| { | |
| "best_metric": 5.7574, | |
| "best_model_checkpoint": "./koja_reverse_13p_run1/checkpoint-3200", | |
| "epoch": 2.9083354972734354, | |
| "eval_steps": 800, | |
| "global_step": 5600, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.26, | |
| "learning_rate": 4.9212598425196856e-05, | |
| "loss": 1.9693, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.42, | |
| "eval_bleu": 4.3096, | |
| "eval_gen_len": 20.2019, | |
| "eval_loss": 1.544534683227539, | |
| "eval_runtime": 598.1788, | |
| "eval_samples_per_second": 12.874, | |
| "eval_steps_per_second": 0.806, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 0.52, | |
| "learning_rate": 4.79002624671916e-05, | |
| "loss": 1.6018, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.78, | |
| "learning_rate": 4.6587926509186354e-05, | |
| "loss": 1.4914, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 0.83, | |
| "eval_bleu": 4.307, | |
| "eval_gen_len": 20.1204, | |
| "eval_loss": 1.4232898950576782, | |
| "eval_runtime": 606.6975, | |
| "eval_samples_per_second": 12.693, | |
| "eval_steps_per_second": 0.794, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 1.04, | |
| "learning_rate": 4.52755905511811e-05, | |
| "loss": 1.4272, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 1.25, | |
| "eval_bleu": 4.8111, | |
| "eval_gen_len": 22.8555, | |
| "eval_loss": 1.3761498928070068, | |
| "eval_runtime": 802.0063, | |
| "eval_samples_per_second": 9.602, | |
| "eval_steps_per_second": 0.601, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 1.3, | |
| "learning_rate": 4.396325459317586e-05, | |
| "loss": 1.2227, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 1.56, | |
| "learning_rate": 4.2650918635170604e-05, | |
| "loss": 1.1571, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 1.66, | |
| "eval_bleu": 5.7574, | |
| "eval_gen_len": 21.151, | |
| "eval_loss": 1.3532484769821167, | |
| "eval_runtime": 674.5674, | |
| "eval_samples_per_second": 11.416, | |
| "eval_steps_per_second": 0.715, | |
| "step": 3200 | |
| }, | |
| { | |
| "epoch": 1.82, | |
| "learning_rate": 4.133858267716536e-05, | |
| "loss": 1.1144, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 2.08, | |
| "learning_rate": 4.00262467191601e-05, | |
| "loss": 1.0833, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 2.08, | |
| "eval_bleu": 5.652, | |
| "eval_gen_len": 21.2158, | |
| "eval_loss": 1.3448729515075684, | |
| "eval_runtime": 674.0234, | |
| "eval_samples_per_second": 11.425, | |
| "eval_steps_per_second": 0.715, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 2.34, | |
| "learning_rate": 3.871391076115486e-05, | |
| "loss": 0.9394, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 2.49, | |
| "eval_bleu": 5.45, | |
| "eval_gen_len": 19.9005, | |
| "eval_loss": 1.3905117511749268, | |
| "eval_runtime": 563.5382, | |
| "eval_samples_per_second": 13.665, | |
| "eval_steps_per_second": 0.855, | |
| "step": 4800 | |
| }, | |
| { | |
| "epoch": 2.6, | |
| "learning_rate": 3.740157480314961e-05, | |
| "loss": 0.8889, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 2.86, | |
| "learning_rate": 3.608923884514436e-05, | |
| "loss": 0.8633, | |
| "step": 5500 | |
| }, | |
| { | |
| "epoch": 2.91, | |
| "eval_bleu": 4.9973, | |
| "eval_gen_len": 19.7844, | |
| "eval_loss": 1.3719298839569092, | |
| "eval_runtime": 568.6725, | |
| "eval_samples_per_second": 13.542, | |
| "eval_steps_per_second": 0.848, | |
| "step": 5600 | |
| }, | |
| { | |
| "epoch": 2.91, | |
| "step": 5600, | |
| "total_flos": 3.883496544141312e+17, | |
| "train_loss": 1.2439973626817975, | |
| "train_runtime": 14074.9837, | |
| "train_samples_per_second": 43.776, | |
| "train_steps_per_second": 1.368 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 19250, | |
| "num_train_epochs": 10, | |
| "save_steps": 800, | |
| "total_flos": 3.883496544141312e+17, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |