| { | |
| "best_metric": 0.012702367268502712, | |
| "best_model_checkpoint": "./vit-base-beans/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": 0.15, | |
| "grad_norm": 1.5160256624221802, | |
| "learning_rate": 0.00019230769230769233, | |
| "loss": 0.8461, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.31, | |
| "grad_norm": 1.1838645935058594, | |
| "learning_rate": 0.00018461538461538463, | |
| "loss": 0.3715, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.46, | |
| "grad_norm": 1.8937193155288696, | |
| "learning_rate": 0.00017692307692307693, | |
| "loss": 0.354, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.62, | |
| "grad_norm": 1.1941872835159302, | |
| "learning_rate": 0.00016923076923076923, | |
| "loss": 0.2516, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.77, | |
| "grad_norm": 1.9662144184112549, | |
| "learning_rate": 0.00016153846153846155, | |
| "loss": 0.1664, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.92, | |
| "grad_norm": 8.811502456665039, | |
| "learning_rate": 0.00015384615384615385, | |
| "loss": 0.2074, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 1.08, | |
| "grad_norm": 0.2282760590314865, | |
| "learning_rate": 0.00014615384615384615, | |
| "loss": 0.0974, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 1.23, | |
| "grad_norm": 0.1650390475988388, | |
| "learning_rate": 0.00013846153846153847, | |
| "loss": 0.1168, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 1.38, | |
| "grad_norm": 2.3822529315948486, | |
| "learning_rate": 0.00013076923076923077, | |
| "loss": 0.0462, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 1.54, | |
| "grad_norm": 0.11838462203741074, | |
| "learning_rate": 0.0001230769230769231, | |
| "loss": 0.0666, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 1.54, | |
| "eval_accuracy": 0.9924812030075187, | |
| "eval_loss": 0.032394178211688995, | |
| "eval_runtime": 1.7903, | |
| "eval_samples_per_second": 74.288, | |
| "eval_steps_per_second": 9.496, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 1.69, | |
| "grad_norm": 3.3664941787719727, | |
| "learning_rate": 0.00011538461538461538, | |
| "loss": 0.0599, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 1.85, | |
| "grad_norm": 0.16068458557128906, | |
| "learning_rate": 0.0001076923076923077, | |
| "loss": 0.0993, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "grad_norm": 0.11501188576221466, | |
| "learning_rate": 0.0001, | |
| "loss": 0.071, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 2.15, | |
| "grad_norm": 0.16083480417728424, | |
| "learning_rate": 9.230769230769232e-05, | |
| "loss": 0.0664, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 2.31, | |
| "grad_norm": 0.5593019127845764, | |
| "learning_rate": 8.461538461538461e-05, | |
| "loss": 0.0469, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 2.46, | |
| "grad_norm": 1.1808232069015503, | |
| "learning_rate": 7.692307692307693e-05, | |
| "loss": 0.0184, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 2.62, | |
| "grad_norm": 0.12332643568515778, | |
| "learning_rate": 6.923076923076924e-05, | |
| "loss": 0.0301, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 2.77, | |
| "grad_norm": 0.06919372081756592, | |
| "learning_rate": 6.153846153846155e-05, | |
| "loss": 0.0414, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 2.92, | |
| "grad_norm": 0.06933612376451492, | |
| "learning_rate": 5.384615384615385e-05, | |
| "loss": 0.0137, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 3.08, | |
| "grad_norm": 0.06432632356882095, | |
| "learning_rate": 4.615384615384616e-05, | |
| "loss": 0.0164, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 3.08, | |
| "eval_accuracy": 1.0, | |
| "eval_loss": 0.012702367268502712, | |
| "eval_runtime": 1.7615, | |
| "eval_samples_per_second": 75.505, | |
| "eval_steps_per_second": 9.651, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 3.23, | |
| "grad_norm": 0.05967758223414421, | |
| "learning_rate": 3.846153846153846e-05, | |
| "loss": 0.0122, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 3.38, | |
| "grad_norm": 0.06129758059978485, | |
| "learning_rate": 3.0769230769230774e-05, | |
| "loss": 0.0118, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 3.54, | |
| "grad_norm": 0.06324906647205353, | |
| "learning_rate": 2.307692307692308e-05, | |
| "loss": 0.0114, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 3.69, | |
| "grad_norm": 0.0768633708357811, | |
| "learning_rate": 1.5384615384615387e-05, | |
| "loss": 0.0114, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 3.85, | |
| "grad_norm": 0.05912179872393608, | |
| "learning_rate": 7.692307692307694e-06, | |
| "loss": 0.011, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "grad_norm": 0.05486920475959778, | |
| "learning_rate": 0.0, | |
| "loss": 0.0111, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "step": 260, | |
| "total_flos": 3.205097416476426e+17, | |
| "train_loss": 0.11755287733215553, | |
| "train_runtime": 95.4129, | |
| "train_samples_per_second": 43.348, | |
| "train_steps_per_second": 2.725 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 260, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 4, | |
| "save_steps": 100, | |
| "total_flos": 3.205097416476426e+17, | |
| "train_batch_size": 16, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |