| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 8.04289544235925, | |
| "eval_steps": 500, | |
| "global_step": 6000, | |
| "is_hyper_param_search": true, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.6702412868632708, | |
| "grad_norm": 2.054403066635132, | |
| "learning_rate": 6.75599008099931e-05, | |
| "loss": 0.6856, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_accuracy": 0.4924572706222534, | |
| "eval_loss": 0.6931472420692444, | |
| "eval_runtime": 8.9909, | |
| "eval_samples_per_second": 331.779, | |
| "eval_steps_per_second": 20.799, | |
| "step": 746 | |
| }, | |
| { | |
| "epoch": 1.3404825737265416, | |
| "grad_norm": 1.88927161693573, | |
| "learning_rate": 6.270645965984992e-05, | |
| "loss": 0.7049, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_accuracy": 0.5129064917564392, | |
| "eval_loss": 0.6931472420692444, | |
| "eval_runtime": 9.0209, | |
| "eval_samples_per_second": 330.678, | |
| "eval_steps_per_second": 20.73, | |
| "step": 1492 | |
| }, | |
| { | |
| "epoch": 2.0107238605898123, | |
| "grad_norm": 1.888950228691101, | |
| "learning_rate": 5.7853018509706724e-05, | |
| "loss": 0.7041, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 2.680965147453083, | |
| "grad_norm": 1.8515383005142212, | |
| "learning_rate": 5.299957735956355e-05, | |
| "loss": 0.7017, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_accuracy": 0.5085484385490417, | |
| "eval_loss": 0.6931472420692444, | |
| "eval_runtime": 9.0308, | |
| "eval_samples_per_second": 330.315, | |
| "eval_steps_per_second": 20.707, | |
| "step": 2238 | |
| }, | |
| { | |
| "epoch": 3.351206434316354, | |
| "grad_norm": 1.8233052492141724, | |
| "learning_rate": 4.8146136209420366e-05, | |
| "loss": 0.7036, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_accuracy": 0.49782097339630127, | |
| "eval_loss": 0.6931472420692444, | |
| "eval_runtime": 9.0208, | |
| "eval_samples_per_second": 330.68, | |
| "eval_steps_per_second": 20.73, | |
| "step": 2984 | |
| }, | |
| { | |
| "epoch": 4.021447721179625, | |
| "grad_norm": 1.7395379543304443, | |
| "learning_rate": 4.329269505927718e-05, | |
| "loss": 0.7043, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 4.6916890080428955, | |
| "grad_norm": 1.853223204612732, | |
| "learning_rate": 3.8439253909134e-05, | |
| "loss": 0.7023, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_accuracy": 0.4968152940273285, | |
| "eval_loss": 0.6931472420692444, | |
| "eval_runtime": 9.0187, | |
| "eval_samples_per_second": 330.757, | |
| "eval_steps_per_second": 20.735, | |
| "step": 3730 | |
| }, | |
| { | |
| "epoch": 5.361930294906166, | |
| "grad_norm": 1.8915830850601196, | |
| "learning_rate": 3.358581275899082e-05, | |
| "loss": 0.6996, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_accuracy": 0.5125712156295776, | |
| "eval_loss": 0.6931472420692444, | |
| "eval_runtime": 8.9826, | |
| "eval_samples_per_second": 332.088, | |
| "eval_steps_per_second": 20.818, | |
| "step": 4476 | |
| }, | |
| { | |
| "epoch": 6.032171581769437, | |
| "grad_norm": 1.8517926931381226, | |
| "learning_rate": 2.8732371608847635e-05, | |
| "loss": 0.699, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 6.702412868632708, | |
| "grad_norm": 1.761203408241272, | |
| "learning_rate": 2.3878930458704453e-05, | |
| "loss": 0.6999, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "eval_accuracy": 0.5018438100814819, | |
| "eval_loss": 0.6931472420692444, | |
| "eval_runtime": 8.9692, | |
| "eval_samples_per_second": 332.582, | |
| "eval_steps_per_second": 20.849, | |
| "step": 5222 | |
| }, | |
| { | |
| "epoch": 7.372654155495979, | |
| "grad_norm": 1.7442079782485962, | |
| "learning_rate": 1.9025489308561274e-05, | |
| "loss": 0.7017, | |
| "step": 5500 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_accuracy": 0.5021790266036987, | |
| "eval_loss": 0.6931472420692444, | |
| "eval_runtime": 9.1231, | |
| "eval_samples_per_second": 326.971, | |
| "eval_steps_per_second": 20.497, | |
| "step": 5968 | |
| }, | |
| { | |
| "epoch": 8.04289544235925, | |
| "grad_norm": 2.0343399047851562, | |
| "learning_rate": 1.417204815841809e-05, | |
| "loss": 0.7009, | |
| "step": 6000 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 7460, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 10, | |
| "save_steps": 500, | |
| "total_flos": 1.563011071614504e+16, | |
| "train_batch_size": 16, | |
| "trial_name": null, | |
| "trial_params": { | |
| "learning_rate": 7.241334196013627e-05, | |
| "per_device_train_batch_size": 16 | |
| } | |
| } | |