End of training

README.md

@@ -17,7 +17,7 @@ should probably proofread and complete it, then remove this comment. -->
 
 This model is a fine-tuned version of [microsoft/swinv2-tiny-patch4-window8-256](https://huggingface.co/microsoft/swinv2-tiny-patch4-window8-256) on the imagefolder dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.0237
+- Loss: 0.0190
 
 ## Model description
 

all_results.json

@@ -1,12 +1,12 @@
 {
-    "epoch": 8.42,
-    "eval_loss": 0.041566625237464905,
-    "eval_runtime": 3.1481,
-    "eval_samples_per_second": 42.248,
-    "eval_steps_per_second": 0.953,
-    "total_flos": 3.274254821086986e+17,
-    "train_loss": 0.2961744487285614,
-    "train_runtime": 328.2191,
-    "train_samples_per_second": 36.378,
-    "train_steps_per_second": 0.122
+    "epoch": 12.63,
+    "eval_loss": 0.018954848870635033,
+    "eval_runtime": 3.0784,
+    "eval_samples_per_second": 43.204,
+    "eval_steps_per_second": 0.975,
+    "total_flos": 4.9113822316304794e+17,
+    "train_loss": 0.24218196471532186,
+    "train_runtime": 426.607,
+    "train_samples_per_second": 41.982,
+    "train_steps_per_second": 0.141
 }

eval_results.json

@@ -1,7 +1,7 @@
 {
-    "epoch": 8.42,
-    "eval_loss": 0.041566625237464905,
-    "eval_runtime": 3.1481,
-    "eval_samples_per_second": 42.248,
-    "eval_steps_per_second": 0.953
+    "epoch": 12.63,
+    "eval_loss": 0.018954848870635033,
+    "eval_runtime": 3.0784,
+    "eval_samples_per_second": 43.204,
+    "eval_steps_per_second": 0.975
 }

runs/Apr23_14-49-28_a738cf30c8c2/events.out.tfevents.1713884210.a738cf30c8c2.34.3

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0238fbc2bc869ec3780fa4756152e727be11c581cb4a2c733b805587b82e92dc
+size 354

train_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 8.42,
-    "total_flos": 3.274254821086986e+17,
-    "train_loss": 0.2961744487285614,
-    "train_runtime": 328.2191,
-    "train_samples_per_second": 36.378,
-    "train_steps_per_second": 0.122
+    "epoch": 12.63,
+    "total_flos": 4.9113822316304794e+17,
+    "train_loss": 0.24218196471532186,
+    "train_runtime": 426.607,
+    "train_samples_per_second": 41.982,
+    "train_steps_per_second": 0.141
 }

trainer_state.json

@@ -1,129 +1,175 @@
 {
-  "best_metric": 0.041566625237464905,
-  "best_model_checkpoint": "swinv2-tiny-patch4-window8-256-Kontur-competition-1.3K/checkpoint-33",
-  "epoch": 8.421052631578947,
+  "best_metric": 0.018954848870635033,
+  "best_model_checkpoint": "swinv2-tiny-patch4-window8-256-Kontur-competition-1.3K/checkpoint-52",
+  "epoch": 12.631578947368421,
   "eval_steps": 500,
-  "global_step": 40,
+  "global_step": 60,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
     {
       "epoch": 0.84,
-      "eval_loss": 0.6306216716766357,
-      "eval_runtime": 5.0675,
-      "eval_samples_per_second": 26.246,
-      "eval_steps_per_second": 0.592,
+      "eval_loss": 0.6833776235580444,
+      "eval_runtime": 3.0114,
+      "eval_samples_per_second": 44.165,
+      "eval_steps_per_second": 0.996,
       "step": 4
     },
     {
       "epoch": 1.89,
-      "eval_loss": 0.41851410269737244,
-      "eval_runtime": 3.8343,
-      "eval_samples_per_second": 34.687,
-      "eval_steps_per_second": 0.782,
+      "eval_loss": 0.4898586869239807,
+      "eval_runtime": 3.4038,
+      "eval_samples_per_second": 39.074,
+      "eval_steps_per_second": 0.881,
       "step": 9
     },
     {
       "epoch": 2.11,
-      "grad_norm": 2.0625433921813965,
-      "learning_rate": 4.166666666666667e-05,
-      "loss": 0.6216,
+      "grad_norm": 2.64339280128479,
+      "learning_rate": 4.62962962962963e-05,
+      "loss": 0.6773,
       "step": 10
     },
     {
       "epoch": 2.95,
-      "eval_loss": 0.23511576652526855,
-      "eval_runtime": 3.8543,
-      "eval_samples_per_second": 34.507,
-      "eval_steps_per_second": 0.778,
+      "eval_loss": 0.2752000093460083,
+      "eval_runtime": 2.6262,
+      "eval_samples_per_second": 50.644,
+      "eval_steps_per_second": 1.142,
       "step": 14
     },
     {
       "epoch": 4.0,
-      "eval_loss": 0.12351791560649872,
-      "eval_runtime": 2.7637,
-      "eval_samples_per_second": 48.123,
-      "eval_steps_per_second": 1.085,
+      "eval_loss": 0.13112162053585052,
+      "eval_runtime": 2.6149,
+      "eval_samples_per_second": 50.862,
+      "eval_steps_per_second": 1.147,
       "step": 19
     },
     {
       "epoch": 4.21,
-      "grad_norm": 2.8277649879455566,
-      "learning_rate": 2.777777777777778e-05,
-      "loss": 0.3128,
+      "grad_norm": 2.5273399353027344,
+      "learning_rate": 3.7037037037037037e-05,
+      "loss": 0.3528,
       "step": 20
     },
     {
       "epoch": 4.84,
-      "eval_loss": 0.07665743678808212,
-      "eval_runtime": 2.651,
-      "eval_samples_per_second": 50.171,
-      "eval_steps_per_second": 1.132,
+      "eval_loss": 0.08120357245206833,
+      "eval_runtime": 3.1533,
+      "eval_samples_per_second": 42.178,
+      "eval_steps_per_second": 0.951,
       "step": 23
     },
     {
       "epoch": 5.89,
-      "eval_loss": 0.06760228425264359,
-      "eval_runtime": 2.7021,
-      "eval_samples_per_second": 49.22,
-      "eval_steps_per_second": 1.11,
+      "eval_loss": 0.04769232124090195,
+      "eval_runtime": 2.61,
+      "eval_samples_per_second": 50.958,
+      "eval_steps_per_second": 1.149,
       "step": 28
     },
     {
       "epoch": 6.32,
-      "grad_norm": 4.639766216278076,
-      "learning_rate": 1.388888888888889e-05,
-      "loss": 0.1492,
+      "grad_norm": 13.101208686828613,
+      "learning_rate": 2.777777777777778e-05,
+      "loss": 0.1595,
       "step": 30
     },
     {
       "epoch": 6.95,
-      "eval_loss": 0.041566625237464905,
-      "eval_runtime": 2.6918,
-      "eval_samples_per_second": 49.41,
-      "eval_steps_per_second": 1.115,
+      "eval_loss": 0.028914544731378555,
+      "eval_runtime": 3.2525,
+      "eval_samples_per_second": 40.891,
+      "eval_steps_per_second": 0.922,
       "step": 33
     },
     {
       "epoch": 8.0,
-      "eval_loss": 0.04927726835012436,
-      "eval_runtime": 2.6173,
-      "eval_samples_per_second": 50.815,
-      "eval_steps_per_second": 1.146,
+      "eval_loss": 0.022866781800985336,
+      "eval_runtime": 2.636,
+      "eval_samples_per_second": 50.456,
+      "eval_steps_per_second": 1.138,
       "step": 38
     },
     {
       "epoch": 8.42,
-      "grad_norm": 3.769456624984741,
-      "learning_rate": 0.0,
-      "loss": 0.1011,
+      "grad_norm": 7.744855880737305,
+      "learning_rate": 1.8518518518518518e-05,
+      "loss": 0.1054,
       "step": 40
     },
     {
-      "epoch": 8.42,
-      "eval_loss": 0.05041909217834473,
-      "eval_runtime": 3.2185,
-      "eval_samples_per_second": 41.323,
-      "eval_steps_per_second": 0.932,
-      "step": 40
+      "epoch": 8.84,
+      "eval_loss": 0.02127755992114544,
+      "eval_runtime": 2.7818,
+      "eval_samples_per_second": 47.812,
+      "eval_steps_per_second": 1.078,
+      "step": 42
     },
     {
-      "epoch": 8.42,
-      "step": 40,
-      "total_flos": 3.274254821086986e+17,
-      "train_loss": 0.2961744487285614,
-      "train_runtime": 328.2191,
-      "train_samples_per_second": 36.378,
-      "train_steps_per_second": 0.122
+      "epoch": 9.89,
+      "eval_loss": 0.028130069375038147,
+      "eval_runtime": 3.1605,
+      "eval_samples_per_second": 42.083,
+      "eval_steps_per_second": 0.949,
+      "step": 47
+    },
+    {
+      "epoch": 10.53,
+      "grad_norm": 4.180393695831299,
+      "learning_rate": 9.259259259259259e-06,
+      "loss": 0.0822,
+      "step": 50
+    },
+    {
+      "epoch": 10.95,
+      "eval_loss": 0.018954848870635033,
+      "eval_runtime": 3.079,
+      "eval_samples_per_second": 43.196,
+      "eval_steps_per_second": 0.974,
+      "step": 52
+    },
+    {
+      "epoch": 12.0,
+      "eval_loss": 0.023994147777557373,
+      "eval_runtime": 2.9444,
+      "eval_samples_per_second": 45.17,
+      "eval_steps_per_second": 1.019,
+      "step": 57
+    },
+    {
+      "epoch": 12.63,
+      "grad_norm": 2.7027509212493896,
+      "learning_rate": 0.0,
+      "loss": 0.076,
+      "step": 60
+    },
+    {
+      "epoch": 12.63,
+      "eval_loss": 0.023665426298975945,
+      "eval_runtime": 3.0873,
+      "eval_samples_per_second": 43.079,
+      "eval_steps_per_second": 0.972,
+      "step": 60
+    },
+    {
+      "epoch": 12.63,
+      "step": 60,
+      "total_flos": 4.9113822316304794e+17,
+      "train_loss": 0.24218196471532186,
+      "train_runtime": 426.607,
+      "train_samples_per_second": 41.982,
+      "train_steps_per_second": 0.141
     }
   ],
   "logging_steps": 10,
-  "max_steps": 40,
+  "max_steps": 60,
   "num_input_tokens_seen": 0,
-  "num_train_epochs": 10,
+  "num_train_epochs": 15,
   "save_steps": 500,
-  "total_flos": 3.274254821086986e+17,
+  "total_flos": 4.9113822316304794e+17,
   "train_batch_size": 64,
   "trial_name": null,
   "trial_params": null