| { | |
| "best_metric": 0.6644416567883901, | |
| "best_model_checkpoint": "output/pretraining/vihealthbert-w_unsup-SynPD/lr3e-5_wr0.1_wd0.0/checkpoint-25000", | |
| "epoch": 4.30811649146993, | |
| "eval_steps": 5000, | |
| "global_step": 25000, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.00017232465965879716, | |
| "grad_norm": 41.827266693115234, | |
| "learning_rate": 5.169739789763915e-09, | |
| "loss": 13.7242, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.8616232982939859, | |
| "grad_norm": 5.6151957511901855, | |
| "learning_rate": 2.584869894881958e-05, | |
| "loss": 7.0234, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 0.8616232982939859, | |
| "eval_accuracy": 0.5576033171446898, | |
| "eval_loss": 2.590928077697754, | |
| "eval_runtime": 37.589, | |
| "eval_samples_per_second": 260.023, | |
| "eval_steps_per_second": 16.255, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 1.7232465965879717, | |
| "grad_norm": 4.410743713378906, | |
| "learning_rate": 2.758917801137343e-05, | |
| "loss": 5.2736, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 1.7232465965879717, | |
| "eval_accuracy": 0.596153624823046, | |
| "eval_loss": 2.188957452774048, | |
| "eval_runtime": 37.6297, | |
| "eval_samples_per_second": 259.741, | |
| "eval_steps_per_second": 16.237, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 2.5848698948819577, | |
| "grad_norm": 4.754213809967041, | |
| "learning_rate": 2.4717100350393475e-05, | |
| "loss": 4.9126, | |
| "step": 15000 | |
| }, | |
| { | |
| "epoch": 2.5848698948819577, | |
| "eval_accuracy": 0.6381037526075694, | |
| "eval_loss": 1.909491777420044, | |
| "eval_runtime": 36.8556, | |
| "eval_samples_per_second": 265.197, | |
| "eval_steps_per_second": 16.578, | |
| "step": 15000 | |
| }, | |
| { | |
| "epoch": 3.4464931931759435, | |
| "grad_norm": 4.2047200202941895, | |
| "learning_rate": 2.184502268941352e-05, | |
| "loss": 4.791, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 3.4464931931759435, | |
| "eval_accuracy": 0.6468589272593681, | |
| "eval_loss": 1.8286069631576538, | |
| "eval_runtime": 35.9342, | |
| "eval_samples_per_second": 271.997, | |
| "eval_steps_per_second": 17.003, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 4.30811649146993, | |
| "grad_norm": 4.629519462585449, | |
| "learning_rate": 1.897294502843357e-05, | |
| "loss": 4.6538, | |
| "step": 25000 | |
| }, | |
| { | |
| "epoch": 4.30811649146993, | |
| "eval_accuracy": 0.6644416567883901, | |
| "eval_loss": 1.714424967765808, | |
| "eval_runtime": 35.6122, | |
| "eval_samples_per_second": 274.456, | |
| "eval_steps_per_second": 17.157, | |
| "step": 25000 | |
| } | |
| ], | |
| "logging_steps": 5000, | |
| "max_steps": 58030, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 10, | |
| "save_steps": 5000, | |
| "total_flos": 0.0, | |
| "train_batch_size": 32, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |