| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 6.289308176100629, | |
| "eval_steps": 500, | |
| "global_step": 2000, | |
| "is_hyper_param_search": true, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 1.0, | |
| "eval_accuracy": 0.5906451612903226, | |
| "eval_loss": 0.20296397805213928, | |
| "eval_runtime": 14.4832, | |
| "eval_samples_per_second": 214.041, | |
| "eval_steps_per_second": 26.79, | |
| "step": 318 | |
| }, | |
| { | |
| "epoch": 1.5723270440251573, | |
| "grad_norm": 0.5446047782897949, | |
| "learning_rate": 1.606918238993711e-05, | |
| "loss": 0.3239, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_accuracy": 0.8229032258064516, | |
| "eval_loss": 0.09900839626789093, | |
| "eval_runtime": 15.5671, | |
| "eval_samples_per_second": 199.138, | |
| "eval_steps_per_second": 24.924, | |
| "step": 636 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_accuracy": 0.8835483870967742, | |
| "eval_loss": 0.06544147431850433, | |
| "eval_runtime": 15.1517, | |
| "eval_samples_per_second": 204.597, | |
| "eval_steps_per_second": 25.608, | |
| "step": 954 | |
| }, | |
| { | |
| "epoch": 3.1446540880503147, | |
| "grad_norm": 0.5175336003303528, | |
| "learning_rate": 1.2138364779874214e-05, | |
| "loss": 0.1141, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_accuracy": 0.8996774193548387, | |
| "eval_loss": 0.05073303356766701, | |
| "eval_runtime": 15.6964, | |
| "eval_samples_per_second": 197.497, | |
| "eval_steps_per_second": 24.719, | |
| "step": 1272 | |
| }, | |
| { | |
| "epoch": 4.716981132075472, | |
| "grad_norm": 0.4490315616130829, | |
| "learning_rate": 8.207547169811321e-06, | |
| "loss": 0.0735, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_accuracy": 0.9045161290322581, | |
| "eval_loss": 0.042617324739694595, | |
| "eval_runtime": 16.182, | |
| "eval_samples_per_second": 191.571, | |
| "eval_steps_per_second": 23.977, | |
| "step": 1590 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_accuracy": 0.9096774193548387, | |
| "eval_loss": 0.03792766109108925, | |
| "eval_runtime": 15.6374, | |
| "eval_samples_per_second": 198.242, | |
| "eval_steps_per_second": 24.812, | |
| "step": 1908 | |
| }, | |
| { | |
| "epoch": 6.289308176100629, | |
| "grad_norm": 0.22402913868427277, | |
| "learning_rate": 4.276729559748428e-06, | |
| "loss": 0.059, | |
| "step": 2000 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 2544, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 8, | |
| "save_steps": 500, | |
| "total_flos": 520991326672152.0, | |
| "train_batch_size": 48, | |
| "trial_name": null, | |
| "trial_params": { | |
| "alpha": 0.7554572962469277, | |
| "num_train_epochs": 8, | |
| "temperature": 10 | |
| } | |
| } | |