| { | |
| "best_metric": 7.865168539325842, | |
| "best_model_checkpoint": "./models/simple_tamil/checkpoint-1000", | |
| "epoch": 272.0016, | |
| "global_step": 3000, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 22.0, | |
| "learning_rate": 4.9000000000000005e-06, | |
| "loss": 1.6002, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 45.0, | |
| "learning_rate": 9.9e-06, | |
| "loss": 0.1883, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 68.0, | |
| "learning_rate": 9.455555555555557e-06, | |
| "loss": 0.1957, | |
| "step": 750 | |
| }, | |
| { | |
| "epoch": 90.0, | |
| "learning_rate": 8.900000000000001e-06, | |
| "loss": 0.1363, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 90.0, | |
| "eval_loss": 0.019793609157204628, | |
| "eval_runtime": 21.4626, | |
| "eval_samples_per_second": 1.351, | |
| "eval_steps_per_second": 0.093, | |
| "eval_wer": 7.865168539325842, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 113.0, | |
| "learning_rate": 8.344444444444445e-06, | |
| "loss": 0.1124, | |
| "step": 1250 | |
| }, | |
| { | |
| "epoch": 136.0, | |
| "learning_rate": 7.788888888888889e-06, | |
| "loss": 0.1005, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 159.0, | |
| "learning_rate": 7.235555555555556e-06, | |
| "loss": 0.1857, | |
| "step": 1750 | |
| }, | |
| { | |
| "epoch": 181.0, | |
| "learning_rate": 6.680000000000001e-06, | |
| "loss": 0.1145, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 181.0, | |
| "eval_loss": 0.027455436065793037, | |
| "eval_runtime": 27.2923, | |
| "eval_samples_per_second": 1.063, | |
| "eval_steps_per_second": 0.073, | |
| "eval_wer": 7.865168539325842, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 204.0, | |
| "learning_rate": 6.124444444444445e-06, | |
| "loss": 0.0713, | |
| "step": 2250 | |
| }, | |
| { | |
| "epoch": 227.0, | |
| "learning_rate": 5.56888888888889e-06, | |
| "loss": 0.0588, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 249.0, | |
| "learning_rate": 5.013333333333333e-06, | |
| "loss": 0.2421, | |
| "step": 2750 | |
| }, | |
| { | |
| "epoch": 272.0, | |
| "learning_rate": 4.457777777777778e-06, | |
| "loss": 0.0472, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 272.0, | |
| "eval_loss": 0.05207383260130882, | |
| "eval_runtime": 20.2088, | |
| "eval_samples_per_second": 1.435, | |
| "eval_steps_per_second": 0.099, | |
| "eval_wer": 14.606741573033707, | |
| "step": 3000 | |
| } | |
| ], | |
| "max_steps": 5000, | |
| "num_train_epochs": 9223372036854775807, | |
| "total_flos": 3.46302480384e+18, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |