| { | |
| "best_metric": 0.03475998714566231, | |
| "best_model_checkpoint": "t5/checkpoint-785", | |
| "epoch": 8.0, | |
| "eval_steps": 500, | |
| "global_step": 1256, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 1.0, | |
| "grad_norm": 0.17618757486343384, | |
| "learning_rate": 1.9e-05, | |
| "loss": 0.0618, | |
| "step": 157 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_accuracy": 0.6860511590727418, | |
| "eval_loss": 0.041351497173309326, | |
| "eval_runtime": 60.885, | |
| "eval_samples_per_second": 82.188, | |
| "eval_steps_per_second": 0.328, | |
| "step": 157 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "grad_norm": 0.18060798943042755, | |
| "learning_rate": 1.8e-05, | |
| "loss": 0.0423, | |
| "step": 314 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_accuracy": 0.707234212629896, | |
| "eval_loss": 0.03738662600517273, | |
| "eval_runtime": 60.1479, | |
| "eval_samples_per_second": 83.195, | |
| "eval_steps_per_second": 0.333, | |
| "step": 314 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "grad_norm": 0.20145682990550995, | |
| "learning_rate": 1.7e-05, | |
| "loss": 0.0368, | |
| "step": 471 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_accuracy": 0.7208233413269385, | |
| "eval_loss": 0.03572960942983627, | |
| "eval_runtime": 59.3756, | |
| "eval_samples_per_second": 84.277, | |
| "eval_steps_per_second": 0.337, | |
| "step": 471 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "grad_norm": 0.20047585666179657, | |
| "learning_rate": 1.6000000000000003e-05, | |
| "loss": 0.0326, | |
| "step": 628 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_accuracy": 0.7262190247801759, | |
| "eval_loss": 0.03503531962633133, | |
| "eval_runtime": 59.034, | |
| "eval_samples_per_second": 84.765, | |
| "eval_steps_per_second": 0.339, | |
| "step": 628 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "grad_norm": 0.1549680233001709, | |
| "learning_rate": 1.5000000000000002e-05, | |
| "loss": 0.0293, | |
| "step": 785 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_accuracy": 0.7294164668265388, | |
| "eval_loss": 0.03475998714566231, | |
| "eval_runtime": 59.2862, | |
| "eval_samples_per_second": 84.404, | |
| "eval_steps_per_second": 0.337, | |
| "step": 785 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "grad_norm": 0.19214752316474915, | |
| "learning_rate": 1.4e-05, | |
| "loss": 0.0266, | |
| "step": 942 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_accuracy": 0.7352118305355716, | |
| "eval_loss": 0.0347641222178936, | |
| "eval_runtime": 59.4378, | |
| "eval_samples_per_second": 84.189, | |
| "eval_steps_per_second": 0.336, | |
| "step": 942 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "grad_norm": 0.1793193817138672, | |
| "learning_rate": 1.3000000000000001e-05, | |
| "loss": 0.0243, | |
| "step": 1099 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "eval_accuracy": 0.7328137490007993, | |
| "eval_loss": 0.03488365188241005, | |
| "eval_runtime": 59.8565, | |
| "eval_samples_per_second": 83.6, | |
| "eval_steps_per_second": 0.334, | |
| "step": 1099 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "grad_norm": 0.19961176812648773, | |
| "learning_rate": 1.2e-05, | |
| "loss": 0.0223, | |
| "step": 1256 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_accuracy": 0.7356115107913669, | |
| "eval_loss": 0.03529239445924759, | |
| "eval_runtime": 58.7625, | |
| "eval_samples_per_second": 85.156, | |
| "eval_steps_per_second": 0.34, | |
| "step": 1256 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 3140, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 20, | |
| "save_steps": 500, | |
| "total_flos": 4.333072619596493e+16, | |
| "train_batch_size": 256, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |