| { | |
| "best_metric": 0.044119928032159805, | |
| "best_model_checkpoint": "./vit-base-beans-demo-v5/checkpoint-100", | |
| "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": 2.110915422439575, | |
| "learning_rate": 0.00019230769230769233, | |
| "loss": 0.8903, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.3076923076923077, | |
| "grad_norm": 1.4865622520446777, | |
| "learning_rate": 0.00018461538461538463, | |
| "loss": 0.3935, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.46153846153846156, | |
| "grad_norm": 0.5153194665908813, | |
| "learning_rate": 0.00017692307692307693, | |
| "loss": 0.197, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.6153846153846154, | |
| "grad_norm": 3.097721815109253, | |
| "learning_rate": 0.00016923076923076923, | |
| "loss": 0.2258, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.7692307692307693, | |
| "grad_norm": 4.940190315246582, | |
| "learning_rate": 0.00016153846153846155, | |
| "loss": 0.2628, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.9230769230769231, | |
| "grad_norm": 0.6224154829978943, | |
| "learning_rate": 0.00015384615384615385, | |
| "loss": 0.2031, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 1.0769230769230769, | |
| "grad_norm": 1.554422378540039, | |
| "learning_rate": 0.00014615384615384615, | |
| "loss": 0.248, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 1.2307692307692308, | |
| "grad_norm": 0.161035418510437, | |
| "learning_rate": 0.00013846153846153847, | |
| "loss": 0.1004, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 1.3846153846153846, | |
| "grad_norm": 7.808058261871338, | |
| "learning_rate": 0.00013076923076923077, | |
| "loss": 0.0656, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 1.5384615384615383, | |
| "grad_norm": 1.3628555536270142, | |
| "learning_rate": 0.0001230769230769231, | |
| "loss": 0.0757, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 1.5384615384615383, | |
| "eval_accuracy": 0.9924812030075187, | |
| "eval_loss": 0.044119928032159805, | |
| "eval_runtime": 1.5577, | |
| "eval_samples_per_second": 85.381, | |
| "eval_steps_per_second": 10.913, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 1.6923076923076923, | |
| "grad_norm": 2.1813511848449707, | |
| "learning_rate": 0.00011538461538461538, | |
| "loss": 0.0681, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 1.8461538461538463, | |
| "grad_norm": 0.24909372627735138, | |
| "learning_rate": 0.0001076923076923077, | |
| "loss": 0.0947, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "grad_norm": 0.14786876738071442, | |
| "learning_rate": 0.0001, | |
| "loss": 0.1183, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 2.1538461538461537, | |
| "grad_norm": 0.26466748118400574, | |
| "learning_rate": 9.230769230769232e-05, | |
| "loss": 0.1728, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 2.3076923076923075, | |
| "grad_norm": 0.2627667784690857, | |
| "learning_rate": 8.461538461538461e-05, | |
| "loss": 0.063, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 2.4615384615384617, | |
| "grad_norm": 0.10506220906972885, | |
| "learning_rate": 7.692307692307693e-05, | |
| "loss": 0.0257, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 2.6153846153846154, | |
| "grad_norm": 2.347315549850464, | |
| "learning_rate": 6.923076923076924e-05, | |
| "loss": 0.0367, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 2.769230769230769, | |
| "grad_norm": 0.6472500562667847, | |
| "learning_rate": 6.153846153846155e-05, | |
| "loss": 0.0342, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 2.9230769230769234, | |
| "grad_norm": 4.727846622467041, | |
| "learning_rate": 5.384615384615385e-05, | |
| "loss": 0.0425, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 3.076923076923077, | |
| "grad_norm": 0.06820247322320938, | |
| "learning_rate": 4.615384615384616e-05, | |
| "loss": 0.0131, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 3.076923076923077, | |
| "eval_accuracy": 0.9624060150375939, | |
| "eval_loss": 0.14681094884872437, | |
| "eval_runtime": 1.805, | |
| "eval_samples_per_second": 73.685, | |
| "eval_steps_per_second": 9.418, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 3.230769230769231, | |
| "grad_norm": 3.413127899169922, | |
| "learning_rate": 3.846153846153846e-05, | |
| "loss": 0.0233, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 3.3846153846153846, | |
| "grad_norm": 3.6428942680358887, | |
| "learning_rate": 3.0769230769230774e-05, | |
| "loss": 0.0145, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 3.5384615384615383, | |
| "grad_norm": 0.4084061086177826, | |
| "learning_rate": 2.307692307692308e-05, | |
| "loss": 0.012, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 3.6923076923076925, | |
| "grad_norm": 0.06784854084253311, | |
| "learning_rate": 1.5384615384615387e-05, | |
| "loss": 0.0114, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 3.8461538461538463, | |
| "grad_norm": 0.05810296908020973, | |
| "learning_rate": 7.692307692307694e-06, | |
| "loss": 0.017, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "grad_norm": 0.057864416390657425, | |
| "learning_rate": 0.0, | |
| "loss": 0.011, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "step": 260, | |
| "total_flos": 3.205097416476426e+17, | |
| "train_loss": 0.13155597568704533, | |
| "train_runtime": 107.1574, | |
| "train_samples_per_second": 38.597, | |
| "train_steps_per_second": 2.426 | |
| } | |
| ], | |
| "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 | |
| } | |