{ "best_metric": null, "best_model_checkpoint": null, "epoch": 59.67567567567568, "eval_steps": 276, "global_step": 2760, "is_hyper_param_search": false, "is_local_process_zero": true, "is_world_process_zero": true, "log_history": [ { "epoch": 5.967567567567568, "grad_norm": 2.231717109680176, "learning_rate": 9.007246376811595e-06, "loss": 1.1843, "step": 276 }, { "epoch": 5.967567567567568, "eval_accuracy": 0.3670169765561843, "eval_loss": 2.2262628078460693, "eval_runtime": 3.9032, "eval_samples_per_second": 316.921, "eval_steps_per_second": 15.884, "step": 276 }, { "epoch": 11.935135135135136, "grad_norm": 3.9312663078308105, "learning_rate": 8.007246376811595e-06, "loss": 0.9309, "step": 552 }, { "epoch": 11.935135135135136, "eval_accuracy": 0.3427647534357316, "eval_loss": 2.3462769985198975, "eval_runtime": 3.9045, "eval_samples_per_second": 316.817, "eval_steps_per_second": 15.879, "step": 552 }, { "epoch": 17.9027027027027, "grad_norm": 4.585145950317383, "learning_rate": 7.007246376811595e-06, "loss": 0.7822, "step": 828 }, { "epoch": 17.9027027027027, "eval_accuracy": 0.33441120991646456, "eval_loss": 2.4977803230285645, "eval_runtime": 3.829, "eval_samples_per_second": 323.058, "eval_steps_per_second": 16.192, "step": 828 }, { "epoch": 23.87027027027027, "grad_norm": 4.77891206741333, "learning_rate": 6.007246376811595e-06, "loss": 0.7077, "step": 1104 }, { "epoch": 23.87027027027027, "eval_accuracy": 0.332053354890865, "eval_loss": 2.5808465480804443, "eval_runtime": 3.9331, "eval_samples_per_second": 314.51, "eval_steps_per_second": 15.764, "step": 1104 }, { "epoch": 29.83783783783784, "grad_norm": 3.8401949405670166, "learning_rate": 5.0072463768115946e-06, "loss": 0.6579, "step": 1380 }, { "epoch": 29.83783783783784, "eval_accuracy": 0.32966855295068714, "eval_loss": 2.588106632232666, "eval_runtime": 3.8661, "eval_samples_per_second": 319.964, "eval_steps_per_second": 16.037, "step": 1380 }, { "epoch": 35.8054054054054, "grad_norm": 5.049342155456543, "learning_rate": 4.007246376811594e-06, "loss": 0.6308, "step": 1656 }, { "epoch": 35.8054054054054, "eval_accuracy": 0.3283481541363514, "eval_loss": 2.6445364952087402, "eval_runtime": 3.9013, "eval_samples_per_second": 317.073, "eval_steps_per_second": 15.892, "step": 1656 }, { "epoch": 41.77297297297297, "grad_norm": 4.533051013946533, "learning_rate": 3.0072463768115946e-06, "loss": 0.6117, "step": 1932 }, { "epoch": 41.77297297297297, "eval_accuracy": 0.3280979327866959, "eval_loss": 2.662590980529785, "eval_runtime": 3.9106, "eval_samples_per_second": 316.317, "eval_steps_per_second": 15.854, "step": 1932 }, { "epoch": 47.74054054054054, "grad_norm": 3.6275293827056885, "learning_rate": 2.0108695652173916e-06, "loss": 0.6045, "step": 2208 }, { "epoch": 47.74054054054054, "eval_accuracy": 0.32841552142279706, "eval_loss": 2.6373841762542725, "eval_runtime": 3.8804, "eval_samples_per_second": 318.779, "eval_steps_per_second": 15.978, "step": 2208 }, { "epoch": 53.70810810810811, "grad_norm": 5.026205539703369, "learning_rate": 1.0108695652173914e-06, "loss": 0.5889, "step": 2484 }, { "epoch": 53.70810810810811, "eval_accuracy": 0.32803377346627144, "eval_loss": 2.6702210903167725, "eval_runtime": 3.8688, "eval_samples_per_second": 319.736, "eval_steps_per_second": 16.026, "step": 2484 }, { "epoch": 59.67567567567568, "grad_norm": 3.773498773574829, "learning_rate": 1.0869565217391305e-08, "loss": 0.5908, "step": 2760 }, { "epoch": 59.67567567567568, "eval_accuracy": 0.3279708973322555, "eval_loss": 2.662411689758301, "eval_runtime": 3.9189, "eval_samples_per_second": 315.65, "eval_steps_per_second": 15.821, "step": 2760 }, { "epoch": 59.67567567567568, "step": 2760, "total_flos": 1.4462122119191654e+17, "train_loss": 0.7289675836977751, "train_runtime": 3803.6526, "train_samples_per_second": 145.628, "train_steps_per_second": 0.726 } ], "logging_steps": 276, "max_steps": 2760, "num_input_tokens_seen": 0, "num_train_epochs": 60, "save_steps": 500, "stateful_callbacks": { "TrainerControl": { "args": { "should_epoch_stop": false, "should_evaluate": false, "should_log": false, "should_save": true, "should_training_stop": true }, "attributes": {} } }, "total_flos": 1.4462122119191654e+17, "train_batch_size": 25, "trial_name": null, "trial_params": null }