| { | |
| "best_metric": 0.8729119896888733, | |
| "best_model_checkpoint": "seq2seq_model_handwritten/checkpoint-1656", | |
| "epoch": 10.0, | |
| "eval_steps": 500, | |
| "global_step": 2070, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 1.0, | |
| "grad_norm": 91.65796661376953, | |
| "learning_rate": 4.512077294685991e-05, | |
| "loss": 2.3731, | |
| "step": 207 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_cer": 0.5382716049382716, | |
| "eval_loss": 1.7942910194396973, | |
| "eval_runtime": 7.573, | |
| "eval_samples_per_second": 3.961, | |
| "eval_steps_per_second": 1.056, | |
| "step": 207 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "grad_norm": 28.777713775634766, | |
| "learning_rate": 4.012077294685991e-05, | |
| "loss": 1.5318, | |
| "step": 414 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_cer": 1.9481481481481482, | |
| "eval_loss": 2.347219944000244, | |
| "eval_runtime": 10.932, | |
| "eval_samples_per_second": 2.744, | |
| "eval_steps_per_second": 0.732, | |
| "step": 414 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "grad_norm": 9.893929481506348, | |
| "learning_rate": 3.51207729468599e-05, | |
| "loss": 1.1272, | |
| "step": 621 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_cer": 0.3925925925925926, | |
| "eval_loss": 1.6664316654205322, | |
| "eval_runtime": 5.6669, | |
| "eval_samples_per_second": 5.294, | |
| "eval_steps_per_second": 1.412, | |
| "step": 621 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "grad_norm": 12.895403861999512, | |
| "learning_rate": 3.0120772946859905e-05, | |
| "loss": 0.8353, | |
| "step": 828 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_cer": 0.35555555555555557, | |
| "eval_loss": 1.744159460067749, | |
| "eval_runtime": 5.9067, | |
| "eval_samples_per_second": 5.079, | |
| "eval_steps_per_second": 1.354, | |
| "step": 828 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "grad_norm": 10.83328914642334, | |
| "learning_rate": 2.5120772946859905e-05, | |
| "loss": 0.6815, | |
| "step": 1035 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_cer": 0.2617283950617284, | |
| "eval_loss": 1.6649540662765503, | |
| "eval_runtime": 6.0195, | |
| "eval_samples_per_second": 4.984, | |
| "eval_steps_per_second": 1.329, | |
| "step": 1035 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "grad_norm": 39.549861907958984, | |
| "learning_rate": 2.0120772946859902e-05, | |
| "loss": 0.4102, | |
| "step": 1242 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_cer": 0.2419753086419753, | |
| "eval_loss": 1.126643180847168, | |
| "eval_runtime": 5.5303, | |
| "eval_samples_per_second": 5.425, | |
| "eval_steps_per_second": 1.447, | |
| "step": 1242 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "grad_norm": 33.302223205566406, | |
| "learning_rate": 1.5120772946859905e-05, | |
| "loss": 0.3097, | |
| "step": 1449 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "eval_cer": 0.2617283950617284, | |
| "eval_loss": 1.1228829622268677, | |
| "eval_runtime": 5.9786, | |
| "eval_samples_per_second": 5.018, | |
| "eval_steps_per_second": 1.338, | |
| "step": 1449 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "grad_norm": 0.6234323382377625, | |
| "learning_rate": 1.0120772946859903e-05, | |
| "loss": 0.2095, | |
| "step": 1656 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_cer": 0.2074074074074074, | |
| "eval_loss": 0.8729119896888733, | |
| "eval_runtime": 6.0046, | |
| "eval_samples_per_second": 4.996, | |
| "eval_steps_per_second": 1.332, | |
| "step": 1656 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "grad_norm": 3.2294392585754395, | |
| "learning_rate": 5.120772946859904e-06, | |
| "loss": 0.1344, | |
| "step": 1863 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "eval_cer": 0.1580246913580247, | |
| "eval_loss": 0.9647266268730164, | |
| "eval_runtime": 6.0304, | |
| "eval_samples_per_second": 4.975, | |
| "eval_steps_per_second": 1.327, | |
| "step": 1863 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "grad_norm": 0.18482227623462677, | |
| "learning_rate": 1.2077294685990338e-07, | |
| "loss": 0.0616, | |
| "step": 2070 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "eval_cer": 0.1308641975308642, | |
| "eval_loss": 0.9429413080215454, | |
| "eval_runtime": 6.0359, | |
| "eval_samples_per_second": 4.97, | |
| "eval_steps_per_second": 1.325, | |
| "step": 2070 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 2070, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 10, | |
| "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": 1.1125309068426609e+19, | |
| "train_batch_size": 4, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |