whisper-vietnamese / trainer_state.json
vikas85's picture
Upload 8 files
d3f9d3f
raw
history blame
8.03 kB
{
"best_metric": 0.16838560303094086,
"best_model_checkpoint": "./whisper-all\\checkpoint-10500",
"epoch": 5.839822024471635,
"global_step": 10500,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.28,
"learning_rate": 9.721382697393274e-07,
"loss": 0.9517,
"step": 500
},
{
"epoch": 0.28,
"eval_loss": 0.23364929854869843,
"eval_runtime": 2055.8229,
"eval_samples_per_second": 1.388,
"eval_steps_per_second": 0.174,
"eval_wer": 12.261746065155874,
"step": 500
},
{
"epoch": 0.56,
"learning_rate": 9.249149981110691e-07,
"loss": 0.4857,
"step": 1000
},
{
"epoch": 0.56,
"eval_loss": 0.19377797842025757,
"eval_runtime": 2058.47,
"eval_samples_per_second": 1.386,
"eval_steps_per_second": 0.173,
"eval_wer": 10.500713300123952,
"step": 1000
},
{
"epoch": 0.83,
"learning_rate": 8.777861730260671e-07,
"loss": 0.4437,
"step": 1500
},
{
"epoch": 0.83,
"eval_loss": 0.15837284922599792,
"eval_runtime": 2059.2478,
"eval_samples_per_second": 1.386,
"eval_steps_per_second": 0.173,
"eval_wer": 8.982904183914497,
"step": 1500
},
{
"epoch": 1.11,
"learning_rate": 8.305629013978088e-07,
"loss": 0.3544,
"step": 2000
},
{
"epoch": 1.11,
"eval_loss": 0.1296420842409134,
"eval_runtime": 2050.2346,
"eval_samples_per_second": 1.392,
"eval_steps_per_second": 0.174,
"eval_wer": 7.067517949437545,
"step": 2000
},
{
"epoch": 1.39,
"learning_rate": 7.833396297695503e-07,
"loss": 0.2743,
"step": 2500
},
{
"epoch": 1.39,
"eval_loss": 0.10552296042442322,
"eval_runtime": 2046.8577,
"eval_samples_per_second": 1.394,
"eval_steps_per_second": 0.174,
"eval_wer": 6.368249959072944,
"step": 2500
},
{
"epoch": 1.67,
"learning_rate": 7.362108046845486e-07,
"loss": 0.2819,
"step": 3000
},
{
"epoch": 1.67,
"eval_loss": 0.08085598796606064,
"eval_runtime": 2023.4637,
"eval_samples_per_second": 1.41,
"eval_steps_per_second": 0.176,
"eval_wer": 4.448186346734021,
"step": 3000
},
{
"epoch": 1.95,
"learning_rate": 6.889875330562901e-07,
"loss": 0.2629,
"step": 3500
},
{
"epoch": 1.95,
"eval_loss": 0.06196921318769455,
"eval_runtime": 2016.6651,
"eval_samples_per_second": 1.415,
"eval_steps_per_second": 0.177,
"eval_wer": 3.309244837344185,
"step": 3500
},
{
"epoch": 2.22,
"learning_rate": 6.417642614280318e-07,
"loss": 0.166,
"step": 4000
},
{
"epoch": 2.22,
"eval_loss": 0.049665648490190506,
"eval_runtime": 2026.4551,
"eval_samples_per_second": 1.408,
"eval_steps_per_second": 0.176,
"eval_wer": 2.70586309314998,
"step": 4000
},
{
"epoch": 2.5,
"learning_rate": 5.945409897997733e-07,
"loss": 0.1443,
"step": 4500
},
{
"epoch": 2.5,
"eval_loss": 0.043939754366874695,
"eval_runtime": 3073.8688,
"eval_samples_per_second": 0.928,
"eval_steps_per_second": 0.116,
"eval_wer": 2.4345751771556867,
"step": 4500
},
{
"epoch": 2.78,
"learning_rate": 5.47317718171515e-07,
"loss": 0.1403,
"step": 5000
},
{
"epoch": 2.78,
"eval_loss": 0.03313903510570526,
"eval_runtime": 2041.5077,
"eval_samples_per_second": 1.398,
"eval_steps_per_second": 0.175,
"eval_wer": 1.725952431067144,
"step": 5000
},
{
"epoch": 3.06,
"learning_rate": 5.000944465432565e-07,
"loss": 0.1229,
"step": 5500
},
{
"epoch": 3.06,
"eval_loss": 0.025085289031267166,
"eval_runtime": 2036.5578,
"eval_samples_per_second": 1.401,
"eval_steps_per_second": 0.175,
"eval_wer": 1.2979723566968357,
"step": 5500
},
{
"epoch": 3.34,
"learning_rate": 4.528711749149981e-07,
"loss": 0.0666,
"step": 6000
},
{
"epoch": 3.34,
"eval_loss": 0.02561613917350769,
"eval_runtime": 2036.521,
"eval_samples_per_second": 1.401,
"eval_steps_per_second": 0.175,
"eval_wer": 1.4336163146939827,
"step": 6000
},
{
"epoch": 3.62,
"learning_rate": 4.056479032867397e-07,
"loss": 0.0681,
"step": 6500
},
{
"epoch": 3.62,
"eval_loss": 0.018902545794844627,
"eval_runtime": 2042.0741,
"eval_samples_per_second": 1.398,
"eval_steps_per_second": 0.175,
"eval_wer": 1.0687808414602773,
"step": 6500
},
{
"epoch": 3.89,
"learning_rate": 3.584246316584813e-07,
"loss": 0.063,
"step": 7000
},
{
"epoch": 3.89,
"eval_loss": 0.011477218009531498,
"eval_runtime": 2040.8765,
"eval_samples_per_second": 1.398,
"eval_steps_per_second": 0.175,
"eval_wer": 0.8068476811899249,
"step": 7000
},
{
"epoch": 4.17,
"learning_rate": 3.112013600302229e-07,
"loss": 0.0427,
"step": 7500
},
{
"epoch": 4.17,
"eval_loss": 0.010536368936300278,
"eval_runtime": 2038.2156,
"eval_samples_per_second": 1.4,
"eval_steps_per_second": 0.175,
"eval_wer": 0.47943123085198436,
"step": 7500
},
{
"epoch": 4.45,
"learning_rate": 2.6397808840196447e-07,
"loss": 0.0277,
"step": 8000
},
{
"epoch": 4.45,
"eval_loss": 0.009207101538777351,
"eval_runtime": 2049.3201,
"eval_samples_per_second": 1.393,
"eval_steps_per_second": 0.174,
"eval_wer": 0.4817699197829697,
"step": 8000
},
{
"epoch": 4.73,
"learning_rate": 2.1675481677370608e-07,
"loss": 0.0289,
"step": 8500
},
{
"epoch": 4.73,
"eval_loss": 0.007379856426268816,
"eval_runtime": 2045.1755,
"eval_samples_per_second": 1.395,
"eval_steps_per_second": 0.175,
"eval_wer": 0.35314202857877874,
"step": 8500
},
{
"epoch": 5.01,
"learning_rate": 1.6953154514544768e-07,
"loss": 0.0253,
"step": 9000
},
{
"epoch": 5.01,
"eval_loss": 0.005186205264180899,
"eval_runtime": 2053.6472,
"eval_samples_per_second": 1.39,
"eval_steps_per_second": 0.174,
"eval_wer": 0.21282069271966134,
"step": 9000
},
{
"epoch": 5.28,
"learning_rate": 1.2230827351718926e-07,
"loss": 0.0133,
"step": 9500
},
{
"epoch": 5.28,
"eval_loss": 0.005096600856631994,
"eval_runtime": 2024.6952,
"eval_samples_per_second": 1.41,
"eval_steps_per_second": 0.176,
"eval_wer": 0.2268528263055731,
"step": 9500
},
{
"epoch": 5.56,
"learning_rate": 7.508500188893087e-08,
"loss": 0.0118,
"step": 10000
},
{
"epoch": 5.56,
"eval_loss": 0.004013683646917343,
"eval_runtime": 2018.9797,
"eval_samples_per_second": 1.414,
"eval_steps_per_second": 0.177,
"eval_wer": 0.1800790476858673,
"step": 10000
},
{
"epoch": 5.84,
"learning_rate": 2.786173026067246e-08,
"loss": 0.0095,
"step": 10500
},
{
"epoch": 5.84,
"eval_loss": 0.0037340286653488874,
"eval_runtime": 2042.3697,
"eval_samples_per_second": 1.397,
"eval_steps_per_second": 0.175,
"eval_wer": 0.16838560303094086,
"step": 10500
}
],
"max_steps": 10788,
"num_train_epochs": 6,
"total_flos": 3.566927314944e+20,
"trial_name": null,
"trial_params": null
}