| { | |
| "best_metric": 0.86832522726694, | |
| "best_model_checkpoint": "./save_models/mnli/roberta-base_lr1e-05/checkpoint-8286", | |
| "epoch": 10.0, | |
| "eval_steps": 500, | |
| "global_step": 13810, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.36, | |
| "learning_rate": 6.031363088057901e-06, | |
| "loss": 0.8769, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.72, | |
| "learning_rate": 9.86826900855096e-06, | |
| "loss": 0.4908, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_accuracy": 0.8439306358381503, | |
| "eval_loss": 0.40705999732017517, | |
| "eval_runtime": 30.2499, | |
| "eval_samples_per_second": 1298.219, | |
| "eval_steps_per_second": 5.091, | |
| "step": 1381 | |
| }, | |
| { | |
| "epoch": 1.09, | |
| "learning_rate": 9.483090670980664e-06, | |
| "loss": 0.4345, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 1.45, | |
| "learning_rate": 9.09791233341037e-06, | |
| "loss": 0.3978, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 1.81, | |
| "learning_rate": 8.712733995840074e-06, | |
| "loss": 0.3855, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_accuracy": 0.8618318861246212, | |
| "eval_loss": 0.3706605136394501, | |
| "eval_runtime": 30.5079, | |
| "eval_samples_per_second": 1287.241, | |
| "eval_steps_per_second": 5.048, | |
| "step": 2762 | |
| }, | |
| { | |
| "epoch": 2.17, | |
| "learning_rate": 8.32755565826978e-06, | |
| "loss": 0.3572, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 2.53, | |
| "learning_rate": 7.942377320699485e-06, | |
| "loss": 0.3368, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 2.9, | |
| "learning_rate": 7.557198983129189e-06, | |
| "loss": 0.3354, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_accuracy": 0.8657533548929235, | |
| "eval_loss": 0.36279234290122986, | |
| "eval_runtime": 30.3356, | |
| "eval_samples_per_second": 1294.551, | |
| "eval_steps_per_second": 5.077, | |
| "step": 4143 | |
| }, | |
| { | |
| "epoch": 3.26, | |
| "learning_rate": 7.172020645558895e-06, | |
| "loss": 0.3097, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 3.62, | |
| "learning_rate": 6.7868423079885995e-06, | |
| "loss": 0.2998, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 3.98, | |
| "learning_rate": 6.401663970418303e-06, | |
| "loss": 0.3005, | |
| "step": 5500 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_accuracy": 0.8673066639504978, | |
| "eval_loss": 0.36057987809181213, | |
| "eval_runtime": 30.2539, | |
| "eval_samples_per_second": 1298.049, | |
| "eval_steps_per_second": 5.09, | |
| "step": 5524 | |
| }, | |
| { | |
| "epoch": 4.34, | |
| "learning_rate": 6.016485632848009e-06, | |
| "loss": 0.2734, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 4.71, | |
| "learning_rate": 5.631307295277714e-06, | |
| "loss": 0.2724, | |
| "step": 6500 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_accuracy": 0.8679178019403632, | |
| "eval_loss": 0.3738739490509033, | |
| "eval_runtime": 30.6701, | |
| "eval_samples_per_second": 1280.434, | |
| "eval_steps_per_second": 5.021, | |
| "step": 6905 | |
| }, | |
| { | |
| "epoch": 5.07, | |
| "learning_rate": 5.2461289577074194e-06, | |
| "loss": 0.2667, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 5.43, | |
| "learning_rate": 4.860950620137123e-06, | |
| "loss": 0.2466, | |
| "step": 7500 | |
| }, | |
| { | |
| "epoch": 5.79, | |
| "learning_rate": 4.475772282566829e-06, | |
| "loss": 0.2481, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_accuracy": 0.86832522726694, | |
| "eval_loss": 0.3865111470222473, | |
| "eval_runtime": 30.6524, | |
| "eval_samples_per_second": 1281.174, | |
| "eval_steps_per_second": 5.024, | |
| "step": 8286 | |
| }, | |
| { | |
| "epoch": 6.15, | |
| "learning_rate": 4.090593944996534e-06, | |
| "loss": 0.2394, | |
| "step": 8500 | |
| }, | |
| { | |
| "epoch": 6.52, | |
| "learning_rate": 3.705415607426239e-06, | |
| "loss": 0.2269, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 6.88, | |
| "learning_rate": 3.3202372698559437e-06, | |
| "loss": 0.2264, | |
| "step": 9500 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "eval_accuracy": 0.8672557357846757, | |
| "eval_loss": 0.3937914967536926, | |
| "eval_runtime": 30.6252, | |
| "eval_samples_per_second": 1282.311, | |
| "eval_steps_per_second": 5.029, | |
| "step": 9667 | |
| }, | |
| { | |
| "epoch": 7.24, | |
| "learning_rate": 2.935058932285649e-06, | |
| "loss": 0.2181, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 7.6, | |
| "learning_rate": 2.5498805947153533e-06, | |
| "loss": 0.2134, | |
| "step": 10500 | |
| }, | |
| { | |
| "epoch": 7.97, | |
| "learning_rate": 2.164702257145058e-06, | |
| "loss": 0.2145, | |
| "step": 11000 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_accuracy": 0.8681469786865627, | |
| "eval_loss": 0.4055633544921875, | |
| "eval_runtime": 30.7036, | |
| "eval_samples_per_second": 1279.035, | |
| "eval_steps_per_second": 5.016, | |
| "step": 11048 | |
| }, | |
| { | |
| "epoch": 8.33, | |
| "learning_rate": 1.7795239195747632e-06, | |
| "loss": 0.2017, | |
| "step": 11500 | |
| }, | |
| { | |
| "epoch": 8.69, | |
| "learning_rate": 1.3943455820044682e-06, | |
| "loss": 0.2008, | |
| "step": 12000 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "eval_accuracy": 0.8672811998675868, | |
| "eval_loss": 0.41524767875671387, | |
| "eval_runtime": 30.8769, | |
| "eval_samples_per_second": 1271.856, | |
| "eval_steps_per_second": 4.988, | |
| "step": 12429 | |
| }, | |
| { | |
| "epoch": 9.05, | |
| "learning_rate": 1.0091672444341732e-06, | |
| "loss": 0.2013, | |
| "step": 12500 | |
| }, | |
| { | |
| "epoch": 9.41, | |
| "learning_rate": 6.239889068638781e-07, | |
| "loss": 0.1909, | |
| "step": 13000 | |
| }, | |
| { | |
| "epoch": 9.78, | |
| "learning_rate": 2.3881056929358295e-07, | |
| "loss": 0.1936, | |
| "step": 13500 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "eval_accuracy": 0.8680960505207405, | |
| "eval_loss": 0.42063575983047485, | |
| "eval_runtime": 30.2907, | |
| "eval_samples_per_second": 1296.472, | |
| "eval_steps_per_second": 5.084, | |
| "step": 13810 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "step": 13810, | |
| "total_flos": 2.1610790039821677e+17, | |
| "train_loss": 0.29975197396357794, | |
| "train_runtime": 5291.5699, | |
| "train_samples_per_second": 667.913, | |
| "train_steps_per_second": 2.61 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 13810, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 10, | |
| "save_steps": 500, | |
| "total_flos": 2.1610790039821677e+17, | |
| "train_batch_size": 256, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |