End of training

README.md

@@ -32,7 +32,7 @@ should probably proofread and complete it, then remove this comment. -->
 
 This model is a fine-tuned version of [facebook/deit-tiny-patch16-224](https://huggingface.co/facebook/deit-tiny-patch16-224) on the imagefolder dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.3409
+- Loss: 0.3421
 - Accuracy: 0.9118
 
 ## Model description

all_results.json

@@ -1,13 +1,13 @@
 {
     "epoch": 14.805194805194805,
-    "eval_accuracy": 0.8529411764705882,
-    "eval_loss": 0.4389486610889435,
-    "eval_runtime": 28.0984,
-    "eval_samples_per_second": 4.84,
-    "eval_steps_per_second": 0.32,
+    "eval_accuracy": 0.9117647058823529,
+    "eval_loss": 0.34208229184150696,
+    "eval_runtime": 28.5471,
+    "eval_samples_per_second": 4.764,
+    "eval_steps_per_second": 0.315,
     "total_flos": 9.04681758989353e+16,
-    "train_loss": 0.6850269221422965,
-    "train_runtime": 4202.2024,
-    "train_samples_per_second": 4.369,
-    "train_steps_per_second": 0.068
+    "train_loss": 0.2430994153022766,
+    "train_runtime": 4010.606,
+    "train_samples_per_second": 4.578,
+    "train_steps_per_second": 0.071
 }

eval_results.json

@@ -1,8 +1,8 @@
 {
     "epoch": 14.805194805194805,
-    "eval_accuracy": 0.8529411764705882,
-    "eval_loss": 0.4389486610889435,
-    "eval_runtime": 28.0984,
-    "eval_samples_per_second": 4.84,
-    "eval_steps_per_second": 0.32
+    "eval_accuracy": 0.9117647058823529,
+    "eval_loss": 0.34208229184150696,
+    "eval_runtime": 28.5471,
+    "eval_samples_per_second": 4.764,
+    "eval_steps_per_second": 0.315
 }

runs/Sep18_05-34-03_0a55b5166c22/events.out.tfevents.1726646257.0a55b5166c22.36.2

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:77503f0fca795d6bb2c8c46ca193331e9cf704748c14caa88f8697dc64e57bb8
+size 411

train_results.json

@@ -1,8 +1,8 @@
 {
     "epoch": 14.805194805194805,
     "total_flos": 9.04681758989353e+16,
-    "train_loss": 0.6850269221422965,
-    "train_runtime": 4202.2024,
-    "train_samples_per_second": 4.369,
-    "train_steps_per_second": 0.068
+    "train_loss": 0.2430994153022766,
+    "train_runtime": 4010.606,
+    "train_samples_per_second": 4.578,
+    "train_steps_per_second": 0.071
 }

trainer_state.json

@@ -1,5 +1,5 @@
 {
-  "best_metric": 0.8529411764705882,
+  "best_metric": 0.9117647058823529,
   "best_model_checkpoint": "deit-tiny-patch16-224-finetuned-papsmear/checkpoint-269",
   "epoch": 14.805194805194805,
   "eval_steps": 500,
@@ -10,343 +10,343 @@
   "log_history": [
     {
       "epoch": 0.5194805194805194,
-      "grad_norm": 5.887332439422607,
+      "grad_norm": 11.024985313415527,
       "learning_rate": 1.7241379310344828e-05,
-      "loss": 1.8247,
+      "loss": 0.3521,
       "step": 10
     },
     {
       "epoch": 0.987012987012987,
-      "eval_accuracy": 0.3014705882352941,
-      "eval_loss": 1.6199148893356323,
-      "eval_runtime": 27.9963,
-      "eval_samples_per_second": 4.858,
-      "eval_steps_per_second": 0.321,
+      "eval_accuracy": 0.8235294117647058,
+      "eval_loss": 0.4595417082309723,
+      "eval_runtime": 28.5401,
+      "eval_samples_per_second": 4.765,
+      "eval_steps_per_second": 0.315,
       "step": 19
     },
     {
       "epoch": 1.0389610389610389,
-      "grad_norm": 7.520359039306641,
+      "grad_norm": 14.813636779785156,
       "learning_rate": 3.4482758620689657e-05,
-      "loss": 1.6044,
+      "loss": 0.3115,
       "step": 20
     },
     {
       "epoch": 1.5584415584415585,
-      "grad_norm": 8.020193099975586,
+      "grad_norm": 16.087106704711914,
       "learning_rate": 4.9804687500000004e-05,
-      "loss": 1.415,
+      "loss": 0.335,
       "step": 30
     },
     {
       "epoch": 1.974025974025974,
-      "eval_accuracy": 0.5147058823529411,
-      "eval_loss": 1.2593767642974854,
-      "eval_runtime": 28.328,
-      "eval_samples_per_second": 4.801,
-      "eval_steps_per_second": 0.318,
+      "eval_accuracy": 0.8602941176470589,
+      "eval_loss": 0.4491410553455353,
+      "eval_runtime": 27.9762,
+      "eval_samples_per_second": 4.861,
+      "eval_steps_per_second": 0.322,
       "step": 38
     },
     {
       "epoch": 2.0779220779220777,
-      "grad_norm": 9.726988792419434,
+      "grad_norm": 19.62188720703125,
       "learning_rate": 4.78515625e-05,
-      "loss": 1.2695,
+      "loss": 0.4021,
       "step": 40
     },
     {
       "epoch": 2.5974025974025974,
-      "grad_norm": 10.028353691101074,
+      "grad_norm": 12.998429298400879,
       "learning_rate": 4.58984375e-05,
-      "loss": 1.06,
+      "loss": 0.3248,
       "step": 50
     },
     {
       "epoch": 2.961038961038961,
-      "eval_accuracy": 0.6470588235294118,
-      "eval_loss": 1.0316276550292969,
-      "eval_runtime": 28.6395,
-      "eval_samples_per_second": 4.749,
-      "eval_steps_per_second": 0.314,
+      "eval_accuracy": 0.875,
+      "eval_loss": 0.4195605218410492,
+      "eval_runtime": 28.3986,
+      "eval_samples_per_second": 4.789,
+      "eval_steps_per_second": 0.317,
       "step": 57
     },
     {
       "epoch": 3.116883116883117,
-      "grad_norm": 11.8433837890625,
+      "grad_norm": 8.330426216125488,
       "learning_rate": 4.3945312500000005e-05,
-      "loss": 1.0001,
+      "loss": 0.3613,
       "step": 60
     },
     {
       "epoch": 3.6363636363636362,
-      "grad_norm": 9.538862228393555,
+      "grad_norm": 10.42120361328125,
       "learning_rate": 4.1992187500000003e-05,
-      "loss": 0.8808,
+      "loss": 0.3271,
       "step": 70
     },
     {
       "epoch": 4.0,
-      "eval_accuracy": 0.625,
-      "eval_loss": 1.0088493824005127,
-      "eval_runtime": 28.6257,
-      "eval_samples_per_second": 4.751,
-      "eval_steps_per_second": 0.314,
+      "eval_accuracy": 0.8014705882352942,
+      "eval_loss": 0.5466737151145935,
+      "eval_runtime": 28.1607,
+      "eval_samples_per_second": 4.829,
+      "eval_steps_per_second": 0.32,
       "step": 77
     },
     {
       "epoch": 4.1558441558441555,
-      "grad_norm": 13.59125804901123,
+      "grad_norm": 14.403099060058594,
       "learning_rate": 4.00390625e-05,
-      "loss": 0.9073,
+      "loss": 0.3451,
       "step": 80
     },
     {
       "epoch": 4.675324675324675,
-      "grad_norm": 10.73096752166748,
+      "grad_norm": 18.443952560424805,
       "learning_rate": 3.80859375e-05,
-      "loss": 0.7646,
+      "loss": 0.3286,
       "step": 90
     },
     {
       "epoch": 4.987012987012987,
-      "eval_accuracy": 0.6985294117647058,
-      "eval_loss": 0.8210764527320862,
-      "eval_runtime": 28.1022,
-      "eval_samples_per_second": 4.839,
-      "eval_steps_per_second": 0.32,
+      "eval_accuracy": 0.8161764705882353,
+      "eval_loss": 0.4768396317958832,
+      "eval_runtime": 28.7873,
+      "eval_samples_per_second": 4.724,
+      "eval_steps_per_second": 0.313,
       "step": 96
     },
     {
       "epoch": 5.194805194805195,
-      "grad_norm": 9.227697372436523,
+      "grad_norm": 8.146430969238281,
       "learning_rate": 3.6132812500000005e-05,
-      "loss": 0.7113,
+      "loss": 0.2577,
       "step": 100
     },
     {
       "epoch": 5.714285714285714,
-      "grad_norm": 9.764923095703125,
+      "grad_norm": 7.037234306335449,
       "learning_rate": 3.41796875e-05,
-      "loss": 0.6798,
+      "loss": 0.2854,
       "step": 110
     },
     {
       "epoch": 5.974025974025974,
-      "eval_accuracy": 0.7132352941176471,
-      "eval_loss": 0.7382610440254211,
-      "eval_runtime": 28.5068,
-      "eval_samples_per_second": 4.771,
-      "eval_steps_per_second": 0.316,
+      "eval_accuracy": 0.8676470588235294,
+      "eval_loss": 0.41474443674087524,
+      "eval_runtime": 28.6719,
+      "eval_samples_per_second": 4.743,
+      "eval_steps_per_second": 0.314,
       "step": 115
     },
     {
       "epoch": 6.233766233766234,
-      "grad_norm": 6.136936187744141,
+      "grad_norm": 10.498587608337402,
       "learning_rate": 3.22265625e-05,
-      "loss": 0.6017,
+      "loss": 0.2295,
       "step": 120
     },
     {
       "epoch": 6.753246753246753,
-      "grad_norm": 14.086496353149414,
+      "grad_norm": 13.865105628967285,
       "learning_rate": 3.02734375e-05,
-      "loss": 0.554,
+      "loss": 0.2291,
       "step": 130
     },
     {
       "epoch": 6.961038961038961,
-      "eval_accuracy": 0.7573529411764706,
-      "eval_loss": 0.6476972103118896,
-      "eval_runtime": 28.4917,
-      "eval_samples_per_second": 4.773,
-      "eval_steps_per_second": 0.316,
+      "eval_accuracy": 0.8676470588235294,
+      "eval_loss": 0.4321264624595642,
+      "eval_runtime": 28.1356,
+      "eval_samples_per_second": 4.834,
+      "eval_steps_per_second": 0.32,
       "step": 134
     },
     {
       "epoch": 7.2727272727272725,
-      "grad_norm": 10.119462966918945,
+      "grad_norm": 14.981769561767578,
       "learning_rate": 2.83203125e-05,
-      "loss": 0.581,
+      "loss": 0.2356,
       "step": 140
     },
     {
       "epoch": 7.792207792207792,
-      "grad_norm": 9.049328804016113,
+      "grad_norm": 17.98302459716797,
       "learning_rate": 2.63671875e-05,
-      "loss": 0.5358,
+      "loss": 0.2619,
       "step": 150
     },
     {
       "epoch": 8.0,
-      "eval_accuracy": 0.7647058823529411,
-      "eval_loss": 0.5824333429336548,
-      "eval_runtime": 28.7788,
-      "eval_samples_per_second": 4.726,
-      "eval_steps_per_second": 0.313,
+      "eval_accuracy": 0.8235294117647058,
+      "eval_loss": 0.5726210474967957,
+      "eval_runtime": 28.5909,
+      "eval_samples_per_second": 4.757,
+      "eval_steps_per_second": 0.315,
       "step": 154
     },
     {
       "epoch": 8.311688311688311,
-      "grad_norm": 14.884687423706055,
+      "grad_norm": 17.48947525024414,
       "learning_rate": 2.44140625e-05,
-      "loss": 0.4674,
+      "loss": 0.2315,
       "step": 160
     },
     {
       "epoch": 8.831168831168831,
-      "grad_norm": 11.825162887573242,
+      "grad_norm": 18.059572219848633,
       "learning_rate": 2.24609375e-05,
-      "loss": 0.4689,
+      "loss": 0.2196,
       "step": 170
     },
     {
       "epoch": 8.987012987012987,
-      "eval_accuracy": 0.7794117647058824,
-      "eval_loss": 0.5570951104164124,
-      "eval_runtime": 28.5562,
-      "eval_samples_per_second": 4.763,
-      "eval_steps_per_second": 0.315,
+      "eval_accuracy": 0.8676470588235294,
+      "eval_loss": 0.4344201982021332,
+      "eval_runtime": 28.8842,
+      "eval_samples_per_second": 4.708,
+      "eval_steps_per_second": 0.312,
       "step": 173
     },
     {
       "epoch": 9.35064935064935,
-      "grad_norm": 9.966752052307129,
+      "grad_norm": 13.211594581604004,
       "learning_rate": 2.05078125e-05,
-      "loss": 0.414,
+      "loss": 0.22,
       "step": 180
     },
     {
       "epoch": 9.87012987012987,
-      "grad_norm": 10.21314525604248,
+      "grad_norm": 14.067967414855957,
       "learning_rate": 1.85546875e-05,
-      "loss": 0.4217,
+      "loss": 0.2116,
       "step": 190
     },
     {
       "epoch": 9.974025974025974,
-      "eval_accuracy": 0.7867647058823529,
-      "eval_loss": 0.5505570769309998,
-      "eval_runtime": 28.6499,
-      "eval_samples_per_second": 4.747,
-      "eval_steps_per_second": 0.314,
+      "eval_accuracy": 0.875,
+      "eval_loss": 0.38092589378356934,
+      "eval_runtime": 28.2538,
+      "eval_samples_per_second": 4.814,
+      "eval_steps_per_second": 0.319,
       "step": 192
     },
     {
       "epoch": 10.38961038961039,
-      "grad_norm": 8.933082580566406,
+      "grad_norm": 8.458534240722656,
       "learning_rate": 1.66015625e-05,
-      "loss": 0.3979,
+      "loss": 0.1843,
       "step": 200
     },
     {
       "epoch": 10.909090909090908,
-      "grad_norm": 12.334419250488281,
+      "grad_norm": 11.62593936920166,
       "learning_rate": 1.4648437500000001e-05,
-      "loss": 0.4063,
+      "loss": 0.1913,
       "step": 210
     },
     {
       "epoch": 10.96103896103896,
-      "eval_accuracy": 0.8235294117647058,
-      "eval_loss": 0.4987229108810425,
-      "eval_runtime": 28.2377,
-      "eval_samples_per_second": 4.816,
-      "eval_steps_per_second": 0.319,
+      "eval_accuracy": 0.8602941176470589,
+      "eval_loss": 0.37573114037513733,
+      "eval_runtime": 28.2614,
+      "eval_samples_per_second": 4.812,
+      "eval_steps_per_second": 0.318,
       "step": 211
     },
     {
       "epoch": 11.428571428571429,
-      "grad_norm": 10.463237762451172,
+      "grad_norm": 15.429533958435059,
       "learning_rate": 1.2695312500000001e-05,
-      "loss": 0.3799,
+      "loss": 0.1654,
       "step": 220
     },
     {
       "epoch": 11.948051948051948,
-      "grad_norm": 9.576591491699219,
+      "grad_norm": 8.531671524047852,
       "learning_rate": 1.0742187500000001e-05,
-      "loss": 0.3827,
+      "loss": 0.1604,
       "step": 230
     },
     {
       "epoch": 12.0,
-      "eval_accuracy": 0.8088235294117647,
-      "eval_loss": 0.47928401827812195,
-      "eval_runtime": 29.299,
-      "eval_samples_per_second": 4.642,
-      "eval_steps_per_second": 0.307,
+      "eval_accuracy": 0.8897058823529411,
+      "eval_loss": 0.35508498549461365,
+      "eval_runtime": 28.7628,
+      "eval_samples_per_second": 4.728,
+      "eval_steps_per_second": 0.313,
       "step": 231
     },
     {
       "epoch": 12.467532467532468,
-      "grad_norm": 7.557727813720703,
+      "grad_norm": 7.988156318664551,
       "learning_rate": 8.789062500000001e-06,
-      "loss": 0.3419,
+      "loss": 0.1558,
       "step": 240
     },
     {
       "epoch": 12.987012987012987,
-      "grad_norm": 7.865088939666748,
+      "grad_norm": 6.887962341308594,
       "learning_rate": 6.8359375e-06,
-      "loss": 0.3095,
+      "loss": 0.1307,
       "step": 250
     },
     {
       "epoch": 12.987012987012987,
-      "eval_accuracy": 0.8014705882352942,
-      "eval_loss": 0.47241097688674927,
-      "eval_runtime": 28.3959,
-      "eval_samples_per_second": 4.789,
-      "eval_steps_per_second": 0.317,
+      "eval_accuracy": 0.8970588235294118,
+      "eval_loss": 0.3330402374267578,
+      "eval_runtime": 29.009,
+      "eval_samples_per_second": 4.688,
+      "eval_steps_per_second": 0.31,
       "step": 250
     },
     {
       "epoch": 13.506493506493506,
-      "grad_norm": 15.5108060836792,
+      "grad_norm": 7.4578070640563965,
       "learning_rate": 4.8828125e-06,
-      "loss": 0.3521,
+      "loss": 0.1425,
       "step": 260
     },
     {
       "epoch": 13.974025974025974,
-      "eval_accuracy": 0.8529411764705882,
-      "eval_loss": 0.4389486610889435,
-      "eval_runtime": 28.274,
-      "eval_samples_per_second": 4.81,
-      "eval_steps_per_second": 0.318,
+      "eval_accuracy": 0.9117647058823529,
+      "eval_loss": 0.34208229184150696,
+      "eval_runtime": 28.868,
+      "eval_samples_per_second": 4.711,
+      "eval_steps_per_second": 0.312,
       "step": 269
     },
     {
       "epoch": 14.025974025974026,
-      "grad_norm": 9.100951194763184,
+      "grad_norm": 13.509991645812988,
       "learning_rate": 2.9296875e-06,
-      "loss": 0.3252,
+      "loss": 0.1368,
       "step": 270
     },
     {
       "epoch": 14.545454545454545,
-      "grad_norm": 11.608354568481445,
+      "grad_norm": 9.896185874938965,
       "learning_rate": 9.765625e-07,
-      "loss": 0.3397,
+      "loss": 0.141,
       "step": 280
     },
     {
       "epoch": 14.805194805194805,
-      "eval_accuracy": 0.8455882352941176,
-      "eval_loss": 0.43828412890434265,
-      "eval_runtime": 28.4264,
-      "eval_samples_per_second": 4.784,
-      "eval_steps_per_second": 0.317,
+      "eval_accuracy": 0.9117647058823529,
+      "eval_loss": 0.3409328758716583,
+      "eval_runtime": 28.479,
+      "eval_samples_per_second": 4.775,
+      "eval_steps_per_second": 0.316,
       "step": 285
     },
     {
       "epoch": 14.805194805194805,
       "step": 285,
       "total_flos": 9.04681758989353e+16,
-      "train_loss": 0.6850269221422965,
-      "train_runtime": 4202.2024,
-      "train_samples_per_second": 4.369,
-      "train_steps_per_second": 0.068
+      "train_loss": 0.2430994153022766,
+      "train_runtime": 4010.606,
+      "train_samples_per_second": 4.578,
+      "train_steps_per_second": 0.071
     }
   ],
   "logging_steps": 10,