Files
llama_finetune/outputs/checkpoint-120/trainer_state.json
2025-06-16 16:28:00 +02:00

875 lines
21 KiB
JSON

{
"best_global_step": null,
"best_metric": null,
"best_model_checkpoint": null,
"epoch": 3.2482758620689656,
"eval_steps": 500,
"global_step": 120,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.027586206896551724,
"grad_norm": 0.8273730278015137,
"learning_rate": 0.0,
"loss": 2.5179,
"step": 1
},
{
"epoch": 0.05517241379310345,
"grad_norm": 0.8761172890663147,
"learning_rate": 4e-05,
"loss": 2.8481,
"step": 2
},
{
"epoch": 0.08275862068965517,
"grad_norm": 0.7226160764694214,
"learning_rate": 8e-05,
"loss": 2.6317,
"step": 3
},
{
"epoch": 0.1103448275862069,
"grad_norm": 0.7074100375175476,
"learning_rate": 0.00012,
"loss": 2.691,
"step": 4
},
{
"epoch": 0.13793103448275862,
"grad_norm": 0.8948147296905518,
"learning_rate": 0.00016,
"loss": 2.5618,
"step": 5
},
{
"epoch": 0.16551724137931034,
"grad_norm": 0.6676309108734131,
"learning_rate": 0.0002,
"loss": 1.7508,
"step": 6
},
{
"epoch": 0.19310344827586207,
"grad_norm": 0.7051960825920105,
"learning_rate": 0.0001982608695652174,
"loss": 2.1303,
"step": 7
},
{
"epoch": 0.2206896551724138,
"grad_norm": 0.9678179025650024,
"learning_rate": 0.0001965217391304348,
"loss": 1.8536,
"step": 8
},
{
"epoch": 0.2482758620689655,
"grad_norm": 0.9203357696533203,
"learning_rate": 0.00019478260869565218,
"loss": 2.055,
"step": 9
},
{
"epoch": 0.27586206896551724,
"grad_norm": 0.8674907684326172,
"learning_rate": 0.00019304347826086958,
"loss": 2.1066,
"step": 10
},
{
"epoch": 0.30344827586206896,
"grad_norm": 1.1215510368347168,
"learning_rate": 0.00019130434782608697,
"loss": 2.1572,
"step": 11
},
{
"epoch": 0.3310344827586207,
"grad_norm": 0.9455694556236267,
"learning_rate": 0.00018956521739130436,
"loss": 1.9253,
"step": 12
},
{
"epoch": 0.3586206896551724,
"grad_norm": 1.1137198209762573,
"learning_rate": 0.00018782608695652175,
"loss": 2.1465,
"step": 13
},
{
"epoch": 0.38620689655172413,
"grad_norm": 1.0382691621780396,
"learning_rate": 0.00018608695652173914,
"loss": 1.8506,
"step": 14
},
{
"epoch": 0.41379310344827586,
"grad_norm": 0.827080249786377,
"learning_rate": 0.00018434782608695653,
"loss": 1.3841,
"step": 15
},
{
"epoch": 0.4413793103448276,
"grad_norm": 0.9159547686576843,
"learning_rate": 0.00018260869565217392,
"loss": 1.8481,
"step": 16
},
{
"epoch": 0.4689655172413793,
"grad_norm": 0.8442085385322571,
"learning_rate": 0.00018086956521739132,
"loss": 1.7424,
"step": 17
},
{
"epoch": 0.496551724137931,
"grad_norm": 0.864258348941803,
"learning_rate": 0.0001791304347826087,
"loss": 1.6131,
"step": 18
},
{
"epoch": 0.5241379310344828,
"grad_norm": 1.0820664167404175,
"learning_rate": 0.0001773913043478261,
"loss": 1.6129,
"step": 19
},
{
"epoch": 0.5517241379310345,
"grad_norm": 0.9673957824707031,
"learning_rate": 0.0001756521739130435,
"loss": 1.5257,
"step": 20
},
{
"epoch": 0.5793103448275863,
"grad_norm": 0.9112041592597961,
"learning_rate": 0.00017391304347826088,
"loss": 1.2875,
"step": 21
},
{
"epoch": 0.6068965517241379,
"grad_norm": 0.9094520211219788,
"learning_rate": 0.00017217391304347827,
"loss": 1.3703,
"step": 22
},
{
"epoch": 0.6344827586206897,
"grad_norm": 0.9438947439193726,
"learning_rate": 0.00017043478260869566,
"loss": 1.3091,
"step": 23
},
{
"epoch": 0.6620689655172414,
"grad_norm": 1.3191273212432861,
"learning_rate": 0.00016869565217391306,
"loss": 1.7087,
"step": 24
},
{
"epoch": 0.6896551724137931,
"grad_norm": 0.9469236135482788,
"learning_rate": 0.00016695652173913042,
"loss": 1.4919,
"step": 25
},
{
"epoch": 0.7172413793103448,
"grad_norm": 1.0983434915542603,
"learning_rate": 0.00016521739130434784,
"loss": 1.464,
"step": 26
},
{
"epoch": 0.7448275862068966,
"grad_norm": 0.9698247313499451,
"learning_rate": 0.00016347826086956523,
"loss": 1.4398,
"step": 27
},
{
"epoch": 0.7724137931034483,
"grad_norm": 0.8902468085289001,
"learning_rate": 0.00016173913043478262,
"loss": 1.3191,
"step": 28
},
{
"epoch": 0.8,
"grad_norm": 0.8863650560379028,
"learning_rate": 0.00016,
"loss": 1.2048,
"step": 29
},
{
"epoch": 0.8275862068965517,
"grad_norm": 1.0257900953292847,
"learning_rate": 0.0001582608695652174,
"loss": 1.2085,
"step": 30
},
{
"epoch": 0.8551724137931035,
"grad_norm": 0.9826428294181824,
"learning_rate": 0.0001565217391304348,
"loss": 1.1733,
"step": 31
},
{
"epoch": 0.8827586206896552,
"grad_norm": 0.9123853445053101,
"learning_rate": 0.0001547826086956522,
"loss": 1.4226,
"step": 32
},
{
"epoch": 0.9103448275862069,
"grad_norm": 0.8653205633163452,
"learning_rate": 0.00015304347826086958,
"loss": 1.3998,
"step": 33
},
{
"epoch": 0.9379310344827586,
"grad_norm": 1.2209527492523193,
"learning_rate": 0.00015130434782608694,
"loss": 1.2225,
"step": 34
},
{
"epoch": 0.9655172413793104,
"grad_norm": 0.977463960647583,
"learning_rate": 0.00014956521739130436,
"loss": 1.186,
"step": 35
},
{
"epoch": 0.993103448275862,
"grad_norm": 0.8854506611824036,
"learning_rate": 0.00014782608695652173,
"loss": 0.9478,
"step": 36
},
{
"epoch": 1.0,
"grad_norm": 1.660280704498291,
"learning_rate": 0.00014608695652173914,
"loss": 0.7476,
"step": 37
},
{
"epoch": 1.0275862068965518,
"grad_norm": 0.9172261953353882,
"learning_rate": 0.00014434782608695654,
"loss": 0.959,
"step": 38
},
{
"epoch": 1.0551724137931036,
"grad_norm": 0.9950329661369324,
"learning_rate": 0.00014260869565217393,
"loss": 1.087,
"step": 39
},
{
"epoch": 1.0827586206896551,
"grad_norm": 0.9052255749702454,
"learning_rate": 0.00014086956521739132,
"loss": 1.0335,
"step": 40
},
{
"epoch": 1.110344827586207,
"grad_norm": 0.8859487771987915,
"learning_rate": 0.0001391304347826087,
"loss": 1.0489,
"step": 41
},
{
"epoch": 1.1379310344827587,
"grad_norm": 0.9165846705436707,
"learning_rate": 0.0001373913043478261,
"loss": 1.0135,
"step": 42
},
{
"epoch": 1.1655172413793102,
"grad_norm": 1.2192325592041016,
"learning_rate": 0.00013565217391304347,
"loss": 1.1084,
"step": 43
},
{
"epoch": 1.193103448275862,
"grad_norm": 1.2101364135742188,
"learning_rate": 0.00013391304347826088,
"loss": 1.1635,
"step": 44
},
{
"epoch": 1.2206896551724138,
"grad_norm": 1.099292516708374,
"learning_rate": 0.00013217391304347825,
"loss": 1.1804,
"step": 45
},
{
"epoch": 1.2482758620689656,
"grad_norm": 0.9990763068199158,
"learning_rate": 0.00013043478260869567,
"loss": 0.8802,
"step": 46
},
{
"epoch": 1.2758620689655173,
"grad_norm": 0.9451124668121338,
"learning_rate": 0.00012869565217391303,
"loss": 0.9852,
"step": 47
},
{
"epoch": 1.303448275862069,
"grad_norm": 0.96523118019104,
"learning_rate": 0.00012695652173913045,
"loss": 1.0243,
"step": 48
},
{
"epoch": 1.3310344827586207,
"grad_norm": 1.0256421566009521,
"learning_rate": 0.00012521739130434784,
"loss": 0.8908,
"step": 49
},
{
"epoch": 1.3586206896551725,
"grad_norm": 1.0647811889648438,
"learning_rate": 0.00012347826086956523,
"loss": 0.7224,
"step": 50
},
{
"epoch": 1.386206896551724,
"grad_norm": 1.042438268661499,
"learning_rate": 0.00012173913043478263,
"loss": 0.5707,
"step": 51
},
{
"epoch": 1.4137931034482758,
"grad_norm": 1.0345197916030884,
"learning_rate": 0.00012,
"loss": 0.7433,
"step": 52
},
{
"epoch": 1.4413793103448276,
"grad_norm": 1.0194092988967896,
"learning_rate": 0.00011826086956521741,
"loss": 0.8686,
"step": 53
},
{
"epoch": 1.4689655172413794,
"grad_norm": 1.0849523544311523,
"learning_rate": 0.00011652173913043479,
"loss": 0.9063,
"step": 54
},
{
"epoch": 1.4965517241379311,
"grad_norm": 1.3685775995254517,
"learning_rate": 0.00011478260869565218,
"loss": 0.8633,
"step": 55
},
{
"epoch": 1.524137931034483,
"grad_norm": 1.2180424928665161,
"learning_rate": 0.00011304347826086956,
"loss": 0.8131,
"step": 56
},
{
"epoch": 1.5517241379310345,
"grad_norm": 1.027662992477417,
"learning_rate": 0.00011130434782608696,
"loss": 0.8072,
"step": 57
},
{
"epoch": 1.5793103448275863,
"grad_norm": 1.0541893243789673,
"learning_rate": 0.00010956521739130434,
"loss": 0.5513,
"step": 58
},
{
"epoch": 1.6068965517241378,
"grad_norm": 0.9840919375419617,
"learning_rate": 0.00010782608695652174,
"loss": 0.6515,
"step": 59
},
{
"epoch": 1.6344827586206896,
"grad_norm": 1.1880444288253784,
"learning_rate": 0.00010608695652173915,
"loss": 0.9452,
"step": 60
},
{
"epoch": 1.6620689655172414,
"grad_norm": 1.1577789783477783,
"learning_rate": 0.00010434782608695653,
"loss": 0.5324,
"step": 61
},
{
"epoch": 1.6896551724137931,
"grad_norm": 1.4066375494003296,
"learning_rate": 0.00010260869565217393,
"loss": 0.7674,
"step": 62
},
{
"epoch": 1.717241379310345,
"grad_norm": 1.5101147890090942,
"learning_rate": 0.00010086956521739131,
"loss": 0.7358,
"step": 63
},
{
"epoch": 1.7448275862068967,
"grad_norm": 1.2288732528686523,
"learning_rate": 9.91304347826087e-05,
"loss": 0.762,
"step": 64
},
{
"epoch": 1.7724137931034483,
"grad_norm": 1.1810815334320068,
"learning_rate": 9.739130434782609e-05,
"loss": 0.824,
"step": 65
},
{
"epoch": 1.8,
"grad_norm": 1.0823071002960205,
"learning_rate": 9.565217391304348e-05,
"loss": 0.683,
"step": 66
},
{
"epoch": 1.8275862068965516,
"grad_norm": 1.1553919315338135,
"learning_rate": 9.391304347826087e-05,
"loss": 0.737,
"step": 67
},
{
"epoch": 1.8551724137931034,
"grad_norm": 1.3099501132965088,
"learning_rate": 9.217391304347827e-05,
"loss": 0.4993,
"step": 68
},
{
"epoch": 1.8827586206896552,
"grad_norm": 1.3969764709472656,
"learning_rate": 9.043478260869566e-05,
"loss": 0.6857,
"step": 69
},
{
"epoch": 1.910344827586207,
"grad_norm": 1.2558094263076782,
"learning_rate": 8.869565217391305e-05,
"loss": 0.5347,
"step": 70
},
{
"epoch": 1.9379310344827587,
"grad_norm": 1.2341969013214111,
"learning_rate": 8.695652173913044e-05,
"loss": 0.6651,
"step": 71
},
{
"epoch": 1.9655172413793105,
"grad_norm": 1.2917416095733643,
"learning_rate": 8.521739130434783e-05,
"loss": 0.6177,
"step": 72
},
{
"epoch": 1.993103448275862,
"grad_norm": 1.2867687940597534,
"learning_rate": 8.347826086956521e-05,
"loss": 0.6304,
"step": 73
},
{
"epoch": 2.0,
"grad_norm": 2.8276941776275635,
"learning_rate": 8.173913043478262e-05,
"loss": 0.829,
"step": 74
},
{
"epoch": 2.027586206896552,
"grad_norm": 1.1930606365203857,
"learning_rate": 8e-05,
"loss": 0.3571,
"step": 75
},
{
"epoch": 2.0551724137931036,
"grad_norm": 1.364702820777893,
"learning_rate": 7.82608695652174e-05,
"loss": 0.4993,
"step": 76
},
{
"epoch": 2.0827586206896553,
"grad_norm": 1.2684059143066406,
"learning_rate": 7.652173913043479e-05,
"loss": 0.4305,
"step": 77
},
{
"epoch": 2.110344827586207,
"grad_norm": 1.1678532361984253,
"learning_rate": 7.478260869565218e-05,
"loss": 0.4332,
"step": 78
},
{
"epoch": 2.1379310344827585,
"grad_norm": 1.3142938613891602,
"learning_rate": 7.304347826086957e-05,
"loss": 0.4618,
"step": 79
},
{
"epoch": 2.1655172413793102,
"grad_norm": 1.359118938446045,
"learning_rate": 7.130434782608696e-05,
"loss": 0.5795,
"step": 80
},
{
"epoch": 2.193103448275862,
"grad_norm": 1.6325678825378418,
"learning_rate": 6.956521739130436e-05,
"loss": 0.4021,
"step": 81
},
{
"epoch": 2.220689655172414,
"grad_norm": 1.2178771495819092,
"learning_rate": 6.782608695652173e-05,
"loss": 0.4262,
"step": 82
},
{
"epoch": 2.2482758620689656,
"grad_norm": 1.0997027158737183,
"learning_rate": 6.608695652173912e-05,
"loss": 0.3029,
"step": 83
},
{
"epoch": 2.2758620689655173,
"grad_norm": 1.1487294435501099,
"learning_rate": 6.434782608695652e-05,
"loss": 0.3831,
"step": 84
},
{
"epoch": 2.303448275862069,
"grad_norm": 1.3247836828231812,
"learning_rate": 6.260869565217392e-05,
"loss": 0.4692,
"step": 85
},
{
"epoch": 2.3310344827586205,
"grad_norm": 1.1617599725723267,
"learning_rate": 6.086956521739131e-05,
"loss": 0.3617,
"step": 86
},
{
"epoch": 2.3586206896551722,
"grad_norm": 1.2517313957214355,
"learning_rate": 5.9130434782608704e-05,
"loss": 0.2729,
"step": 87
},
{
"epoch": 2.386206896551724,
"grad_norm": 1.1272892951965332,
"learning_rate": 5.739130434782609e-05,
"loss": 0.3707,
"step": 88
},
{
"epoch": 2.413793103448276,
"grad_norm": 1.196664571762085,
"learning_rate": 5.565217391304348e-05,
"loss": 0.4032,
"step": 89
},
{
"epoch": 2.4413793103448276,
"grad_norm": 1.4257408380508423,
"learning_rate": 5.391304347826087e-05,
"loss": 0.3194,
"step": 90
},
{
"epoch": 2.4689655172413794,
"grad_norm": 1.798063039779663,
"learning_rate": 5.217391304347826e-05,
"loss": 0.5721,
"step": 91
},
{
"epoch": 2.496551724137931,
"grad_norm": 1.449183702468872,
"learning_rate": 5.0434782608695655e-05,
"loss": 0.3841,
"step": 92
},
{
"epoch": 2.524137931034483,
"grad_norm": 1.4660217761993408,
"learning_rate": 4.8695652173913046e-05,
"loss": 0.4134,
"step": 93
},
{
"epoch": 2.5517241379310347,
"grad_norm": 1.3259236812591553,
"learning_rate": 4.695652173913044e-05,
"loss": 0.5053,
"step": 94
},
{
"epoch": 2.5793103448275865,
"grad_norm": 1.1987637281417847,
"learning_rate": 4.521739130434783e-05,
"loss": 0.3203,
"step": 95
},
{
"epoch": 2.606896551724138,
"grad_norm": 1.702609896659851,
"learning_rate": 4.347826086956522e-05,
"loss": 0.5448,
"step": 96
},
{
"epoch": 2.6344827586206896,
"grad_norm": 1.2012200355529785,
"learning_rate": 4.1739130434782605e-05,
"loss": 0.4839,
"step": 97
},
{
"epoch": 2.6620689655172414,
"grad_norm": 1.1377613544464111,
"learning_rate": 4e-05,
"loss": 0.3131,
"step": 98
},
{
"epoch": 2.689655172413793,
"grad_norm": 1.377774953842163,
"learning_rate": 3.8260869565217395e-05,
"loss": 0.3463,
"step": 99
},
{
"epoch": 2.717241379310345,
"grad_norm": 1.1738471984863281,
"learning_rate": 3.6521739130434786e-05,
"loss": 0.2963,
"step": 100
},
{
"epoch": 2.7448275862068967,
"grad_norm": 1.1475613117218018,
"learning_rate": 3.478260869565218e-05,
"loss": 0.2953,
"step": 101
},
{
"epoch": 2.772413793103448,
"grad_norm": 1.5838022232055664,
"learning_rate": 3.304347826086956e-05,
"loss": 0.3852,
"step": 102
},
{
"epoch": 2.8,
"grad_norm": 1.6446831226348877,
"learning_rate": 3.130434782608696e-05,
"loss": 0.5384,
"step": 103
},
{
"epoch": 2.8275862068965516,
"grad_norm": 1.4402813911437988,
"learning_rate": 2.9565217391304352e-05,
"loss": 0.3184,
"step": 104
},
{
"epoch": 2.8551724137931034,
"grad_norm": 1.3366456031799316,
"learning_rate": 2.782608695652174e-05,
"loss": 0.4545,
"step": 105
},
{
"epoch": 2.882758620689655,
"grad_norm": 1.4988086223602295,
"learning_rate": 2.608695652173913e-05,
"loss": 0.2341,
"step": 106
},
{
"epoch": 2.910344827586207,
"grad_norm": 1.35313880443573,
"learning_rate": 2.4347826086956523e-05,
"loss": 0.3555,
"step": 107
},
{
"epoch": 2.9379310344827587,
"grad_norm": 1.1439647674560547,
"learning_rate": 2.2608695652173914e-05,
"loss": 0.299,
"step": 108
},
{
"epoch": 2.9655172413793105,
"grad_norm": 1.2948118448257446,
"learning_rate": 2.0869565217391303e-05,
"loss": 0.4421,
"step": 109
},
{
"epoch": 2.9931034482758623,
"grad_norm": 1.283116340637207,
"learning_rate": 1.9130434782608697e-05,
"loss": 0.368,
"step": 110
},
{
"epoch": 3.0,
"grad_norm": 3.516788959503174,
"learning_rate": 1.739130434782609e-05,
"loss": 0.3088,
"step": 111
},
{
"epoch": 3.027586206896552,
"grad_norm": 1.1082452535629272,
"learning_rate": 1.565217391304348e-05,
"loss": 0.2828,
"step": 112
},
{
"epoch": 3.0551724137931036,
"grad_norm": 1.109584927558899,
"learning_rate": 1.391304347826087e-05,
"loss": 0.2485,
"step": 113
},
{
"epoch": 3.0827586206896553,
"grad_norm": 1.0583851337432861,
"learning_rate": 1.2173913043478261e-05,
"loss": 0.2712,
"step": 114
},
{
"epoch": 3.110344827586207,
"grad_norm": 1.1029939651489258,
"learning_rate": 1.0434782608695651e-05,
"loss": 0.3537,
"step": 115
},
{
"epoch": 3.1379310344827585,
"grad_norm": 1.0736896991729736,
"learning_rate": 8.695652173913044e-06,
"loss": 0.2121,
"step": 116
},
{
"epoch": 3.1655172413793102,
"grad_norm": 1.1432276964187622,
"learning_rate": 6.956521739130435e-06,
"loss": 0.2371,
"step": 117
},
{
"epoch": 3.193103448275862,
"grad_norm": 1.4690179824829102,
"learning_rate": 5.217391304347826e-06,
"loss": 0.3357,
"step": 118
},
{
"epoch": 3.220689655172414,
"grad_norm": 1.240258812904358,
"learning_rate": 3.4782608695652175e-06,
"loss": 0.2727,
"step": 119
},
{
"epoch": 3.2482758620689656,
"grad_norm": 1.2821192741394043,
"learning_rate": 1.7391304347826088e-06,
"loss": 0.3032,
"step": 120
}
],
"logging_steps": 1,
"max_steps": 120,
"num_input_tokens_seen": 0,
"num_train_epochs": 4,
"save_steps": 500,
"stateful_callbacks": {
"TrainerControl": {
"args": {
"should_epoch_stop": false,
"should_evaluate": false,
"should_log": false,
"should_save": true,
"should_training_stop": true
},
"attributes": {}
}
},
"total_flos": 2.068699077260083e+16,
"train_batch_size": 2,
"trial_name": null,
"trial_params": null
}