| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 15.0, | |
| "eval_steps": 30, | |
| "global_step": 90, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.17, | |
| "grad_norm": 1600.5392304887378, | |
| "learning_rate": 1.999390827019096e-05, | |
| "loss": 10.0226, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.83, | |
| "grad_norm": 2032.6698503030166, | |
| "learning_rate": 1.9848077530122083e-05, | |
| "loss": 8.797, | |
| "step": 5 | |
| }, | |
| { | |
| "epoch": 1.67, | |
| "grad_norm": 213.98723674458682, | |
| "learning_rate": 1.9396926207859085e-05, | |
| "loss": 6.3385, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 2.5, | |
| "grad_norm": 704.9269625939797, | |
| "learning_rate": 1.866025403784439e-05, | |
| "loss": 6.4244, | |
| "step": 15 | |
| }, | |
| { | |
| "epoch": 3.33, | |
| "grad_norm": 415.53646633609685, | |
| "learning_rate": 1.766044443118978e-05, | |
| "loss": 4.6473, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 4.17, | |
| "grad_norm": 117.74034120947316, | |
| "learning_rate": 1.6427876096865394e-05, | |
| "loss": 3.1231, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "grad_norm": 35.05307192586602, | |
| "learning_rate": 1.5000000000000002e-05, | |
| "loss": 1.9689, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_loss": 1.464415192604065, | |
| "eval_runtime": 54.045, | |
| "eval_samples_per_second": 18.244, | |
| "eval_steps_per_second": 0.296, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 5.83, | |
| "grad_norm": 17.30445284838731, | |
| "learning_rate": 1.342020143325669e-05, | |
| "loss": 1.6624, | |
| "step": 35 | |
| }, | |
| { | |
| "epoch": 6.67, | |
| "grad_norm": 18.489148704869553, | |
| "learning_rate": 1.1736481776669307e-05, | |
| "loss": 1.4962, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 7.5, | |
| "grad_norm": 11.164738834605398, | |
| "learning_rate": 1e-05, | |
| "loss": 1.3667, | |
| "step": 45 | |
| }, | |
| { | |
| "epoch": 8.33, | |
| "grad_norm": 12.450510814489615, | |
| "learning_rate": 8.263518223330698e-06, | |
| "loss": 1.2545, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 9.17, | |
| "grad_norm": 7.10392485555226, | |
| "learning_rate": 6.579798566743314e-06, | |
| "loss": 1.2288, | |
| "step": 55 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "grad_norm": 7.210838208620658, | |
| "learning_rate": 5.000000000000003e-06, | |
| "loss": 1.1624, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "eval_loss": 1.0362504720687866, | |
| "eval_runtime": 46.757, | |
| "eval_samples_per_second": 21.088, | |
| "eval_steps_per_second": 0.342, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 10.83, | |
| "grad_norm": 5.108017518056073, | |
| "learning_rate": 3.5721239031346067e-06, | |
| "loss": 1.1292, | |
| "step": 65 | |
| }, | |
| { | |
| "epoch": 11.67, | |
| "grad_norm": 4.729793063810578, | |
| "learning_rate": 2.339555568810221e-06, | |
| "loss": 1.1012, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 12.5, | |
| "grad_norm": 4.227960417256771, | |
| "learning_rate": 1.339745962155613e-06, | |
| "loss": 1.0784, | |
| "step": 75 | |
| }, | |
| { | |
| "epoch": 13.33, | |
| "grad_norm": 5.236888613480014, | |
| "learning_rate": 6.030737921409169e-07, | |
| "loss": 1.0902, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 14.17, | |
| "grad_norm": 5.367835987082188, | |
| "learning_rate": 1.519224698779198e-07, | |
| "loss": 1.0574, | |
| "step": 85 | |
| }, | |
| { | |
| "epoch": 15.0, | |
| "grad_norm": 5.098249451357161, | |
| "learning_rate": 0.0, | |
| "loss": 1.0662, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 15.0, | |
| "eval_loss": 1.0137017965316772, | |
| "eval_runtime": 45.4033, | |
| "eval_samples_per_second": 21.716, | |
| "eval_steps_per_second": 0.352, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 15.0, | |
| "step": 90, | |
| "total_flos": 24718476312576.0, | |
| "train_loss": 2.5687779426574706, | |
| "train_runtime": 665.8135, | |
| "train_samples_per_second": 16.649, | |
| "train_steps_per_second": 0.135 | |
| } | |
| ], | |
| "logging_steps": 5, | |
| "max_steps": 90, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 15, | |
| "save_steps": 100, | |
| "total_flos": 24718476312576.0, | |
| "train_batch_size": 4, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |