| { | |
| "best_metric": 0.8456, | |
| "best_model_checkpoint": "./resnet_finetuned_models_dataset/CIFAR100/50_from_100/microsoft_resnet-101/model_idx_0273/checkpoints/checkpoint-1998", | |
| "epoch": 6.0, | |
| "eval_steps": 500, | |
| "global_step": 1998, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 1.0, | |
| "grad_norm": 105.13134002685547, | |
| "learning_rate": 5e-05, | |
| "loss": 3.6555, | |
| "step": 333 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_accuracy": 0.4648, | |
| "eval_loss": 3.0037853717803955, | |
| "eval_runtime": 15.3242, | |
| "eval_samples_per_second": 244.711, | |
| "eval_steps_per_second": 3.85, | |
| "step": 333 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "grad_norm": 192.63848876953125, | |
| "learning_rate": 5e-05, | |
| "loss": 1.8103, | |
| "step": 666 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_accuracy": 0.7133333333333334, | |
| "eval_loss": 1.09019136428833, | |
| "eval_runtime": 15.6661, | |
| "eval_samples_per_second": 239.37, | |
| "eval_steps_per_second": 3.766, | |
| "step": 666 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "grad_norm": 99.28185272216797, | |
| "learning_rate": 5e-05, | |
| "loss": 0.8535, | |
| "step": 999 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_accuracy": 0.7952, | |
| "eval_loss": 0.7170190811157227, | |
| "eval_runtime": 15.393, | |
| "eval_samples_per_second": 243.617, | |
| "eval_steps_per_second": 3.833, | |
| "step": 999 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "grad_norm": 177.78045654296875, | |
| "learning_rate": 5e-05, | |
| "loss": 0.6201, | |
| "step": 1332 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_accuracy": 0.8234666666666667, | |
| "eval_loss": 0.5978338122367859, | |
| "eval_runtime": 15.3831, | |
| "eval_samples_per_second": 243.775, | |
| "eval_steps_per_second": 3.835, | |
| "step": 1332 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "grad_norm": 222.76309204101562, | |
| "learning_rate": 5e-05, | |
| "loss": 0.4866, | |
| "step": 1665 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_accuracy": 0.8330666666666666, | |
| "eval_loss": 0.5556532740592957, | |
| "eval_runtime": 15.3448, | |
| "eval_samples_per_second": 244.383, | |
| "eval_steps_per_second": 3.845, | |
| "step": 1665 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "grad_norm": 250.12911987304688, | |
| "learning_rate": 5e-05, | |
| "loss": 0.3977, | |
| "step": 1998 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_accuracy": 0.8456, | |
| "eval_loss": 0.5213834643363953, | |
| "eval_runtime": 15.404, | |
| "eval_samples_per_second": 243.443, | |
| "eval_steps_per_second": 3.83, | |
| "step": 1998 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 1998, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 6, | |
| "save_steps": 500, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": true | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 4.9058575437312e+18, | |
| "train_batch_size": 64, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |