| { | |
| "best_metric": 0.10638269782066345, | |
| "best_model_checkpoint": "/data/wheld3/mt5-small-pointer-mtop/checkpoint-1400", | |
| "epoch": 99.9795918367347, | |
| "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": 2.0944, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 6.65, | |
| "eval_exact_match": 0.0026845637583892616, | |
| "eval_loss": 0.6548439264297485, | |
| "eval_runtime": 44.8886, | |
| "eval_samples_per_second": 49.79, | |
| "eval_steps_per_second": 3.119, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 13.33, | |
| "learning_rate": 0.0008666666666666667, | |
| "loss": 0.5307, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 13.33, | |
| "eval_exact_match": 0.2778523489932886, | |
| "eval_loss": 0.24562813341617584, | |
| "eval_runtime": 45.1956, | |
| "eval_samples_per_second": 49.452, | |
| "eval_steps_per_second": 3.098, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 19.98, | |
| "learning_rate": 0.0008, | |
| "loss": 0.2388, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 19.98, | |
| "eval_exact_match": 0.4559284116331096, | |
| "eval_loss": 0.14859849214553833, | |
| "eval_runtime": 45.6558, | |
| "eval_samples_per_second": 48.953, | |
| "eval_steps_per_second": 3.066, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 26.65, | |
| "learning_rate": 0.0007333333333333333, | |
| "loss": 0.1459, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 26.65, | |
| "eval_exact_match": 0.5042505592841163, | |
| "eval_loss": 0.11904484033584595, | |
| "eval_runtime": 54.895, | |
| "eval_samples_per_second": 40.714, | |
| "eval_steps_per_second": 2.55, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 33.33, | |
| "learning_rate": 0.0006666666666666666, | |
| "loss": 0.1011, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 33.33, | |
| "eval_exact_match": 0.5230425055928412, | |
| "eval_loss": 0.11172914505004883, | |
| "eval_runtime": 53.9217, | |
| "eval_samples_per_second": 41.449, | |
| "eval_steps_per_second": 2.596, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 39.98, | |
| "learning_rate": 0.0006, | |
| "loss": 0.0774, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 39.98, | |
| "eval_exact_match": 0.5373601789709173, | |
| "eval_loss": 0.10843723267316818, | |
| "eval_runtime": 54.3522, | |
| "eval_samples_per_second": 41.121, | |
| "eval_steps_per_second": 2.576, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 46.65, | |
| "learning_rate": 0.0005333333333333334, | |
| "loss": 0.0598, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 46.65, | |
| "eval_exact_match": 0.5404921700223714, | |
| "eval_loss": 0.10638269782066345, | |
| "eval_runtime": 44.8128, | |
| "eval_samples_per_second": 49.874, | |
| "eval_steps_per_second": 3.124, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 53.33, | |
| "learning_rate": 0.00046666666666666666, | |
| "loss": 0.0478, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 53.33, | |
| "eval_exact_match": 0.545413870246085, | |
| "eval_loss": 0.11469161510467529, | |
| "eval_runtime": 45.8513, | |
| "eval_samples_per_second": 48.745, | |
| "eval_steps_per_second": 3.053, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 59.98, | |
| "learning_rate": 0.0004, | |
| "loss": 0.0397, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 59.98, | |
| "eval_exact_match": 0.5472035794183445, | |
| "eval_loss": 0.11389175802469254, | |
| "eval_runtime": 44.8546, | |
| "eval_samples_per_second": 49.828, | |
| "eval_steps_per_second": 3.121, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 66.65, | |
| "learning_rate": 0.0003333333333333333, | |
| "loss": 0.0337, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 66.65, | |
| "eval_exact_match": 0.5480984340044742, | |
| "eval_loss": 0.117930106818676, | |
| "eval_runtime": 47.7426, | |
| "eval_samples_per_second": 46.814, | |
| "eval_steps_per_second": 2.932, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 73.33, | |
| "learning_rate": 0.0002666666666666667, | |
| "loss": 0.0286, | |
| "step": 2200 | |
| }, | |
| { | |
| "epoch": 73.33, | |
| "eval_exact_match": 0.5498881431767337, | |
| "eval_loss": 0.12433456629514694, | |
| "eval_runtime": 45.3848, | |
| "eval_samples_per_second": 49.246, | |
| "eval_steps_per_second": 3.085, | |
| "step": 2200 | |
| }, | |
| { | |
| "epoch": 79.98, | |
| "learning_rate": 0.0002, | |
| "loss": 0.0251, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 79.98, | |
| "eval_exact_match": 0.5480984340044742, | |
| "eval_loss": 0.12593261897563934, | |
| "eval_runtime": 45.7566, | |
| "eval_samples_per_second": 48.845, | |
| "eval_steps_per_second": 3.06, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 86.65, | |
| "learning_rate": 0.00013333333333333334, | |
| "loss": 0.0218, | |
| "step": 2600 | |
| }, | |
| { | |
| "epoch": 86.65, | |
| "eval_exact_match": 0.5503355704697986, | |
| "eval_loss": 0.12759922444820404, | |
| "eval_runtime": 45.5648, | |
| "eval_samples_per_second": 49.051, | |
| "eval_steps_per_second": 3.073, | |
| "step": 2600 | |
| }, | |
| { | |
| "epoch": 93.33, | |
| "learning_rate": 6.666666666666667e-05, | |
| "loss": 0.0197, | |
| "step": 2800 | |
| }, | |
| { | |
| "epoch": 93.33, | |
| "eval_exact_match": 0.5503355704697986, | |
| "eval_loss": 0.13092631101608276, | |
| "eval_runtime": 45.4393, | |
| "eval_samples_per_second": 49.186, | |
| "eval_steps_per_second": 3.081, | |
| "step": 2800 | |
| }, | |
| { | |
| "epoch": 99.98, | |
| "learning_rate": 0.0, | |
| "loss": 0.0184, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 99.98, | |
| "eval_exact_match": 0.5503355704697986, | |
| "eval_loss": 0.1317095011472702, | |
| "eval_runtime": 45.2758, | |
| "eval_samples_per_second": 49.364, | |
| "eval_steps_per_second": 3.092, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 99.98, | |
| "step": 3000, | |
| "total_flos": 3.831174875254272e+16, | |
| "train_loss": 0.23219555624326071, | |
| "train_runtime": 13657.0516, | |
| "train_samples_per_second": 112.469, | |
| "train_steps_per_second": 0.22 | |
| } | |
| ], | |
| "max_steps": 3000, | |
| "num_train_epochs": 100, | |
| "total_flos": 3.831174875254272e+16, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |