| { | |
| "best_metric": 0.534845769405365, | |
| "best_model_checkpoint": "./vit-epsilon-1e-7/checkpoint-4173", | |
| "epoch": 23.0, | |
| "eval_steps": 500, | |
| "global_step": 7383, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 1.0, | |
| "grad_norm": 26.270666122436523, | |
| "learning_rate": 1.8349682631275247e-05, | |
| "loss": 1.765, | |
| "step": 321 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_accuracy": 0.691747572815534, | |
| "eval_f1": 0.6531054572377663, | |
| "eval_loss": 0.9570140838623047, | |
| "eval_precision": 0.6486719468870186, | |
| "eval_recall": 0.691747572815534, | |
| "eval_runtime": 22.6395, | |
| "eval_samples_per_second": 127.388, | |
| "eval_steps_per_second": 15.946, | |
| "step": 321 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "grad_norm": 23.715213775634766, | |
| "learning_rate": 3.6872475476053084e-05, | |
| "loss": 1.1815, | |
| "step": 642 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_accuracy": 0.7558945908460472, | |
| "eval_f1": 0.7314126926737505, | |
| "eval_loss": 0.6948704123497009, | |
| "eval_precision": 0.7314749711745714, | |
| "eval_recall": 0.7558945908460472, | |
| "eval_runtime": 22.8506, | |
| "eval_samples_per_second": 126.211, | |
| "eval_steps_per_second": 15.798, | |
| "step": 642 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "grad_norm": 5.459173202514648, | |
| "learning_rate": 5.539526832083093e-05, | |
| "loss": 1.0605, | |
| "step": 963 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_accuracy": 0.771497919556172, | |
| "eval_f1": 0.752980330371167, | |
| "eval_loss": 0.6212790012359619, | |
| "eval_precision": 0.7648636119210427, | |
| "eval_recall": 0.771497919556172, | |
| "eval_runtime": 22.9013, | |
| "eval_samples_per_second": 125.932, | |
| "eval_steps_per_second": 15.763, | |
| "step": 963 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "grad_norm": 23.622180938720703, | |
| "learning_rate": 7.391806116560877e-05, | |
| "loss": 1.0147, | |
| "step": 1284 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_accuracy": 0.7732316227461858, | |
| "eval_f1": 0.773441625340043, | |
| "eval_loss": 0.5767605304718018, | |
| "eval_precision": 0.7806580682757648, | |
| "eval_recall": 0.7732316227461858, | |
| "eval_runtime": 22.9855, | |
| "eval_samples_per_second": 125.47, | |
| "eval_steps_per_second": 15.706, | |
| "step": 1284 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "grad_norm": 9.584253311157227, | |
| "learning_rate": 9.244085401038661e-05, | |
| "loss": 0.93, | |
| "step": 1605 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_accuracy": 0.7586685159500693, | |
| "eval_f1": 0.7661547242155936, | |
| "eval_loss": 0.657228946685791, | |
| "eval_precision": 0.7939813909551887, | |
| "eval_recall": 0.7586685159500693, | |
| "eval_runtime": 22.8929, | |
| "eval_samples_per_second": 125.978, | |
| "eval_steps_per_second": 15.769, | |
| "step": 1605 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "grad_norm": 13.967341423034668, | |
| "learning_rate": 9.984553168214161e-05, | |
| "loss": 0.9793, | |
| "step": 1926 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_accuracy": 0.7701109570041609, | |
| "eval_f1": 0.7742181002637638, | |
| "eval_loss": 0.61651611328125, | |
| "eval_precision": 0.7940357585259091, | |
| "eval_recall": 0.7701109570041609, | |
| "eval_runtime": 22.9308, | |
| "eval_samples_per_second": 125.77, | |
| "eval_steps_per_second": 15.743, | |
| "step": 1926 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "grad_norm": 12.87781810760498, | |
| "learning_rate": 9.889061131437471e-05, | |
| "loss": 0.8662, | |
| "step": 2247 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "eval_accuracy": 0.723994452149792, | |
| "eval_f1": 0.745594196112459, | |
| "eval_loss": 0.6534591913223267, | |
| "eval_precision": 0.8097504651538117, | |
| "eval_recall": 0.723994452149792, | |
| "eval_runtime": 24.6973, | |
| "eval_samples_per_second": 116.774, | |
| "eval_steps_per_second": 14.617, | |
| "step": 2247 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "grad_norm": 3.272010326385498, | |
| "learning_rate": 9.707265436104638e-05, | |
| "loss": 0.7767, | |
| "step": 2568 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_accuracy": 0.7565880721220527, | |
| "eval_f1": 0.773329540133459, | |
| "eval_loss": 0.5812659859657288, | |
| "eval_precision": 0.8124371295597024, | |
| "eval_recall": 0.7565880721220527, | |
| "eval_runtime": 24.6374, | |
| "eval_samples_per_second": 117.058, | |
| "eval_steps_per_second": 14.653, | |
| "step": 2568 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "grad_norm": 12.090767860412598, | |
| "learning_rate": 9.44253127296151e-05, | |
| "loss": 0.7572, | |
| "step": 2889 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "eval_accuracy": 0.8144937586685159, | |
| "eval_f1": 0.811409191547864, | |
| "eval_loss": 0.5385076403617859, | |
| "eval_precision": 0.8131138362731425, | |
| "eval_recall": 0.8144937586685159, | |
| "eval_runtime": 23.005, | |
| "eval_samples_per_second": 125.364, | |
| "eval_steps_per_second": 15.692, | |
| "step": 2889 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "grad_norm": 9.961050987243652, | |
| "learning_rate": 9.099523058358976e-05, | |
| "loss": 0.7003, | |
| "step": 3210 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "eval_accuracy": 0.8027045769764216, | |
| "eval_f1": 0.809293314931955, | |
| "eval_loss": 0.5354512929916382, | |
| "eval_precision": 0.8276170557646749, | |
| "eval_recall": 0.8027045769764216, | |
| "eval_runtime": 23.6688, | |
| "eval_samples_per_second": 121.848, | |
| "eval_steps_per_second": 15.252, | |
| "step": 3210 | |
| }, | |
| { | |
| "epoch": 11.0, | |
| "grad_norm": 25.105077743530273, | |
| "learning_rate": 8.684284338417735e-05, | |
| "loss": 0.6316, | |
| "step": 3531 | |
| }, | |
| { | |
| "epoch": 11.0, | |
| "eval_accuracy": 0.765256588072122, | |
| "eval_f1": 0.7816439176174772, | |
| "eval_loss": 0.6284907460212708, | |
| "eval_precision": 0.8322254365589742, | |
| "eval_recall": 0.765256588072122, | |
| "eval_runtime": 23.5967, | |
| "eval_samples_per_second": 122.22, | |
| "eval_steps_per_second": 15.299, | |
| "step": 3531 | |
| }, | |
| { | |
| "epoch": 12.0, | |
| "grad_norm": 10.844632148742676, | |
| "learning_rate": 8.204131306302357e-05, | |
| "loss": 0.5723, | |
| "step": 3852 | |
| }, | |
| { | |
| "epoch": 12.0, | |
| "eval_accuracy": 0.8016643550624133, | |
| "eval_f1": 0.8104537685542633, | |
| "eval_loss": 0.5774895548820496, | |
| "eval_precision": 0.8279128953158427, | |
| "eval_recall": 0.8016643550624133, | |
| "eval_runtime": 23.4985, | |
| "eval_samples_per_second": 122.731, | |
| "eval_steps_per_second": 15.363, | |
| "step": 3852 | |
| }, | |
| { | |
| "epoch": 13.0, | |
| "grad_norm": 11.590338706970215, | |
| "learning_rate": 7.667523896413962e-05, | |
| "loss": 0.4717, | |
| "step": 4173 | |
| }, | |
| { | |
| "epoch": 13.0, | |
| "eval_accuracy": 0.8349514563106796, | |
| "eval_f1": 0.8349682338681513, | |
| "eval_loss": 0.534845769405365, | |
| "eval_precision": 0.8390873801645569, | |
| "eval_recall": 0.8349514563106796, | |
| "eval_runtime": 23.9213, | |
| "eval_samples_per_second": 120.562, | |
| "eval_steps_per_second": 15.091, | |
| "step": 4173 | |
| }, | |
| { | |
| "epoch": 14.0, | |
| "grad_norm": 3.6801199913024902, | |
| "learning_rate": 7.083916726724684e-05, | |
| "loss": 0.4472, | |
| "step": 4494 | |
| }, | |
| { | |
| "epoch": 14.0, | |
| "eval_accuracy": 0.8238557558945908, | |
| "eval_f1": 0.8298970188071358, | |
| "eval_loss": 0.5469394326210022, | |
| "eval_precision": 0.844241279806222, | |
| "eval_recall": 0.8238557558945908, | |
| "eval_runtime": 23.6076, | |
| "eval_samples_per_second": 122.164, | |
| "eval_steps_per_second": 15.292, | |
| "step": 4494 | |
| }, | |
| { | |
| "epoch": 15.0, | |
| "grad_norm": 2.446948528289795, | |
| "learning_rate": 6.46557084486047e-05, | |
| "loss": 0.3764, | |
| "step": 4815 | |
| }, | |
| { | |
| "epoch": 15.0, | |
| "eval_accuracy": 0.8290568654646324, | |
| "eval_f1": 0.8355171455268711, | |
| "eval_loss": 0.5737930536270142, | |
| "eval_precision": 0.8501341497772392, | |
| "eval_recall": 0.8290568654646324, | |
| "eval_runtime": 23.7441, | |
| "eval_samples_per_second": 121.462, | |
| "eval_steps_per_second": 15.204, | |
| "step": 4815 | |
| }, | |
| { | |
| "epoch": 16.0, | |
| "grad_norm": 6.043246269226074, | |
| "learning_rate": 5.8195220793532045e-05, | |
| "loss": 0.3346, | |
| "step": 5136 | |
| }, | |
| { | |
| "epoch": 16.0, | |
| "eval_accuracy": 0.8436199722607489, | |
| "eval_f1": 0.8461452975648701, | |
| "eval_loss": 0.5367853045463562, | |
| "eval_precision": 0.851244402940535, | |
| "eval_recall": 0.8436199722607489, | |
| "eval_runtime": 23.9285, | |
| "eval_samples_per_second": 120.526, | |
| "eval_steps_per_second": 15.087, | |
| "step": 5136 | |
| }, | |
| { | |
| "epoch": 17.0, | |
| "grad_norm": 15.190231323242188, | |
| "learning_rate": 5.15903395270923e-05, | |
| "loss": 0.2826, | |
| "step": 5457 | |
| }, | |
| { | |
| "epoch": 17.0, | |
| "eval_accuracy": 0.8474341192787794, | |
| "eval_f1": 0.8467916215502757, | |
| "eval_loss": 0.5510370135307312, | |
| "eval_precision": 0.8489075342171026, | |
| "eval_recall": 0.8474341192787794, | |
| "eval_runtime": 23.9505, | |
| "eval_samples_per_second": 120.415, | |
| "eval_steps_per_second": 15.073, | |
| "step": 5457 | |
| }, | |
| { | |
| "epoch": 18.0, | |
| "grad_norm": 24.944795608520508, | |
| "learning_rate": 4.4978023449711664e-05, | |
| "loss": 0.2659, | |
| "step": 5778 | |
| }, | |
| { | |
| "epoch": 18.0, | |
| "eval_accuracy": 0.8547156726768377, | |
| "eval_f1": 0.8549199048922214, | |
| "eval_loss": 0.546658456325531, | |
| "eval_precision": 0.8559700691467993, | |
| "eval_recall": 0.8547156726768377, | |
| "eval_runtime": 23.4765, | |
| "eval_samples_per_second": 122.846, | |
| "eval_steps_per_second": 15.377, | |
| "step": 5778 | |
| }, | |
| { | |
| "epoch": 19.0, | |
| "grad_norm": 0.11274675279855728, | |
| "learning_rate": 3.843351058509752e-05, | |
| "loss": 0.2545, | |
| "step": 6099 | |
| }, | |
| { | |
| "epoch": 19.0, | |
| "eval_accuracy": 0.8432732316227461, | |
| "eval_f1": 0.8487015901360936, | |
| "eval_loss": 0.6155611872673035, | |
| "eval_precision": 0.8617407571979913, | |
| "eval_recall": 0.8432732316227461, | |
| "eval_runtime": 22.4381, | |
| "eval_samples_per_second": 128.532, | |
| "eval_steps_per_second": 16.089, | |
| "step": 6099 | |
| }, | |
| { | |
| "epoch": 20.0, | |
| "grad_norm": 0.5206512808799744, | |
| "learning_rate": 3.209279054309716e-05, | |
| "loss": 0.2123, | |
| "step": 6420 | |
| }, | |
| { | |
| "epoch": 20.0, | |
| "eval_accuracy": 0.8429264909847434, | |
| "eval_f1": 0.8426622560864744, | |
| "eval_loss": 0.6871052384376526, | |
| "eval_precision": 0.8498556529498513, | |
| "eval_recall": 0.8429264909847434, | |
| "eval_runtime": 22.4536, | |
| "eval_samples_per_second": 128.443, | |
| "eval_steps_per_second": 16.078, | |
| "step": 6420 | |
| }, | |
| { | |
| "epoch": 21.0, | |
| "grad_norm": 6.716039657592773, | |
| "learning_rate": 2.6067582030088143e-05, | |
| "loss": 0.1655, | |
| "step": 6741 | |
| }, | |
| { | |
| "epoch": 21.0, | |
| "eval_accuracy": 0.8609570041608876, | |
| "eval_f1": 0.856691150022433, | |
| "eval_loss": 0.6139292120933533, | |
| "eval_precision": 0.8551833143301755, | |
| "eval_recall": 0.8609570041608876, | |
| "eval_runtime": 22.975, | |
| "eval_samples_per_second": 125.528, | |
| "eval_steps_per_second": 15.713, | |
| "step": 6741 | |
| }, | |
| { | |
| "epoch": 22.0, | |
| "grad_norm": 108.25263977050781, | |
| "learning_rate": 2.0464044677195966e-05, | |
| "loss": 0.1246, | |
| "step": 7062 | |
| }, | |
| { | |
| "epoch": 22.0, | |
| "eval_accuracy": 0.8675450762829403, | |
| "eval_f1": 0.8676542597333672, | |
| "eval_loss": 0.6129117012023926, | |
| "eval_precision": 0.8680620732919173, | |
| "eval_recall": 0.8675450762829403, | |
| "eval_runtime": 22.6536, | |
| "eval_samples_per_second": 127.309, | |
| "eval_steps_per_second": 15.936, | |
| "step": 7062 | |
| }, | |
| { | |
| "epoch": 23.0, | |
| "grad_norm": 2.2161059379577637, | |
| "learning_rate": 1.538090858763548e-05, | |
| "loss": 0.1394, | |
| "step": 7383 | |
| }, | |
| { | |
| "epoch": 23.0, | |
| "eval_accuracy": 0.8713592233009708, | |
| "eval_f1": 0.867730828379127, | |
| "eval_loss": 0.6522772908210754, | |
| "eval_precision": 0.8674975891372388, | |
| "eval_recall": 0.8713592233009708, | |
| "eval_runtime": 22.7054, | |
| "eval_samples_per_second": 127.018, | |
| "eval_steps_per_second": 15.899, | |
| "step": 7383 | |
| }, | |
| { | |
| "epoch": 23.0, | |
| "step": 7383, | |
| "total_flos": 9.140124804986585e+18, | |
| "train_loss": 0.6221655089083716, | |
| "train_runtime": 2787.435, | |
| "train_samples_per_second": 183.968, | |
| "train_steps_per_second": 11.516 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 32100, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 100, | |
| "save_steps": 500, | |
| "total_flos": 9.140124804986585e+18, | |
| "train_batch_size": 16, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |