| { | |
| "best_global_step": 200, | |
| "best_metric": 1.0, | |
| "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": 1.957282304763794, | |
| "learning_rate": 0.0001930769230769231, | |
| "loss": 0.8039, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.3076923076923077, | |
| "grad_norm": 1.0513660907745361, | |
| "learning_rate": 0.0001853846153846154, | |
| "loss": 0.3269, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.46153846153846156, | |
| "grad_norm": 2.8595218658447266, | |
| "learning_rate": 0.0001776923076923077, | |
| "loss": 0.2376, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.6153846153846154, | |
| "grad_norm": 0.8380834460258484, | |
| "learning_rate": 0.00017, | |
| "loss": 0.2117, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.7692307692307693, | |
| "grad_norm": 4.57429838180542, | |
| "learning_rate": 0.0001623076923076923, | |
| "loss": 0.1837, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.9230769230769231, | |
| "grad_norm": 4.2446722984313965, | |
| "learning_rate": 0.00015461538461538464, | |
| "loss": 0.3205, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 1.0769230769230769, | |
| "grad_norm": 0.22835291922092438, | |
| "learning_rate": 0.00014692307692307693, | |
| "loss": 0.1658, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 1.2307692307692308, | |
| "grad_norm": 2.5722620487213135, | |
| "learning_rate": 0.00013923076923076923, | |
| "loss": 0.0715, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 1.3846153846153846, | |
| "grad_norm": 0.15161257982254028, | |
| "learning_rate": 0.00013153846153846156, | |
| "loss": 0.0435, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 1.5384615384615383, | |
| "grad_norm": 3.3087499141693115, | |
| "learning_rate": 0.00012384615384615385, | |
| "loss": 0.0537, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 1.5384615384615383, | |
| "eval_accuracy": 0.9849624060150376, | |
| "eval_loss": 0.055375177413225174, | |
| "eval_runtime": 2.6053, | |
| "eval_samples_per_second": 51.05, | |
| "eval_steps_per_second": 3.454, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 1.6923076923076923, | |
| "grad_norm": 0.14113253355026245, | |
| "learning_rate": 0.00011615384615384617, | |
| "loss": 0.1055, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 1.8461538461538463, | |
| "grad_norm": 0.1897106021642685, | |
| "learning_rate": 0.00010846153846153846, | |
| "loss": 0.1263, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "grad_norm": 0.11339154094457626, | |
| "learning_rate": 0.00010076923076923077, | |
| "loss": 0.1044, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 2.1538461538461537, | |
| "grad_norm": 0.28828829526901245, | |
| "learning_rate": 9.307692307692309e-05, | |
| "loss": 0.0738, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 2.3076923076923075, | |
| "grad_norm": 0.11714714765548706, | |
| "learning_rate": 8.538461538461538e-05, | |
| "loss": 0.0301, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 2.4615384615384617, | |
| "grad_norm": 0.08417778462171555, | |
| "learning_rate": 7.76923076923077e-05, | |
| "loss": 0.0157, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 2.6153846153846154, | |
| "grad_norm": 0.0785544291138649, | |
| "learning_rate": 7e-05, | |
| "loss": 0.0143, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 2.769230769230769, | |
| "grad_norm": 0.9593146443367004, | |
| "learning_rate": 6.23076923076923e-05, | |
| "loss": 0.0422, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 2.9230769230769234, | |
| "grad_norm": 0.06379880756139755, | |
| "learning_rate": 5.461538461538461e-05, | |
| "loss": 0.0228, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 3.076923076923077, | |
| "grad_norm": 0.06561139225959778, | |
| "learning_rate": 4.692307692307693e-05, | |
| "loss": 0.0129, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 3.076923076923077, | |
| "eval_accuracy": 1.0, | |
| "eval_loss": 0.013629596680402756, | |
| "eval_runtime": 1.2116, | |
| "eval_samples_per_second": 109.77, | |
| "eval_steps_per_second": 7.428, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 3.230769230769231, | |
| "grad_norm": 0.06040000915527344, | |
| "learning_rate": 3.923076923076923e-05, | |
| "loss": 0.032, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 3.3846153846153846, | |
| "grad_norm": 0.0607382096350193, | |
| "learning_rate": 3.153846153846154e-05, | |
| "loss": 0.0114, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 3.5384615384615383, | |
| "grad_norm": 0.06130724027752876, | |
| "learning_rate": 2.384615384615385e-05, | |
| "loss": 0.0111, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 3.6923076923076925, | |
| "grad_norm": 0.06276917457580566, | |
| "learning_rate": 1.6153846153846154e-05, | |
| "loss": 0.0108, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 3.8461538461538463, | |
| "grad_norm": 0.06522306054830551, | |
| "learning_rate": 8.461538461538462e-06, | |
| "loss": 0.0216, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "grad_norm": 0.05368223413825035, | |
| "learning_rate": 7.692307692307694e-07, | |
| "loss": 0.039, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "step": 260, | |
| "total_flos": 3.205097416476426e+17, | |
| "train_loss": 0.11894229616110141, | |
| "train_runtime": 102.8707, | |
| "train_samples_per_second": 40.206, | |
| "train_steps_per_second": 2.527 | |
| } | |
| ], | |
| "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 | |
| } | |