| { | |
| "best_metric": 0.11308582127094269, | |
| "best_model_checkpoint": "/data/wheld3/mt5-base-pointer-mtop/checkpoint-1200", | |
| "epoch": 99.9800918836141, | |
| "global_step": 3000, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 6.65, | |
| "learning_rate": 0.0009333333333333333, | |
| "loss": 1.7749, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 6.65, | |
| "eval_exact_match": 0.003131991051454139, | |
| "eval_loss": 0.5892038345336914, | |
| "eval_runtime": 135.5401, | |
| "eval_samples_per_second": 16.49, | |
| "eval_steps_per_second": 2.066, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 13.33, | |
| "learning_rate": 0.0008666666666666667, | |
| "loss": 0.6021, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 13.33, | |
| "eval_exact_match": 0.013870246085011185, | |
| "eval_loss": 0.5159956216812134, | |
| "eval_runtime": 130.613, | |
| "eval_samples_per_second": 17.112, | |
| "eval_steps_per_second": 2.144, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 19.98, | |
| "learning_rate": 0.0008, | |
| "loss": 0.6044, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 19.98, | |
| "eval_exact_match": 0.053243847874720356, | |
| "eval_loss": 0.40801870822906494, | |
| "eval_runtime": 132.1423, | |
| "eval_samples_per_second": 16.914, | |
| "eval_steps_per_second": 2.119, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 26.65, | |
| "learning_rate": 0.0007333333333333333, | |
| "loss": 0.3302, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 26.65, | |
| "eval_exact_match": 0.36196868008948546, | |
| "eval_loss": 0.18654391169548035, | |
| "eval_runtime": 134.276, | |
| "eval_samples_per_second": 16.645, | |
| "eval_steps_per_second": 2.085, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 33.33, | |
| "learning_rate": 0.0006666666666666666, | |
| "loss": 0.1483, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 33.33, | |
| "eval_exact_match": 0.5105145413870246, | |
| "eval_loss": 0.1267053484916687, | |
| "eval_runtime": 133.0912, | |
| "eval_samples_per_second": 16.793, | |
| "eval_steps_per_second": 2.104, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 39.98, | |
| "learning_rate": 0.0006, | |
| "loss": 0.0768, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 39.98, | |
| "eval_exact_match": 0.5297539149888143, | |
| "eval_loss": 0.11308582127094269, | |
| "eval_runtime": 139.8147, | |
| "eval_samples_per_second": 15.985, | |
| "eval_steps_per_second": 2.003, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 46.65, | |
| "learning_rate": 0.0005333333333333334, | |
| "loss": 0.0525, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 46.65, | |
| "eval_exact_match": 0.5413870246085011, | |
| "eval_loss": 0.12185565382242203, | |
| "eval_runtime": 132.0637, | |
| "eval_samples_per_second": 16.924, | |
| "eval_steps_per_second": 2.12, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 53.33, | |
| "learning_rate": 0.00046666666666666666, | |
| "loss": 0.0801, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 53.33, | |
| "eval_exact_match": 0.5275167785234899, | |
| "eval_loss": 0.11860152333974838, | |
| "eval_runtime": 132.256, | |
| "eval_samples_per_second": 16.899, | |
| "eval_steps_per_second": 2.117, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 59.98, | |
| "learning_rate": 0.0004, | |
| "loss": 0.0331, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 59.98, | |
| "eval_exact_match": 0.5422818791946309, | |
| "eval_loss": 0.13056021928787231, | |
| "eval_runtime": 132.9523, | |
| "eval_samples_per_second": 16.811, | |
| "eval_steps_per_second": 2.106, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 66.65, | |
| "learning_rate": 0.0003333333333333333, | |
| "loss": 0.0254, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 66.65, | |
| "eval_exact_match": 0.5395973154362416, | |
| "eval_loss": 0.13960428535938263, | |
| "eval_runtime": 135.7759, | |
| "eval_samples_per_second": 16.461, | |
| "eval_steps_per_second": 2.062, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 73.33, | |
| "learning_rate": 0.0002666666666666667, | |
| "loss": 0.0168, | |
| "step": 2200 | |
| }, | |
| { | |
| "epoch": 73.33, | |
| "eval_exact_match": 0.5436241610738255, | |
| "eval_loss": 0.15595464408397675, | |
| "eval_runtime": 134.0818, | |
| "eval_samples_per_second": 16.669, | |
| "eval_steps_per_second": 2.088, | |
| "step": 2200 | |
| }, | |
| { | |
| "epoch": 79.98, | |
| "learning_rate": 0.0002, | |
| "loss": 0.0129, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 79.98, | |
| "eval_exact_match": 0.5494407158836689, | |
| "eval_loss": 0.16592496633529663, | |
| "eval_runtime": 133.0433, | |
| "eval_samples_per_second": 16.799, | |
| "eval_steps_per_second": 2.105, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 86.65, | |
| "learning_rate": 0.00013333333333333334, | |
| "loss": 0.0105, | |
| "step": 2600 | |
| }, | |
| { | |
| "epoch": 86.65, | |
| "eval_exact_match": 0.5422818791946309, | |
| "eval_loss": 0.16985595226287842, | |
| "eval_runtime": 137.112, | |
| "eval_samples_per_second": 16.301, | |
| "eval_steps_per_second": 2.042, | |
| "step": 2600 | |
| }, | |
| { | |
| "epoch": 93.33, | |
| "learning_rate": 6.666666666666667e-05, | |
| "loss": 0.0088, | |
| "step": 2800 | |
| }, | |
| { | |
| "epoch": 93.33, | |
| "eval_exact_match": 0.5472035794183445, | |
| "eval_loss": 0.17423103749752045, | |
| "eval_runtime": 132.2768, | |
| "eval_samples_per_second": 16.896, | |
| "eval_steps_per_second": 2.117, | |
| "step": 2800 | |
| }, | |
| { | |
| "epoch": 99.98, | |
| "learning_rate": 0.0, | |
| "loss": 0.0077, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 99.98, | |
| "eval_exact_match": 0.5467561521252796, | |
| "eval_loss": 0.17754317820072174, | |
| "eval_runtime": 132.953, | |
| "eval_samples_per_second": 16.81, | |
| "eval_steps_per_second": 2.106, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 99.98, | |
| "step": 3000, | |
| "total_flos": 7.72413437221586e+16, | |
| "train_loss": 0.2523062037229538, | |
| "train_runtime": 33274.6395, | |
| "train_samples_per_second": 46.161, | |
| "train_steps_per_second": 0.09 | |
| } | |
| ], | |
| "max_steps": 3000, | |
| "num_train_epochs": 100, | |
| "total_flos": 7.72413437221586e+16, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |