🍻 cheers

README.md

@@ -2,6 +2,7 @@
 license: apache-2.0
 base_model: google/vit-base-patch16-224
 tags:
+- image-classification
 - generated_from_trainer
 metrics:
 - accuracy
@@ -18,13 +19,13 @@ should probably proofread and complete it, then remove this comment. -->
 
 # vit-lr-cosine-restarts
 
-This model is a fine-tuned version of [google/vit-base-patch16-224](https://huggingface.co/google/vit-base-patch16-224) on an unknown dataset.
+This model is a fine-tuned version of [google/vit-base-patch16-224](https://huggingface.co/google/vit-base-patch16-224) on the skin-cancer dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.7236
-- Accuracy: 0.8696
-- Precision: 0.8655
-- Recall: 0.8696
-- F1: 0.8661
+- Loss: 0.5117
+- Accuracy: 0.8221
+- Precision: 0.8347
+- Recall: 0.8221
+- F1: 0.8100
 
 ## Model description
 

all_results.json

@@ -1,16 +1,16 @@
 {
     "epoch": 12.0,
-    "eval_accuracy": 0.8262829403606102,
-    "eval_f1": 0.8158028786365703,
-    "eval_loss": 0.49287769198417664,
-    "eval_precision": 0.8255125137698959,
-    "eval_recall": 0.8262829403606102,
-    "eval_runtime": 38.4698,
-    "eval_samples_per_second": 74.968,
-    "eval_steps_per_second": 9.384,
+    "eval_accuracy": 0.8221220527045769,
+    "eval_f1": 0.8099830282273331,
+    "eval_loss": 0.5117006897926331,
+    "eval_precision": 0.8347375649374938,
+    "eval_recall": 0.8221220527045769,
+    "eval_runtime": 38.1667,
+    "eval_samples_per_second": 75.563,
+    "eval_steps_per_second": 9.459,
     "total_flos": 4.768760767819088e+18,
-    "train_loss": 0.17886438471413105,
-    "train_runtime": 1744.4117,
-    "train_samples_per_second": 35.276,
-    "train_steps_per_second": 2.208
+    "train_loss": 0.17854708074838666,
+    "train_runtime": 1730.8935,
+    "train_samples_per_second": 35.552,
+    "train_steps_per_second": 2.225
 }

eval_results.json

@@ -1,11 +1,11 @@
 {
     "epoch": 12.0,
-    "eval_accuracy": 0.8262829403606102,
-    "eval_f1": 0.8158028786365703,
-    "eval_loss": 0.49287769198417664,
-    "eval_precision": 0.8255125137698959,
-    "eval_recall": 0.8262829403606102,
-    "eval_runtime": 38.4698,
-    "eval_samples_per_second": 74.968,
-    "eval_steps_per_second": 9.384
+    "eval_accuracy": 0.8221220527045769,
+    "eval_f1": 0.8099830282273331,
+    "eval_loss": 0.5117006897926331,
+    "eval_precision": 0.8347375649374938,
+    "eval_recall": 0.8221220527045769,
+    "eval_runtime": 38.1667,
+    "eval_samples_per_second": 75.563,
+    "eval_steps_per_second": 9.459
 }

runs/Mar20_17-24-10_f2e1fee5f9b2/events.out.tfevents.1710957259.f2e1fee5f9b2.2315.9

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3c2516b6e3d47288b7b87f5e4159d8e96f87ac82a85f00f034e372df42ea69bb
+size 560

train_results.json

@@ -1,8 +1,8 @@
 {
     "epoch": 12.0,
     "total_flos": 4.768760767819088e+18,
-    "train_loss": 0.17886438471413105,
-    "train_runtime": 1744.4117,
-    "train_samples_per_second": 35.276,
-    "train_steps_per_second": 2.208
+    "train_loss": 0.17854708074838666,
+    "train_runtime": 1730.8935,
+    "train_samples_per_second": 35.552,
+    "train_steps_per_second": 2.225
 }

trainer_state.json

@@ -1,5 +1,5 @@
 {
-  "best_metric": 0.49287769198417664,
+  "best_metric": 0.5117006897926331,
   "best_model_checkpoint": "./vit-lr-cosine-restarts/checkpoint-642",
   "epoch": 12.0,
   "eval_steps": 500,
@@ -10,240 +10,240 @@
   "log_history": [
     {
       "epoch": 1.0,
-      "grad_norm": 6.734107971191406,
+      "grad_norm": 7.814958572387695,
       "learning_rate": 4.1428571428571437e-05,
-      "loss": 0.8832,
+      "loss": 0.86,
       "step": 321
     },
     {
       "epoch": 1.0,
-      "eval_accuracy": 0.7999306518723994,
-      "eval_f1": 0.7874950822007427,
-      "eval_loss": 0.5542528629302979,
-      "eval_precision": 0.8092011027858591,
-      "eval_recall": 0.7999306518723994,
-      "eval_runtime": 37.6681,
-      "eval_samples_per_second": 76.563,
-      "eval_steps_per_second": 9.584,
+      "eval_accuracy": 0.8141470180305131,
+      "eval_f1": 0.8010598763076213,
+      "eval_loss": 0.5250416994094849,
+      "eval_precision": 0.8100096575743447,
+      "eval_recall": 0.8141470180305131,
+      "eval_runtime": 36.2746,
+      "eval_samples_per_second": 79.505,
+      "eval_steps_per_second": 9.952,
       "step": 321
     },
     {
       "epoch": 2.0,
-      "grad_norm": 6.56775426864624,
+      "grad_norm": 6.845722198486328,
       "learning_rate": 8.311688311688312e-05,
-      "loss": 0.4647,
+      "loss": 0.4517,
       "step": 642
     },
     {
       "epoch": 2.0,
-      "eval_accuracy": 0.8262829403606102,
-      "eval_f1": 0.8158028786365703,
-      "eval_loss": 0.49287769198417664,
-      "eval_precision": 0.8255125137698959,
-      "eval_recall": 0.8262829403606102,
-      "eval_runtime": 37.8104,
-      "eval_samples_per_second": 76.275,
-      "eval_steps_per_second": 9.548,
+      "eval_accuracy": 0.8221220527045769,
+      "eval_f1": 0.8099830282273331,
+      "eval_loss": 0.5117006897926331,
+      "eval_precision": 0.8347375649374938,
+      "eval_recall": 0.8221220527045769,
+      "eval_runtime": 36.0052,
+      "eval_samples_per_second": 80.099,
+      "eval_steps_per_second": 10.026,
       "step": 642
     },
     {
       "epoch": 3.0,
-      "grad_norm": 2.852621078491211,
-      "learning_rate": 9.171100544010777e-05,
-      "loss": 0.3465,
+      "grad_norm": 1.1383476257324219,
+      "learning_rate": 9.625711149159094e-05,
+      "loss": 0.3462,
       "step": 963
     },
     {
       "epoch": 3.0,
-      "eval_accuracy": 0.8214285714285714,
-      "eval_f1": 0.8025208154787117,
-      "eval_loss": 0.6518540978431702,
-      "eval_precision": 0.8133669627117774,
-      "eval_recall": 0.8214285714285714,
-      "eval_runtime": 37.046,
-      "eval_samples_per_second": 77.849,
-      "eval_steps_per_second": 9.745,
+      "eval_accuracy": 0.8318307905686546,
+      "eval_f1": 0.8145801957068103,
+      "eval_loss": 0.5491535663604736,
+      "eval_precision": 0.8237881382705559,
+      "eval_recall": 0.8318307905686546,
+      "eval_runtime": 37.0153,
+      "eval_samples_per_second": 77.914,
+      "eval_steps_per_second": 9.753,
       "step": 963
     },
     {
       "epoch": 4.0,
-      "grad_norm": 9.967479705810547,
-      "learning_rate": 5.025483282385923e-05,
-      "loss": 0.1493,
+      "grad_norm": 8.22785472869873,
+      "learning_rate": 7.514698384161923e-05,
+      "loss": 0.1884,
       "step": 1284
     },
     {
       "epoch": 4.0,
-      "eval_accuracy": 0.8401525658807212,
-      "eval_f1": 0.8265360989719667,
-      "eval_loss": 0.6805647611618042,
-      "eval_precision": 0.8385311261178104,
-      "eval_recall": 0.8401525658807212,
-      "eval_runtime": 37.4024,
-      "eval_samples_per_second": 77.107,
-      "eval_steps_per_second": 9.652,
+      "eval_accuracy": 0.8477808599167822,
+      "eval_f1": 0.8424611946030242,
+      "eval_loss": 0.5316428542137146,
+      "eval_precision": 0.8452078508575436,
+      "eval_recall": 0.8477808599167822,
+      "eval_runtime": 37.0521,
+      "eval_samples_per_second": 77.836,
+      "eval_steps_per_second": 9.743,
       "step": 1284
     },
     {
       "epoch": 5.0,
-      "grad_norm": 0.2563266158103943,
-      "learning_rate": 8.572203615936175e-06,
-      "loss": 0.0211,
+      "grad_norm": 0.08933322876691818,
+      "learning_rate": 4.3646376374911565e-05,
+      "loss": 0.0661,
       "step": 1605
     },
     {
       "epoch": 5.0,
-      "eval_accuracy": 0.8762135922330098,
-      "eval_f1": 0.8742202382058613,
-      "eval_loss": 0.5605077743530273,
-      "eval_precision": 0.8749370661972383,
-      "eval_recall": 0.8762135922330098,
-      "eval_runtime": 37.3835,
-      "eval_samples_per_second": 77.146,
-      "eval_steps_per_second": 9.657,
+      "eval_accuracy": 0.8661581137309292,
+      "eval_f1": 0.8631893027855349,
+      "eval_loss": 0.5807815790176392,
+      "eval_precision": 0.8631976574679416,
+      "eval_recall": 0.8661581137309292,
+      "eval_runtime": 36.8683,
+      "eval_samples_per_second": 78.224,
+      "eval_steps_per_second": 9.792,
       "step": 1605
     },
     {
       "epoch": 6.0,
-      "grad_norm": 0.11774127185344696,
-      "learning_rate": 9.629571142901386e-05,
-      "loss": 0.0084,
+      "grad_norm": 0.004284233320504427,
+      "learning_rate": 1.4771022038177957e-05,
+      "loss": 0.0072,
       "step": 1926
     },
     {
       "epoch": 6.0,
-      "eval_accuracy": 0.8283633841886269,
-      "eval_f1": 0.8089134462154245,
-      "eval_loss": 0.9771777391433716,
-      "eval_precision": 0.8382740264211974,
-      "eval_recall": 0.8283633841886269,
-      "eval_runtime": 36.767,
-      "eval_samples_per_second": 78.44,
-      "eval_steps_per_second": 9.819,
+      "eval_accuracy": 0.8713592233009708,
+      "eval_f1": 0.8685210731779419,
+      "eval_loss": 0.6703277230262756,
+      "eval_precision": 0.8676177648822543,
+      "eval_recall": 0.8713592233009708,
+      "eval_runtime": 37.1065,
+      "eval_samples_per_second": 77.722,
+      "eval_steps_per_second": 9.729,
       "step": 1926
     },
     {
       "epoch": 7.0,
-      "grad_norm": 8.762543678283691,
-      "learning_rate": 6.002305101515785e-05,
-      "loss": 0.1058,
+      "grad_norm": 0.02234409935772419,
+      "learning_rate": 4.5192569174565823e-07,
+      "loss": 0.0005,
       "step": 2247
     },
     {
       "epoch": 7.0,
-      "eval_accuracy": 0.8581830790568654,
-      "eval_f1": 0.8533407001043691,
-      "eval_loss": 0.6204692721366882,
-      "eval_precision": 0.8542132893520025,
-      "eval_recall": 0.8581830790568654,
-      "eval_runtime": 37.6784,
-      "eval_samples_per_second": 76.543,
-      "eval_steps_per_second": 9.581,
+      "eval_accuracy": 0.8723994452149791,
+      "eval_f1": 0.8691996055190594,
+      "eval_loss": 0.66409832239151,
+      "eval_precision": 0.8680805567269357,
+      "eval_recall": 0.8723994452149791,
+      "eval_runtime": 37.5216,
+      "eval_samples_per_second": 76.862,
+      "eval_steps_per_second": 9.621,
       "step": 2247
     },
     {
       "epoch": 8.0,
-      "grad_norm": 0.16941487789154053,
-      "learning_rate": 1.484342980756105e-05,
-      "loss": 0.0171,
+      "grad_norm": 16.355737686157227,
+      "learning_rate": 9.339440663254618e-05,
+      "loss": 0.0253,
       "step": 2568
     },
     {
       "epoch": 8.0,
-      "eval_accuracy": 0.8671983356449375,
-      "eval_f1": 0.8617834655772643,
-      "eval_loss": 0.7407661080360413,
-      "eval_precision": 0.8622466117266067,
-      "eval_recall": 0.8671983356449375,
-      "eval_runtime": 37.6949,
-      "eval_samples_per_second": 76.509,
-      "eval_steps_per_second": 9.577,
+      "eval_accuracy": 0.763869625520111,
+      "eval_f1": 0.7811647054797142,
+      "eval_loss": 1.1242671012878418,
+      "eval_precision": 0.8425391875202219,
+      "eval_recall": 0.763869625520111,
+      "eval_runtime": 37.4111,
+      "eval_samples_per_second": 77.089,
+      "eval_steps_per_second": 9.65,
       "step": 2568
     },
     {
       "epoch": 9.0,
-      "grad_norm": 1.5445165634155273,
-      "learning_rate": 9.909438769446831e-05,
-      "loss": 0.0029,
+      "grad_norm": 0.16647577285766602,
+      "learning_rate": 6.93106082853232e-05,
+      "loss": 0.163,
       "step": 2889
     },
     {
       "epoch": 9.0,
-      "eval_accuracy": 0.8564493758668515,
-      "eval_f1": 0.8473856298855159,
-      "eval_loss": 0.9039825797080994,
-      "eval_precision": 0.8486485221307185,
-      "eval_recall": 0.8564493758668515,
-      "eval_runtime": 37.879,
-      "eval_samples_per_second": 76.137,
-      "eval_steps_per_second": 9.53,
+      "eval_accuracy": 0.8446601941747572,
+      "eval_f1": 0.8442509694752051,
+      "eval_loss": 0.7353706955909729,
+      "eval_precision": 0.8560368714680845,
+      "eval_recall": 0.8446601941747572,
+      "eval_runtime": 37.5281,
+      "eval_samples_per_second": 76.849,
+      "eval_steps_per_second": 9.619,
       "step": 2889
     },
     {
       "epoch": 10.0,
-      "grad_norm": 0.17326785624027252,
-      "learning_rate": 6.94045926336125e-05,
-      "loss": 0.1232,
+      "grad_norm": 0.006103373598307371,
+      "learning_rate": 3.724786157191619e-05,
+      "loss": 0.0324,
       "step": 3210
     },
     {
       "epoch": 10.0,
-      "eval_accuracy": 0.829750346740638,
-      "eval_f1": 0.8157566455477707,
-      "eval_loss": 0.9558870792388916,
-      "eval_precision": 0.84362582840677,
-      "eval_recall": 0.829750346740638,
-      "eval_runtime": 37.6559,
-      "eval_samples_per_second": 76.588,
-      "eval_steps_per_second": 9.587,
+      "eval_accuracy": 0.8640776699029126,
+      "eval_f1": 0.860900542767183,
+      "eval_loss": 0.7327257990837097,
+      "eval_precision": 0.8630486238544385,
+      "eval_recall": 0.8640776699029126,
+      "eval_runtime": 35.8668,
+      "eval_samples_per_second": 80.409,
+      "eval_steps_per_second": 10.065,
       "step": 3210
     },
     {
       "epoch": 11.0,
-      "grad_norm": 0.0039984057657420635,
-      "learning_rate": 2.2470951807159123e-05,
-      "loss": 0.0233,
+      "grad_norm": 0.0016283913282677531,
+      "learning_rate": 1.0454169699906635e-05,
+      "loss": 0.0014,
       "step": 3531
     },
     {
       "epoch": 11.0,
-      "eval_accuracy": 0.8564493758668515,
-      "eval_f1": 0.857178695277239,
-      "eval_loss": 0.7810959219932556,
-      "eval_precision": 0.8607102524458344,
-      "eval_recall": 0.8564493758668515,
-      "eval_runtime": 37.2622,
-      "eval_samples_per_second": 77.397,
-      "eval_steps_per_second": 9.688,
+      "eval_accuracy": 0.8678918169209431,
+      "eval_f1": 0.8642538050561708,
+      "eval_loss": 0.7242896556854248,
+      "eval_precision": 0.8636644990951133,
+      "eval_recall": 0.8678918169209431,
+      "eval_runtime": 36.277,
+      "eval_samples_per_second": 79.499,
+      "eval_steps_per_second": 9.951,
       "step": 3531
     },
     {
       "epoch": 12.0,
-      "grad_norm": 0.003946017008274794,
-      "learning_rate": 9.351378430000336e-10,
-      "loss": 0.0009,
+      "grad_norm": 0.0033853575587272644,
+      "learning_rate": 4.156175388514427e-10,
+      "loss": 0.0002,
       "step": 3852
     },
     {
       "epoch": 12.0,
-      "eval_accuracy": 0.8626907073509015,
-      "eval_f1": 0.859450727771903,
-      "eval_loss": 0.799753725528717,
-      "eval_precision": 0.8599141248108649,
-      "eval_recall": 0.8626907073509015,
-      "eval_runtime": 37.2156,
-      "eval_samples_per_second": 77.494,
-      "eval_steps_per_second": 9.7,
+      "eval_accuracy": 0.869625520110957,
+      "eval_f1": 0.8660960395798193,
+      "eval_loss": 0.7235937714576721,
+      "eval_precision": 0.8654974771709472,
+      "eval_recall": 0.869625520110957,
+      "eval_runtime": 36.6203,
+      "eval_samples_per_second": 78.754,
+      "eval_steps_per_second": 9.858,
       "step": 3852
     },
     {
       "epoch": 12.0,
       "step": 3852,
       "total_flos": 4.768760767819088e+18,
-      "train_loss": 0.17886438471413105,
-      "train_runtime": 1744.4117,
-      "train_samples_per_second": 35.276,
-      "train_steps_per_second": 2.208
+      "train_loss": 0.17854708074838666,
+      "train_runtime": 1730.8935,
+      "train_samples_per_second": 35.552,
+      "train_steps_per_second": 2.225
     }
   ],
   "logging_steps": 500,