| { | |
| "best_metric": 0.045289285480976105, | |
| "best_model_checkpoint": "./vit-base-beans-demo-v5/checkpoint-200", | |
| "epoch": 4.0, | |
| "eval_steps": 100, | |
| "global_step": 260, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.15384615384615385, | |
| "grad_norm": 1.3005889654159546, | |
| "learning_rate": 0.00019230769230769233, | |
| "loss": 0.8273, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.3076923076923077, | |
| "grad_norm": 1.6120234727859497, | |
| "learning_rate": 0.00018461538461538463, | |
| "loss": 0.3158, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.46153846153846156, | |
| "grad_norm": 3.364579916000366, | |
| "learning_rate": 0.00017692307692307693, | |
| "loss": 0.1878, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.6153846153846154, | |
| "grad_norm": 0.9248621463775635, | |
| "learning_rate": 0.00016923076923076923, | |
| "loss": 0.1993, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.7692307692307693, | |
| "grad_norm": 0.9253031611442566, | |
| "learning_rate": 0.00016153846153846155, | |
| "loss": 0.1792, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.9230769230769231, | |
| "grad_norm": 2.9216034412384033, | |
| "learning_rate": 0.00015384615384615385, | |
| "loss": 0.2601, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 1.0769230769230769, | |
| "grad_norm": 0.7130853533744812, | |
| "learning_rate": 0.00014615384615384615, | |
| "loss": 0.2204, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 1.2307692307692308, | |
| "grad_norm": 0.3435571491718292, | |
| "learning_rate": 0.00013846153846153847, | |
| "loss": 0.0978, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 1.3846153846153846, | |
| "grad_norm": 0.5589646100997925, | |
| "learning_rate": 0.00013076923076923077, | |
| "loss": 0.1044, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 1.5384615384615383, | |
| "grad_norm": 0.12235202640295029, | |
| "learning_rate": 0.0001230769230769231, | |
| "loss": 0.0573, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 1.5384615384615383, | |
| "eval_accuracy": 0.9924812030075187, | |
| "eval_loss": 0.05181942135095596, | |
| "eval_runtime": 1.8339, | |
| "eval_samples_per_second": 72.522, | |
| "eval_steps_per_second": 9.27, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 1.6923076923076923, | |
| "grad_norm": 1.0851041078567505, | |
| "learning_rate": 0.00011538461538461538, | |
| "loss": 0.0457, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 1.8461538461538463, | |
| "grad_norm": 1.3997029066085815, | |
| "learning_rate": 0.0001076923076923077, | |
| "loss": 0.0794, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "grad_norm": 0.10364062339067459, | |
| "learning_rate": 0.0001, | |
| "loss": 0.0565, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 2.1538461538461537, | |
| "grad_norm": 4.5433855056762695, | |
| "learning_rate": 9.230769230769232e-05, | |
| "loss": 0.0756, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 2.3076923076923075, | |
| "grad_norm": 1.4649441242218018, | |
| "learning_rate": 8.461538461538461e-05, | |
| "loss": 0.0738, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 2.4615384615384617, | |
| "grad_norm": 0.42457565665245056, | |
| "learning_rate": 7.692307692307693e-05, | |
| "loss": 0.0196, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 2.6153846153846154, | |
| "grad_norm": 0.23078908026218414, | |
| "learning_rate": 6.923076923076924e-05, | |
| "loss": 0.0146, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 2.769230769230769, | |
| "grad_norm": 2.5094916820526123, | |
| "learning_rate": 6.153846153846155e-05, | |
| "loss": 0.0417, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 2.9230769230769234, | |
| "grad_norm": 0.06619451940059662, | |
| "learning_rate": 5.384615384615385e-05, | |
| "loss": 0.0334, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 3.076923076923077, | |
| "grad_norm": 0.062100838869810104, | |
| "learning_rate": 4.615384615384616e-05, | |
| "loss": 0.0122, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 3.076923076923077, | |
| "eval_accuracy": 0.9924812030075187, | |
| "eval_loss": 0.045289285480976105, | |
| "eval_runtime": 2.6212, | |
| "eval_samples_per_second": 50.74, | |
| "eval_steps_per_second": 6.486, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 3.230769230769231, | |
| "grad_norm": 0.06120683252811432, | |
| "learning_rate": 3.846153846153846e-05, | |
| "loss": 0.0125, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 3.3846153846153846, | |
| "grad_norm": 0.06342925131320953, | |
| "learning_rate": 3.0769230769230774e-05, | |
| "loss": 0.0116, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 3.5384615384615383, | |
| "grad_norm": 0.06138905882835388, | |
| "learning_rate": 2.307692307692308e-05, | |
| "loss": 0.0109, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 3.6923076923076925, | |
| "grad_norm": 0.05651098117232323, | |
| "learning_rate": 1.5384615384615387e-05, | |
| "loss": 0.0106, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 3.8461538461538463, | |
| "grad_norm": 0.0597892589867115, | |
| "learning_rate": 7.692307692307694e-06, | |
| "loss": 0.0105, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "grad_norm": 0.051263097673654556, | |
| "learning_rate": 0.0, | |
| "loss": 0.0106, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "step": 260, | |
| "total_flos": 3.205097416476426e+17, | |
| "train_loss": 0.11417198760005144, | |
| "train_runtime": 120.8248, | |
| "train_samples_per_second": 34.231, | |
| "train_steps_per_second": 2.152 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 260, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 4, | |
| "save_steps": 100, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": true | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 3.205097416476426e+17, | |
| "train_batch_size": 16, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |