| { | |
| "best_metric": 0.9880821108818054, | |
| "best_model_checkpoint": "th_cl_23epochs_lora_pos_neg/checkpoint-96", | |
| "epoch": 5.0, | |
| "eval_steps": 500, | |
| "global_step": 160, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.3125, | |
| "grad_norm": 16.722179412841797, | |
| "learning_rate": 9.375e-05, | |
| "loss": 0.2879, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.625, | |
| "grad_norm": 26.74497413635254, | |
| "learning_rate": 8.75e-05, | |
| "loss": 0.4505, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.9375, | |
| "grad_norm": 17.404348373413086, | |
| "learning_rate": 8.125000000000001e-05, | |
| "loss": 0.5546, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_accuracy": 0.4888888888888889, | |
| "eval_balanced_accuracy": 0.4910714285714286, | |
| "eval_loss": 1.0123536586761475, | |
| "eval_runtime": 86.3441, | |
| "eval_samples_per_second": 0.521, | |
| "eval_steps_per_second": 0.069, | |
| "step": 32 | |
| }, | |
| { | |
| "epoch": 1.25, | |
| "grad_norm": 25.232460021972656, | |
| "learning_rate": 7.500000000000001e-05, | |
| "loss": 0.2502, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 1.5625, | |
| "grad_norm": 24.478240966796875, | |
| "learning_rate": 6.875e-05, | |
| "loss": 0.2332, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 1.875, | |
| "grad_norm": 24.21991539001465, | |
| "learning_rate": 6.25e-05, | |
| "loss": 0.3157, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_accuracy": 0.4666666666666667, | |
| "eval_balanced_accuracy": 0.4722222222222222, | |
| "eval_loss": 1.0065922737121582, | |
| "eval_runtime": 86.4764, | |
| "eval_samples_per_second": 0.52, | |
| "eval_steps_per_second": 0.069, | |
| "step": 64 | |
| }, | |
| { | |
| "epoch": 2.1875, | |
| "grad_norm": 8.731344223022461, | |
| "learning_rate": 5.6250000000000005e-05, | |
| "loss": 0.207, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 2.5, | |
| "grad_norm": 15.17884349822998, | |
| "learning_rate": 5e-05, | |
| "loss": 0.1905, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 2.8125, | |
| "grad_norm": 8.238539695739746, | |
| "learning_rate": 4.375e-05, | |
| "loss": 0.1355, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_accuracy": 0.6, | |
| "eval_balanced_accuracy": 0.5976890756302521, | |
| "eval_loss": 0.9880821108818054, | |
| "eval_runtime": 86.3822, | |
| "eval_samples_per_second": 0.521, | |
| "eval_steps_per_second": 0.069, | |
| "step": 96 | |
| }, | |
| { | |
| "epoch": 3.125, | |
| "grad_norm": 2.8414692878723145, | |
| "learning_rate": 3.7500000000000003e-05, | |
| "loss": 0.1228, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 3.4375, | |
| "grad_norm": 24.006149291992188, | |
| "learning_rate": 3.125e-05, | |
| "loss": 0.097, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 3.75, | |
| "grad_norm": 7.5499267578125, | |
| "learning_rate": 2.5e-05, | |
| "loss": 0.0512, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_accuracy": 0.5555555555555556, | |
| "eval_balanced_accuracy": 0.5535714285714286, | |
| "eval_loss": 1.1006296873092651, | |
| "eval_runtime": 86.2683, | |
| "eval_samples_per_second": 0.522, | |
| "eval_steps_per_second": 0.07, | |
| "step": 128 | |
| }, | |
| { | |
| "epoch": 4.0625, | |
| "grad_norm": 1.3797688484191895, | |
| "learning_rate": 1.8750000000000002e-05, | |
| "loss": 0.0577, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 4.375, | |
| "grad_norm": 1.6198076009750366, | |
| "learning_rate": 1.25e-05, | |
| "loss": 0.0232, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 4.6875, | |
| "grad_norm": 6.544897556304932, | |
| "learning_rate": 6.25e-06, | |
| "loss": 0.0334, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "grad_norm": 14.765763282775879, | |
| "learning_rate": 0.0, | |
| "loss": 0.0605, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_accuracy": 0.4666666666666667, | |
| "eval_balanced_accuracy": 0.47, | |
| "eval_loss": 1.1120657920837402, | |
| "eval_runtime": 86.5412, | |
| "eval_samples_per_second": 0.52, | |
| "eval_steps_per_second": 0.069, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "step": 160, | |
| "total_flos": 2.696144222158848e+16, | |
| "train_loss": 0.1919396677054465, | |
| "train_runtime": 8173.0352, | |
| "train_samples_per_second": 0.154, | |
| "train_steps_per_second": 0.02 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 160, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 5, | |
| "save_steps": 500, | |
| "total_flos": 2.696144222158848e+16, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |