| { | |
| "best_metric": 0.13407668471336365, | |
| "best_model_checkpoint": "/data/wheld3/mt5-small-pointer-adv-mtop/checkpoint-3000", | |
| "epoch": 16.304012214776485, | |
| "global_step": 3000, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 1.09, | |
| "learning_rate": 0.0009333333333333333, | |
| "loss": 2.1628, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 1.09, | |
| "eval_exact_match": 0.0022371364653243847, | |
| "eval_loss": 0.7204959988594055, | |
| "eval_runtime": 75.5591, | |
| "eval_samples_per_second": 29.58, | |
| "eval_steps_per_second": 3.706, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 2.17, | |
| "learning_rate": 0.0008666666666666667, | |
| "loss": 1.1208, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 2.17, | |
| "eval_exact_match": 0.0013422818791946308, | |
| "eval_loss": 0.63932865858078, | |
| "eval_runtime": 78.5979, | |
| "eval_samples_per_second": 28.436, | |
| "eval_steps_per_second": 3.562, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 3.26, | |
| "learning_rate": 0.0008, | |
| "loss": 0.8675, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 3.26, | |
| "eval_exact_match": 0.0026845637583892616, | |
| "eval_loss": 0.5905107259750366, | |
| "eval_runtime": 71.821, | |
| "eval_samples_per_second": 31.119, | |
| "eval_steps_per_second": 3.899, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 4.35, | |
| "learning_rate": 0.0007333333333333333, | |
| "loss": 1.8729, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 4.35, | |
| "eval_exact_match": 0.003131991051454139, | |
| "eval_loss": 0.5726307034492493, | |
| "eval_runtime": 78.7696, | |
| "eval_samples_per_second": 28.374, | |
| "eval_steps_per_second": 3.555, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 5.43, | |
| "learning_rate": 0.0006666666666666666, | |
| "loss": 3.5417, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 5.43, | |
| "eval_exact_match": 0.006711409395973154, | |
| "eval_loss": 0.5370941758155823, | |
| "eval_runtime": 81.2185, | |
| "eval_samples_per_second": 27.518, | |
| "eval_steps_per_second": 3.447, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 6.52, | |
| "learning_rate": 0.0006, | |
| "loss": 0.9087, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 6.52, | |
| "eval_exact_match": 0.11185682326621924, | |
| "eval_loss": 0.3511998653411865, | |
| "eval_runtime": 82.0065, | |
| "eval_samples_per_second": 27.254, | |
| "eval_steps_per_second": 3.414, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 7.61, | |
| "learning_rate": 0.0005333333333333334, | |
| "loss": 1.2224, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 7.61, | |
| "eval_exact_match": 0.19105145413870245, | |
| "eval_loss": 0.27385014295578003, | |
| "eval_runtime": 82.7127, | |
| "eval_samples_per_second": 27.021, | |
| "eval_steps_per_second": 3.385, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 8.69, | |
| "learning_rate": 0.00046666666666666666, | |
| "loss": 0.7597, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 8.69, | |
| "eval_exact_match": 0.30156599552572705, | |
| "eval_loss": 0.21514081954956055, | |
| "eval_runtime": 82.6161, | |
| "eval_samples_per_second": 27.053, | |
| "eval_steps_per_second": 3.389, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 9.78, | |
| "learning_rate": 0.0004, | |
| "loss": 0.6981, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 9.78, | |
| "eval_exact_match": 0.3749440715883669, | |
| "eval_loss": 0.17362748086452484, | |
| "eval_runtime": 82.0736, | |
| "eval_samples_per_second": 27.232, | |
| "eval_steps_per_second": 3.412, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 10.87, | |
| "learning_rate": 0.0003333333333333333, | |
| "loss": 0.4779, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 10.87, | |
| "eval_exact_match": 0.41655480984340043, | |
| "eval_loss": 0.15482261776924133, | |
| "eval_runtime": 82.6077, | |
| "eval_samples_per_second": 27.056, | |
| "eval_steps_per_second": 3.39, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 11.96, | |
| "learning_rate": 0.0002666666666666667, | |
| "loss": 0.4397, | |
| "step": 2200 | |
| }, | |
| { | |
| "epoch": 11.96, | |
| "eval_exact_match": 0.45100671140939597, | |
| "eval_loss": 0.13771148025989532, | |
| "eval_runtime": 79.9201, | |
| "eval_samples_per_second": 27.965, | |
| "eval_steps_per_second": 3.503, | |
| "step": 2200 | |
| }, | |
| { | |
| "epoch": 13.04, | |
| "learning_rate": 0.0002, | |
| "loss": 0.4101, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 13.04, | |
| "eval_exact_match": 0.4196868008948546, | |
| "eval_loss": 0.14801675081253052, | |
| "eval_runtime": 81.9056, | |
| "eval_samples_per_second": 27.288, | |
| "eval_steps_per_second": 3.419, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 14.13, | |
| "learning_rate": 0.00013333333333333334, | |
| "loss": 0.3323, | |
| "step": 2600 | |
| }, | |
| { | |
| "epoch": 14.13, | |
| "eval_exact_match": 0.43982102908277404, | |
| "eval_loss": 0.13956378400325775, | |
| "eval_runtime": 80.3363, | |
| "eval_samples_per_second": 27.821, | |
| "eval_steps_per_second": 3.485, | |
| "step": 2600 | |
| }, | |
| { | |
| "epoch": 15.22, | |
| "learning_rate": 6.666666666666667e-05, | |
| "loss": 0.2565, | |
| "step": 2800 | |
| }, | |
| { | |
| "epoch": 15.22, | |
| "eval_exact_match": 0.45234899328859063, | |
| "eval_loss": 0.13505251705646515, | |
| "eval_runtime": 80.6724, | |
| "eval_samples_per_second": 27.705, | |
| "eval_steps_per_second": 3.471, | |
| "step": 2800 | |
| }, | |
| { | |
| "epoch": 16.3, | |
| "learning_rate": 0.0, | |
| "loss": 0.2108, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 16.3, | |
| "eval_exact_match": 0.4541387024608501, | |
| "eval_loss": 0.13407668471336365, | |
| "eval_runtime": 80.1776, | |
| "eval_samples_per_second": 27.876, | |
| "eval_steps_per_second": 3.492, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 16.3, | |
| "step": 3000, | |
| "total_flos": 3.3508283441823776e+16, | |
| "train_loss": 1.0188030764261882, | |
| "train_runtime": 23687.0417, | |
| "train_samples_per_second": 64.846, | |
| "train_steps_per_second": 0.127 | |
| } | |
| ], | |
| "max_steps": 3000, | |
| "num_train_epochs": 17, | |
| "total_flos": 3.3508283441823776e+16, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |