| { | |
| "best_metric": 0.025585556402802467, | |
| "best_model_checkpoint": "/data/wheld3/mt5-base-pointer-top_v2/checkpoint-2600", | |
| "epoch": 12.345168539325842, | |
| "global_step": 3000, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.82, | |
| "learning_rate": 0.0009333333333333333, | |
| "loss": 1.4545, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.82, | |
| "eval_exact_match": 0.12942890442890442, | |
| "eval_loss": 0.25418975949287415, | |
| "eval_runtime": 1925.1292, | |
| "eval_samples_per_second": 8.914, | |
| "eval_steps_per_second": 2.228, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 1.65, | |
| "learning_rate": 0.0008666666666666667, | |
| "loss": 0.1878, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 1.65, | |
| "eval_exact_match": 0.21276223776223777, | |
| "eval_loss": 0.06684188544750214, | |
| "eval_runtime": 1936.6986, | |
| "eval_samples_per_second": 8.86, | |
| "eval_steps_per_second": 2.215, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 2.47, | |
| "learning_rate": 0.0008, | |
| "loss": 0.0796, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 2.47, | |
| "eval_exact_match": 0.22756410256410256, | |
| "eval_loss": 0.04656381905078888, | |
| "eval_runtime": 1952.2349, | |
| "eval_samples_per_second": 8.79, | |
| "eval_steps_per_second": 2.197, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 3.29, | |
| "learning_rate": 0.0007333333333333333, | |
| "loss": 0.0536, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 3.29, | |
| "eval_exact_match": 0.23088578088578088, | |
| "eval_loss": 0.03555314987897873, | |
| "eval_runtime": 1939.2428, | |
| "eval_samples_per_second": 8.849, | |
| "eval_steps_per_second": 2.212, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 4.12, | |
| "learning_rate": 0.0006666666666666666, | |
| "loss": 0.0424, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 4.12, | |
| "eval_exact_match": 0.23280885780885782, | |
| "eval_loss": 0.031658221036195755, | |
| "eval_runtime": 1929.3944, | |
| "eval_samples_per_second": 8.894, | |
| "eval_steps_per_second": 2.223, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 4.94, | |
| "learning_rate": 0.0006, | |
| "loss": 0.0356, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 4.94, | |
| "eval_exact_match": 0.23403263403263402, | |
| "eval_loss": 0.02952779084444046, | |
| "eval_runtime": 1929.8018, | |
| "eval_samples_per_second": 8.892, | |
| "eval_steps_per_second": 2.223, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 5.76, | |
| "learning_rate": 0.0005333333333333334, | |
| "loss": 0.0306, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 5.76, | |
| "eval_exact_match": 0.23572261072261072, | |
| "eval_loss": 0.028804348781704903, | |
| "eval_runtime": 1939.2007, | |
| "eval_samples_per_second": 8.849, | |
| "eval_steps_per_second": 2.212, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 6.58, | |
| "learning_rate": 0.00046666666666666666, | |
| "loss": 0.0277, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 6.58, | |
| "eval_exact_match": 0.23513986013986013, | |
| "eval_loss": 0.027136022225022316, | |
| "eval_runtime": 1941.0005, | |
| "eval_samples_per_second": 8.841, | |
| "eval_steps_per_second": 2.21, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 7.41, | |
| "learning_rate": 0.0004, | |
| "loss": 0.0243, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 7.41, | |
| "eval_exact_match": 0.23513986013986013, | |
| "eval_loss": 0.02723020501434803, | |
| "eval_runtime": 1956.0015, | |
| "eval_samples_per_second": 8.773, | |
| "eval_steps_per_second": 2.193, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 8.23, | |
| "learning_rate": 0.0003333333333333333, | |
| "loss": 0.0225, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 8.23, | |
| "eval_exact_match": 0.2353146853146853, | |
| "eval_loss": 0.02720719203352928, | |
| "eval_runtime": 1939.1308, | |
| "eval_samples_per_second": 8.849, | |
| "eval_steps_per_second": 2.212, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 9.05, | |
| "learning_rate": 0.0002666666666666667, | |
| "loss": 0.0206, | |
| "step": 2200 | |
| }, | |
| { | |
| "epoch": 9.05, | |
| "eval_exact_match": 0.23682983682983683, | |
| "eval_loss": 0.026673857122659683, | |
| "eval_runtime": 1950.8619, | |
| "eval_samples_per_second": 8.796, | |
| "eval_steps_per_second": 2.199, | |
| "step": 2200 | |
| }, | |
| { | |
| "epoch": 9.88, | |
| "learning_rate": 0.0002, | |
| "loss": 0.0187, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 9.88, | |
| "eval_exact_match": 0.2367132867132867, | |
| "eval_loss": 0.02599777653813362, | |
| "eval_runtime": 1938.3074, | |
| "eval_samples_per_second": 8.853, | |
| "eval_steps_per_second": 2.213, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 10.7, | |
| "learning_rate": 0.00013333333333333334, | |
| "loss": 0.0173, | |
| "step": 2600 | |
| }, | |
| { | |
| "epoch": 10.7, | |
| "eval_exact_match": 0.23828671328671328, | |
| "eval_loss": 0.025585556402802467, | |
| "eval_runtime": 1929.5665, | |
| "eval_samples_per_second": 8.893, | |
| "eval_steps_per_second": 2.223, | |
| "step": 2600 | |
| }, | |
| { | |
| "epoch": 11.52, | |
| "learning_rate": 6.666666666666667e-05, | |
| "loss": 0.0161, | |
| "step": 2800 | |
| }, | |
| { | |
| "epoch": 11.52, | |
| "eval_exact_match": 0.23828671328671328, | |
| "eval_loss": 0.02604704163968563, | |
| "eval_runtime": 1944.3112, | |
| "eval_samples_per_second": 8.826, | |
| "eval_steps_per_second": 2.206, | |
| "step": 2800 | |
| }, | |
| { | |
| "epoch": 12.35, | |
| "learning_rate": 0.0, | |
| "loss": 0.0153, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 12.35, | |
| "eval_exact_match": 0.2377039627039627, | |
| "eval_loss": 0.025679251179099083, | |
| "eval_runtime": 1944.0793, | |
| "eval_samples_per_second": 8.827, | |
| "eval_steps_per_second": 2.207, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 12.35, | |
| "step": 3000, | |
| "total_flos": 7.256060777948774e+16, | |
| "train_loss": 0.13643742847442628, | |
| "train_runtime": 83186.8691, | |
| "train_samples_per_second": 18.464, | |
| "train_steps_per_second": 0.036 | |
| } | |
| ], | |
| "max_steps": 3000, | |
| "num_train_epochs": 13, | |
| "total_flos": 7.256060777948774e+16, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |