{ "best_metric": 0.4240950047969818, "best_model_checkpoint": "ckpt/origin_vehicle_view/vehicle_appearance/checkpoint-32", "epoch": 4.516129032258064, "eval_steps": 2, "global_step": 35, "is_hyper_param_search": false, "is_local_process_zero": true, "is_world_process_zero": true, "log_history": [ { "epoch": 0.13, "learning_rate": 0.0001, "loss": 0.4736, "step": 1 }, { "epoch": 0.26, "learning_rate": 9.978670881475172e-05, "loss": 0.4764, "step": 2 }, { "epoch": 0.26, "eval_loss": 0.4665876626968384, "eval_runtime": 43.3272, "eval_samples_per_second": 1.962, "eval_steps_per_second": 1.962, "step": 2 }, { "epoch": 0.39, "learning_rate": 9.91486549841951e-05, "loss": 0.3893, "step": 3 }, { "epoch": 0.52, "learning_rate": 9.809128215864097e-05, "loss": 0.4222, "step": 4 }, { "epoch": 0.52, "eval_loss": 0.4527605473995209, "eval_runtime": 43.0079, "eval_samples_per_second": 1.976, "eval_steps_per_second": 1.976, "step": 4 }, { "epoch": 0.65, "learning_rate": 9.662361147021779e-05, "loss": 0.4746, "step": 5 }, { "epoch": 0.77, "learning_rate": 9.475816456775313e-05, "loss": 0.4558, "step": 6 }, { "epoch": 0.77, "eval_loss": 0.43778982758522034, "eval_runtime": 42.8176, "eval_samples_per_second": 1.985, "eval_steps_per_second": 1.985, "step": 6 }, { "epoch": 0.9, "learning_rate": 9.251085678648072e-05, "loss": 0.3981, "step": 7 }, { "epoch": 1.03, "learning_rate": 8.9900861364012e-05, "loss": 0.3561, "step": 8 }, { "epoch": 1.03, "eval_loss": 0.43061670660972595, "eval_runtime": 42.8128, "eval_samples_per_second": 1.985, "eval_steps_per_second": 1.985, "step": 8 }, { "epoch": 1.16, "learning_rate": 8.695044586103296e-05, "loss": 0.3842, "step": 9 }, { "epoch": 1.29, "learning_rate": 8.368478218232787e-05, "loss": 0.3748, "step": 10 }, { "epoch": 1.29, "eval_loss": 0.43047404289245605, "eval_runtime": 42.8408, "eval_samples_per_second": 1.984, "eval_steps_per_second": 1.984, "step": 10 }, { "epoch": 1.42, "learning_rate": 8.013173181896283e-05, "loss": 0.4174, "step": 11 }, { "epoch": 1.55, "learning_rate": 7.63216081438678e-05, "loss": 0.3478, "step": 12 }, { "epoch": 1.55, "eval_loss": 0.4285131096839905, "eval_runtime": 42.8184, "eval_samples_per_second": 1.985, "eval_steps_per_second": 1.985, "step": 12 }, { "epoch": 1.68, "learning_rate": 7.228691778882693e-05, "loss": 0.3117, "step": 13 }, { "epoch": 1.81, "learning_rate": 6.806208330935766e-05, "loss": 0.37, "step": 14 }, { "epoch": 1.81, "eval_loss": 0.4288734793663025, "eval_runtime": 42.8941, "eval_samples_per_second": 1.982, "eval_steps_per_second": 1.982, "step": 14 }, { "epoch": 1.94, "learning_rate": 6.368314950360415e-05, "loss": 0.318, "step": 15 }, { "epoch": 2.06, "learning_rate": 5.918747589082853e-05, "loss": 0.3739, "step": 16 }, { "epoch": 2.06, "eval_loss": 0.4283580183982849, "eval_runtime": 42.8288, "eval_samples_per_second": 1.985, "eval_steps_per_second": 1.985, "step": 16 }, { "epoch": 2.19, "learning_rate": 5.4613417973165106e-05, "loss": 0.3442, "step": 17 }, { "epoch": 2.32, "learning_rate": 5e-05, "loss": 0.3312, "step": 18 }, { "epoch": 2.32, "eval_loss": 0.42809075117111206, "eval_runtime": 42.8738, "eval_samples_per_second": 1.983, "eval_steps_per_second": 1.983, "step": 18 }, { "epoch": 2.45, "learning_rate": 4.5386582026834906e-05, "loss": 0.3589, "step": 19 }, { "epoch": 2.58, "learning_rate": 4.0812524109171476e-05, "loss": 0.3522, "step": 20 }, { "epoch": 2.58, "eval_loss": 0.4269283413887024, "eval_runtime": 42.7612, "eval_samples_per_second": 1.988, "eval_steps_per_second": 1.988, "step": 20 }, { "epoch": 2.71, "learning_rate": 3.631685049639586e-05, "loss": 0.3335, "step": 21 }, { "epoch": 2.84, "learning_rate": 3.1937916690642356e-05, "loss": 0.3236, "step": 22 }, { "epoch": 2.84, "eval_loss": 0.4258747696876526, "eval_runtime": 42.8316, "eval_samples_per_second": 1.985, "eval_steps_per_second": 1.985, "step": 22 }, { "epoch": 2.97, "learning_rate": 2.771308221117309e-05, "loss": 0.3151, "step": 23 }, { "epoch": 3.1, "learning_rate": 2.3678391856132204e-05, "loss": 0.3514, "step": 24 }, { "epoch": 3.1, "eval_loss": 0.42495113611221313, "eval_runtime": 42.8388, "eval_samples_per_second": 1.984, "eval_steps_per_second": 1.984, "step": 24 }, { "epoch": 3.23, "learning_rate": 1.9868268181037185e-05, "loss": 0.3345, "step": 25 }, { "epoch": 3.35, "learning_rate": 1.631521781767214e-05, "loss": 0.3407, "step": 26 }, { "epoch": 3.35, "eval_loss": 0.42438650131225586, "eval_runtime": 42.8465, "eval_samples_per_second": 1.984, "eval_steps_per_second": 1.984, "step": 26 }, { "epoch": 3.48, "learning_rate": 1.3049554138967051e-05, "loss": 0.2774, "step": 27 }, { "epoch": 3.61, "learning_rate": 1.0099138635988026e-05, "loss": 0.3132, "step": 28 }, { "epoch": 3.61, "eval_loss": 0.4248015880584717, "eval_runtime": 42.8857, "eval_samples_per_second": 1.982, "eval_steps_per_second": 1.982, "step": 28 }, { "epoch": 3.74, "learning_rate": 7.489143213519301e-06, "loss": 0.3051, "step": 29 }, { "epoch": 3.87, "learning_rate": 5.241835432246889e-06, "loss": 0.3114, "step": 30 }, { "epoch": 3.87, "eval_loss": 0.424568772315979, "eval_runtime": 42.8327, "eval_samples_per_second": 1.984, "eval_steps_per_second": 1.984, "step": 30 }, { "epoch": 4.0, "learning_rate": 3.376388529782215e-06, "loss": 0.3299, "step": 31 }, { "epoch": 4.13, "learning_rate": 1.908717841359048e-06, "loss": 0.3142, "step": 32 }, { "epoch": 4.13, "eval_loss": 0.4240950047969818, "eval_runtime": 42.8824, "eval_samples_per_second": 1.982, "eval_steps_per_second": 1.982, "step": 32 }, { "epoch": 4.26, "learning_rate": 8.513450158049108e-07, "loss": 0.3234, "step": 33 }, { "epoch": 4.39, "learning_rate": 2.1329118524827662e-07, "loss": 0.3088, "step": 34 }, { "epoch": 4.39, "eval_loss": 0.4241916835308075, "eval_runtime": 42.8268, "eval_samples_per_second": 1.985, "eval_steps_per_second": 1.985, "step": 34 }, { "epoch": 4.52, "learning_rate": 0.0, "loss": 0.339, "step": 35 }, { "epoch": 4.52, "step": 35, "total_flos": 9.3962984585429e+16, "train_loss": 0.358608683518001, "train_runtime": 2610.1034, "train_samples_per_second": 0.709, "train_steps_per_second": 0.013 } ], "logging_steps": 1.0, "max_steps": 35, "num_train_epochs": 5, "save_steps": 4, "total_flos": 9.3962984585429e+16, "trial_name": null, "trial_params": null }