| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 9.433962264150944, | |
| "eval_steps": 500, | |
| "global_step": 3000, | |
| "is_hyper_param_search": true, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 1.0, | |
| "eval_accuracy": 0.6083870967741936, | |
| "eval_loss": 0.20560868084430695, | |
| "eval_runtime": 2.1282, | |
| "eval_samples_per_second": 1456.619, | |
| "eval_steps_per_second": 30.542, | |
| "step": 318 | |
| }, | |
| { | |
| "epoch": 1.57, | |
| "grad_norm": 0.6179726123809814, | |
| "learning_rate": 1.685534591194969e-05, | |
| "loss": 0.3326, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_accuracy": 0.8329032258064516, | |
| "eval_loss": 0.09781355410814285, | |
| "eval_runtime": 5.4647, | |
| "eval_samples_per_second": 567.274, | |
| "eval_steps_per_second": 11.894, | |
| "step": 636 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_accuracy": 0.8767741935483871, | |
| "eval_loss": 0.06353170424699783, | |
| "eval_runtime": 1.915, | |
| "eval_samples_per_second": 1618.783, | |
| "eval_steps_per_second": 33.942, | |
| "step": 954 | |
| }, | |
| { | |
| "epoch": 3.14, | |
| "grad_norm": 0.49803560972213745, | |
| "learning_rate": 1.371069182389937e-05, | |
| "loss": 0.1133, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_accuracy": 0.9025806451612903, | |
| "eval_loss": 0.04788599908351898, | |
| "eval_runtime": 2.0998, | |
| "eval_samples_per_second": 1476.359, | |
| "eval_steps_per_second": 30.956, | |
| "step": 1272 | |
| }, | |
| { | |
| "epoch": 4.72, | |
| "grad_norm": 0.36760082840919495, | |
| "learning_rate": 1.0566037735849058e-05, | |
| "loss": 0.0705, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_accuracy": 0.9125806451612903, | |
| "eval_loss": 0.039302438497543335, | |
| "eval_runtime": 2.0979, | |
| "eval_samples_per_second": 1477.652, | |
| "eval_steps_per_second": 30.983, | |
| "step": 1590 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_accuracy": 0.9170967741935484, | |
| "eval_loss": 0.03421470522880554, | |
| "eval_runtime": 2.0365, | |
| "eval_samples_per_second": 1522.235, | |
| "eval_steps_per_second": 31.918, | |
| "step": 1908 | |
| }, | |
| { | |
| "epoch": 6.29, | |
| "grad_norm": 0.31035614013671875, | |
| "learning_rate": 7.421383647798742e-06, | |
| "loss": 0.0545, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "eval_accuracy": 0.9245161290322581, | |
| "eval_loss": 0.031093453988432884, | |
| "eval_runtime": 2.0547, | |
| "eval_samples_per_second": 1508.719, | |
| "eval_steps_per_second": 31.634, | |
| "step": 2226 | |
| }, | |
| { | |
| "epoch": 7.86, | |
| "grad_norm": 0.19460733234882355, | |
| "learning_rate": 4.276729559748428e-06, | |
| "loss": 0.047, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_accuracy": 0.9280645161290323, | |
| "eval_loss": 0.028909271582961082, | |
| "eval_runtime": 2.0696, | |
| "eval_samples_per_second": 1497.887, | |
| "eval_steps_per_second": 31.407, | |
| "step": 2544 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "eval_accuracy": 0.9283870967741935, | |
| "eval_loss": 0.02790053002536297, | |
| "eval_runtime": 2.0834, | |
| "eval_samples_per_second": 1487.926, | |
| "eval_steps_per_second": 31.198, | |
| "step": 2862 | |
| }, | |
| { | |
| "epoch": 9.43, | |
| "grad_norm": 0.2455333024263382, | |
| "learning_rate": 1.1320754716981133e-06, | |
| "loss": 0.0435, | |
| "step": 3000 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 3180, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 10, | |
| "save_steps": 500, | |
| "total_flos": 792206967613404.0, | |
| "train_batch_size": 48, | |
| "trial_name": null, | |
| "trial_params": { | |
| "alpha": 0.6128903054129909, | |
| "num_train_epochs": 10, | |
| "temperature": 8 | |
| } | |
| } | |