| { | |
| "best_metric": 0.08161530643701553, | |
| "best_model_checkpoint": "/data/wheld3/mt5-small-pointer-cstop_artificial/checkpoint-800", | |
| "epoch": 428.5019607843137, | |
| "global_step": 3000, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 28.5, | |
| "learning_rate": 0.0009333333333333333, | |
| "loss": 2.08, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 28.5, | |
| "eval_exact_match": 0.03756708407871199, | |
| "eval_loss": 0.331952303647995, | |
| "eval_runtime": 10.6308, | |
| "eval_samples_per_second": 52.583, | |
| "eval_steps_per_second": 3.292, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 57.13, | |
| "learning_rate": 0.0008666666666666667, | |
| "loss": 0.272, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 57.13, | |
| "eval_exact_match": 0.2629695885509839, | |
| "eval_loss": 0.10836012661457062, | |
| "eval_runtime": 10.9563, | |
| "eval_samples_per_second": 51.021, | |
| "eval_steps_per_second": 3.194, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 85.63, | |
| "learning_rate": 0.0008, | |
| "loss": 0.0789, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 85.63, | |
| "eval_exact_match": 0.3184257602862254, | |
| "eval_loss": 0.08299708366394043, | |
| "eval_runtime": 11.1629, | |
| "eval_samples_per_second": 50.077, | |
| "eval_steps_per_second": 3.135, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 114.25, | |
| "learning_rate": 0.0007333333333333333, | |
| "loss": 0.0355, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 114.25, | |
| "eval_exact_match": 0.3363148479427549, | |
| "eval_loss": 0.08161530643701553, | |
| "eval_runtime": 11.0139, | |
| "eval_samples_per_second": 50.754, | |
| "eval_steps_per_second": 3.178, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 142.75, | |
| "learning_rate": 0.0006666666666666666, | |
| "loss": 0.0207, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 142.75, | |
| "eval_exact_match": 0.3291592128801431, | |
| "eval_loss": 0.08680771291255951, | |
| "eval_runtime": 11.1063, | |
| "eval_samples_per_second": 50.332, | |
| "eval_steps_per_second": 3.151, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 171.38, | |
| "learning_rate": 0.0006, | |
| "loss": 0.014, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 171.38, | |
| "eval_exact_match": 0.33989266547406083, | |
| "eval_loss": 0.09518682211637497, | |
| "eval_runtime": 10.9222, | |
| "eval_samples_per_second": 51.18, | |
| "eval_steps_per_second": 3.204, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 199.88, | |
| "learning_rate": 0.0005333333333333334, | |
| "loss": 0.0099, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 199.88, | |
| "eval_exact_match": 0.33810375670840787, | |
| "eval_loss": 0.1089245155453682, | |
| "eval_runtime": 10.85, | |
| "eval_samples_per_second": 51.521, | |
| "eval_steps_per_second": 3.226, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 228.5, | |
| "learning_rate": 0.00046666666666666666, | |
| "loss": 0.0076, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 228.5, | |
| "eval_exact_match": 0.33810375670840787, | |
| "eval_loss": 0.11035683006048203, | |
| "eval_runtime": 11.2414, | |
| "eval_samples_per_second": 49.727, | |
| "eval_steps_per_second": 3.113, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 257.13, | |
| "learning_rate": 0.0004, | |
| "loss": 0.0057, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 257.13, | |
| "eval_exact_match": 0.3291592128801431, | |
| "eval_loss": 0.11526591330766678, | |
| "eval_runtime": 10.9866, | |
| "eval_samples_per_second": 50.88, | |
| "eval_steps_per_second": 3.186, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 285.63, | |
| "learning_rate": 0.0003333333333333333, | |
| "loss": 0.0048, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 285.63, | |
| "eval_exact_match": 0.33273703041144903, | |
| "eval_loss": 0.11532965302467346, | |
| "eval_runtime": 11.0349, | |
| "eval_samples_per_second": 50.658, | |
| "eval_steps_per_second": 3.172, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 314.25, | |
| "learning_rate": 0.0002666666666666667, | |
| "loss": 0.004, | |
| "step": 2200 | |
| }, | |
| { | |
| "epoch": 314.25, | |
| "eval_exact_match": 0.3363148479427549, | |
| "eval_loss": 0.1206129789352417, | |
| "eval_runtime": 10.904, | |
| "eval_samples_per_second": 51.266, | |
| "eval_steps_per_second": 3.21, | |
| "step": 2200 | |
| }, | |
| { | |
| "epoch": 342.75, | |
| "learning_rate": 0.0002, | |
| "loss": 0.0032, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 342.75, | |
| "eval_exact_match": 0.3363148479427549, | |
| "eval_loss": 0.1229204535484314, | |
| "eval_runtime": 10.9966, | |
| "eval_samples_per_second": 50.834, | |
| "eval_steps_per_second": 3.183, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 371.38, | |
| "learning_rate": 0.00013333333333333334, | |
| "loss": 0.0028, | |
| "step": 2600 | |
| }, | |
| { | |
| "epoch": 371.38, | |
| "eval_exact_match": 0.33810375670840787, | |
| "eval_loss": 0.12684644758701324, | |
| "eval_runtime": 10.9769, | |
| "eval_samples_per_second": 50.925, | |
| "eval_steps_per_second": 3.189, | |
| "step": 2600 | |
| }, | |
| { | |
| "epoch": 399.88, | |
| "learning_rate": 6.666666666666667e-05, | |
| "loss": 0.0023, | |
| "step": 2800 | |
| }, | |
| { | |
| "epoch": 399.88, | |
| "eval_exact_match": 0.33989266547406083, | |
| "eval_loss": 0.12882176041603088, | |
| "eval_runtime": 10.9618, | |
| "eval_samples_per_second": 50.995, | |
| "eval_steps_per_second": 3.193, | |
| "step": 2800 | |
| }, | |
| { | |
| "epoch": 428.5, | |
| "learning_rate": 0.0, | |
| "loss": 0.002, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 428.5, | |
| "eval_exact_match": 0.33989266547406083, | |
| "eval_loss": 0.12921586632728577, | |
| "eval_runtime": 11.4552, | |
| "eval_samples_per_second": 48.799, | |
| "eval_steps_per_second": 3.055, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 428.5, | |
| "step": 3000, | |
| "total_flos": 3.4731986474409984e+16, | |
| "train_loss": 0.16957045908768972, | |
| "train_runtime": 14253.2821, | |
| "train_samples_per_second": 107.765, | |
| "train_steps_per_second": 0.21 | |
| } | |
| ], | |
| "max_steps": 3000, | |
| "num_train_epochs": 429, | |
| "total_flos": 3.4731986474409984e+16, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |