| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 8.02139037433155, | |
| "eval_steps": 500, | |
| "global_step": 1500, | |
| "is_hyper_param_search": true, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 1.0, | |
| "eval_accuracy": 0.7532685399055481, | |
| "eval_loss": 0.4793490469455719, | |
| "eval_runtime": 9.3602, | |
| "eval_samples_per_second": 318.688, | |
| "eval_steps_per_second": 19.978, | |
| "step": 187 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_accuracy": 0.7790814638137817, | |
| "eval_loss": 0.4668562114238739, | |
| "eval_runtime": 9.4916, | |
| "eval_samples_per_second": 314.277, | |
| "eval_steps_per_second": 19.702, | |
| "step": 374 | |
| }, | |
| { | |
| "epoch": 2.6737967914438503, | |
| "grad_norm": 9.970394134521484, | |
| "learning_rate": 1.0116255710136231e-05, | |
| "loss": 0.4413, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_accuracy": 0.7790814638137817, | |
| "eval_loss": 0.5445067882537842, | |
| "eval_runtime": 9.4939, | |
| "eval_samples_per_second": 314.202, | |
| "eval_steps_per_second": 19.697, | |
| "step": 561 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_accuracy": 0.7700302004814148, | |
| "eval_loss": 0.6839850544929504, | |
| "eval_runtime": 9.4969, | |
| "eval_samples_per_second": 314.101, | |
| "eval_steps_per_second": 19.691, | |
| "step": 748 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_accuracy": 0.7680187821388245, | |
| "eval_loss": 0.7934393286705017, | |
| "eval_runtime": 9.5009, | |
| "eval_samples_per_second": 313.97, | |
| "eval_steps_per_second": 19.682, | |
| "step": 935 | |
| }, | |
| { | |
| "epoch": 5.347593582887701, | |
| "grad_norm": 6.287047863006592, | |
| "learning_rate": 6.424191582349286e-06, | |
| "loss": 0.1515, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_accuracy": 0.7730472683906555, | |
| "eval_loss": 0.7747542262077332, | |
| "eval_runtime": 9.4572, | |
| "eval_samples_per_second": 315.421, | |
| "eval_steps_per_second": 19.773, | |
| "step": 1122 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "eval_accuracy": 0.7700302004814148, | |
| "eval_loss": 0.8394743800163269, | |
| "eval_runtime": 9.6279, | |
| "eval_samples_per_second": 309.83, | |
| "eval_steps_per_second": 19.423, | |
| "step": 1309 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_accuracy": 0.7717063426971436, | |
| "eval_loss": 0.8363062143325806, | |
| "eval_runtime": 9.6181, | |
| "eval_samples_per_second": 310.145, | |
| "eval_steps_per_second": 19.443, | |
| "step": 1496 | |
| }, | |
| { | |
| "epoch": 8.02139037433155, | |
| "grad_norm": 7.763587474822998, | |
| "learning_rate": 2.73212745456234e-06, | |
| "loss": 0.0648, | |
| "step": 1500 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 1870, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 10, | |
| "save_steps": 500, | |
| "total_flos": 1.7577443886948468e+16, | |
| "train_batch_size": 64, | |
| "trial_name": null, | |
| "trial_params": { | |
| "learning_rate": 1.3808319837923177e-05, | |
| "per_device_train_batch_size": 64 | |
| } | |
| } | |