runtime": 11.1451,
"eval_samples_per_second": 11.126,
"eval_steps_per_second": 0.718,
"step": 29
},
{
"epoch": 0.8108108108108109,
"grad_norm": 0.09667662531137466,
"learning_rate": 0.0003272727272727273,
"loss": 0.9947,
"step": 30
},
{
"epoch": 0.8108108108108109,
"eval_loss": 1.6455894708633423,
"eval_runtime": 11.163,
"eval_samples_per_second": 11.108,
"eval_steps_per_second": 0.717,
"step": 30
},
{
"epoch": 0.8378378378378378,
"grad_norm": 0.07943412661552429,
"learning_rate": 0.00032323232323232324,
"loss": 0.9682,
"step": 31
},
{
"epoch": 0.8378378378378378,
"eval_loss": 1.6466931104660034,
"eval_runtime": 11.1798,
"eval_samples_per_second": 11.091,
"eval_steps_per_second": 0.716,
"step": 31
},
{
"epoch": 0.8648648648648649,
"grad_norm": 0.08821859955787659,
"learning_rate": 0.0003191919191919192,
"loss": 0.9828,
"step": 32
},
{
"epoch": 0.8648648648648649,
"eval_loss": 1.6367429494857788,
"eval_runtime": 11.1174,
"eval_samples_per_second": 11.154,
"eval_steps_per_second": 0.72,
"step": 32
},
{
"epoch": 0.8918918918918919,
"grad_norm": 0.07129044830799103,
"learning_rate": 0.00031515151515151515,
"loss": 0.9315,
"step": 33
},
{
"epoch": 0.8918918918918919,
"eval_loss": 1.632204294204712,
"eval_runtime": 11.1509,
"eval_samples_per_second": 11.12,
"eval_steps_per_second": 0.717,
"step": 33
},
{
"epoch": 0.918918918918919,
"grad_norm": 0.06505230814218521,
"learning_rate": 0.0003111111111111111,
"loss": 0.9541,
"step": 34
},
{
"epoch": 0.918918918918919,
"eval_loss": 1.6281696557998657,
"eval_runtime": 11.2284,
"eval_samples_per_second": 11.043,
"eval_steps_per_second": 0.712,
"step": 34
},
{
"epoch": 0.9459459459459459,
"grad_norm": 0.0691852867603302,
"learning_rate": 0.00030707070707070705,
"loss": 0.9753,
"step": 35
},
{
"epoch": 0.9459459459459459,
"eval_loss": 1.6272518634796143,
"eval_runtime": 11.1646,
"eval_samples_per_second": 11.106,
"eval_steps_per_second": 0.717,
"step": 35
},
{
"epoch": 0.972972972972973,
"grad_norm": 0.07006494700908661,
"learning_rate": 0.00030303030303030303,
"loss": 0.9399,
"step": 36
},
{
"epoch": 0.972972972972973,
"eval_loss": 1.6321260929107666,
"eval_runtime": 11.2226,
"eval_samples_per_second": 11.049,
"eval_steps_per_second": 0.713,
"step": 36
},
{
"epoch": 1.0,
"grad_norm": 0.08096865564584732,
"learning_rate": 0.000298989898989899,
"loss": 0.9456,
"step": 37
},
{
"epoch": 1.0,
"eval_loss": 1.6329066753387451,
"eval_runtime": 11.1546,
"eval_samples_per_second": 11.116,
"eval_steps_per_second": 0.717,
"step": 37
},
{
"epoch": 1.027027027027027,
"grad_norm": 0.06808632612228394,
"learning_rate": 0.00029494949494949493,
"loss": 0.9535,
"step": 38
},
{
"epoch": 1.027027027027027,
"eval_loss": 1.6349523067474365,
"eval_runtime": 11.1118,
"eval_samples_per_second": 11.159,
"eval_steps_per_second": 0.72,
"step": 38
},
{
"epoch": 1.054054054054054,
"grad_norm": 0.06536230444908142,
"learning_rate": 0.0002909090909090909,
"loss": 0.913,
"step": 39
},
{
"epoch": 1.054054054054054,
"eval_loss": 1.6410040855407715,
"eval_runtime": 11.252,
"eval_samples_per_second": 11.02,
"eval_steps_per_second": 0.711,
"step": 39
},
{
"epoch": 1.0810810810810811,
"grad_norm": 0.06929512321949005,
"learning_rate": 0.00028686868686868684,
"loss": 0.8924,
"step": 40
},
{
"epoch": 1.0810810810810811,
"eval_loss": 1.6510140895843506,
"eval_runtime": 11.1385,
"eval_samples_per_second": 11.133,
"eval_steps_per_second": 0.718,
"step": 40
},
{
"epoch": 1.1081081081081081,
"grad_norm": 0.06684686243534088,
"learning_rate": 0.0002828282828282828,
"loss": 0.901,
"step": 41
},
{
"epoch": 1.1081081081081081,
"eval_loss": 1.6621599197387695,
"eval_runtime": 11.1957,
"eval_samples_per_second": 11.076,
"eval_steps_per_second": 0.715,
"step": 41
},
{
"epoch": 1.135135135135135,
"grad_norm": 0.07049503922462463,
"learning_rate": 0.0002787878787878788,
"loss": 0.9256,
"step": 42
},
{
"epoch": 1.135135135135135,
"eval_loss": 1.6734611988067627,
"eval_runtime": 11.1581,
"eval_samples_per_second": 11.113,
"eval_steps_per_second": 0.717,
"step": 42
},
{
"epoch": 1.1621621621621623,
"grad_norm": 0.07425494492053986,
"learning_rate": 0.0002747474747474748,
"loss": 0.9002,
"step": 43
},
{
"epoch": 1.1621621621621623,
"eval_loss": 1.6794347763061523,
"eval_runtime": 11.0906,
"eval_samples_per_second": 11.181,
"eval_steps_per_second": 0.721,
"step": 43
},
{
"epoch": 1.1891891891891893,
"grad_norm": 0.07607400417327881,
"learning_rate": 0.00027070707070707075,
"loss": 0.8958,
"step": 44
},
{
"epoch": 1.1891891891891893,
"eval_loss": 1.6791983842849731,
"eval_runtime": 11.9226,
"eval_samples_per_second": 10.4,
"eval_steps_per_second": 0.671,
"step": 44
},
{
"epoch": 1.2162162162162162,
"grad_norm": 0.08622392266988754,
"learning_rate": 0.0002666666666666667,
"loss": 0.9143,
"step": 45
},
{
"epoch": 1.2162162162162162,
"eval_loss": 1.671402096748352,
"eval_runtime": 11.1349,
"eval_samples_per_second": 11.136,
"eval_steps_per_second": 0.718,
"step": 45
},
{
"epoch": 1.2432432432432432,
"grad_norm": 0.07751356065273285,
"learning_rate": 0.00026262626262626266,
"loss": 0.9092,
"step": 46
},
{
"epoch": 1.2432432432432432,
"eval_loss": 1.6581931114196777,
"eval_runtime": 11.1166,
"eval_samples_per_second": 11.154,
"eval_steps_per_second": 0.72,
"step": 46
},
{
"epoch": 1.2702702702702702,
"grad_norm": 0.07954052835702896,
"learning_rate": 0.00025858585858585864,
"loss": 0.9116,
"step": 47
},
{
"epoch": 1.2702702702702702,
"eval_loss": 1.6473214626312256,
"eval_runtime": 11.1774,
"eval_samples_per_second": 11.094,
"eval_steps_per_second": 0.716,
"step": 47
},
{
"epoch": 1.2972972972972974,
"grad_norm": 0.08470755070447922,
"learning_rate": 0.00025454545454545456,
"loss": 0.9039,
"step": 48
},
{
"epoch": 1.2972972972972974,
"eval_loss": 1.6490856409072876,
"eval_runtime": 11.1926,
"eval_samples_per_second": 11.079,
"eval_steps_per_second": 0.715,
"step": 48
},
{
"epoch": 1.3243243243243243,
"grad_norm": 0.08538611233234406,
"learning_rate": 0.00025050505050505054,
"loss": 0.8945,
"step": 49
},
{
"epoch": 1.3243243243243243,
"eval_loss": 1.6652594804763794,
"eval_runtime": 11.189,
"eval_samples_per_second": 11.082,
"eval_steps_per_second": 0.715,
"step": 49
},
{
"epoch": 1.3513513513513513,
"grad_norm": 0.08962593227624893,
"learning_rate": 0.00024646464646464647,
"loss": 0.8878,
"step": 50
},
{
"epoch": 1.3513513513513513,
"eval_loss": 1.6737427711486816,
"eval_runtime": 11.194,
"eval_samples_per_second": 11.077,
"eval_steps_per_second": 0.715,
"step": 50
},
{
"epoch": 1.3783783783783785,
"grad_norm": 0.09344533085823059,
"learning_rate": 0.00024242424242424245,
"loss": 0.8921,
"step": 51
},
{
"epoch": 1.3783783783783785,
"eval_loss": 1.6767627000808716,
"eval_runtime": 11.9154,
"eval_samples_per_second": 10.407,
"eval_steps_per_second": 0.671,
"step": 51
},
{
"epoch": 1.4054054054054055,
"grad_norm": 0.08334413915872574,
"learning_rate": 0.00023838383838383837,
"loss": 0.8879,
"step": 52
},
{
"epoch": 1.4054054054054055,
"eval_loss": 1.6724966764450073,
"eval_runtime": 11.0787,
"eval_samples_per_second": 11.193,
"eval_steps_per_second": 0.722,
"step": 52
},
{
"epoch": 1.4324324324324325,
"grad_norm": 0.0849594995379448,
"learning_rate": 0.00023434343434343435,
"loss": 0.9021,
"step": 53
},
{
"epoch": 1.4324324324324325,
"eval_loss": 1.6656012535095215,
"eval_runtime": 11.1933,
"eval_samples_per_second": 11.078,
"eval_steps_per_second": 0.715,
"step": 53
},
{
"epoch": 1.4594594594594594,
"grad_norm": 0.09131031483411789,
"learning_rate": 0.00023030303030303033,
"loss": 0.8975,
"step": 54
},
{
"epoch": 1.4594594594594594,
"eval_loss": 1.6548563241958618,
"eval_runtime": 11.2025,
"eval_samples_per_second": 11.069,
"eval_steps_per_second": 0.714,
"step": 54
},
{
"epoch": 1.4864864864864864,
"grad_norm": 0.0867348462343216,
"learning_rate": 0.00022626262626262625,
"loss": 0.8746,
"step": 55
},
{
"epoch": 1.4864864864864864,
"eval_loss": 1.6583256721496582,
"eval_runtime": 11.1258,
"eval_samples_per_second": 11.145,
"eval_steps_per_second": 0.719,
"step": 55
},
{
"epoch": 1.5135135135135136,
"grad_norm": 0.0924694687128067,
"learning_rate": 0.00022222222222222223,
"loss": 0.8818,
"step": 56
},
{
"epoch": 1.5135135135135136,
"eval_loss": 1.6754804849624634,
"eval_runtime": 11.1671,
"eval_samples_per_second": 11.104,
"eval_steps_per_second": 0.716,
"step": 56
},
{
"epoch": 1.5405405405405406,
"grad_norm": 0.09178721904754639,
"learning_rate": 0.00021818181818181818,
"loss": 0.9022,
"step": 57
},
{
"epoch": 1.5405405405405406,
"eval_loss": 1.690301775932312,
"eval_runtime": 11.0856,
"eval_samples_per_second": 11.186,
"eval_steps_per_second": 0.722,
"step": 57
},
{
"epoch": 1.5675675675675675,
"grad_norm": 0.08697674423456192,
"learning_rate": 0.00021414141414141416,
"loss": 0.8596,
"step": 58
},
{
"epoch": 1.5675675675675675,
"eval_loss": 1.7041698694229126,
"eval_runtime": 11.2652,
"eval_samples_per_second": 11.007,
"eval_steps_per_second": 0.71,
"step": 58
},
{
"epoch": 1.5945945945945947,
"grad_norm": 0.08802448958158493,
"learning_rate": 0.00021010101010101014,
"loss": 0.8572,
"step": 59
},
{
"epoch": 1.5945945945945947,
"eval_loss": 1.7124710083007812,
"eval_runtime": 11.152,
"eval_samples_per_second": 11.119,
"eval_steps_per_second": 0.717,
"step": 59
},
{
"epoch": 1.6216216216216215,
"grad_norm": 0.10100586712360382,
"learning_rate": 0.00020606060606060607,
"loss": 0.8883,
"step": 60
},
{
"epoch": 1.6216216216216215,
"eval_loss": 1.7054810523986816,
"eval_runtime": 11.1538,
"eval_samples_per_second": 11.117,
"eval_steps_per_second": 0.717,
"step": 60
},
{
"epoch": 1.6486486486486487,
"grad_norm": 0.11259682476520538,
"learning_rate": 0.00020202020202020205,
"loss": 0.8968,
"step": 61
},
{
"epoch": 1.6486486486486487,
"eval_loss": 1.683681845664978,
"eval_runtime": 11.1309,
"eval_samples_per_second": 11.14,
"eval_steps_per_second": 0.719,
"step": 61
},
{
"epoch": 1.6756756756756757,
"grad_norm": 0.0911870002746582,
"learning_rate": 0.000197979797979798,
"loss": 0.8287,
"step": 62
},
{
"epoch": 1.6756756756756757,
"eval_loss": 1.6670082807540894,
"eval_runtime": 11.1469,
"eval_samples_per_second": 11.124,
"eval_steps_per_second": 0.718,
"step": 62
},
{
"epoch": 1.7027027027027026,
"grad_norm": 0.08852899819612503,
"learning_rate": 0.00019393939393939395,
"loss": 0.8699,
"step": 63
},
{
"epoch": 1.7027027027027026,
"eval_loss": 1.6602628231048584,
"eval_runtime": 11.2189,
"eval_samples_per_second": 11.053,
"eval_steps_per_second": 0.713,
"step": 63
},
{
"epoch": 1.7297297297297298,
"grad_norm": 0.0944983959197998,
"learning_rate": 0.0001898989898989899,
"loss": 0.8812,
"step": 64
},
{
"epoch": 1.7297297297297298,
"eval_loss": 1.664315104484558,
"eval_runtime": 11.1484,
"eval_samples_per_second": 11.123,
"eval_steps_per_second": 0.718,
"step": 64
},
{
"epoch": 1.7567567567567568,
"grad_norm": 0.09581159055233002,
"learning_rate": 0.00018585858585858586,
"loss": 0.8671,
"step": 65
},
{
"epoch": 1.7567567567567568,
"eval_loss": 1.6748771667480469,
"eval_runtime": 11.2306,
"eval_samples_per_second": 11.041,
"eval_steps_per_second": 0.712,
"step": 65
},
{
"epoch": 1.7837837837837838,
"grad_norm": 0.09136416763067245,
"learning_rate": 0.00018181818181818183,
"loss": 0.8857,
"step": 66
},
{
"epoch": 1.7837837837837838,
"eval_loss": 1.6853251457214355,
"eval_runtime": 11.1546,
"eval_samples_per_second": 11.116,
"eval_steps_per_second": 0.717,
"step": 66
},
{
"epoch": 1.810810810810811,
"grad_norm": 0.09602421522140503,
"learning_rate": 0.00017777777777777779,
"loss": 0.8578,
"step": 67
},
{
"epoch": 1.810810810810811,
"eval_loss": 1.6973600387573242,
"eval_runtime": 11.2294,
"eval_samples_per_second": 11.042,
"eval_steps_per_second": 0.712,
"step": 67
},
{
"epoch": 1.8378378378378377,
"grad_norm": 0.10458902269601822,
"learning_rate": 0.00017373737373737377,
"loss": 0.8767,
"step": 68
},
{
"epoch": 1.8378378378378377,
"eval_loss": 1.7042148113250732,
"eval_runtime": 11.1244,
"eval_samples_per_second": 11.147,
"eval_steps_per_second": 0.719,
"step": 68
},
{
"epoch": 1.864864864864865,
"grad_norm": 0.12431007623672485,
"learning_rate": 0.00016969696969696972,
"loss": 0.8914,
"step": 69
},
{
"epoch": 1.864864864864865,
"eval_loss": 1.702877402305603,
"eval_runtime": 11.144,
"eval_samples_per_second": 11.127,
"eval_steps_per_second": 0.718,
"step": 69
},
{
"epoch": 1.8918918918918919,
"grad_norm": 0.11106838285923004,
"learning_rate": 0.00016565656565656567,
"loss": 0.8843,
"step": 70
},
{
"epoch": 1.8918918918918919,
"eval_loss": 1.691541314125061,
"eval_runtime": 11.1751,
"eval_samples_per_second": 11.096,
"eval_steps_per_second": 0.716,
"step": 70
},
{
"epoch": 1.9189189189189189,
"grad_norm": 0.0916941687464714,
"learning_rate": 0.00016161616161616162,
"loss": 0.8576,
"step": 71
},
{
"epoch": 1.9189189189189189,
"eval_loss": 1.68495774269104,
"eval_runtime": 11.208,
"eval_samples_per_second": 11.063,
"eval_steps_per_second": 0.714,
"step": 71
},
{
"epoch": 1.945945945945946,
"grad_norm": 0.10394521802663803,
"learning_rate": 0.00015757575757575757,
"loss": 0.853,
"step": 72
},
{
"epoch": 1.945945945945946,
"eval_loss": 1.6894781589508057,
"eval_runtime": 11.1546,
"eval_samples_per_second": 11.117,
"eval_steps_per_second": 0.717,
"step": 72
},
{
"epoch": 1.972972972972973,
"grad_norm": 0.09884631633758545,
"learning_rate": 0.00015353535353535353,
"loss": 0.8346,
"step": 73
},
{
"epoch": 1.972972972972973,
"eval_loss": 1.7004313468933105,
"eval_runtime": 11.107,
"eval_samples_per_second": 11.164,
"eval_steps_per_second": 0.72,
"step": 73
},
{
"epoch": 2.0,
"grad_norm": 0.11277178674936295,
"learning_rate": 0.0001494949494949495,
"loss": 0.8524,
"step": 74
},
{
"epoch": 2.0,
"eval_loss": 1.704146385192871,
"eval_runtime": 11.1746,
"eval_samples_per_second": 11.097,
"eval_steps_per_second": 0.716,
"step": 74
},
{
"epoch": 2.027027027027027,
"grad_norm": 0.0982976108789444,
"learning_rate": 0.00014545454545454546,
"loss": 0.8267,
"step": 75
},
{
"epoch": 2.027027027027027,
"eval_loss": 1.7128726243972778,
"eval_runtime": 11.1189,
"eval_samples_per_second": 11.152,
"eval_steps_per_second": 0.719,
"step": 75
},
{
"epoch": 2.054054054054054,
"grad_norm": 0.09641794860363007,
"learning_rate": 0.0001414141414141414,
"loss": 0.8157,
"step": 76
},
{
"epoch": 2.054054054054054,
"eval_loss": 1.7283172607421875,
"eval_runtime": 11.1515,
"eval_samples_per_second": 11.12,
"eval_steps_per_second": 0.717,
"step": 76
},
{
"epoch": 2.081081081081081,
"grad_norm": 0.10205162316560745,
"learning_rate": 0.0001373737373737374,
"loss": 0.7977,
"step": 77
},
{
"epoch": 2.081081081081081,
"eval_loss": 1.7365342378616333,
"eval_runtime": 11.0683,
"eval_samples_per_second": 11.203,
"eval_steps_per_second": 0.723,
"step": 77
},
{
"epoch": 2.108108108108108,
"grad_norm": 0.1005796566605568,
"learning_rate": 0.00013333333333333334,
"loss": 0.815,
"step": 78
},
{
"epoch": 2.108108108108108,
"eval_loss": 1.7438631057739258,
"eval_runtime": 11.1586,
"eval_samples_per_second": 11.113,
"eval_steps_per_second": 0.717,
"step": 78
},
{
"epoch": 2.135135135135135,
"grad_norm": 0.111351877450943,
"learning_rate": 0.00012929292929292932,
"loss": 0.7987,
"step": 79
},
{
"epoch": 2.135135135135135,
"eval_loss": 1.7426629066467285,
"eval_runtime": 11.1248,
"eval_samples_per_second": 11.146,
"eval_steps_per_second": 0.719,
"step": 79
},
{
"epoch": 2.1621621621621623,
"grad_norm": 0.10649854689836502,
"learning_rate": 0.00012525252525252527,
"loss": 0.8072,
"step": 80
},
{
"epoch": 2.1621621621621623,
"eval_loss": 1.7457417249679565,
"eval_runtime": 11.1907,
"eval_samples_per_second": 11.081,
"eval_steps_per_second": 0.715,
"step": 80
},
{
"epoch": 2.189189189189189,
"grad_norm": 0.09818455576896667,
"learning_rate": 0.00012121212121212122,
"loss": 0.8024,
"step": 81
},
{
"epoch": 2.189189189189189,
"eval_loss": 1.7508054971694946,
"eval_runtime": 11.1556,
"eval_samples_per_second": 11.115,
"eval_steps_per_second": 0.717,
"step": 81
},
{
"epoch": 2.2162162162162162,
"grad_norm": 0.1122928112745285,
"learning_rate": 0.00011717171717171717,
"loss": 0.8208,
"step": 82
},
{
"epoch": 2.2162162162162162,
"eval_loss": 1.7618576288223267,
"eval_runtime": 11.2048,
"eval_samples_per_second": 11.067,
"eval_steps_per_second": 0.714,
"step": 82
},
{
"epoch": 2.2432432432432434,
"grad_norm": 0.10680953413248062,
"learning_rate": 0.00011313131313131313,
"loss": 0.7858,
"step": 83
},
{
"epoch": 2.2432432432432434,
"eval_loss": 1.7693935632705688,
"eval_runtime": 11.1465,
"eval_samples_per_second": 11.125,
"eval_steps_per_second": 0.718,
"step": 83
},
{
"epoch": 2.27027027027027,
"grad_norm": 0.11239504814147949,
"learning_rate": 0.00010909090909090909,
"loss": 0.8082,
"step": 84
},
{
"epoch": 2.27027027027027,
"eval_loss": 1.780279517173767,
"eval_runtime": 11.1781,
"eval_samples_per_second": 11.093,
"eval_steps_per_second": 0.716,
"step": 84
},
{
"epoch": 2.2972972972972974,
"grad_norm": 0.12527017295360565,
"learning_rate": 0.00010505050505050507,
"loss": 0.7943,
"step": 85
},
{
"epoch": 2.2972972972972974,
"eval_loss": 1.7797414064407349,
"eval_runtime": 11.1811,
"eval_samples_per_second": 11.09,
"eval_steps_per_second": 0.715,
"step": 85
},
{
"epoch": 2.3243243243243246,
"grad_norm": 0.13715100288391113,
"learning_rate": 0.00010101010101010102,
"loss": 0.8338,
"step": 86
},
{
"epoch": 2.3243243243243246,
"eval_loss": 1.7673975229263306,
"eval_runtime": 11.217,
"eval_samples_per_second": 11.055,
"eval_steps_per_second": 0.713,
"step": 86
},
{
"epoch": 2.3513513513513513,
"grad_norm": 0.11464574187994003,
"learning_rate": 9.696969696969698e-05,
"loss": 0.8038,
"step": 87
},
{
"epoch": 2.3513513513513513,
"eval_loss": 1.75497305393219,
"eval_runtime": 11.1384,
"eval_samples_per_second": 11.133,
"eval_steps_per_second": 0.718,
"step": 87
},
{
"epoch": 2.3783783783783785,
"grad_norm": 0.11703913658857346,
"learning_rate": 9.292929292929293e-05,
"loss": 0.812,
"step": 88
},
{
"epoch": 2.3783783783783785,
"eval_loss": 1.7427462339401245,
"eval_runtime": 11.1796,
"eval_samples_per_second": 11.092,
"eval_steps_per_second": 0.716,
"step": 88
},
{
"epoch": 2.4054054054054053,
"grad_norm": 0.1178952306509018,
"learning_rate": 8.888888888888889e-05,
"loss": 0.8117,
"step": 89
},
{
"epoch": 2.4054054054054053,
"eval_loss": 1.7358033657073975,
"eval_runtime": 11.1223,
"eval_samples_per_second": 11.149,
"eval_steps_per_second": 0.719,
"step": 89
},
{
"epoch": 2.4324324324324325,
"grad_norm": 0.1348622441291809,
"learning_rate": 8.484848484848486e-05,
"loss": 0.8172,
"step": 90
},
{
"epoch": 2.4324324324324325,
"eval_loss": 1.7397310733795166,
"eval_runtime": 11.1841,
"eval_samples_per_second": 11.087,
"eval_steps_per_second": 0.715,
"step": 90
},
{
"epoch": 2.4594594594594597,
"grad_norm": 0.12706562876701355,
"learning_rate": 8.080808080808081e-05,
"loss": 0.808,
"step": 91
},
{
"epoch": 2.4594594594594597,
"eval_loss": 1.751862645149231,
"eval_runtime": 11.0963,
"eval_samples_per_second": 11.175,
"eval_steps_per_second": 0.721,
"step": 91
},
{
"epoch": 2.4864864864864864,
"grad_norm": 0.11583199352025986,
"learning_rate": 7.676767676767676e-05,
"loss": 0.8108,
"step": 92
},
{
"epoch": 2.4864864864864864,
"eval_loss": 1.764027714729309,
"eval_runtime": 11.1926,
"eval_samples_per_second": 11.079,
"eval_steps_per_second": 0.715,
"step": 92
},
{
"epoch": 2.5135135135135136,
"grad_norm": 0.12227907031774521,
"learning_rate": 7.272727272727273e-05,
"loss": 0.8086,
"step": 93
},
{
"epoch": 2.5135135135135136,
"eval_loss": 1.7744636535644531,
"eval_runtime": 11.1394,
"eval_samples_per_second": 11.132,
"eval_steps_per_second": 0.718,
"step": 93
},
{
"epoch": 2.5405405405405403,
"grad_norm": 0.1270534247159958,
"learning_rate": 6.86868686868687e-05,
"loss": 0.8079,
"step": 94
},
{
"epoch": 2.5405405405405403,
"eval_loss": 1.7796293497085571,
"eval_runtime": 11.1924,
"eval_samples_per_second": 11.079,
"eval_steps_per_second": 0.715,
"step": 94
},
{
"epoch": 2.5675675675675675,
"grad_norm": 0.1266915649175644,
"learning_rate": 6.464646464646466e-05,
"loss": 0.7876,
"step": 95
},
{
"epoch": 2.5675675675675675,
"eval_loss": 1.7799245119094849,
"eval_runtime": 11.1921,
"eval_samples_per_second": 11.079,
"eval_steps_per_second": 0.715,
"step": 95
},
{
"epoch": 2.5945945945945947,
"grad_norm": 0.13456638157367706,
"learning_rate": 6.060606060606061e-05,
"loss": 0.7756,
"step": 96
},
{
"epoch": 2.5945945945945947,
"eval_loss": 1.7748299837112427,
"eval_runtime": 11.1592,
"eval_samples_per_second": 11.112,
"eval_steps_per_second": 0.717,
"step": 96
},
{
"epoch": 2.6216216216216215,
"grad_norm": 0.12409545481204987,
"learning_rate": 5.6565656565656563e-05,
"loss": 0.795,
"step": 97
},
{
"epoch": 2.6216216216216215,
"eval_loss": 1.7676475048065186,
"eval_runtime": 11.1442,
"eval_samples_per_second": 11.127,
"eval_steps_per_second": 0.718,
"step": 97
},
{
"epoch": 2.6486486486486487,
"grad_norm": 0.11815664917230606,
"learning_rate": 5.2525252525252536e-05,
"loss": 0.7706,
"step": 98
},
{
"epoch": 2.6486486486486487,
"eval_loss": 1.7634869813919067,
"eval_runtime": 11.2593,
"eval_samples_per_second": 11.013,
"eval_steps_per_second": 0.711,
"step": 98
},
{
"epoch": 2.6756756756756754,
"grad_norm": 0.11971355229616165,
"learning_rate": 4.848484848484849e-05,
"loss": 0.792,
"step": 99
},
{
"epoch": 2.6756756756756754,
"eval_loss": 1.7610678672790527,
"eval_runtime": 11.1535,
"eval_samples_per_second": 11.118,
"eval_steps_per_second": 0.717,
"step": 99
},
{
"epoch": 2.7027027027027026,
"grad_norm": 0.12585307657718658,
"learning_rate": 4.4444444444444447e-05,
"loss": 0.7942,
"step": 100
},
{
"epoch": 2.7027027027027026,
"eval_loss": 1.7610435485839844,
"eval_runtime": 11.0721,
"eval_samples_per_second": 11.199,
"eval_steps_per_second": 0.723,
"step": 100
},
{
"epoch": 2.72972972972973,
"grad_norm": 0.11809434741735458,
"learning_rate": 4.0404040404040405e-05,
"loss": 0.7697,
"step": 101
},
{
"epoch": 2.72972972972973,
"eval_loss": 1.7632513046264648,
"eval_runtime": 11.1597,
"eval_samples_per_second": 11.111,
"eval_steps_per_second": 0.717,
"step": 101
},
{
"epoch": 2.756756756756757,
"grad_norm": 0.12419259548187256,
"learning_rate": 3.6363636363636364e-05,
"loss": 0.7789,
"step": 102
},
{
"epoch": 2.756756756756757,
"eval_loss": 1.7658323049545288,
"eval_runtime": 11.1861,
"eval_samples_per_second": 11.085,
"eval_steps_per_second": 0.715,
"step": 102
},
{
"epoch": 2.7837837837837838,
"grad_norm": 0.12964174151420593,
"learning_rate": 3.232323232323233e-05,
"loss": 0.8047,
"step": 103
},
{
"epoch": 2.7837837837837838,
"eval_loss": 1.7687405347824097,
"eval_runtime": 11.0868,
"eval_samples_per_second": 11.184,
"eval_steps_per_second": 0.722,
"step": 103
},
{
"epoch": 2.810810810810811,
"grad_norm": 0.1257963329553604,
"learning_rate": 2.8282828282828282e-05,
"loss": 0.7653,
"step": 104
},
{
"epoch": 2.810810810810811,
"eval_loss": 1.7723119258880615,
"eval_runtime": 11.1173,
"eval_samples_per_second": 11.154,
"eval_steps_per_second": 0.72,
"step": 104
},
{
"epoch": 2.8378378378378377,
"grad_norm": 0.12039487063884735,
"learning_rate": 2.4242424242424244e-05,
"loss": 0.7753,
"step": 105
},
{
"epoch": 2.8378378378378377,
"eval_loss": 1.7749425172805786,
"eval_runtime": 11.157,
"eval_samples_per_second": 11.114,
"eval_steps_per_second": 0.717,
"step": 105
},
{
"epoch": 2.864864864864865,
"grad_norm": 0.13161887228488922,
"learning_rate": 2.0202020202020203e-05,
"loss": 0.7631,
"step": 106
},
{
"epoch": 2.864864864864865,
"eval_loss": 1.7760539054870605,
"eval_runtime": 11.1933,
"eval_samples_per_second": 11.078,
"eval_steps_per_second": 0.715,
"step": 106
},
{
"epoch": 2.891891891891892,
"grad_norm": 0.12347672879695892,
"learning_rate": 1.6161616161616165e-05,
"loss": 0.7821,
"step": 107
},
{
"epoch": 2.891891891891892,
"eval_loss": 1.7769430875778198,
"eval_runtime": 11.169,
"eval_samples_per_second": 11.102,
"eval_steps_per_second": 0.716,
"step": 107
},
{
"epoch": 2.918918918918919,
"grad_norm": 0.12539447844028473,
"learning_rate": 1.2121212121212122e-05,
"loss": 0.7682,
"step": 108
},
{
"epoch": 2.918918918918919,
"eval_loss": 1.77769136428833,
"eval_runtime": 11.2241,
"eval_samples_per_second": 11.048,
"eval_steps_per_second": 0.713,
"step": 108
},
{
"epoch": 2.945945945945946,
"grad_norm": 0.12666372954845428,
"learning_rate": 8.080808080808082e-06,
"loss": 0.7893,
"step": 109
},
{
"epoch": 2.945945945945946,
"eval_loss": 1.7779371738433838,
"eval_runtime": 11.1071,
"eval_samples_per_second": 11.164,
"eval_steps_per_second": 0.72,
"step": 109
},
{
"epoch": 2.972972972972973,
"grad_norm": 0.11765623837709427,
"learning_rate": 4.040404040404041e-06,
"loss": 0.7729,
"step": 110
},
{
"epoch": 2.972972972972973,
"eval_loss": 1.7782450914382935,
"eval_runtime": 11.121,
"eval_samples_per_second": 11.15,
"eval_steps_per_second": 0.719,
"step": 110
},
{
"epoch": 3.0,
"grad_norm": 0.14280778169631958,
"learning_rate": 0.0,
"loss": 0.7783,
"step": 111
},
{
"epoch": 3.0,
"eval_loss": 1.7782894372940063,
"eval_runtime": 11.0882,
"eval_samples_per_second": 11.183,
"eval_steps_per_second": 0.721,
"step": 111
},
{
"epoch": 3.0,
"step": 111,
"total_flos": 4.092563650668134e+16,
"train_loss": 1.148636352371525,
"train_runtime": 2293.4445,
"train_samples_per_second": 1.54,
"train_steps_per_second": 0.048
}
],
"logging_steps": 1,
"max_steps": 111,
"num_input_tokens_seen": 0,
"num_train_epochs": 3,
"save_steps": 10,
"stateful_callbacks": {
"TrainerControl": {
"args": {
"should_epoch_stop": false,
"should_evaluate": false,
"should_log": false,
"should_save": true,
"should_training_stop": false
},
"attributes": {}
}
},
"total_flos": 4.092563650668134e+16,
"train_batch_size": 32,
"trial_name": null,
"trial_params": null
}