whisper-large-v3-ft-cy / trainer_state.json
DewiBrynJones's picture
Training in progress, step 1000
123ecf5 verified
raw
history blame
3.89 kB
{
"best_metric": 18.70514087587693,
"best_model_checkpoint": "/models/hfhub/whisper-large-v3-ft-cy/checkpoint-3000",
"epoch": 1.005593614480548,
"eval_steps": 1000,
"global_step": 4000,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.06,
"learning_rate": 5e-06,
"loss": 0.4589,
"step": 250
},
{
"epoch": 0.13,
"learning_rate": 1e-05,
"loss": 0.2548,
"step": 500
},
{
"epoch": 0.19,
"learning_rate": 9.285714285714288e-06,
"loss": 0.2358,
"step": 750
},
{
"epoch": 0.25,
"learning_rate": 8.571428571428571e-06,
"loss": 0.2078,
"step": 1000
},
{
"epoch": 0.25,
"eval_loss": 0.21982243657112122,
"eval_runtime": 30133.785,
"eval_samples_per_second": 0.87,
"eval_steps_per_second": 0.109,
"eval_wer": 28.755609922068665,
"step": 1000
},
{
"epoch": 0.31,
"learning_rate": 7.857142857142858e-06,
"loss": 0.1921,
"step": 1250
},
{
"epoch": 0.38,
"learning_rate": 7.1428571428571436e-06,
"loss": 0.1803,
"step": 1500
},
{
"epoch": 0.44,
"learning_rate": 6.4285714285714295e-06,
"loss": 0.1714,
"step": 1750
},
{
"epoch": 0.5,
"learning_rate": 5.7142857142857145e-06,
"loss": 0.1623,
"step": 2000
},
{
"epoch": 0.5,
"eval_loss": 0.1799924671649933,
"eval_runtime": 31181.9539,
"eval_samples_per_second": 0.841,
"eval_steps_per_second": 0.105,
"eval_wer": 31.369793101213528,
"step": 2000
},
{
"epoch": 0.57,
"learning_rate": 5e-06,
"loss": 0.1592,
"step": 2250
},
{
"epoch": 0.63,
"learning_rate": 4.2857142857142855e-06,
"loss": 0.1451,
"step": 2500
},
{
"epoch": 0.69,
"learning_rate": 3.5714285714285718e-06,
"loss": 0.1447,
"step": 2750
},
{
"epoch": 0.75,
"learning_rate": 2.8571428571428573e-06,
"loss": 0.1417,
"step": 3000
},
{
"epoch": 0.75,
"eval_loss": 0.15845398604869843,
"eval_runtime": 32690.2238,
"eval_samples_per_second": 0.802,
"eval_steps_per_second": 0.1,
"eval_wer": 18.70514087587693,
"step": 3000
},
{
"epoch": 0.82,
"learning_rate": 2.1428571428571427e-06,
"loss": 0.1342,
"step": 3250
},
{
"epoch": 0.88,
"learning_rate": 1.4285714285714286e-06,
"loss": 0.1315,
"step": 3500
},
{
"epoch": 0.94,
"learning_rate": 7.142857142857143e-07,
"loss": 0.1285,
"step": 3750
},
{
"epoch": 1.01,
"learning_rate": 0.0,
"loss": 0.1188,
"step": 4000
},
{
"epoch": 1.01,
"eval_loss": 0.14796172082424164,
"eval_runtime": 32177.5808,
"eval_samples_per_second": 0.815,
"eval_steps_per_second": 0.102,
"eval_wer": 25.134071061713193,
"step": 4000
},
{
"epoch": 1.01,
"step": 4000,
"total_flos": 4.3487638574137344e+20,
"train_loss": 0.18544992589950562,
"train_runtime": 248926.3333,
"train_samples_per_second": 0.514,
"train_steps_per_second": 0.016
},
{
"epoch": 1.01,
"eval_loss": 0.14796172082424164,
"eval_runtime": 30969.981,
"eval_samples_per_second": 0.847,
"eval_steps_per_second": 0.106,
"eval_wer": 25.134071061713193,
"step": 4000
}
],
"logging_steps": 250,
"max_steps": 4000,
"num_input_tokens_seen": 0,
"num_train_epochs": 2,
"save_steps": 1000,
"total_flos": 4.3487638574137344e+20,
"train_batch_size": 4,
"trial_name": null,
"trial_params": null
}