| { | |
| "best_metric": 0.023472091183066368, | |
| "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": 1.6673494577407837, | |
| "learning_rate": 0.00019230769230769233, | |
| "loss": 0.854, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.3076923076923077, | |
| "grad_norm": 2.448685646057129, | |
| "learning_rate": 0.00018461538461538463, | |
| "loss": 0.3317, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.46153846153846156, | |
| "grad_norm": 2.043360471725464, | |
| "learning_rate": 0.00017692307692307693, | |
| "loss": 0.1775, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.6153846153846154, | |
| "grad_norm": 1.6278131008148193, | |
| "learning_rate": 0.00016923076923076923, | |
| "loss": 0.2664, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.7692307692307693, | |
| "grad_norm": 2.8160293102264404, | |
| "learning_rate": 0.00016153846153846155, | |
| "loss": 0.1958, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.9230769230769231, | |
| "grad_norm": 3.9445712566375732, | |
| "learning_rate": 0.00015384615384615385, | |
| "loss": 0.3197, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 1.0769230769230769, | |
| "grad_norm": 1.7708736658096313, | |
| "learning_rate": 0.00014615384615384615, | |
| "loss": 0.17, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 1.2307692307692308, | |
| "grad_norm": 0.18868741393089294, | |
| "learning_rate": 0.00013846153846153847, | |
| "loss": 0.1075, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 1.3846153846153846, | |
| "grad_norm": 0.14158321917057037, | |
| "learning_rate": 0.00013076923076923077, | |
| "loss": 0.0441, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 1.5384615384615383, | |
| "grad_norm": 0.9749258160591125, | |
| "learning_rate": 0.0001230769230769231, | |
| "loss": 0.1108, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 1.5384615384615383, | |
| "eval_accuracy": 1.0, | |
| "eval_loss": 0.02870349958539009, | |
| "eval_runtime": 1.8021, | |
| "eval_samples_per_second": 73.804, | |
| "eval_steps_per_second": 9.434, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 1.6923076923076923, | |
| "grad_norm": 0.22034430503845215, | |
| "learning_rate": 0.00011538461538461538, | |
| "loss": 0.0436, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 1.8461538461538463, | |
| "grad_norm": 0.13325901329517365, | |
| "learning_rate": 0.0001076923076923077, | |
| "loss": 0.1215, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "grad_norm": 0.2772440016269684, | |
| "learning_rate": 0.0001, | |
| "loss": 0.0707, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 2.1538461538461537, | |
| "grad_norm": 0.9475046992301941, | |
| "learning_rate": 9.230769230769232e-05, | |
| "loss": 0.0471, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 2.3076923076923075, | |
| "grad_norm": 1.3417366743087769, | |
| "learning_rate": 8.461538461538461e-05, | |
| "loss": 0.0655, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 2.4615384615384617, | |
| "grad_norm": 0.09867988526821136, | |
| "learning_rate": 7.692307692307693e-05, | |
| "loss": 0.0296, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 2.6153846153846154, | |
| "grad_norm": 0.10215840488672256, | |
| "learning_rate": 6.923076923076924e-05, | |
| "loss": 0.0161, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 2.769230769230769, | |
| "grad_norm": 0.11826897412538528, | |
| "learning_rate": 6.153846153846155e-05, | |
| "loss": 0.0345, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 2.9230769230769234, | |
| "grad_norm": 0.06787172704935074, | |
| "learning_rate": 5.384615384615385e-05, | |
| "loss": 0.0145, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 3.076923076923077, | |
| "grad_norm": 0.06440696865320206, | |
| "learning_rate": 4.615384615384616e-05, | |
| "loss": 0.0128, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 3.076923076923077, | |
| "eval_accuracy": 1.0, | |
| "eval_loss": 0.023472091183066368, | |
| "eval_runtime": 1.2523, | |
| "eval_samples_per_second": 106.201, | |
| "eval_steps_per_second": 13.575, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 3.230769230769231, | |
| "grad_norm": 0.06286267191171646, | |
| "learning_rate": 3.846153846153846e-05, | |
| "loss": 0.0123, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 3.3846153846153846, | |
| "grad_norm": 0.06109225004911423, | |
| "learning_rate": 3.0769230769230774e-05, | |
| "loss": 0.0121, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 3.5384615384615383, | |
| "grad_norm": 0.06259910762310028, | |
| "learning_rate": 2.307692307692308e-05, | |
| "loss": 0.0117, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 3.6923076923076925, | |
| "grad_norm": 0.060203827917575836, | |
| "learning_rate": 1.5384615384615387e-05, | |
| "loss": 0.0114, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 3.8461538461538463, | |
| "grad_norm": 0.05893297493457794, | |
| "learning_rate": 7.692307692307694e-06, | |
| "loss": 0.0114, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "grad_norm": 0.05691501870751381, | |
| "learning_rate": 0.0, | |
| "loss": 0.0114, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "step": 260, | |
| "total_flos": 3.205097416476426e+17, | |
| "train_loss": 0.11937046449344892, | |
| "train_runtime": 97.651, | |
| "train_samples_per_second": 42.355, | |
| "train_steps_per_second": 2.663 | |
| } | |
| ], | |
| "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": false | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 3.205097416476426e+17, | |
| "train_batch_size": 16, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |