| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 0.0005850775959161584, | |
| "eval_steps": 8, | |
| "global_step": 30, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 1.9502586530538612e-05, | |
| "eval_loss": 3.45263934135437, | |
| "eval_runtime": 2111.6522, | |
| "eval_samples_per_second": 10.224, | |
| "eval_steps_per_second": 5.112, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 5.850775959161584e-05, | |
| "grad_norm": 28.103641510009766, | |
| "learning_rate": 6e-05, | |
| "loss": 15.4458, | |
| "step": 3 | |
| }, | |
| { | |
| "epoch": 0.00011701551918323167, | |
| "grad_norm": 16.842132568359375, | |
| "learning_rate": 0.00012, | |
| "loss": 12.578, | |
| "step": 6 | |
| }, | |
| { | |
| "epoch": 0.0001560206922443089, | |
| "eval_loss": 2.9562416076660156, | |
| "eval_runtime": 2118.5998, | |
| "eval_samples_per_second": 10.191, | |
| "eval_steps_per_second": 5.095, | |
| "step": 8 | |
| }, | |
| { | |
| "epoch": 0.00017552327877484752, | |
| "grad_norm": 12.545882225036621, | |
| "learning_rate": 0.00018, | |
| "loss": 11.6659, | |
| "step": 9 | |
| }, | |
| { | |
| "epoch": 0.00023403103836646335, | |
| "grad_norm": 17.450605392456055, | |
| "learning_rate": 0.00019510565162951537, | |
| "loss": 11.7232, | |
| "step": 12 | |
| }, | |
| { | |
| "epoch": 0.0002925387979580792, | |
| "grad_norm": 14.034260749816895, | |
| "learning_rate": 0.00017071067811865476, | |
| "loss": 11.5954, | |
| "step": 15 | |
| }, | |
| { | |
| "epoch": 0.0003120413844886178, | |
| "eval_loss": 2.7472758293151855, | |
| "eval_runtime": 2118.2647, | |
| "eval_samples_per_second": 10.192, | |
| "eval_steps_per_second": 5.096, | |
| "step": 16 | |
| }, | |
| { | |
| "epoch": 0.00035104655754969504, | |
| "grad_norm": 14.112640380859375, | |
| "learning_rate": 0.00013090169943749476, | |
| "loss": 10.7891, | |
| "step": 18 | |
| }, | |
| { | |
| "epoch": 0.00040955431714131084, | |
| "grad_norm": 16.250656127929688, | |
| "learning_rate": 8.435655349597689e-05, | |
| "loss": 10.661, | |
| "step": 21 | |
| }, | |
| { | |
| "epoch": 0.0004680620767329267, | |
| "grad_norm": 14.20734691619873, | |
| "learning_rate": 4.12214747707527e-05, | |
| "loss": 11.2675, | |
| "step": 24 | |
| }, | |
| { | |
| "epoch": 0.0004680620767329267, | |
| "eval_loss": 2.685302257537842, | |
| "eval_runtime": 2119.1554, | |
| "eval_samples_per_second": 10.188, | |
| "eval_steps_per_second": 5.094, | |
| "step": 24 | |
| }, | |
| { | |
| "epoch": 0.0005265698363245426, | |
| "grad_norm": 16.34672737121582, | |
| "learning_rate": 1.0899347581163221e-05, | |
| "loss": 10.9929, | |
| "step": 27 | |
| }, | |
| { | |
| "epoch": 0.0005850775959161584, | |
| "grad_norm": 13.728306770324707, | |
| "learning_rate": 0.0, | |
| "loss": 11.1346, | |
| "step": 30 | |
| } | |
| ], | |
| "logging_steps": 3, | |
| "max_steps": 30, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "save_steps": 10, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": true | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 1.054694248022016e+16, | |
| "train_batch_size": 2, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |