breeze-dsw-small-id / trainer_state.json
hanasim's picture
End of training
9b8c818 verified
{
"best_metric": 17.70632072867789,
"best_model_checkpoint": "/cosmos/home/sp-operator/ai/training/models/huggingface/scripts/../breeze-dsw-small-id/checkpoint-200",
"epoch": 6.064,
"eval_steps": 100,
"global_step": 1000,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.03,
"learning_rate": 4.898977360288234e-06,
"loss": 1.0704,
"step": 25
},
{
"epoch": 0.05,
"learning_rate": 6.160712527409633e-06,
"loss": 0.5908,
"step": 50
},
{
"epoch": 0.07,
"learning_rate": 6.85912902234906e-06,
"loss": 0.4936,
"step": 75
},
{
"epoch": 0.1,
"learning_rate": 7.344547104469332e-06,
"loss": 0.4199,
"step": 100
},
{
"epoch": 0.1,
"eval_loss": 0.349365234375,
"eval_runtime": 6608.5627,
"eval_samples_per_second": 0.551,
"eval_steps_per_second": 0.035,
"eval_wer": 19.15999631980863,
"step": 100
},
{
"epoch": 0.12,
"learning_rate": 7.716963756434345e-06,
"loss": 0.3139,
"step": 125
},
{
"epoch": 0.15,
"learning_rate": 8.019180844200955e-06,
"loss": 0.2397,
"step": 150
},
{
"epoch": 1.02,
"learning_rate": 8.27351214279797e-06,
"loss": 0.2943,
"step": 175
},
{
"epoch": 1.04,
"learning_rate": 8.49307723936858e-06,
"loss": 0.282,
"step": 200
},
{
"epoch": 1.04,
"eval_loss": 0.31591796875,
"eval_runtime": 6606.4412,
"eval_samples_per_second": 0.551,
"eval_steps_per_second": 0.035,
"eval_wer": 17.70632072867789,
"step": 200
},
{
"epoch": 1.07,
"learning_rate": 8.686247975778677e-06,
"loss": 0.3122,
"step": 225
},
{
"epoch": 1.09,
"learning_rate": 8.858694625217149e-06,
"loss": 0.2207,
"step": 250
},
{
"epoch": 1.12,
"learning_rate": 9.014436199608479e-06,
"loss": 0.1808,
"step": 275
},
{
"epoch": 1.14,
"learning_rate": 9.156425255148058e-06,
"loss": 0.1241,
"step": 300
},
{
"epoch": 1.14,
"eval_loss": 0.3291015625,
"eval_runtime": 6608.4009,
"eval_samples_per_second": 0.551,
"eval_steps_per_second": 0.035,
"eval_wer": 18.598767135891066,
"step": 300
},
{
"epoch": 2.01,
"learning_rate": 9.28689473531776e-06,
"loss": 0.1342,
"step": 325
},
{
"epoch": 2.04,
"learning_rate": 9.407574351377137e-06,
"loss": 0.1435,
"step": 350
},
{
"epoch": 2.06,
"learning_rate": 9.519831289296397e-06,
"loss": 0.1539,
"step": 375
},
{
"epoch": 2.09,
"learning_rate": 9.624764935335318e-06,
"loss": 0.1232,
"step": 400
},
{
"epoch": 2.09,
"eval_loss": 0.327880859375,
"eval_runtime": 6854.1492,
"eval_samples_per_second": 0.531,
"eval_steps_per_second": 0.033,
"eval_wer": 18.888582206274727,
"step": 400
},
{
"epoch": 2.11,
"learning_rate": 9.723272550712454e-06,
"loss": 0.1001,
"step": 425
},
{
"epoch": 2.14,
"learning_rate": 9.816095971633122e-06,
"loss": 0.066,
"step": 450
},
{
"epoch": 3.01,
"learning_rate": 9.90385555539545e-06,
"loss": 0.0587,
"step": 475
},
{
"epoch": 3.03,
"learning_rate": 9.987075336738768e-06,
"loss": 0.064,
"step": 500
},
{
"epoch": 3.03,
"eval_loss": 0.3388671875,
"eval_runtime": 6629.0714,
"eval_samples_per_second": 0.549,
"eval_steps_per_second": 0.034,
"eval_wer": 17.91333149323765,
"step": 500
},
{
"epoch": 3.06,
"learning_rate": 9.600000000000001e-06,
"loss": 0.0723,
"step": 525
},
{
"epoch": 3.08,
"learning_rate": 9.100000000000001e-06,
"loss": 0.0575,
"step": 550
},
{
"epoch": 3.11,
"learning_rate": 8.6e-06,
"loss": 0.0444,
"step": 575
},
{
"epoch": 3.13,
"learning_rate": 8.1e-06,
"loss": 0.0305,
"step": 600
},
{
"epoch": 3.13,
"eval_loss": 0.355712890625,
"eval_runtime": 6595.7112,
"eval_samples_per_second": 0.552,
"eval_steps_per_second": 0.035,
"eval_wer": 18.47916091636765,
"step": 600
},
{
"epoch": 4.0,
"learning_rate": 7.600000000000001e-06,
"loss": 0.0245,
"step": 625
},
{
"epoch": 4.03,
"learning_rate": 7.100000000000001e-06,
"loss": 0.0281,
"step": 650
},
{
"epoch": 4.05,
"learning_rate": 6.600000000000001e-06,
"loss": 0.0298,
"step": 675
},
{
"epoch": 4.08,
"learning_rate": 6.1e-06,
"loss": 0.0282,
"step": 700
},
{
"epoch": 4.08,
"eval_loss": 0.362548828125,
"eval_runtime": 6513.7808,
"eval_samples_per_second": 0.559,
"eval_steps_per_second": 0.035,
"eval_wer": 18.055938908823258,
"step": 700
},
{
"epoch": 4.1,
"learning_rate": 5.600000000000001e-06,
"loss": 0.0215,
"step": 725
},
{
"epoch": 4.13,
"learning_rate": 5.1e-06,
"loss": 0.0138,
"step": 750
},
{
"epoch": 4.15,
"learning_rate": 4.600000000000001e-06,
"loss": 0.011,
"step": 775
},
{
"epoch": 5.02,
"learning_rate": 4.1e-06,
"loss": 0.0117,
"step": 800
},
{
"epoch": 5.02,
"eval_loss": 0.369873046875,
"eval_runtime": 6724.4574,
"eval_samples_per_second": 0.542,
"eval_steps_per_second": 0.034,
"eval_wer": 18.29055110865765,
"step": 800
},
{
"epoch": 5.04,
"learning_rate": 3.6000000000000003e-06,
"loss": 0.013,
"step": 825
},
{
"epoch": 5.07,
"learning_rate": 3.1000000000000004e-06,
"loss": 0.0142,
"step": 850
},
{
"epoch": 5.09,
"learning_rate": 2.6e-06,
"loss": 0.0094,
"step": 875
},
{
"epoch": 5.12,
"learning_rate": 2.1000000000000002e-06,
"loss": 0.0079,
"step": 900
},
{
"epoch": 5.12,
"eval_loss": 0.37939453125,
"eval_runtime": 6640.5093,
"eval_samples_per_second": 0.548,
"eval_steps_per_second": 0.034,
"eval_wer": 18.359554696844235,
"step": 900
},
{
"epoch": 5.14,
"learning_rate": 1.6000000000000001e-06,
"loss": 0.0064,
"step": 925
},
{
"epoch": 6.01,
"learning_rate": 1.1e-06,
"loss": 0.0061,
"step": 950
},
{
"epoch": 6.04,
"learning_rate": 6.000000000000001e-07,
"loss": 0.0073,
"step": 975
},
{
"epoch": 6.06,
"learning_rate": 1.0000000000000001e-07,
"loss": 0.0081,
"step": 1000
},
{
"epoch": 6.06,
"eval_loss": 0.382568359375,
"eval_runtime": 6601.76,
"eval_samples_per_second": 0.552,
"eval_steps_per_second": 0.035,
"eval_wer": 18.29055110865765,
"step": 1000
},
{
"epoch": 6.06,
"step": 1000,
"total_flos": 9.194908120353079e+18,
"train_loss": 0.14578299713134765,
"train_runtime": 145825.9025,
"train_samples_per_second": 0.219,
"train_steps_per_second": 0.007
}
],
"logging_steps": 25,
"max_steps": 1000,
"num_input_tokens_seen": 0,
"num_train_epochs": 9223372036854775807,
"save_steps": 100,
"total_flos": 9.194908120353079e+18,
"train_batch_size": 32,
"trial_name": null,
"trial_params": null
}