| { | |
| "best_metric": 0.12814190983772278, | |
| "best_model_checkpoint": "/data/wheld3/mt5-base-pointer-adv-mtop/checkpoint-1000", | |
| "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": 1.7704, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 1.09, | |
| "eval_exact_match": 0.13154362416107382, | |
| "eval_loss": 0.3664160668849945, | |
| "eval_runtime": 115.2843, | |
| "eval_samples_per_second": 19.387, | |
| "eval_steps_per_second": 2.429, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 2.17, | |
| "learning_rate": 0.0008666666666666667, | |
| "loss": 1.9751, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 2.17, | |
| "eval_exact_match": 0.3400447427293065, | |
| "eval_loss": 0.20914442837238312, | |
| "eval_runtime": 116.9108, | |
| "eval_samples_per_second": 19.117, | |
| "eval_steps_per_second": 2.395, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 3.26, | |
| "learning_rate": 0.0008, | |
| "loss": 1.0019, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 3.26, | |
| "eval_exact_match": 0.45861297539149887, | |
| "eval_loss": 0.14529764652252197, | |
| "eval_runtime": 115.0721, | |
| "eval_samples_per_second": 19.423, | |
| "eval_steps_per_second": 2.433, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 4.35, | |
| "learning_rate": 0.0007333333333333333, | |
| "loss": 1.313, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 4.35, | |
| "eval_exact_match": 0.5064876957494407, | |
| "eval_loss": 0.13125699758529663, | |
| "eval_runtime": 116.3825, | |
| "eval_samples_per_second": 19.204, | |
| "eval_steps_per_second": 2.406, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 5.43, | |
| "learning_rate": 0.0006666666666666666, | |
| "loss": 0.6593, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 5.43, | |
| "eval_exact_match": 0.5266219239373602, | |
| "eval_loss": 0.12814190983772278, | |
| "eval_runtime": 116.3912, | |
| "eval_samples_per_second": 19.202, | |
| "eval_steps_per_second": 2.406, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 6.52, | |
| "learning_rate": 0.0006, | |
| "loss": 0.3216, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 6.52, | |
| "eval_exact_match": 0.5252796420581656, | |
| "eval_loss": 0.13165239989757538, | |
| "eval_runtime": 116.4297, | |
| "eval_samples_per_second": 19.196, | |
| "eval_steps_per_second": 2.405, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 7.61, | |
| "learning_rate": 0.0005333333333333334, | |
| "loss": 0.4614, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 7.61, | |
| "eval_exact_match": 0.5261744966442953, | |
| "eval_loss": 0.1507694572210312, | |
| "eval_runtime": 116.8619, | |
| "eval_samples_per_second": 19.125, | |
| "eval_steps_per_second": 2.396, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 8.69, | |
| "learning_rate": 0.00046666666666666666, | |
| "loss": 0.3577, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 8.69, | |
| "eval_exact_match": 0.5360178970917227, | |
| "eval_loss": 0.1421622931957245, | |
| "eval_runtime": 121.8344, | |
| "eval_samples_per_second": 18.345, | |
| "eval_steps_per_second": 2.298, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 9.78, | |
| "learning_rate": 0.0004, | |
| "loss": 0.3748, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 9.78, | |
| "eval_exact_match": 0.5458612975391499, | |
| "eval_loss": 0.14189742505550385, | |
| "eval_runtime": 119.5737, | |
| "eval_samples_per_second": 18.691, | |
| "eval_steps_per_second": 2.342, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 10.87, | |
| "learning_rate": 0.0003333333333333333, | |
| "loss": 0.2422, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 10.87, | |
| "eval_exact_match": 0.5355704697986577, | |
| "eval_loss": 0.16032171249389648, | |
| "eval_runtime": 125.8745, | |
| "eval_samples_per_second": 17.756, | |
| "eval_steps_per_second": 2.224, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 11.96, | |
| "learning_rate": 0.0002666666666666667, | |
| "loss": 0.4443, | |
| "step": 2200 | |
| }, | |
| { | |
| "epoch": 11.96, | |
| "eval_exact_match": 0.5472035794183445, | |
| "eval_loss": 0.15260477364063263, | |
| "eval_runtime": 118.267, | |
| "eval_samples_per_second": 18.898, | |
| "eval_steps_per_second": 2.368, | |
| "step": 2200 | |
| }, | |
| { | |
| "epoch": 13.04, | |
| "learning_rate": 0.0002, | |
| "loss": 0.2671, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 13.04, | |
| "eval_exact_match": 0.5480984340044742, | |
| "eval_loss": 0.16060054302215576, | |
| "eval_runtime": 118.1286, | |
| "eval_samples_per_second": 18.92, | |
| "eval_steps_per_second": 2.37, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 14.13, | |
| "learning_rate": 0.00013333333333333334, | |
| "loss": 0.227, | |
| "step": 2600 | |
| }, | |
| { | |
| "epoch": 14.13, | |
| "eval_exact_match": 0.5440715883668904, | |
| "eval_loss": 0.1774316281080246, | |
| "eval_runtime": 119.688, | |
| "eval_samples_per_second": 18.674, | |
| "eval_steps_per_second": 2.339, | |
| "step": 2600 | |
| }, | |
| { | |
| "epoch": 15.22, | |
| "learning_rate": 6.666666666666667e-05, | |
| "loss": 0.2053, | |
| "step": 2800 | |
| }, | |
| { | |
| "epoch": 15.22, | |
| "eval_exact_match": 0.5440715883668904, | |
| "eval_loss": 0.1752384901046753, | |
| "eval_runtime": 119.0856, | |
| "eval_samples_per_second": 18.768, | |
| "eval_steps_per_second": 2.351, | |
| "step": 2800 | |
| }, | |
| { | |
| "epoch": 16.3, | |
| "learning_rate": 0.0, | |
| "loss": 0.1517, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 16.3, | |
| "eval_exact_match": 0.5480984340044742, | |
| "eval_loss": 0.1770186424255371, | |
| "eval_runtime": 119.4908, | |
| "eval_samples_per_second": 18.704, | |
| "eval_steps_per_second": 2.343, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 16.3, | |
| "step": 3000, | |
| "total_flos": 7.59685289063286e+16, | |
| "train_loss": 0.6515167045593262, | |
| "train_runtime": 35508.9264, | |
| "train_samples_per_second": 43.257, | |
| "train_steps_per_second": 0.084 | |
| } | |
| ], | |
| "max_steps": 3000, | |
| "num_train_epochs": 17, | |
| "total_flos": 7.59685289063286e+16, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |