| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 10.0, | |
| "eval_steps": 500, | |
| "global_step": 70, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.14285714285714285, | |
| "grad_norm": 2.171875, | |
| "learning_rate": 2.857142857142857e-05, | |
| "loss": 3.0133, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.7142857142857143, | |
| "grad_norm": 1.96875, | |
| "learning_rate": 0.00014285714285714287, | |
| "loss": 2.9598, | |
| "step": 5 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_loss": 3.0818610191345215, | |
| "eval_runtime": 0.5063, | |
| "eval_samples_per_second": 19.752, | |
| "eval_steps_per_second": 1.975, | |
| "step": 7 | |
| }, | |
| { | |
| "epoch": 1.4285714285714286, | |
| "grad_norm": 1.4765625, | |
| "learning_rate": 0.00019888308262251285, | |
| "loss": 2.584, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_loss": 2.779895067214966, | |
| "eval_runtime": 0.53, | |
| "eval_samples_per_second": 18.868, | |
| "eval_steps_per_second": 1.887, | |
| "step": 14 | |
| }, | |
| { | |
| "epoch": 2.142857142857143, | |
| "grad_norm": 1.2578125, | |
| "learning_rate": 0.00019214762118704076, | |
| "loss": 2.1727, | |
| "step": 15 | |
| }, | |
| { | |
| "epoch": 2.857142857142857, | |
| "grad_norm": 0.9609375, | |
| "learning_rate": 0.00017971325072229226, | |
| "loss": 1.8743, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_loss": 2.6390492916107178, | |
| "eval_runtime": 0.5412, | |
| "eval_samples_per_second": 18.478, | |
| "eval_steps_per_second": 1.848, | |
| "step": 21 | |
| }, | |
| { | |
| "epoch": 3.571428571428571, | |
| "grad_norm": 0.609375, | |
| "learning_rate": 0.00016234898018587337, | |
| "loss": 1.6598, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_loss": 2.6230721473693848, | |
| "eval_runtime": 0.5473, | |
| "eval_samples_per_second": 18.272, | |
| "eval_steps_per_second": 1.827, | |
| "step": 28 | |
| }, | |
| { | |
| "epoch": 4.285714285714286, | |
| "grad_norm": 0.455078125, | |
| "learning_rate": 0.00014112871031306119, | |
| "loss": 1.5423, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "grad_norm": 0.5703125, | |
| "learning_rate": 0.00011736481776669306, | |
| "loss": 1.4458, | |
| "step": 35 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_loss": 2.5861191749572754, | |
| "eval_runtime": 0.5498, | |
| "eval_samples_per_second": 18.188, | |
| "eval_steps_per_second": 1.819, | |
| "step": 35 | |
| }, | |
| { | |
| "epoch": 5.714285714285714, | |
| "grad_norm": 0.404296875, | |
| "learning_rate": 9.252699064135758e-05, | |
| "loss": 1.3889, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_loss": 2.5646462440490723, | |
| "eval_runtime": 0.5474, | |
| "eval_samples_per_second": 18.269, | |
| "eval_steps_per_second": 1.827, | |
| "step": 42 | |
| }, | |
| { | |
| "epoch": 6.428571428571429, | |
| "grad_norm": 0.34765625, | |
| "learning_rate": 6.815133497483157e-05, | |
| "loss": 1.3527, | |
| "step": 45 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "eval_loss": 2.556933879852295, | |
| "eval_runtime": 0.5441, | |
| "eval_samples_per_second": 18.378, | |
| "eval_steps_per_second": 1.838, | |
| "step": 49 | |
| }, | |
| { | |
| "epoch": 7.142857142857143, | |
| "grad_norm": 0.345703125, | |
| "learning_rate": 4.574537361342407e-05, | |
| "loss": 1.3118, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 7.857142857142857, | |
| "grad_norm": 0.44921875, | |
| "learning_rate": 2.669481281701739e-05, | |
| "loss": 1.2936, | |
| "step": 55 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_loss": 2.5531744956970215, | |
| "eval_runtime": 0.5391, | |
| "eval_samples_per_second": 18.55, | |
| "eval_steps_per_second": 1.855, | |
| "step": 56 | |
| }, | |
| { | |
| "epoch": 8.571428571428571, | |
| "grad_norm": 0.32421875, | |
| "learning_rate": 1.2177842662977135e-05, | |
| "loss": 1.2855, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "eval_loss": 2.551410436630249, | |
| "eval_runtime": 0.5405, | |
| "eval_samples_per_second": 18.501, | |
| "eval_steps_per_second": 1.85, | |
| "step": 63 | |
| }, | |
| { | |
| "epoch": 9.285714285714286, | |
| "grad_norm": 0.244140625, | |
| "learning_rate": 3.092271377092215e-06, | |
| "loss": 1.2767, | |
| "step": 65 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "grad_norm": 0.2431640625, | |
| "learning_rate": 0.0, | |
| "loss": 1.2802, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "eval_loss": 2.5520482063293457, | |
| "eval_runtime": 0.5385, | |
| "eval_samples_per_second": 18.569, | |
| "eval_steps_per_second": 1.857, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "step": 70, | |
| "total_flos": 5.48215128588288e+16, | |
| "train_loss": 1.6741911002567835, | |
| "train_runtime": 247.913, | |
| "train_samples_per_second": 17.062, | |
| "train_steps_per_second": 0.282 | |
| } | |
| ], | |
| "logging_steps": 5, | |
| "max_steps": 70, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 10, | |
| "save_steps": 100, | |
| "total_flos": 5.48215128588288e+16, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |