| { | |
| "best_metric": 0.07756359130144119, | |
| "best_model_checkpoint": "/data/wheld3/mt5-base-pointer-cstop_artificial/checkpoint-400", | |
| "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": 1.7482, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 28.5, | |
| "eval_exact_match": 0.10196779964221825, | |
| "eval_loss": 0.2504979968070984, | |
| "eval_runtime": 32.6293, | |
| "eval_samples_per_second": 17.132, | |
| "eval_steps_per_second": 2.145, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 57.13, | |
| "learning_rate": 0.0008666666666666667, | |
| "loss": 0.1366, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 57.13, | |
| "eval_exact_match": 0.32379248658318427, | |
| "eval_loss": 0.07756359130144119, | |
| "eval_runtime": 32.4871, | |
| "eval_samples_per_second": 17.207, | |
| "eval_steps_per_second": 2.155, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 85.63, | |
| "learning_rate": 0.0008, | |
| "loss": 0.0275, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 85.63, | |
| "eval_exact_match": 0.33810375670840787, | |
| "eval_loss": 0.0881282389163971, | |
| "eval_runtime": 32.966, | |
| "eval_samples_per_second": 16.957, | |
| "eval_steps_per_second": 2.123, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 114.25, | |
| "learning_rate": 0.0007333333333333333, | |
| "loss": 0.0114, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 114.25, | |
| "eval_exact_match": 0.33989266547406083, | |
| "eval_loss": 0.0990389883518219, | |
| "eval_runtime": 33.8803, | |
| "eval_samples_per_second": 16.499, | |
| "eval_steps_per_second": 2.066, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 142.75, | |
| "learning_rate": 0.0006666666666666666, | |
| "loss": 0.0064, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 142.75, | |
| "eval_exact_match": 0.3416815742397138, | |
| "eval_loss": 0.11197585612535477, | |
| "eval_runtime": 32.7339, | |
| "eval_samples_per_second": 17.077, | |
| "eval_steps_per_second": 2.138, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 171.38, | |
| "learning_rate": 0.0006, | |
| "loss": 0.0045, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 171.38, | |
| "eval_exact_match": 0.3434704830053667, | |
| "eval_loss": 0.10809481889009476, | |
| "eval_runtime": 32.4483, | |
| "eval_samples_per_second": 17.227, | |
| "eval_steps_per_second": 2.157, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 199.88, | |
| "learning_rate": 0.0005333333333333334, | |
| "loss": 0.0036, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 199.88, | |
| "eval_exact_match": 0.3434704830053667, | |
| "eval_loss": 0.12303601205348969, | |
| "eval_runtime": 33.761, | |
| "eval_samples_per_second": 16.558, | |
| "eval_steps_per_second": 2.073, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 228.5, | |
| "learning_rate": 0.00046666666666666666, | |
| "loss": 0.0025, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 228.5, | |
| "eval_exact_match": 0.33989266547406083, | |
| "eval_loss": 0.12110502272844315, | |
| "eval_runtime": 39.4879, | |
| "eval_samples_per_second": 14.156, | |
| "eval_steps_per_second": 1.773, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 257.13, | |
| "learning_rate": 0.0004, | |
| "loss": 0.002, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 257.13, | |
| "eval_exact_match": 0.33989266547406083, | |
| "eval_loss": 0.13667185604572296, | |
| "eval_runtime": 33.6552, | |
| "eval_samples_per_second": 16.61, | |
| "eval_steps_per_second": 2.08, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 285.63, | |
| "learning_rate": 0.0003333333333333333, | |
| "loss": 0.0016, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 285.63, | |
| "eval_exact_match": 0.3434704830053667, | |
| "eval_loss": 0.13241663575172424, | |
| "eval_runtime": 33.4579, | |
| "eval_samples_per_second": 16.708, | |
| "eval_steps_per_second": 2.092, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 314.25, | |
| "learning_rate": 0.0002666666666666667, | |
| "loss": 0.0013, | |
| "step": 2200 | |
| }, | |
| { | |
| "epoch": 314.25, | |
| "eval_exact_match": 0.3470483005366726, | |
| "eval_loss": 0.13395807147026062, | |
| "eval_runtime": 34.1281, | |
| "eval_samples_per_second": 16.379, | |
| "eval_steps_per_second": 2.051, | |
| "step": 2200 | |
| }, | |
| { | |
| "epoch": 342.75, | |
| "learning_rate": 0.0002, | |
| "loss": 0.001, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 342.75, | |
| "eval_exact_match": 0.3434704830053667, | |
| "eval_loss": 0.13736827671527863, | |
| "eval_runtime": 33.0567, | |
| "eval_samples_per_second": 16.91, | |
| "eval_steps_per_second": 2.118, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 371.38, | |
| "learning_rate": 0.00013333333333333334, | |
| "loss": 0.0009, | |
| "step": 2600 | |
| }, | |
| { | |
| "epoch": 371.38, | |
| "eval_exact_match": 0.3416815742397138, | |
| "eval_loss": 0.13843561708927155, | |
| "eval_runtime": 33.1328, | |
| "eval_samples_per_second": 16.871, | |
| "eval_steps_per_second": 2.113, | |
| "step": 2600 | |
| }, | |
| { | |
| "epoch": 399.88, | |
| "learning_rate": 6.666666666666667e-05, | |
| "loss": 0.0007, | |
| "step": 2800 | |
| }, | |
| { | |
| "epoch": 399.88, | |
| "eval_exact_match": 0.3434704830053667, | |
| "eval_loss": 0.1421527862548828, | |
| "eval_runtime": 32.5079, | |
| "eval_samples_per_second": 17.196, | |
| "eval_steps_per_second": 2.153, | |
| "step": 2800 | |
| }, | |
| { | |
| "epoch": 428.5, | |
| "learning_rate": 0.0, | |
| "loss": 0.0006, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 428.5, | |
| "eval_exact_match": 0.3416815742397138, | |
| "eval_loss": 0.1451684683561325, | |
| "eval_runtime": 32.385, | |
| "eval_samples_per_second": 17.261, | |
| "eval_steps_per_second": 2.161, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 428.5, | |
| "step": 3000, | |
| "total_flos": 7.330479471426355e+16, | |
| "train_loss": 0.12990600236753622, | |
| "train_runtime": 34726.3796, | |
| "train_samples_per_second": 44.232, | |
| "train_steps_per_second": 0.086 | |
| } | |
| ], | |
| "max_steps": 3000, | |
| "num_train_epochs": 429, | |
| "total_flos": 7.330479471426355e+16, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |