NMM_Classification_base_V10 / trainer_state.json
djbp's picture
End of training
594a617 verified
raw
history blame
9.04 kB
{
"best_metric": 0.8677685950413223,
"best_model_checkpoint": "swin-base-patch4-window7-224-in22k-MM/NMM_Classification_base_V10/checkpoint-76",
"epoch": 6.885245901639344,
"eval_steps": 500,
"global_step": 105,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.6557377049180327,
"grad_norm": 4.473873615264893,
"learning_rate": 4.545454545454546e-05,
"loss": 0.8448,
"step": 10
},
{
"epoch": 0.9836065573770492,
"step": 15,
"train_accuracy": 0.8035137579124144,
"train_auc_class_0": 0.9088436086296741,
"train_auc_class_1": 0.8379446511466175,
"train_auc_class_2": 0.8758602364377229,
"train_auc_overall": 0.8742161654046715,
"train_loss": 0.49683675169944763,
"train_runtime": 777.5048,
"train_samples_per_second": 9.956,
"train_steps_per_second": 0.078
},
{
"epoch": 0.9836065573770492,
"eval_accuracy": 0.8362603305785123,
"eval_auc_class_0": 0.9403517055860806,
"eval_auc_class_1": 0.8605572313411073,
"eval_auc_class_2": 0.8997462510986076,
"eval_auc_overall": 0.9002183960085985,
"eval_loss": 0.44267916679382324,
"eval_runtime": 267.5178,
"eval_samples_per_second": 7.237,
"eval_steps_per_second": 0.06,
"step": 15
},
{
"epoch": 1.3114754098360657,
"grad_norm": 2.1386613845825195,
"learning_rate": 4.5212765957446815e-05,
"loss": 0.512,
"step": 20
},
{
"epoch": 1.9672131147540983,
"grad_norm": 1.435038685798645,
"learning_rate": 3.9893617021276594e-05,
"loss": 0.4434,
"step": 30
},
{
"epoch": 1.9672131147540983,
"step": 30,
"train_accuracy": 0.840201524350859,
"train_auc_class_0": 0.9414490637347536,
"train_auc_class_1": 0.8997057391727663,
"train_auc_class_2": 0.9139346212557102,
"train_auc_overall": 0.9183631413877434,
"train_loss": 0.4013007879257202,
"train_runtime": 714.2886,
"train_samples_per_second": 10.837,
"train_steps_per_second": 0.085
},
{
"epoch": 1.9672131147540983,
"eval_accuracy": 0.8445247933884298,
"eval_auc_class_0": 0.9560139079670331,
"eval_auc_class_1": 0.9042625890669138,
"eval_auc_class_2": 0.9195875061825728,
"eval_auc_overall": 0.9266213344055066,
"eval_loss": 0.3872886896133423,
"eval_runtime": 214.551,
"eval_samples_per_second": 9.023,
"eval_steps_per_second": 0.075,
"step": 30
},
{
"epoch": 2.6229508196721314,
"grad_norm": 2.210986852645874,
"learning_rate": 3.4574468085106386e-05,
"loss": 0.3979,
"step": 40
},
{
"epoch": 2.9508196721311473,
"step": 45,
"train_accuracy": 0.8537656633509882,
"train_auc_class_0": 0.9565036613124893,
"train_auc_class_1": 0.9187875975798461,
"train_auc_class_2": 0.9285740636485835,
"train_auc_overall": 0.9346217741803063,
"train_loss": 0.3611737787723541,
"train_runtime": 704.8242,
"train_samples_per_second": 10.983,
"train_steps_per_second": 0.087
},
{
"epoch": 2.9508196721311473,
"eval_accuracy": 0.8584710743801653,
"eval_auc_class_0": 0.9593907394688643,
"eval_auc_class_1": 0.9080973087319308,
"eval_auc_class_2": 0.9256975615561118,
"eval_auc_overall": 0.931061869918969,
"eval_loss": 0.36146315932273865,
"eval_runtime": 207.9281,
"eval_samples_per_second": 9.311,
"eval_steps_per_second": 0.077,
"step": 45
},
{
"epoch": 3.278688524590164,
"grad_norm": 2.5246520042419434,
"learning_rate": 2.925531914893617e-05,
"loss": 0.3733,
"step": 50
},
{
"epoch": 3.9344262295081966,
"grad_norm": 2.381304979324341,
"learning_rate": 2.393617021276596e-05,
"loss": 0.3593,
"step": 60
},
{
"epoch": 4.0,
"step": 61,
"train_accuracy": 0.865779615036817,
"train_auc_class_0": 0.9630790814865717,
"train_auc_class_1": 0.9318475950588521,
"train_auc_class_2": 0.9373109719652593,
"train_auc_overall": 0.9440792161702277,
"train_loss": 0.33730149269104004,
"train_runtime": 698.4574,
"train_samples_per_second": 11.083,
"train_steps_per_second": 0.087
},
{
"epoch": 4.0,
"eval_accuracy": 0.862603305785124,
"eval_auc_class_0": 0.9629407051282051,
"eval_auc_class_1": 0.9201237904963373,
"eval_auc_class_2": 0.9316191956643276,
"eval_auc_overall": 0.9382278970962901,
"eval_loss": 0.34607622027397156,
"eval_runtime": 206.0574,
"eval_samples_per_second": 9.395,
"eval_steps_per_second": 0.078,
"step": 61
},
{
"epoch": 4.590163934426229,
"grad_norm": 2.1240530014038086,
"learning_rate": 1.8617021276595745e-05,
"loss": 0.3473,
"step": 70
},
{
"epoch": 4.983606557377049,
"step": 76,
"train_accuracy": 0.8688799896654179,
"train_auc_class_0": 0.9654619767508447,
"train_auc_class_1": 0.9391473390475861,
"train_auc_class_2": 0.9405490167752112,
"train_auc_overall": 0.9483861108578807,
"train_loss": 0.32362109422683716,
"train_runtime": 706.7637,
"train_samples_per_second": 10.953,
"train_steps_per_second": 0.086
},
{
"epoch": 4.983606557377049,
"eval_accuracy": 0.8677685950413223,
"eval_auc_class_0": 0.9631310096153847,
"eval_auc_class_1": 0.9189268008165645,
"eval_auc_class_2": 0.9322997435239486,
"eval_auc_overall": 0.938119184651966,
"eval_loss": 0.34411147236824036,
"eval_runtime": 207.2323,
"eval_samples_per_second": 9.342,
"eval_steps_per_second": 0.077,
"step": 76
},
{
"epoch": 5.245901639344262,
"grad_norm": 2.395565986633301,
"learning_rate": 1.3297872340425532e-05,
"loss": 0.3381,
"step": 80
},
{
"epoch": 5.901639344262295,
"grad_norm": 1.6295080184936523,
"learning_rate": 7.97872340425532e-06,
"loss": 0.3318,
"step": 90
},
{
"epoch": 5.967213114754099,
"step": 91,
"train_accuracy": 0.8766309262369203,
"train_auc_class_0": 0.9675600734409895,
"train_auc_class_1": 0.9360009866648122,
"train_auc_class_2": 0.9450547274524197,
"train_auc_overall": 0.9495385958527405,
"train_loss": 0.31677091121673584,
"train_runtime": 712.6133,
"train_samples_per_second": 10.863,
"train_steps_per_second": 0.086
},
{
"epoch": 5.967213114754099,
"eval_accuracy": 0.8667355371900827,
"eval_auc_class_0": 0.9631009615384616,
"eval_auc_class_1": 0.916550232215998,
"eval_auc_class_2": 0.9317034893882878,
"eval_auc_overall": 0.9371182277142491,
"eval_loss": 0.34614038467407227,
"eval_runtime": 209.8829,
"eval_samples_per_second": 9.224,
"eval_steps_per_second": 0.076,
"step": 91
},
{
"epoch": 6.557377049180328,
"grad_norm": 3.0411829948425293,
"learning_rate": 2.6595744680851065e-06,
"loss": 0.3147,
"step": 100
},
{
"epoch": 6.885245901639344,
"step": 105,
"train_accuracy": 0.8770184730654954,
"train_auc_class_0": 0.9688894018782569,
"train_auc_class_1": 0.9353519937583671,
"train_auc_class_2": 0.9460088191382243,
"train_auc_overall": 0.9500834049249495,
"train_loss": 0.3142475485801697,
"train_runtime": 690.6226,
"train_samples_per_second": 11.209,
"train_steps_per_second": 0.088
},
{
"epoch": 6.885245901639344,
"eval_accuracy": 0.862603305785124,
"eval_auc_class_0": 0.9628863324175823,
"eval_auc_class_1": 0.9189616223345215,
"eval_auc_class_2": 0.93194521403788,
"eval_auc_overall": 0.937931056263328,
"eval_loss": 0.3451487123966217,
"eval_runtime": 207.6543,
"eval_samples_per_second": 9.323,
"eval_steps_per_second": 0.077,
"step": 105
},
{
"epoch": 6.885245901639344,
"step": 105,
"total_flos": 4.180411553219629e+18,
"train_loss": 0.4214494603020804,
"train_runtime": 12181.9025,
"train_samples_per_second": 4.448,
"train_steps_per_second": 0.009
}
],
"logging_steps": 10,
"max_steps": 105,
"num_input_tokens_seen": 0,
"num_train_epochs": 7,
"save_steps": 500,
"stateful_callbacks": {
"TrainerControl": {
"args": {
"should_epoch_stop": false,
"should_evaluate": false,
"should_log": false,
"should_save": true,
"should_training_stop": true
},
"attributes": {}
}
},
"total_flos": 4.180411553219629e+18,
"train_batch_size": 128,
"trial_name": null,
"trial_params": null
}