| { | |
| "best_metric": 0.4671708345413208, | |
| "best_model_checkpoint": "./Vit-GPT2-COCO2017Flickr-80k-08/checkpoint-3000", | |
| "epoch": 0.6, | |
| "eval_steps": 500, | |
| "global_step": 3000, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.1, | |
| "grad_norm": 0.9981909990310669, | |
| "learning_rate": 4.8333333333333334e-05, | |
| "loss": 0.3691, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.1, | |
| "eval_gen_len": 11.77575, | |
| "eval_loss": 0.4729623794555664, | |
| "eval_rouge1": 39.8086, | |
| "eval_rouge2": 14.7674, | |
| "eval_rougeL": 36.1546, | |
| "eval_rougeLsum": 36.1739, | |
| "eval_runtime": 459.5392, | |
| "eval_samples_per_second": 8.704, | |
| "eval_steps_per_second": 2.176, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.2, | |
| "grad_norm": 0.8220647573471069, | |
| "learning_rate": 4.666666666666667e-05, | |
| "loss": 0.3706, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.2, | |
| "eval_gen_len": 11.59775, | |
| "eval_loss": 0.47387805581092834, | |
| "eval_rouge1": 39.8972, | |
| "eval_rouge2": 14.9064, | |
| "eval_rougeL": 36.1193, | |
| "eval_rougeLsum": 36.138, | |
| "eval_runtime": 446.0956, | |
| "eval_samples_per_second": 8.967, | |
| "eval_steps_per_second": 2.242, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.3, | |
| "grad_norm": 1.052058219909668, | |
| "learning_rate": 4.5e-05, | |
| "loss": 0.3709, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 0.3, | |
| "eval_gen_len": 11.71025, | |
| "eval_loss": 0.47592678666114807, | |
| "eval_rouge1": 39.9874, | |
| "eval_rouge2": 14.8528, | |
| "eval_rougeL": 36.3155, | |
| "eval_rougeLsum": 36.3317, | |
| "eval_runtime": 466.5502, | |
| "eval_samples_per_second": 8.574, | |
| "eval_steps_per_second": 2.143, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 0.4, | |
| "grad_norm": 0.8642559051513672, | |
| "learning_rate": 4.3333333333333334e-05, | |
| "loss": 0.3721, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 0.4, | |
| "eval_gen_len": 12.175, | |
| "eval_loss": 0.4677760601043701, | |
| "eval_rouge1": 39.7192, | |
| "eval_rouge2": 14.5844, | |
| "eval_rougeL": 35.8447, | |
| "eval_rougeLsum": 35.8728, | |
| "eval_runtime": 455.6895, | |
| "eval_samples_per_second": 8.778, | |
| "eval_steps_per_second": 2.194, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 0.5, | |
| "grad_norm": 0.8769400715827942, | |
| "learning_rate": 4.166666666666667e-05, | |
| "loss": 0.3655, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 0.5, | |
| "eval_gen_len": 11.90025, | |
| "eval_loss": 0.4683643877506256, | |
| "eval_rouge1": 40.3132, | |
| "eval_rouge2": 15.1157, | |
| "eval_rougeL": 36.5749, | |
| "eval_rougeLsum": 36.5823, | |
| "eval_runtime": 449.7304, | |
| "eval_samples_per_second": 8.894, | |
| "eval_steps_per_second": 2.224, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 0.6, | |
| "grad_norm": 0.8613843321800232, | |
| "learning_rate": 4e-05, | |
| "loss": 0.3623, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 0.6, | |
| "eval_gen_len": 12.025, | |
| "eval_loss": 0.4671708345413208, | |
| "eval_rouge1": 40.1643, | |
| "eval_rouge2": 14.978, | |
| "eval_rougeL": 36.3002, | |
| "eval_rougeLsum": 36.3232, | |
| "eval_runtime": 451.1814, | |
| "eval_samples_per_second": 8.866, | |
| "eval_steps_per_second": 2.216, | |
| "step": 3000 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 15000, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 3, | |
| "save_steps": 500, | |
| "total_flos": 8.662266773766144e+18, | |
| "train_batch_size": 4, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |