zephyr-7b-sft-lora / trainer_state.json
wenzw's picture
Model save
9cd6a29
raw
history blame
7.48 kB
{
"best_metric": null,
"best_model_checkpoint": null,
"epoch": 0.6699636307656782,
"eval_steps": 500,
"global_step": 272,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.0,
"learning_rate": 1.999969914479768e-05,
"loss": 1.1295,
"step": 1
},
{
"epoch": 0.01,
"learning_rate": 1.9992479525042305e-05,
"loss": 1.1199,
"step": 5
},
{
"epoch": 0.02,
"learning_rate": 1.996992941167792e-05,
"loss": 1.1063,
"step": 10
},
{
"epoch": 0.04,
"learning_rate": 1.9932383577419432e-05,
"loss": 1.0914,
"step": 15
},
{
"epoch": 0.05,
"learning_rate": 1.9879898494768093e-05,
"loss": 1.094,
"step": 20
},
{
"epoch": 0.06,
"learning_rate": 1.9812553106273848e-05,
"loss": 1.0658,
"step": 25
},
{
"epoch": 0.07,
"learning_rate": 1.973044870579824e-05,
"loss": 1.0527,
"step": 30
},
{
"epoch": 0.09,
"learning_rate": 1.9633708786158803e-05,
"loss": 1.0503,
"step": 35
},
{
"epoch": 0.1,
"learning_rate": 1.9522478853384154e-05,
"loss": 1.0305,
"step": 40
},
{
"epoch": 0.11,
"learning_rate": 1.9396926207859085e-05,
"loss": 1.0268,
"step": 45
},
{
"epoch": 0.12,
"learning_rate": 1.9257239692688907e-05,
"loss": 1.0291,
"step": 50
},
{
"epoch": 0.14,
"learning_rate": 1.9103629409661468e-05,
"loss": 1.02,
"step": 55
},
{
"epoch": 0.15,
"learning_rate": 1.8936326403234125e-05,
"loss": 1.0127,
"step": 60
},
{
"epoch": 0.16,
"learning_rate": 1.8755582313020912e-05,
"loss": 1.0078,
"step": 65
},
{
"epoch": 0.17,
"learning_rate": 1.8561668995302668e-05,
"loss": 1.0003,
"step": 70
},
{
"epoch": 0.18,
"learning_rate": 1.8354878114129368e-05,
"loss": 1.0081,
"step": 75
},
{
"epoch": 0.2,
"learning_rate": 1.8135520702629677e-05,
"loss": 1.0157,
"step": 80
},
{
"epoch": 0.21,
"learning_rate": 1.7903926695187595e-05,
"loss": 1.0118,
"step": 85
},
{
"epoch": 0.22,
"learning_rate": 1.766044443118978e-05,
"loss": 1.0011,
"step": 90
},
{
"epoch": 0.23,
"learning_rate": 1.740544013109005e-05,
"loss": 0.9919,
"step": 95
},
{
"epoch": 0.25,
"learning_rate": 1.7139297345578992e-05,
"loss": 1.0025,
"step": 100
},
{
"epoch": 0.26,
"learning_rate": 1.686241637868734e-05,
"loss": 1.0187,
"step": 105
},
{
"epoch": 0.27,
"learning_rate": 1.657521368569064e-05,
"loss": 0.9959,
"step": 110
},
{
"epoch": 0.28,
"learning_rate": 1.627812124672099e-05,
"loss": 0.9936,
"step": 115
},
{
"epoch": 0.3,
"learning_rate": 1.5971585917027864e-05,
"loss": 0.9981,
"step": 120
},
{
"epoch": 0.31,
"learning_rate": 1.5656068754865388e-05,
"loss": 0.9954,
"step": 125
},
{
"epoch": 0.32,
"learning_rate": 1.5332044328016916e-05,
"loss": 1.0044,
"step": 130
},
{
"epoch": 0.33,
"learning_rate": 1.5000000000000002e-05,
"loss": 1.0032,
"step": 135
},
{
"epoch": 0.34,
"learning_rate": 1.4660435197025391e-05,
"loss": 0.9901,
"step": 140
},
{
"epoch": 0.36,
"learning_rate": 1.4313860656812537e-05,
"loss": 0.9965,
"step": 145
},
{
"epoch": 0.37,
"learning_rate": 1.396079766039157e-05,
"loss": 0.9915,
"step": 150
},
{
"epoch": 0.38,
"learning_rate": 1.3601777248047105e-05,
"loss": 0.9861,
"step": 155
},
{
"epoch": 0.39,
"learning_rate": 1.3237339420583213e-05,
"loss": 0.9757,
"step": 160
},
{
"epoch": 0.41,
"learning_rate": 1.2868032327110904e-05,
"loss": 0.9944,
"step": 165
},
{
"epoch": 0.42,
"learning_rate": 1.2494411440579814e-05,
"loss": 0.9963,
"step": 170
},
{
"epoch": 0.43,
"learning_rate": 1.211703872229411e-05,
"loss": 0.9817,
"step": 175
},
{
"epoch": 0.44,
"learning_rate": 1.1736481776669307e-05,
"loss": 0.9881,
"step": 180
},
{
"epoch": 0.46,
"learning_rate": 1.1353312997501313e-05,
"loss": 0.9965,
"step": 185
},
{
"epoch": 0.47,
"learning_rate": 1.0968108707031792e-05,
"loss": 0.9921,
"step": 190
},
{
"epoch": 0.48,
"learning_rate": 1.0581448289104759e-05,
"loss": 0.9934,
"step": 195
},
{
"epoch": 0.49,
"learning_rate": 1.0193913317718245e-05,
"loss": 0.9823,
"step": 200
},
{
"epoch": 0.5,
"learning_rate": 9.806086682281759e-06,
"loss": 0.988,
"step": 205
},
{
"epoch": 0.52,
"learning_rate": 9.418551710895243e-06,
"loss": 0.9866,
"step": 210
},
{
"epoch": 0.53,
"learning_rate": 9.03189129296821e-06,
"loss": 0.9983,
"step": 215
},
{
"epoch": 0.54,
"learning_rate": 8.646687002498692e-06,
"loss": 0.9866,
"step": 220
},
{
"epoch": 0.55,
"learning_rate": 8.263518223330698e-06,
"loss": 0.9918,
"step": 225
},
{
"epoch": 0.57,
"learning_rate": 7.882961277705897e-06,
"loss": 0.9901,
"step": 230
},
{
"epoch": 0.58,
"learning_rate": 7.505588559420188e-06,
"loss": 0.987,
"step": 235
},
{
"epoch": 0.59,
"learning_rate": 7.131967672889101e-06,
"loss": 0.9941,
"step": 240
},
{
"epoch": 0.6,
"learning_rate": 6.762660579416791e-06,
"loss": 0.9968,
"step": 245
},
{
"epoch": 0.62,
"learning_rate": 6.3982227519528986e-06,
"loss": 0.9842,
"step": 250
},
{
"epoch": 0.63,
"learning_rate": 6.039202339608432e-06,
"loss": 0.9788,
"step": 255
},
{
"epoch": 0.64,
"learning_rate": 5.686139343187468e-06,
"loss": 0.9929,
"step": 260
},
{
"epoch": 0.65,
"learning_rate": 5.339564802974615e-06,
"loss": 0.9834,
"step": 265
},
{
"epoch": 0.67,
"learning_rate": 5.000000000000003e-06,
"loss": 0.9866,
"step": 270
},
{
"epoch": 0.67,
"eval_loss": 0.9899678230285645,
"eval_runtime": 3893.4345,
"eval_samples_per_second": 5.936,
"eval_steps_per_second": 0.742,
"step": 272
},
{
"epoch": 0.67,
"step": 272,
"total_flos": 1.2275875027210994e+19,
"train_loss": 1.0095771827241953,
"train_runtime": 116282.0771,
"train_samples_per_second": 1.788,
"train_steps_per_second": 0.003
}
],
"logging_steps": 5,
"max_steps": 405,
"num_train_epochs": 1,
"save_steps": 500,
"total_flos": 1.2275875027210994e+19,
"trial_name": null,
"trial_params": null
}