| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 0.056100981767180924, | |
| "eval_steps": 5, | |
| "global_step": 30, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.0018700327255726976, | |
| "eval_loss": 9.52480411529541, | |
| "eval_runtime": 1.3214, | |
| "eval_samples_per_second": 171.036, | |
| "eval_steps_per_second": 85.518, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.005610098176718092, | |
| "grad_norm": 28202.841796875, | |
| "learning_rate": 4e-05, | |
| "loss": 36.7618, | |
| "step": 3 | |
| }, | |
| { | |
| "epoch": 0.009350163627863487, | |
| "eval_loss": 9.48367691040039, | |
| "eval_runtime": 1.3071, | |
| "eval_samples_per_second": 172.904, | |
| "eval_steps_per_second": 86.452, | |
| "step": 5 | |
| }, | |
| { | |
| "epoch": 0.011220196353436185, | |
| "grad_norm": 7682.89697265625, | |
| "learning_rate": 8e-05, | |
| "loss": 38.6061, | |
| "step": 6 | |
| }, | |
| { | |
| "epoch": 0.016830294530154277, | |
| "grad_norm": 5112.90234375, | |
| "learning_rate": 0.00012, | |
| "loss": 37.4866, | |
| "step": 9 | |
| }, | |
| { | |
| "epoch": 0.018700327255726974, | |
| "eval_loss": 9.333919525146484, | |
| "eval_runtime": 1.3034, | |
| "eval_samples_per_second": 173.397, | |
| "eval_steps_per_second": 86.698, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.02244039270687237, | |
| "grad_norm": 6471.3271484375, | |
| "learning_rate": 0.00016, | |
| "loss": 38.0092, | |
| "step": 12 | |
| }, | |
| { | |
| "epoch": 0.028050490883590462, | |
| "grad_norm": 9522.9501953125, | |
| "learning_rate": 0.0002, | |
| "loss": 37.7697, | |
| "step": 15 | |
| }, | |
| { | |
| "epoch": 0.028050490883590462, | |
| "eval_loss": 9.143020629882812, | |
| "eval_runtime": 1.3124, | |
| "eval_samples_per_second": 172.201, | |
| "eval_steps_per_second": 86.101, | |
| "step": 15 | |
| }, | |
| { | |
| "epoch": 0.033660589060308554, | |
| "grad_norm": 4783.828125, | |
| "learning_rate": 0.00018090169943749476, | |
| "loss": 35.8871, | |
| "step": 18 | |
| }, | |
| { | |
| "epoch": 0.03740065451145395, | |
| "eval_loss": 9.025113105773926, | |
| "eval_runtime": 1.3432, | |
| "eval_samples_per_second": 168.25, | |
| "eval_steps_per_second": 84.125, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.03927068723702665, | |
| "grad_norm": 6039.88037109375, | |
| "learning_rate": 0.00013090169943749476, | |
| "loss": 36.1402, | |
| "step": 21 | |
| }, | |
| { | |
| "epoch": 0.04488078541374474, | |
| "grad_norm": 3826.15283203125, | |
| "learning_rate": 6.909830056250527e-05, | |
| "loss": 35.2476, | |
| "step": 24 | |
| }, | |
| { | |
| "epoch": 0.04675081813931744, | |
| "eval_loss": 8.821419715881348, | |
| "eval_runtime": 1.3129, | |
| "eval_samples_per_second": 172.135, | |
| "eval_steps_per_second": 86.068, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 0.05049088359046283, | |
| "grad_norm": 3266.89892578125, | |
| "learning_rate": 1.9098300562505266e-05, | |
| "loss": 35.5308, | |
| "step": 27 | |
| }, | |
| { | |
| "epoch": 0.056100981767180924, | |
| "grad_norm": 6962.6591796875, | |
| "learning_rate": 0.0, | |
| "loss": 35.3495, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.056100981767180924, | |
| "eval_loss": 8.779716491699219, | |
| "eval_runtime": 1.3096, | |
| "eval_samples_per_second": 172.573, | |
| "eval_steps_per_second": 86.286, | |
| "step": 30 | |
| } | |
| ], | |
| "logging_steps": 3, | |
| "max_steps": 30, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "save_steps": 15, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": true | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 11394078474240.0, | |
| "train_batch_size": 2, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |