End of training

README.md

@@ -17,8 +17,8 @@ should probably proofread and complete it, then remove this comment. -->
 
 This model is a fine-tuned version of [microsoft/swin-tiny-patch4-window7-224](https://huggingface.co/microsoft/swin-tiny-patch4-window7-224) on an unknown dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.3035
-- Accuracy: 0.8793
+- Loss: 0.3718
+- Accuracy: 0.8966
 
 ## Model description
 

all_results.json

@@ -0,0 +1,13 @@
+{
+ "epoch": 37.65,
+ "eval_accuracy": 0.896551724137931,
+ "eval_loss": 0.3717654049396515,
+ "eval_runtime": 0.6072,
+ "eval_samples_per_second": 95.521,
+ "eval_steps_per_second": 3.294,
+ "total_flos": 4.870057509170381e+17,
+ "train_loss": 0.5200165078043938,
+ "train_runtime": 456.0431,
+ "train_samples_per_second": 45.61,
+ "train_steps_per_second": 0.351
+}

eval_results.json

@@ -0,0 +1,8 @@
+{
+ "epoch": 37.65,
+ "eval_accuracy": 0.896551724137931,
+ "eval_loss": 0.3717654049396515,
+ "eval_runtime": 0.6072,
+ "eval_samples_per_second": 95.521,
+ "eval_steps_per_second": 3.294
+}

runs/Dec05_23-14-42_2e18f523b17c/events.out.tfevents.1701818581.2e18f523b17c.3067.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:33793eb5623af834896cf2a2d62087852de48c71ca53564b07fd56aa1ab2a832
+size 411

train_results.json

@@ -0,0 +1,8 @@
+{
+ "epoch": 37.65,
+ "total_flos": 4.870057509170381e+17,
+ "train_loss": 0.5200165078043938,
+ "train_runtime": 456.0431,
+ "train_samples_per_second": 45.61,
+ "train_steps_per_second": 0.351
+}

trainer_state.json

@@ -0,0 +1,466 @@
+{
+ "best_metric": 0.896551724137931,
+ "best_model_checkpoint": "swin-tiny-patch4-window7-224-dmae-va-da2-40/checkpoint-110",
+ "epoch": 37.64705882352941,
+ "eval_steps": 500,
+ "global_step": 160,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "epoch": 0.94,
+ "eval_accuracy": 0.3620689655172414,
+ "eval_loss": 1.3102844953536987,
+ "eval_runtime": 0.5859,
+ "eval_samples_per_second": 99.0,
+ "eval_steps_per_second": 3.414,
+ "step": 4
+ },
+ {
+ "epoch": 1.88,
+ "eval_accuracy": 0.3275862068965517,
+ "eval_loss": 1.2844481468200684,
+ "eval_runtime": 0.541,
+ "eval_samples_per_second": 107.217,
+ "eval_steps_per_second": 3.697,
+ "step": 8
+ },
+ {
+ "epoch": 2.35,
+ "learning_rate": 3.125e-05,
+ "loss": 1.3144,
+ "step": 10
+ },
+ {
+ "epoch": 2.82,
+ "eval_accuracy": 0.41379310344827586,
+ "eval_loss": 1.2574777603149414,
+ "eval_runtime": 0.5361,
+ "eval_samples_per_second": 108.186,
+ "eval_steps_per_second": 3.731,
+ "step": 12
+ },
+ {
+ "epoch": 4.0,
+ "eval_accuracy": 0.5,
+ "eval_loss": 1.1180709600448608,
+ "eval_runtime": 0.5531,
+ "eval_samples_per_second": 104.867,
+ "eval_steps_per_second": 3.616,
+ "step": 17
+ },
+ {
+ "epoch": 4.71,
+ "learning_rate": 4.8611111111111115e-05,
+ "loss": 1.1723,
+ "step": 20
+ },
+ {
+ "epoch": 4.94,
+ "eval_accuracy": 0.5344827586206896,
+ "eval_loss": 1.0054141283035278,
+ "eval_runtime": 0.7262,
+ "eval_samples_per_second": 79.866,
+ "eval_steps_per_second": 2.754,
+ "step": 21
+ },
+ {
+ "epoch": 5.88,
+ "eval_accuracy": 0.5689655172413793,
+ "eval_loss": 0.8771402835845947,
+ "eval_runtime": 0.632,
+ "eval_samples_per_second": 91.766,
+ "eval_steps_per_second": 3.164,
+ "step": 25
+ },
+ {
+ "epoch": 6.82,
+ "eval_accuracy": 0.7068965517241379,
+ "eval_loss": 0.794106125831604,
+ "eval_runtime": 0.5449,
+ "eval_samples_per_second": 106.436,
+ "eval_steps_per_second": 3.67,
+ "step": 29
+ },
+ {
+ "epoch": 7.06,
+ "learning_rate": 4.5138888888888894e-05,
+ "loss": 0.9243,
+ "step": 30
+ },
+ {
+ "epoch": 8.0,
+ "eval_accuracy": 0.6724137931034483,
+ "eval_loss": 0.7037971615791321,
+ "eval_runtime": 0.5421,
+ "eval_samples_per_second": 106.985,
+ "eval_steps_per_second": 3.689,
+ "step": 34
+ },
+ {
+ "epoch": 8.94,
+ "eval_accuracy": 0.7241379310344828,
+ "eval_loss": 0.7341203093528748,
+ "eval_runtime": 0.5497,
+ "eval_samples_per_second": 105.515,
+ "eval_steps_per_second": 3.638,
+ "step": 38
+ },
+ {
+ "epoch": 9.41,
+ "learning_rate": 4.166666666666667e-05,
+ "loss": 0.7394,
+ "step": 40
+ },
+ {
+ "epoch": 9.88,
+ "eval_accuracy": 0.7413793103448276,
+ "eval_loss": 0.6256024837493896,
+ "eval_runtime": 0.544,
+ "eval_samples_per_second": 106.617,
+ "eval_steps_per_second": 3.676,
+ "step": 42
+ },
+ {
+ "epoch": 10.82,
+ "eval_accuracy": 0.7586206896551724,
+ "eval_loss": 0.6058707237243652,
+ "eval_runtime": 0.6912,
+ "eval_samples_per_second": 83.916,
+ "eval_steps_per_second": 2.894,
+ "step": 46
+ },
+ {
+ "epoch": 11.76,
+ "learning_rate": 3.8194444444444444e-05,
+ "loss": 0.627,
+ "step": 50
+ },
+ {
+ "epoch": 12.0,
+ "eval_accuracy": 0.7413793103448276,
+ "eval_loss": 0.5662473440170288,
+ "eval_runtime": 0.7222,
+ "eval_samples_per_second": 80.313,
+ "eval_steps_per_second": 2.769,
+ "step": 51
+ },
+ {
+ "epoch": 12.94,
+ "eval_accuracy": 0.7931034482758621,
+ "eval_loss": 0.5383967757225037,
+ "eval_runtime": 0.5429,
+ "eval_samples_per_second": 106.842,
+ "eval_steps_per_second": 3.684,
+ "step": 55
+ },
+ {
+ "epoch": 13.88,
+ "eval_accuracy": 0.7758620689655172,
+ "eval_loss": 0.5630860328674316,
+ "eval_runtime": 0.5402,
+ "eval_samples_per_second": 107.371,
+ "eval_steps_per_second": 3.702,
+ "step": 59
+ },
+ {
+ "epoch": 14.12,
+ "learning_rate": 3.472222222222222e-05,
+ "loss": 0.5026,
+ "step": 60
+ },
+ {
+ "epoch": 14.82,
+ "eval_accuracy": 0.8103448275862069,
+ "eval_loss": 0.49379998445510864,
+ "eval_runtime": 0.5445,
+ "eval_samples_per_second": 106.521,
+ "eval_steps_per_second": 3.673,
+ "step": 63
+ },
+ {
+ "epoch": 16.0,
+ "eval_accuracy": 0.8448275862068966,
+ "eval_loss": 0.4768355190753937,
+ "eval_runtime": 0.5424,
+ "eval_samples_per_second": 106.928,
+ "eval_steps_per_second": 3.687,
+ "step": 68
+ },
+ {
+ "epoch": 16.47,
+ "learning_rate": 3.125e-05,
+ "loss": 0.4284,
+ "step": 70
+ },
+ {
+ "epoch": 16.94,
+ "eval_accuracy": 0.8620689655172413,
+ "eval_loss": 0.4082733988761902,
+ "eval_runtime": 0.5537,
+ "eval_samples_per_second": 104.745,
+ "eval_steps_per_second": 3.612,
+ "step": 72
+ },
+ {
+ "epoch": 17.88,
+ "eval_accuracy": 0.8275862068965517,
+ "eval_loss": 0.3942280411720276,
+ "eval_runtime": 0.5427,
+ "eval_samples_per_second": 106.871,
+ "eval_steps_per_second": 3.685,
+ "step": 76
+ },
+ {
+ "epoch": 18.82,
+ "learning_rate": 2.777777777777778e-05,
+ "loss": 0.3922,
+ "step": 80
+ },
+ {
+ "epoch": 18.82,
+ "eval_accuracy": 0.8620689655172413,
+ "eval_loss": 0.37597107887268066,
+ "eval_runtime": 0.6952,
+ "eval_samples_per_second": 83.423,
+ "eval_steps_per_second": 2.877,
+ "step": 80
+ },
+ {
+ "epoch": 20.0,
+ "eval_accuracy": 0.8275862068965517,
+ "eval_loss": 0.4109302759170532,
+ "eval_runtime": 0.548,
+ "eval_samples_per_second": 105.841,
+ "eval_steps_per_second": 3.65,
+ "step": 85
+ },
+ {
+ "epoch": 20.94,
+ "eval_accuracy": 0.8620689655172413,
+ "eval_loss": 0.38523542881011963,
+ "eval_runtime": 0.5465,
+ "eval_samples_per_second": 106.125,
+ "eval_steps_per_second": 3.659,
+ "step": 89
+ },
+ {
+ "epoch": 21.18,
+ "learning_rate": 2.4305555555555558e-05,
+ "loss": 0.3927,
+ "step": 90
+ },
+ {
+ "epoch": 21.88,
+ "eval_accuracy": 0.8620689655172413,
+ "eval_loss": 0.3660654127597809,
+ "eval_runtime": 0.5512,
+ "eval_samples_per_second": 105.229,
+ "eval_steps_per_second": 3.629,
+ "step": 93
+ },
+ {
+ "epoch": 22.82,
+ "eval_accuracy": 0.8793103448275862,
+ "eval_loss": 0.3930301368236542,
+ "eval_runtime": 0.5482,
+ "eval_samples_per_second": 105.792,
+ "eval_steps_per_second": 3.648,
+ "step": 97
+ },
+ {
+ "epoch": 23.53,
+ "learning_rate": 2.0833333333333336e-05,
+ "loss": 0.3338,
+ "step": 100
+ },
+ {
+ "epoch": 24.0,
+ "eval_accuracy": 0.8620689655172413,
+ "eval_loss": 0.35855162143707275,
+ "eval_runtime": 0.5507,
+ "eval_samples_per_second": 105.323,
+ "eval_steps_per_second": 3.632,
+ "step": 102
+ },
+ {
+ "epoch": 24.94,
+ "eval_accuracy": 0.8620689655172413,
+ "eval_loss": 0.36067575216293335,
+ "eval_runtime": 0.7414,
+ "eval_samples_per_second": 78.233,
+ "eval_steps_per_second": 2.698,
+ "step": 106
+ },
+ {
+ "epoch": 25.88,
+ "learning_rate": 1.736111111111111e-05,
+ "loss": 0.2698,
+ "step": 110
+ },
+ {
+ "epoch": 25.88,
+ "eval_accuracy": 0.896551724137931,
+ "eval_loss": 0.3717654049396515,
+ "eval_runtime": 0.619,
+ "eval_samples_per_second": 93.702,
+ "eval_steps_per_second": 3.231,
+ "step": 110
+ },
+ {
+ "epoch": 26.82,
+ "eval_accuracy": 0.8793103448275862,
+ "eval_loss": 0.35024696588516235,
+ "eval_runtime": 0.5475,
+ "eval_samples_per_second": 105.944,
+ "eval_steps_per_second": 3.653,
+ "step": 114
+ },
+ {
+ "epoch": 28.0,
+ "eval_accuracy": 0.8448275862068966,
+ "eval_loss": 0.3614311218261719,
+ "eval_runtime": 0.5691,
+ "eval_samples_per_second": 101.924,
+ "eval_steps_per_second": 3.515,
+ "step": 119
+ },
+ {
+ "epoch": 28.24,
+ "learning_rate": 1.388888888888889e-05,
+ "loss": 0.2707,
+ "step": 120
+ },
+ {
+ "epoch": 28.94,
+ "eval_accuracy": 0.8620689655172413,
+ "eval_loss": 0.3537382185459137,
+ "eval_runtime": 0.5442,
+ "eval_samples_per_second": 106.586,
+ "eval_steps_per_second": 3.675,
+ "step": 123
+ },
+ {
+ "epoch": 29.88,
+ "eval_accuracy": 0.8793103448275862,
+ "eval_loss": 0.354471892118454,
+ "eval_runtime": 0.5491,
+ "eval_samples_per_second": 105.623,
+ "eval_steps_per_second": 3.642,
+ "step": 127
+ },
+ {
+ "epoch": 30.59,
+ "learning_rate": 1.0416666666666668e-05,
+ "loss": 0.2622,
+ "step": 130
+ },
+ {
+ "epoch": 30.82,
+ "eval_accuracy": 0.8793103448275862,
+ "eval_loss": 0.32017067074775696,
+ "eval_runtime": 0.6084,
+ "eval_samples_per_second": 95.331,
+ "eval_steps_per_second": 3.287,
+ "step": 131
+ },
+ {
+ "epoch": 32.0,
+ "eval_accuracy": 0.8793103448275862,
+ "eval_loss": 0.3125433623790741,
+ "eval_runtime": 0.6933,
+ "eval_samples_per_second": 83.663,
+ "eval_steps_per_second": 2.885,
+ "step": 136
+ },
+ {
+ "epoch": 32.94,
+ "learning_rate": 6.944444444444445e-06,
+ "loss": 0.2499,
+ "step": 140
+ },
+ {
+ "epoch": 32.94,
+ "eval_accuracy": 0.8620689655172413,
+ "eval_loss": 0.3215100169181824,
+ "eval_runtime": 0.5383,
+ "eval_samples_per_second": 107.744,
+ "eval_steps_per_second": 3.715,
+ "step": 140
+ },
+ {
+ "epoch": 33.88,
+ "eval_accuracy": 0.8620689655172413,
+ "eval_loss": 0.3194180428981781,
+ "eval_runtime": 0.5395,
+ "eval_samples_per_second": 107.502,
+ "eval_steps_per_second": 3.707,
+ "step": 144
+ },
+ {
+ "epoch": 34.82,
+ "eval_accuracy": 0.8793103448275862,
+ "eval_loss": 0.3065814971923828,
+ "eval_runtime": 0.5319,
+ "eval_samples_per_second": 109.036,
+ "eval_steps_per_second": 3.76,
+ "step": 148
+ },
+ {
+ "epoch": 35.29,
+ "learning_rate": 3.4722222222222224e-06,
+ "loss": 0.2157,
+ "step": 150
+ },
+ {
+ "epoch": 36.0,
+ "eval_accuracy": 0.8793103448275862,
+ "eval_loss": 0.3023557662963867,
+ "eval_runtime": 0.5521,
+ "eval_samples_per_second": 105.047,
+ "eval_steps_per_second": 3.622,
+ "step": 153
+ },
+ {
+ "epoch": 36.94,
+ "eval_accuracy": 0.8793103448275862,
+ "eval_loss": 0.30368945002555847,
+ "eval_runtime": 0.5271,
+ "eval_samples_per_second": 110.045,
+ "eval_steps_per_second": 3.795,
+ "step": 157
+ },
+ {
+ "epoch": 37.65,
+ "learning_rate": 0.0,
+ "loss": 0.2248,
+ "step": 160
+ },
+ {
+ "epoch": 37.65,
+ "eval_accuracy": 0.8793103448275862,
+ "eval_loss": 0.303451269865036,
+ "eval_runtime": 0.5525,
+ "eval_samples_per_second": 104.979,
+ "eval_steps_per_second": 3.62,
+ "step": 160
+ },
+ {
+ "epoch": 37.65,
+ "step": 160,
+ "total_flos": 4.870057509170381e+17,
+ "train_loss": 0.5200165078043938,
+ "train_runtime": 456.0431,
+ "train_samples_per_second": 45.61,
+ "train_steps_per_second": 0.351
+ }
+ ],
+ "logging_steps": 10,
+ "max_steps": 160,
+ "num_train_epochs": 40,
+ "save_steps": 500,
+ "total_flos": 4.870057509170381e+17,
+ "trial_name": null,
+ "trial_params": null
+}