| { | |
| "best_global_step": 6136, | |
| "best_metric": 0.8316934704780579, | |
| "best_model_checkpoint": "distilbert_km_100_v2_mnli/checkpoint-6136", | |
| "epoch": 9.0, | |
| "eval_steps": 500, | |
| "global_step": 13806, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 1.0, | |
| "grad_norm": 1.1543691158294678, | |
| "learning_rate": 4.9e-05, | |
| "loss": 0.9951, | |
| "step": 1534 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_accuracy": 0.5552725420275089, | |
| "eval_loss": 0.9284037947654724, | |
| "eval_runtime": 4.0887, | |
| "eval_samples_per_second": 2400.513, | |
| "eval_steps_per_second": 9.538, | |
| "step": 1534 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "grad_norm": 1.5014262199401855, | |
| "learning_rate": 4.8e-05, | |
| "loss": 0.8931, | |
| "step": 3068 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_accuracy": 0.5958227203260316, | |
| "eval_loss": 0.8693681955337524, | |
| "eval_runtime": 3.9911, | |
| "eval_samples_per_second": 2459.247, | |
| "eval_steps_per_second": 9.772, | |
| "step": 3068 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "grad_norm": 1.612996220588684, | |
| "learning_rate": 4.7e-05, | |
| "loss": 0.8222, | |
| "step": 4602 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_accuracy": 0.6183392766174223, | |
| "eval_loss": 0.8401850461959839, | |
| "eval_runtime": 3.9915, | |
| "eval_samples_per_second": 2458.984, | |
| "eval_steps_per_second": 9.771, | |
| "step": 4602 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "grad_norm": 2.094791889190674, | |
| "learning_rate": 4.600000000000001e-05, | |
| "loss": 0.7558, | |
| "step": 6136 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_accuracy": 0.6360672440142638, | |
| "eval_loss": 0.8316934704780579, | |
| "eval_runtime": 3.9821, | |
| "eval_samples_per_second": 2464.772, | |
| "eval_steps_per_second": 9.794, | |
| "step": 6136 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "grad_norm": 2.091482400894165, | |
| "learning_rate": 4.5e-05, | |
| "loss": 0.6891, | |
| "step": 7670 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_accuracy": 0.64381049414162, | |
| "eval_loss": 0.8356432914733887, | |
| "eval_runtime": 3.977, | |
| "eval_samples_per_second": 2467.941, | |
| "eval_steps_per_second": 9.806, | |
| "step": 7670 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "grad_norm": 2.5932934284210205, | |
| "learning_rate": 4.4000000000000006e-05, | |
| "loss": 0.621, | |
| "step": 9204 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_accuracy": 0.6444218033622007, | |
| "eval_loss": 0.8813253045082092, | |
| "eval_runtime": 3.9808, | |
| "eval_samples_per_second": 2465.558, | |
| "eval_steps_per_second": 9.797, | |
| "step": 9204 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "grad_norm": 2.725822687149048, | |
| "learning_rate": 4.3e-05, | |
| "loss": 0.552, | |
| "step": 10738 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "eval_accuracy": 0.6444218033622007, | |
| "eval_loss": 0.9221841096878052, | |
| "eval_runtime": 3.9717, | |
| "eval_samples_per_second": 2471.227, | |
| "eval_steps_per_second": 9.819, | |
| "step": 10738 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "grad_norm": 3.178081512451172, | |
| "learning_rate": 4.2e-05, | |
| "loss": 0.4855, | |
| "step": 12272 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_accuracy": 0.6479877738155884, | |
| "eval_loss": 0.9823774099349976, | |
| "eval_runtime": 4.057, | |
| "eval_samples_per_second": 2419.277, | |
| "eval_steps_per_second": 9.613, | |
| "step": 12272 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "grad_norm": 3.798142671585083, | |
| "learning_rate": 4.1e-05, | |
| "loss": 0.4229, | |
| "step": 13806 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "eval_accuracy": 0.6417727967396841, | |
| "eval_loss": 1.114707589149475, | |
| "eval_runtime": 4.0382, | |
| "eval_samples_per_second": 2430.561, | |
| "eval_steps_per_second": 9.658, | |
| "step": 13806 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "step": 13806, | |
| "total_flos": 2.340951304214108e+17, | |
| "train_loss": 0.6929499982457809, | |
| "train_runtime": 3245.2005, | |
| "train_samples_per_second": 6050.504, | |
| "train_steps_per_second": 23.635 | |
| } | |
| ], | |
| "logging_steps": 1, | |
| "max_steps": 76700, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 50, | |
| "save_steps": 500, | |
| "stateful_callbacks": { | |
| "EarlyStoppingCallback": { | |
| "args": { | |
| "early_stopping_patience": 5, | |
| "early_stopping_threshold": 0.0 | |
| }, | |
| "attributes": { | |
| "early_stopping_patience_counter": 5 | |
| } | |
| }, | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": true | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 2.340951304214108e+17, | |
| "train_batch_size": 256, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |