| { | |
| "best_metric": 0.2866075932979584, | |
| "best_model_checkpoint": "./convnext-nano-1e-4-augment/checkpoint-2750", | |
| "epoch": 10.0, | |
| "eval_steps": 500, | |
| "global_step": 2750, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.36, | |
| "grad_norm": 11.396906852722168, | |
| "learning_rate": 9.967408676742751e-05, | |
| "loss": 1.7369, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.73, | |
| "grad_norm": 14.72342300415039, | |
| "learning_rate": 9.870059584711668e-05, | |
| "loss": 0.8282, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_accuracy": 0.8536779324055666, | |
| "eval_loss": 0.5148155689239502, | |
| "eval_runtime": 61.4101, | |
| "eval_samples_per_second": 40.954, | |
| "eval_steps_per_second": 0.651, | |
| "step": 275 | |
| }, | |
| { | |
| "epoch": 1.09, | |
| "grad_norm": 16.179250717163086, | |
| "learning_rate": 9.709221818197624e-05, | |
| "loss": 0.6581, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 1.45, | |
| "grad_norm": 10.215620994567871, | |
| "learning_rate": 9.486992143456792e-05, | |
| "loss": 0.5141, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 1.82, | |
| "grad_norm": 8.687200546264648, | |
| "learning_rate": 9.206267664155907e-05, | |
| "loss": 0.5209, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_accuracy": 0.88389662027833, | |
| "eval_loss": 0.4150846004486084, | |
| "eval_runtime": 61.5346, | |
| "eval_samples_per_second": 40.871, | |
| "eval_steps_per_second": 0.65, | |
| "step": 550 | |
| }, | |
| { | |
| "epoch": 2.18, | |
| "grad_norm": 10.711565971374512, | |
| "learning_rate": 8.870708053195413e-05, | |
| "loss": 0.4456, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 2.55, | |
| "grad_norm": 9.642666816711426, | |
| "learning_rate": 8.484687843276469e-05, | |
| "loss": 0.4067, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 2.91, | |
| "grad_norm": 11.507894515991211, | |
| "learning_rate": 8.053239398177191e-05, | |
| "loss": 0.3867, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_accuracy": 0.9009940357852882, | |
| "eval_loss": 0.3642739951610565, | |
| "eval_runtime": 61.3025, | |
| "eval_samples_per_second": 41.026, | |
| "eval_steps_per_second": 0.653, | |
| "step": 825 | |
| }, | |
| { | |
| "epoch": 3.27, | |
| "grad_norm": 13.279052734375, | |
| "learning_rate": 7.58198730819481e-05, | |
| "loss": 0.3452, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 3.64, | |
| "grad_norm": 9.233952522277832, | |
| "learning_rate": 7.077075065009433e-05, | |
| "loss": 0.3216, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "grad_norm": 9.883641242980957, | |
| "learning_rate": 6.545084971874738e-05, | |
| "loss": 0.3183, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_accuracy": 0.9049701789264414, | |
| "eval_loss": 0.3240828812122345, | |
| "eval_runtime": 61.2041, | |
| "eval_samples_per_second": 41.092, | |
| "eval_steps_per_second": 0.654, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 4.36, | |
| "grad_norm": 10.317498207092285, | |
| "learning_rate": 5.992952333228728e-05, | |
| "loss": 0.2789, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 4.73, | |
| "grad_norm": 6.958377838134766, | |
| "learning_rate": 5.427875042394199e-05, | |
| "loss": 0.2679, | |
| "step": 1300 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_accuracy": 0.904572564612326, | |
| "eval_loss": 0.32900604605674744, | |
| "eval_runtime": 61.4874, | |
| "eval_samples_per_second": 40.903, | |
| "eval_steps_per_second": 0.651, | |
| "step": 1375 | |
| }, | |
| { | |
| "epoch": 5.09, | |
| "grad_norm": 7.9765119552612305, | |
| "learning_rate": 4.85721974603152e-05, | |
| "loss": 0.261, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 5.45, | |
| "grad_norm": 6.655781269073486, | |
| "learning_rate": 4.288425808633575e-05, | |
| "loss": 0.2243, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 5.82, | |
| "grad_norm": 9.778780937194824, | |
| "learning_rate": 3.728908329032567e-05, | |
| "loss": 0.2364, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_accuracy": 0.9137176938369781, | |
| "eval_loss": 0.30884459614753723, | |
| "eval_runtime": 61.4411, | |
| "eval_samples_per_second": 40.934, | |
| "eval_steps_per_second": 0.651, | |
| "step": 1650 | |
| }, | |
| { | |
| "epoch": 6.18, | |
| "grad_norm": 6.904848098754883, | |
| "learning_rate": 3.1859614732467954e-05, | |
| "loss": 0.2176, | |
| "step": 1700 | |
| }, | |
| { | |
| "epoch": 6.55, | |
| "grad_norm": 6.360904216766357, | |
| "learning_rate": 2.6666633838716314e-05, | |
| "loss": 0.1943, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 6.91, | |
| "grad_norm": 9.572250366210938, | |
| "learning_rate": 2.1777839056661554e-05, | |
| "loss": 0.1981, | |
| "step": 1900 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "eval_accuracy": 0.9137176938369781, | |
| "eval_loss": 0.29816073179244995, | |
| "eval_runtime": 62.0018, | |
| "eval_samples_per_second": 40.563, | |
| "eval_steps_per_second": 0.645, | |
| "step": 1925 | |
| }, | |
| { | |
| "epoch": 7.27, | |
| "grad_norm": 4.384122371673584, | |
| "learning_rate": 1.725696330273575e-05, | |
| "loss": 0.1778, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 7.64, | |
| "grad_norm": 7.344178199768066, | |
| "learning_rate": 1.3162943106179749e-05, | |
| "loss": 0.1692, | |
| "step": 2100 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "grad_norm": 5.066224575042725, | |
| "learning_rate": 9.549150281252633e-06, | |
| "loss": 0.1704, | |
| "step": 2200 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_accuracy": 0.9168986083499006, | |
| "eval_loss": 0.2899409532546997, | |
| "eval_runtime": 61.3116, | |
| "eval_samples_per_second": 41.02, | |
| "eval_steps_per_second": 0.652, | |
| "step": 2200 | |
| }, | |
| { | |
| "epoch": 8.36, | |
| "grad_norm": 7.729918003082275, | |
| "learning_rate": 6.462696144011149e-06, | |
| "loss": 0.1627, | |
| "step": 2300 | |
| }, | |
| { | |
| "epoch": 8.73, | |
| "grad_norm": 9.76986312866211, | |
| "learning_rate": 3.9438173442575e-06, | |
| "loss": 0.1572, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "eval_accuracy": 0.920079522862823, | |
| "eval_loss": 0.2868165671825409, | |
| "eval_runtime": 61.3101, | |
| "eval_samples_per_second": 41.021, | |
| "eval_steps_per_second": 0.652, | |
| "step": 2475 | |
| }, | |
| { | |
| "epoch": 9.09, | |
| "grad_norm": 7.546455383300781, | |
| "learning_rate": 2.0253513192751373e-06, | |
| "loss": 0.1548, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 9.45, | |
| "grad_norm": 8.111244201660156, | |
| "learning_rate": 7.323082076153509e-07, | |
| "loss": 0.1519, | |
| "step": 2600 | |
| }, | |
| { | |
| "epoch": 9.82, | |
| "grad_norm": 8.917200088500977, | |
| "learning_rate": 8.15448036932176e-08, | |
| "loss": 0.168, | |
| "step": 2700 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "eval_accuracy": 0.9204771371769384, | |
| "eval_loss": 0.2866075932979584, | |
| "eval_runtime": 61.2534, | |
| "eval_samples_per_second": 41.059, | |
| "eval_steps_per_second": 0.653, | |
| "step": 2750 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "step": 2750, | |
| "total_flos": 7.000491898906214e+18, | |
| "train_loss": 0.3542752295407382, | |
| "train_runtime": 6594.6916, | |
| "train_samples_per_second": 26.659, | |
| "train_steps_per_second": 0.417 | |
| } | |
| ], | |
| "logging_steps": 100, | |
| "max_steps": 2750, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 10, | |
| "save_steps": 500, | |
| "total_flos": 7.000491898906214e+18, | |
| "train_batch_size": 64, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |