| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 0.9948186528497409, | |
| "eval_steps": 500, | |
| "global_step": 96, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.01, | |
| "grad_norm": 0.407313588435244, | |
| "learning_rate": 2e-05, | |
| "loss": 1.4393, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.05, | |
| "grad_norm": 0.44021568832651153, | |
| "learning_rate": 0.0001, | |
| "loss": 1.4151, | |
| "step": 5 | |
| }, | |
| { | |
| "epoch": 0.1, | |
| "grad_norm": 0.4536750312789675, | |
| "learning_rate": 0.0002, | |
| "loss": 1.3352, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.16, | |
| "grad_norm": 0.18522669739061845, | |
| "learning_rate": 0.00019833656768294662, | |
| "loss": 1.1436, | |
| "step": 15 | |
| }, | |
| { | |
| "epoch": 0.21, | |
| "grad_norm": 0.11940000702730413, | |
| "learning_rate": 0.0001934016108732548, | |
| "loss": 1.0861, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.26, | |
| "grad_norm": 0.11067706553060182, | |
| "learning_rate": 0.00018535930890373466, | |
| "loss": 1.0546, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 0.31, | |
| "grad_norm": 0.08326696985890841, | |
| "learning_rate": 0.0001744772182743782, | |
| "loss": 1.0452, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.36, | |
| "grad_norm": 0.07551662936007832, | |
| "learning_rate": 0.00016111737140978494, | |
| "loss": 1.0374, | |
| "step": 35 | |
| }, | |
| { | |
| "epoch": 0.41, | |
| "grad_norm": 0.06940303607219772, | |
| "learning_rate": 0.00014572423233046386, | |
| "loss": 1.0203, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.47, | |
| "grad_norm": 0.0646380507229549, | |
| "learning_rate": 0.00012880990993652377, | |
| "loss": 1.0025, | |
| "step": 45 | |
| }, | |
| { | |
| "epoch": 0.52, | |
| "grad_norm": 0.06370005100951007, | |
| "learning_rate": 0.00011093712083778746, | |
| "loss": 1.0186, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.57, | |
| "grad_norm": 0.06034880535670624, | |
| "learning_rate": 9.270046853390925e-05, | |
| "loss": 1.0299, | |
| "step": 55 | |
| }, | |
| { | |
| "epoch": 0.62, | |
| "grad_norm": 0.06311665876872108, | |
| "learning_rate": 7.470666176083192e-05, | |
| "loss": 1.0156, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.67, | |
| "grad_norm": 0.06131237795687744, | |
| "learning_rate": 5.755433011241851e-05, | |
| "loss": 1.0074, | |
| "step": 65 | |
| }, | |
| { | |
| "epoch": 0.73, | |
| "grad_norm": 0.06522472527006118, | |
| "learning_rate": 4.181410844420474e-05, | |
| "loss": 1.0452, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.78, | |
| "grad_norm": 0.06426335241789467, | |
| "learning_rate": 2.800965262420043e-05, | |
| "loss": 1.0238, | |
| "step": 75 | |
| }, | |
| { | |
| "epoch": 0.83, | |
| "grad_norm": 0.06843282189243624, | |
| "learning_rate": 1.660021821101222e-05, | |
| "loss": 1.0167, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.88, | |
| "grad_norm": 0.06475830798907048, | |
| "learning_rate": 7.96538164308407e-06, | |
| "loss": 1.0045, | |
| "step": 85 | |
| }, | |
| { | |
| "epoch": 0.93, | |
| "grad_norm": 0.06274448627075681, | |
| "learning_rate": 2.392412244407294e-06, | |
| "loss": 1.0133, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.98, | |
| "grad_norm": 0.06343293325060437, | |
| "learning_rate": 6.671516297606095e-08, | |
| "loss": 1.0275, | |
| "step": 95 | |
| }, | |
| { | |
| "epoch": 0.99, | |
| "eval_loss": NaN, | |
| "eval_runtime": 238.4177, | |
| "eval_samples_per_second": 9.689, | |
| "eval_steps_per_second": 2.424, | |
| "step": 96 | |
| }, | |
| { | |
| "epoch": 0.99, | |
| "step": 96, | |
| "total_flos": 3308484214915072.0, | |
| "train_loss": 1.0698106847703457, | |
| "train_runtime": 2647.5008, | |
| "train_samples_per_second": 8.156, | |
| "train_steps_per_second": 0.036 | |
| } | |
| ], | |
| "logging_steps": 5, | |
| "max_steps": 96, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "save_steps": 100, | |
| "total_flos": 3308484214915072.0, | |
| "train_batch_size": 14, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |