| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 15.0, | |
| "eval_steps": 500, | |
| "global_step": 4770, | |
| "is_hyper_param_search": true, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.9968553459119497, | |
| "grad_norm": 0.3300945460796356, | |
| "learning_rate": 0.00026832335597634286, | |
| "loss": 0.2135, | |
| "step": 317 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_accuracy": 0.8893548387096775, | |
| "eval_loss": 0.056028928607702255, | |
| "eval_runtime": 2.6543, | |
| "eval_samples_per_second": 1167.921, | |
| "eval_steps_per_second": 24.489, | |
| "step": 318 | |
| }, | |
| { | |
| "epoch": 1.9937106918238994, | |
| "grad_norm": 0.7173739671707153, | |
| "learning_rate": 0.0002492219627932077, | |
| "loss": 0.0478, | |
| "step": 634 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_accuracy": 0.9193548387096774, | |
| "eval_loss": 0.03729024901986122, | |
| "eval_runtime": 2.6136, | |
| "eval_samples_per_second": 1186.114, | |
| "eval_steps_per_second": 24.87, | |
| "step": 636 | |
| }, | |
| { | |
| "epoch": 2.990566037735849, | |
| "grad_norm": 0.135478213429451, | |
| "learning_rate": 0.0002301205696100726, | |
| "loss": 0.0305, | |
| "step": 951 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_accuracy": 0.9312903225806451, | |
| "eval_loss": 0.031044812873005867, | |
| "eval_runtime": 2.6685, | |
| "eval_samples_per_second": 1161.69, | |
| "eval_steps_per_second": 24.358, | |
| "step": 954 | |
| }, | |
| { | |
| "epoch": 3.9874213836477987, | |
| "grad_norm": 0.11118420213460922, | |
| "learning_rate": 0.00021101917642693747, | |
| "loss": 0.0241, | |
| "step": 1268 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_accuracy": 0.937741935483871, | |
| "eval_loss": 0.030774397775530815, | |
| "eval_runtime": 2.6746, | |
| "eval_samples_per_second": 1159.032, | |
| "eval_steps_per_second": 24.302, | |
| "step": 1272 | |
| }, | |
| { | |
| "epoch": 4.984276729559748, | |
| "grad_norm": 0.07805101573467255, | |
| "learning_rate": 0.00019191778324380236, | |
| "loss": 0.0223, | |
| "step": 1585 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_accuracy": 0.9367741935483871, | |
| "eval_loss": 0.027621902525424957, | |
| "eval_runtime": 2.6474, | |
| "eval_samples_per_second": 1170.97, | |
| "eval_steps_per_second": 24.553, | |
| "step": 1590 | |
| }, | |
| { | |
| "epoch": 5.981132075471698, | |
| "grad_norm": 0.1515282243490219, | |
| "learning_rate": 0.00017281639006066723, | |
| "loss": 0.0199, | |
| "step": 1902 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_accuracy": 0.9403225806451613, | |
| "eval_loss": 0.025732239708304405, | |
| "eval_runtime": 2.6367, | |
| "eval_samples_per_second": 1175.705, | |
| "eval_steps_per_second": 24.652, | |
| "step": 1908 | |
| }, | |
| { | |
| "epoch": 6.977987421383648, | |
| "grad_norm": 0.05811043828725815, | |
| "learning_rate": 0.00015371499687753213, | |
| "loss": 0.018, | |
| "step": 2219 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "eval_accuracy": 0.9454838709677419, | |
| "eval_loss": 0.024622950702905655, | |
| "eval_runtime": 2.6832, | |
| "eval_samples_per_second": 1155.323, | |
| "eval_steps_per_second": 24.225, | |
| "step": 2226 | |
| }, | |
| { | |
| "epoch": 7.9748427672955975, | |
| "grad_norm": 0.06495410203933716, | |
| "learning_rate": 0.000134613603694397, | |
| "loss": 0.0166, | |
| "step": 2536 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_accuracy": 0.9461290322580646, | |
| "eval_loss": 0.022791748866438866, | |
| "eval_runtime": 2.6439, | |
| "eval_samples_per_second": 1172.489, | |
| "eval_steps_per_second": 24.584, | |
| "step": 2544 | |
| }, | |
| { | |
| "epoch": 8.971698113207546, | |
| "grad_norm": 0.06378670036792755, | |
| "learning_rate": 0.00011551221051126189, | |
| "loss": 0.0158, | |
| "step": 2853 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "eval_accuracy": 0.9454838709677419, | |
| "eval_loss": 0.021935414522886276, | |
| "eval_runtime": 2.6438, | |
| "eval_samples_per_second": 1172.55, | |
| "eval_steps_per_second": 24.586, | |
| "step": 2862 | |
| }, | |
| { | |
| "epoch": 9.968553459119496, | |
| "grad_norm": 0.05568454787135124, | |
| "learning_rate": 9.641081732812678e-05, | |
| "loss": 0.015, | |
| "step": 3170 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "eval_accuracy": 0.9461290322580646, | |
| "eval_loss": 0.02108229137957096, | |
| "eval_runtime": 2.6559, | |
| "eval_samples_per_second": 1167.227, | |
| "eval_steps_per_second": 24.474, | |
| "step": 3180 | |
| }, | |
| { | |
| "epoch": 10.965408805031446, | |
| "grad_norm": 0.09788110107183456, | |
| "learning_rate": 7.730942414499166e-05, | |
| "loss": 0.0141, | |
| "step": 3487 | |
| }, | |
| { | |
| "epoch": 11.0, | |
| "eval_accuracy": 0.944516129032258, | |
| "eval_loss": 0.02056371048092842, | |
| "eval_runtime": 2.6614, | |
| "eval_samples_per_second": 1164.798, | |
| "eval_steps_per_second": 24.423, | |
| "step": 3498 | |
| }, | |
| { | |
| "epoch": 11.962264150943396, | |
| "grad_norm": 0.04933185130357742, | |
| "learning_rate": 5.820803096185654e-05, | |
| "loss": 0.0135, | |
| "step": 3804 | |
| }, | |
| { | |
| "epoch": 12.0, | |
| "eval_accuracy": 0.9454838709677419, | |
| "eval_loss": 0.020129289478063583, | |
| "eval_runtime": 2.663, | |
| "eval_samples_per_second": 1164.083, | |
| "eval_steps_per_second": 24.408, | |
| "step": 3816 | |
| }, | |
| { | |
| "epoch": 12.959119496855346, | |
| "grad_norm": 0.04687652736902237, | |
| "learning_rate": 3.9106637778721424e-05, | |
| "loss": 0.0129, | |
| "step": 4121 | |
| }, | |
| { | |
| "epoch": 13.0, | |
| "eval_accuracy": 0.9464516129032258, | |
| "eval_loss": 0.019344650208950043, | |
| "eval_runtime": 2.6542, | |
| "eval_samples_per_second": 1167.98, | |
| "eval_steps_per_second": 24.49, | |
| "step": 4134 | |
| }, | |
| { | |
| "epoch": 13.955974842767295, | |
| "grad_norm": 0.0457344576716423, | |
| "learning_rate": 2.0005244595586304e-05, | |
| "loss": 0.0124, | |
| "step": 4438 | |
| }, | |
| { | |
| "epoch": 14.0, | |
| "eval_accuracy": 0.9458064516129032, | |
| "eval_loss": 0.01906457170844078, | |
| "eval_runtime": 2.658, | |
| "eval_samples_per_second": 1166.279, | |
| "eval_steps_per_second": 24.454, | |
| "step": 4452 | |
| }, | |
| { | |
| "epoch": 14.952830188679245, | |
| "grad_norm": 0.046884529292583466, | |
| "learning_rate": 9.038514124511885e-07, | |
| "loss": 0.012, | |
| "step": 4755 | |
| } | |
| ], | |
| "logging_steps": 317, | |
| "max_steps": 4770, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 15, | |
| "save_steps": 1000000000.0, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": true | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 1236796014635076.0, | |
| "train_batch_size": 48, | |
| "trial_name": null, | |
| "trial_params": { | |
| "alpha": 0.951089869492584, | |
| "fp16": false, | |
| "learning_rate": 0.00028188112716311066, | |
| "lr_scheduler": "cosine_with_restarts", | |
| "num_train_epochs": 15, | |
| "temperature": 13, | |
| "warmup_steps": 92, | |
| "weight_decay": 0.05924052651721969 | |
| } | |
| } | |