{ "best_metric": 0.024307304993271828, "best_model_checkpoint": "./dinov2_LoRA_Liveness_detection_v1.1/checkpoint-448", "epoch": 3.0, "eval_steps": 64, "global_step": 471, "is_hyper_param_search": false, "is_local_process_zero": true, "is_world_process_zero": true, "log_history": [ { "epoch": 0.10191082802547771, "grad_norm": 3.3717641830444336, "learning_rate": 4.840764331210191e-05, "loss": 0.3564, "step": 16 }, { "epoch": 0.20382165605095542, "grad_norm": 1.4053750038146973, "learning_rate": 4.6709129511677285e-05, "loss": 0.1026, "step": 32 }, { "epoch": 0.3057324840764331, "grad_norm": 1.2090115547180176, "learning_rate": 4.501061571125266e-05, "loss": 0.0606, "step": 48 }, { "epoch": 0.40764331210191085, "grad_norm": 2.0192923545837402, "learning_rate": 4.331210191082803e-05, "loss": 0.0501, "step": 64 }, { "epoch": 0.40764331210191085, "eval_accuracy": 0.9841349785948124, "eval_f1": 0.9841073630518083, "eval_loss": 0.04576801881194115, "eval_precision": 0.9842311607904581, "eval_recall": 0.9841349785948124, "eval_runtime": 317.7291, "eval_samples_per_second": 62.947, "eval_steps_per_second": 7.868, "step": 64 }, { "epoch": 0.5095541401273885, "grad_norm": 1.1693896055221558, "learning_rate": 4.1613588110403404e-05, "loss": 0.0437, "step": 80 }, { "epoch": 0.6114649681528662, "grad_norm": 1.2890028953552246, "learning_rate": 3.991507430997877e-05, "loss": 0.0472, "step": 96 }, { "epoch": 0.7133757961783439, "grad_norm": 1.012406826019287, "learning_rate": 3.821656050955414e-05, "loss": 0.0388, "step": 112 }, { "epoch": 0.8152866242038217, "grad_norm": 1.3089163303375244, "learning_rate": 3.651804670912951e-05, "loss": 0.0355, "step": 128 }, { "epoch": 0.8152866242038217, "eval_accuracy": 0.9879627297909847, "eval_f1": 0.9879658772630169, "eval_loss": 0.034234367311000824, "eval_precision": 0.9879573976357167, "eval_recall": 0.9879627297909847, "eval_runtime": 309.1521, "eval_samples_per_second": 64.693, "eval_steps_per_second": 8.087, "step": 128 }, { "epoch": 0.9171974522292994, "grad_norm": 1.0827258825302124, "learning_rate": 3.481953290870488e-05, "loss": 0.0327, "step": 144 }, { "epoch": 1.019108280254777, "grad_norm": 1.2000535726547241, "learning_rate": 3.3121019108280256e-05, "loss": 0.035, "step": 160 }, { "epoch": 1.1210191082802548, "grad_norm": 2.487323045730591, "learning_rate": 3.142250530785563e-05, "loss": 0.0275, "step": 176 }, { "epoch": 1.2229299363057324, "grad_norm": 2.1451828479766846, "learning_rate": 2.9723991507431e-05, "loss": 0.0253, "step": 192 }, { "epoch": 1.2229299363057324, "eval_accuracy": 0.990078065978343, "eval_f1": 0.9900620320034099, "eval_loss": 0.030784847214818, "eval_precision": 0.9901457841504492, "eval_recall": 0.990078065978343, "eval_runtime": 313.1272, "eval_samples_per_second": 63.872, "eval_steps_per_second": 7.984, "step": 192 }, { "epoch": 1.3248407643312101, "grad_norm": 1.699660062789917, "learning_rate": 2.802547770700637e-05, "loss": 0.0241, "step": 208 }, { "epoch": 1.426751592356688, "grad_norm": 1.5516241788864136, "learning_rate": 2.632696390658174e-05, "loss": 0.0186, "step": 224 }, { "epoch": 1.5286624203821657, "grad_norm": 1.0598669052124023, "learning_rate": 2.4628450106157114e-05, "loss": 0.0198, "step": 240 }, { "epoch": 1.6305732484076434, "grad_norm": 1.8472939729690552, "learning_rate": 2.2929936305732484e-05, "loss": 0.0226, "step": 256 }, { "epoch": 1.6305732484076434, "eval_accuracy": 0.990984638630068, "eval_f1": 0.9909855466235222, "eval_loss": 0.026686813682317734, "eval_precision": 0.9909829656263511, "eval_recall": 0.990984638630068, "eval_runtime": 311.8552, "eval_samples_per_second": 64.132, "eval_steps_per_second": 8.017, "step": 256 }, { "epoch": 1.732484076433121, "grad_norm": 1.6193630695343018, "learning_rate": 2.1231422505307857e-05, "loss": 0.021, "step": 272 }, { "epoch": 1.8343949044585988, "grad_norm": 1.1963945627212524, "learning_rate": 1.953290870488323e-05, "loss": 0.0196, "step": 288 }, { "epoch": 1.9363057324840764, "grad_norm": 1.3636715412139893, "learning_rate": 1.78343949044586e-05, "loss": 0.0231, "step": 304 }, { "epoch": 2.038216560509554, "grad_norm": 1.2037672996520996, "learning_rate": 1.613588110403397e-05, "loss": 0.0264, "step": 320 }, { "epoch": 2.038216560509554, "eval_accuracy": 0.9911860992193402, "eval_f1": 0.9911791189750626, "eval_loss": 0.02550842985510826, "eval_precision": 0.9912087608299792, "eval_recall": 0.9911860992193402, "eval_runtime": 310.9636, "eval_samples_per_second": 64.316, "eval_steps_per_second": 8.04, "step": 320 }, { "epoch": 2.140127388535032, "grad_norm": 1.120186448097229, "learning_rate": 1.4437367303609342e-05, "loss": 0.0143, "step": 336 }, { "epoch": 2.2420382165605095, "grad_norm": 1.6843891143798828, "learning_rate": 1.2738853503184714e-05, "loss": 0.0148, "step": 352 }, { "epoch": 2.343949044585987, "grad_norm": 1.4458714723587036, "learning_rate": 1.1040339702760085e-05, "loss": 0.0179, "step": 368 }, { "epoch": 2.445859872611465, "grad_norm": 1.239098310470581, "learning_rate": 9.341825902335457e-06, "loss": 0.0168, "step": 384 }, { "epoch": 2.445859872611465, "eval_accuracy": 0.9921934021657013, "eval_f1": 0.9921865805615873, "eval_loss": 0.02449687570333481, "eval_precision": 0.992217636683327, "eval_recall": 0.9921934021657013, "eval_runtime": 309.071, "eval_samples_per_second": 64.71, "eval_steps_per_second": 8.089, "step": 384 }, { "epoch": 2.5477707006369426, "grad_norm": 1.4529588222503662, "learning_rate": 7.643312101910828e-06, "loss": 0.0152, "step": 400 }, { "epoch": 2.6496815286624202, "grad_norm": 1.1592313051223755, "learning_rate": 5.9447983014862e-06, "loss": 0.0127, "step": 416 }, { "epoch": 2.7515923566878984, "grad_norm": 1.4602242708206177, "learning_rate": 4.246284501061571e-06, "loss": 0.0166, "step": 432 }, { "epoch": 2.853503184713376, "grad_norm": 1.3949599266052246, "learning_rate": 2.547770700636943e-06, "loss": 0.0144, "step": 448 }, { "epoch": 2.853503184713376, "eval_accuracy": 0.9917401158398388, "eval_f1": 0.9917381130215306, "eval_loss": 0.024307304993271828, "eval_precision": 0.9917449458573591, "eval_recall": 0.9917401158398388, "eval_runtime": 304.7867, "eval_samples_per_second": 65.62, "eval_steps_per_second": 8.202, "step": 448 }, { "epoch": 2.9554140127388537, "grad_norm": 0.5842750668525696, "learning_rate": 8.492569002123142e-07, "loss": 0.0177, "step": 464 }, { "epoch": 3.0, "step": 471, "total_flos": 4.878503731310727e+18, "train_loss": 0.040008016563398315, "train_runtime": 8568.9568, "train_samples_per_second": 28.008, "train_steps_per_second": 0.055 } ], "logging_steps": 16, "max_steps": 471, "num_input_tokens_seen": 0, "num_train_epochs": 3, "save_steps": 64, "stateful_callbacks": { "TrainerControl": { "args": { "should_epoch_stop": false, "should_evaluate": false, "should_log": false, "should_save": true, "should_training_stop": true }, "attributes": {} } }, "total_flos": 4.878503731310727e+18, "train_batch_size": 512, "trial_name": null, "trial_params": null }