🍻 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: 0.7956
-- Accuracy: 0.8731
-- Precision: 0.8708
-- Recall: 0.8731
-- F1: 0.8699
+- Loss: 0.5117
+- Accuracy: 0.8221
+- Precision: 0.8347
+- Recall: 0.8221
+- F1: 0.8100
 
 ## Model description
 

all_results.json

@@ -5,12 +5,12 @@
     "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,
+    "eval_runtime": 39.2908,
+    "eval_samples_per_second": 73.401,
+    "eval_steps_per_second": 9.188,
     "total_flos": 4.768760767819088e+18,
-    "train_loss": 0.17854708074838666,
-    "train_runtime": 1730.8935,
-    "train_samples_per_second": 35.552,
-    "train_steps_per_second": 2.225
+    "train_loss": 0.17688380913259627,
+    "train_runtime": 1754.6284,
+    "train_samples_per_second": 35.071,
+    "train_steps_per_second": 2.195
 }

eval_results.json

@@ -5,7 +5,7 @@
     "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
+    "eval_runtime": 39.2908,
+    "eval_samples_per_second": 73.401,
+    "eval_steps_per_second": 9.188
 }

runs/Mar20_17-59-47_f2e1fee5f9b2/events.out.tfevents.1710959425.f2e1fee5f9b2.2315.11

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3be1208b535e4827dd86d24e37080384e4cb37c51273299fff83e4fafd0b63cb
+size 560

train_results.json

@@ -1,8 +1,8 @@
 {
     "epoch": 12.0,
     "total_flos": 4.768760767819088e+18,
-    "train_loss": 0.17854708074838666,
-    "train_runtime": 1730.8935,
-    "train_samples_per_second": 35.552,
-    "train_steps_per_second": 2.225
+    "train_loss": 0.17688380913259627,
+    "train_runtime": 1754.6284,
+    "train_samples_per_second": 35.071,
+    "train_steps_per_second": 2.195
 }

trainer_state.json

@@ -9,11 +9,25 @@
   "is_world_process_zero": true,
   "log_history": [
     {
-      "epoch": 1.0,
-      "grad_norm": 7.814958572387695,
-      "learning_rate": 4.1428571428571437e-05,
-      "loss": 0.86,
-      "step": 321
+      "epoch": 0.31,
+      "grad_norm": 5.410258769989014,
+      "learning_rate": 1.2727272727272727e-05,
+      "loss": 1.3284,
+      "step": 100
+    },
+    {
+      "epoch": 0.62,
+      "grad_norm": 5.756007671356201,
+      "learning_rate": 2.5714285714285714e-05,
+      "loss": 0.7111,
+      "step": 200
+    },
+    {
+      "epoch": 0.93,
+      "grad_norm": 5.29420804977417,
+      "learning_rate": 3.87012987012987e-05,
+      "loss": 0.5999,
+      "step": 300
     },
     {
       "epoch": 1.0,
@@ -22,17 +36,31 @@
       "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,
+      "eval_runtime": 37.3322,
+      "eval_samples_per_second": 77.252,
+      "eval_steps_per_second": 9.67,
       "step": 321
     },
     {
-      "epoch": 2.0,
-      "grad_norm": 6.845722198486328,
-      "learning_rate": 8.311688311688312e-05,
-      "loss": 0.4517,
-      "step": 642
+      "epoch": 1.25,
+      "grad_norm": 4.180977821350098,
+      "learning_rate": 5.1688311688311686e-05,
+      "loss": 0.478,
+      "step": 400
+    },
+    {
+      "epoch": 1.56,
+      "grad_norm": 4.176523208618164,
+      "learning_rate": 6.467532467532468e-05,
+      "loss": 0.4691,
+      "step": 500
+    },
+    {
+      "epoch": 1.87,
+      "grad_norm": 3.4944210052490234,
+      "learning_rate": 7.766233766233767e-05,
+      "loss": 0.4483,
+      "step": 600
     },
     {
       "epoch": 2.0,
@@ -41,212 +69,366 @@
       "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,
+      "eval_runtime": 37.3595,
+      "eval_samples_per_second": 77.196,
+      "eval_steps_per_second": 9.663,
       "step": 642
     },
     {
-      "epoch": 3.0,
-      "grad_norm": 1.1383476257324219,
-      "learning_rate": 9.625711149159094e-05,
-      "loss": 0.3462,
-      "step": 963
+      "epoch": 2.18,
+      "grad_norm": 4.858165740966797,
+      "learning_rate": 9.064935064935065e-05,
+      "loss": 0.3437,
+      "step": 700
+    },
+    {
+      "epoch": 2.49,
+      "grad_norm": 3.893717050552368,
+      "learning_rate": 9.981682436521149e-05,
+      "loss": 0.374,
+      "step": 800
+    },
+    {
+      "epoch": 2.8,
+      "grad_norm": 6.913142681121826,
+      "learning_rate": 9.621831930129997e-05,
+      "loss": 0.3425,
+      "step": 900
     },
     {
       "epoch": 3.0,
-      "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,
+      "eval_accuracy": 0.8269764216366158,
+      "eval_f1": 0.810684439200049,
+      "eval_loss": 0.5708675384521484,
+      "eval_precision": 0.8207951483368479,
+      "eval_recall": 0.8269764216366158,
+      "eval_runtime": 37.569,
+      "eval_samples_per_second": 76.765,
+      "eval_steps_per_second": 9.609,
       "step": 963
     },
     {
-      "epoch": 4.0,
-      "grad_norm": 8.22785472869873,
-      "learning_rate": 7.514698384161923e-05,
-      "loss": 0.1884,
-      "step": 1284
+      "epoch": 3.12,
+      "grad_norm": 5.459534645080566,
+      "learning_rate": 8.833132715666827e-05,
+      "loss": 0.2761,
+      "step": 1000
+    },
+    {
+      "epoch": 3.43,
+      "grad_norm": 3.252346992492676,
+      "learning_rate": 7.68876630303792e-05,
+      "loss": 0.1709,
+      "step": 1100
+    },
+    {
+      "epoch": 3.74,
+      "grad_norm": 5.587292194366455,
+      "learning_rate": 6.294915709072077e-05,
+      "loss": 0.176,
+      "step": 1200
     },
     {
       "epoch": 4.0,
-      "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,
+      "eval_accuracy": 0.8574895977808599,
+      "eval_f1": 0.854413257501997,
+      "eval_loss": 0.5578957200050354,
+      "eval_precision": 0.8549085651193107,
+      "eval_recall": 0.8574895977808599,
+      "eval_runtime": 37.9199,
+      "eval_samples_per_second": 76.055,
+      "eval_steps_per_second": 9.52,
       "step": 1284
     },
     {
-      "epoch": 5.0,
-      "grad_norm": 0.08933322876691818,
-      "learning_rate": 4.3646376374911565e-05,
-      "loss": 0.0661,
-      "step": 1605
+      "epoch": 4.05,
+      "grad_norm": 0.42479750514030457,
+      "learning_rate": 4.78091298986829e-05,
+      "loss": 0.1181,
+      "step": 1300
+    },
+    {
+      "epoch": 4.36,
+      "grad_norm": 1.4105452299118042,
+      "learning_rate": 3.2872388292686876e-05,
+      "loss": 0.0404,
+      "step": 1400
+    },
+    {
+      "epoch": 4.67,
+      "grad_norm": 0.6475068926811218,
+      "learning_rate": 1.952487672882131e-05,
+      "loss": 0.014,
+      "step": 1500
+    },
+    {
+      "epoch": 4.98,
+      "grad_norm": 0.14880895614624023,
+      "learning_rate": 9.005078816328771e-06,
+      "loss": 0.0183,
+      "step": 1600
     },
     {
       "epoch": 5.0,
-      "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,
+      "eval_accuracy": 0.8772538141470181,
+      "eval_f1": 0.87586423765656,
+      "eval_loss": 0.5555465817451477,
+      "eval_precision": 0.8761682759483852,
+      "eval_recall": 0.8772538141470181,
+      "eval_runtime": 37.4554,
+      "eval_samples_per_second": 76.998,
+      "eval_steps_per_second": 9.638,
       "step": 1605
     },
     {
-      "epoch": 6.0,
-      "grad_norm": 0.004284233320504427,
-      "learning_rate": 1.4771022038177957e-05,
-      "loss": 0.0072,
-      "step": 1926
+      "epoch": 5.3,
+      "grad_norm": 0.15054507553577423,
+      "learning_rate": 2.2891013916653993e-06,
+      "loss": 0.0022,
+      "step": 1700
+    },
+    {
+      "epoch": 5.61,
+      "grad_norm": 0.09130716323852539,
+      "learning_rate": 9.999989609550733e-05,
+      "loss": 0.0012,
+      "step": 1800
+    },
+    {
+      "epoch": 5.92,
+      "grad_norm": 0.693184494972229,
+      "learning_rate": 9.7649523054187e-05,
+      "loss": 0.0273,
+      "step": 1900
     },
     {
       "epoch": 6.0,
-      "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,
+      "eval_accuracy": 0.8415395284327323,
+      "eval_f1": 0.8328026992097984,
+      "eval_loss": 0.8165751099586487,
+      "eval_precision": 0.8368720433613618,
+      "eval_recall": 0.8415395284327323,
+      "eval_runtime": 37.8432,
+      "eval_samples_per_second": 76.209,
+      "eval_steps_per_second": 9.539,
       "step": 1926
     },
     {
-      "epoch": 7.0,
-      "grad_norm": 0.02234409935772419,
-      "learning_rate": 4.5192569174565823e-07,
-      "loss": 0.0005,
-      "step": 2247
+      "epoch": 6.23,
+      "grad_norm": 0.32119280099868774,
+      "learning_rate": 9.087786496595214e-05,
+      "loss": 0.1061,
+      "step": 2000
+    },
+    {
+      "epoch": 6.54,
+      "grad_norm": 9.584900856018066,
+      "learning_rate": 8.03132477557923e-05,
+      "loss": 0.1256,
+      "step": 2100
+    },
+    {
+      "epoch": 6.85,
+      "grad_norm": 9.815910339355469,
+      "learning_rate": 6.693593693537977e-05,
+      "loss": 0.1186,
+      "step": 2200
     },
     {
       "epoch": 7.0,
-      "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,
+      "eval_accuracy": 0.8616504854368932,
+      "eval_f1": 0.8590594832634068,
+      "eval_loss": 0.6379634737968445,
+      "eval_precision": 0.8598260577161764,
+      "eval_recall": 0.8616504854368932,
+      "eval_runtime": 37.9305,
+      "eval_samples_per_second": 76.034,
+      "eval_steps_per_second": 9.517,
       "step": 2247
     },
     {
-      "epoch": 8.0,
-      "grad_norm": 16.355737686157227,
-      "learning_rate": 9.339440663254618e-05,
-      "loss": 0.0253,
-      "step": 2568
+      "epoch": 7.17,
+      "grad_norm": 0.01356012187898159,
+      "learning_rate": 5.198718111555498e-05,
+      "loss": 0.0602,
+      "step": 2300
+    },
+    {
+      "epoch": 7.48,
+      "grad_norm": 0.6182898283004761,
+      "learning_rate": 3.685403952251429e-05,
+      "loss": 0.0212,
+      "step": 2400
+    },
+    {
+      "epoch": 7.79,
+      "grad_norm": 0.0050924597308039665,
+      "learning_rate": 2.294068009645023e-05,
+      "loss": 0.0166,
+      "step": 2500
     },
     {
       "epoch": 8.0,
-      "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,
+      "eval_accuracy": 0.8730929264909847,
+      "eval_f1": 0.8708719242638395,
+      "eval_loss": 0.6608479619026184,
+      "eval_precision": 0.8716446126885153,
+      "eval_recall": 0.8730929264909847,
+      "eval_runtime": 37.8013,
+      "eval_samples_per_second": 76.294,
+      "eval_steps_per_second": 9.55,
       "step": 2568
     },
     {
-      "epoch": 9.0,
-      "grad_norm": 0.16647577285766602,
-      "learning_rate": 6.93106082853232e-05,
-      "loss": 0.163,
-      "step": 2889
+      "epoch": 8.1,
+      "grad_norm": 0.0024975521955639124,
+      "learning_rate": 1.1538090113622485e-05,
+      "loss": 0.0115,
+      "step": 2600
+    },
+    {
+      "epoch": 8.41,
+      "grad_norm": 0.03202090784907341,
+      "learning_rate": 3.704288570986153e-06,
+      "loss": 0.0013,
+      "step": 2700
+    },
+    {
+      "epoch": 8.72,
+      "grad_norm": 0.011940655298531055,
+      "learning_rate": 1.6615513912184854e-07,
+      "loss": 0.0003,
+      "step": 2800
     },
     {
       "epoch": 9.0,
-      "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,
+      "eval_accuracy": 0.8460471567267683,
+      "eval_f1": 0.8306037539812486,
+      "eval_loss": 1.0044565200805664,
+      "eval_precision": 0.8438100295092479,
+      "eval_recall": 0.8460471567267683,
+      "eval_runtime": 37.8956,
+      "eval_samples_per_second": 76.104,
+      "eval_steps_per_second": 9.526,
       "step": 2889
     },
     {
-      "epoch": 10.0,
-      "grad_norm": 0.006103373598307371,
-      "learning_rate": 3.724786157191619e-05,
-      "loss": 0.0324,
-      "step": 3210
+      "epoch": 9.03,
+      "grad_norm": 0.005646233446896076,
+      "learning_rate": 9.87480152603507e-05,
+      "loss": 0.0011,
+      "step": 2900
+    },
+    {
+      "epoch": 9.35,
+      "grad_norm": 12.50674819946289,
+      "learning_rate": 9.313897415375534e-05,
+      "loss": 0.069,
+      "step": 3000
+    },
+    {
+      "epoch": 9.66,
+      "grad_norm": 0.29825684428215027,
+      "learning_rate": 8.352717102269369e-05,
+      "loss": 0.1307,
+      "step": 3100
+    },
+    {
+      "epoch": 9.97,
+      "grad_norm": 0.08965686708688736,
+      "learning_rate": 7.080446204396154e-05,
+      "loss": 0.1258,
+      "step": 3200
     },
     {
       "epoch": 10.0,
-      "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,
+      "eval_accuracy": 0.8498613037447988,
+      "eval_f1": 0.8489972660614399,
+      "eval_loss": 0.671244740486145,
+      "eval_precision": 0.8526962090532623,
+      "eval_recall": 0.8498613037447988,
+      "eval_runtime": 37.4226,
+      "eval_samples_per_second": 77.066,
+      "eval_steps_per_second": 9.647,
       "step": 3210
     },
     {
-      "epoch": 11.0,
-      "grad_norm": 0.0016283913282677531,
-      "learning_rate": 1.0454169699906635e-05,
-      "loss": 0.0014,
-      "step": 3531
+      "epoch": 10.28,
+      "grad_norm": 0.010914159007370472,
+      "learning_rate": 5.615135689233456e-05,
+      "loss": 0.0472,
+      "step": 3300
+    },
+    {
+      "epoch": 10.59,
+      "grad_norm": 0.03197154030203819,
+      "learning_rate": 4.092748207706606e-05,
+      "loss": 0.0183,
+      "step": 3400
+    },
+    {
+      "epoch": 10.9,
+      "grad_norm": 0.0029819884803146124,
+      "learning_rate": 2.6545424456587097e-05,
+      "loss": 0.0137,
+      "step": 3500
     },
     {
       "epoch": 11.0,
-      "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,
+      "eval_accuracy": 0.8737864077669902,
+      "eval_f1": 0.8740056608487029,
+      "eval_loss": 0.795246958732605,
+      "eval_precision": 0.8765383771673093,
+      "eval_recall": 0.8737864077669902,
+      "eval_runtime": 38.3355,
+      "eval_samples_per_second": 75.23,
+      "eval_steps_per_second": 9.417,
       "step": 3531
     },
     {
-      "epoch": 12.0,
-      "grad_norm": 0.0033853575587272644,
-      "learning_rate": 4.156175388514427e-10,
-      "loss": 0.0002,
-      "step": 3852
+      "epoch": 11.21,
+      "grad_norm": 0.0016925123054534197,
+      "learning_rate": 1.4339660689330769e-05,
+      "loss": 0.006,
+      "step": 3600
+    },
+    {
+      "epoch": 11.53,
+      "grad_norm": 0.0018662018701434135,
+      "learning_rate": 5.44273434281552e-06,
+      "loss": 0.0004,
+      "step": 3700
+    },
+    {
+      "epoch": 11.84,
+      "grad_norm": 0.01316988654434681,
+      "learning_rate": 6.801698902793418e-07,
+      "loss": 0.0004,
+      "step": 3800
     },
     {
       "epoch": 12.0,
-      "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,
+      "eval_accuracy": 0.8730929264909847,
+      "eval_f1": 0.8699086463488187,
+      "eval_loss": 0.7956326007843018,
+      "eval_precision": 0.8707501942119029,
+      "eval_recall": 0.8730929264909847,
+      "eval_runtime": 38.1198,
+      "eval_samples_per_second": 75.656,
+      "eval_steps_per_second": 9.47,
       "step": 3852
     },
     {
       "epoch": 12.0,
       "step": 3852,
       "total_flos": 4.768760767819088e+18,
-      "train_loss": 0.17854708074838666,
-      "train_runtime": 1730.8935,
-      "train_samples_per_second": 35.552,
-      "train_steps_per_second": 2.225
+      "train_loss": 0.17688380913259627,
+      "train_runtime": 1754.6284,
+      "train_samples_per_second": 35.071,
+      "train_steps_per_second": 2.195
     }
   ],
-  "logging_steps": 500,
+  "logging_steps": 100,
   "max_steps": 3852,
   "num_input_tokens_seen": 0,
   "num_train_epochs": 12,