Upload folder using huggingface_hub

.gitattributes

@@ -36,3 +36,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 efficientdetv2/aug_model/model.keras filter=lfs diff=lfs merge=lfs -text
 efficientdetv2/noaug_model/model.keras filter=lfs diff=lfs merge=lfs -text
 efficientdetv2_cbam/noaug_model/model.keras filter=lfs diff=lfs merge=lfs -text
+efficientdetv2_cbam/aug_model/model.keras filter=lfs diff=lfs merge=lfs -text

Spectrogram.ipynb

@@ -997,7 +997,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 21,
+   "execution_count": 27,
    "id": "5fff93e3-1468-4623-930e-095d192d19c0",
    "metadata": {},
    "outputs": [
@@ -1012,7 +1012,7 @@
      "name": "stderr",
      "output_type": "stream",
      "text": [
-      "2025-04-19 08:36:26.446942: I external/local_xla/xla/stream_executor/cuda/subprocess_compilation.cc:346] ptxas warning : Registers are spilled to local memory in function 'input_multiply_reduce_fusion_47', 124 bytes spill stores, 116 bytes spill loads\n",
+      "2025-04-19 09:42:04.484751: I external/local_xla/xla/stream_executor/cuda/subprocess_compilation.cc:346] ptxas warning : Registers are spilled to local memory in function 'input_multiply_reduce_fusion_47', 124 bytes spill stores, 116 bytes spill loads\n",
       "\n"
      ]
     },
@@ -1020,14 +1020,14 @@
      "name": "stdout",
      "output_type": "stream",
      "text": [
-      "\u001b[1m16/17\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m━━\u001b[0m \u001b[1m0s\u001b[0m 147ms/step - accuracy: 0.0330 - loss: 5.9164     "
+      "\u001b[1m16/17\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m━━\u001b[0m \u001b[1m0s\u001b[0m 151ms/step - accuracy: 0.0309 - loss: 6.2074     "
      ]
     },
     {
      "name": "stderr",
      "output_type": "stream",
      "text": [
-      "2025-04-19 08:36:53.116794: I external/local_xla/xla/stream_executor/cuda/subprocess_compilation.cc:346] ptxas warning : Registers are spilled to local memory in function 'input_multiply_reduce_fusion_24', 124 bytes spill stores, 116 bytes spill loads\n",
+      "2025-04-19 09:42:31.060623: I external/local_xla/xla/stream_executor/cuda/subprocess_compilation.cc:346] ptxas warning : Registers are spilled to local memory in function 'input_multiply_reduce_fusion_24', 124 bytes spill stores, 116 bytes spill loads\n",
       "\n"
      ]
     },
@@ -1035,115 +1035,75 @@
      "name": "stdout",
      "output_type": "stream",
      "text": [
-      "\u001b[1m17/17\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m85s\u001b[0m 2s/step - accuracy: 0.0349 - loss: 5.7857 - val_accuracy: 0.0000e+00 - val_loss: 7.4727 - learning_rate: 0.0010\n",
+      "\u001b[1m17/17\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m85s\u001b[0m 2s/step - accuracy: 0.0328 - loss: 6.0940 - val_accuracy: 0.0429 - val_loss: 6.4578 - learning_rate: 0.0010\n",
       "Epoch 2/100\n",
-      "\u001b[1m17/17\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 169ms/step - accuracy: 0.0876 - loss: 3.0671 - val_accuracy: 0.0458 - val_loss: 5.8671 - learning_rate: 0.0010\n",
+      "\u001b[1m17/17\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 172ms/step - accuracy: 0.1411 - loss: 2.8682 - val_accuracy: 0.0429 - val_loss: 5.2527 - learning_rate: 0.0010\n",
       "Epoch 3/100\n",
-      "\u001b[1m17/17\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 161ms/step - accuracy: 0.1480 - loss: 2.7784 - val_accuracy: 0.0429 - val_loss: 5.1126 - learning_rate: 0.0010\n",
+      "\u001b[1m17/17\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 171ms/step - accuracy: 0.3634 - loss: 2.0056 - val_accuracy: 0.0000e+00 - val_loss: 8.8178 - learning_rate: 0.0010\n",
       "Epoch 4/100\n",
-      "\u001b[1m17/17\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 166ms/step - accuracy: 0.2415 - loss: 2.3438 - val_accuracy: 0.0409 - val_loss: 5.5744 - learning_rate: 0.0010\n",
+      "\u001b[1m17/17\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 161ms/step - accuracy: 0.5757 - loss: 1.2645 - val_accuracy: 0.0000e+00 - val_loss: 9.1331 - learning_rate: 0.0010\n",
       "Epoch 5/100\n",
-      "\u001b[1m17/17\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 168ms/step - accuracy: 0.3587 - loss: 1.9786 - val_accuracy: 0.0000e+00 - val_loss: 8.7867 - learning_rate: 0.0010\n",
+      "\u001b[1m16/17\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m━━\u001b[0m \u001b[1m0s\u001b[0m 153ms/step - accuracy: 0.7069 - loss: 0.8994\n",
+      "Epoch 5: ReduceLROnPlateau reducing learning rate to 0.0005000000237487257.\n",
+      "\u001b[1m17/17\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 169ms/step - accuracy: 0.7093 - loss: 0.8910 - val_accuracy: 0.0000e+00 - val_loss: 7.4638 - learning_rate: 0.0010\n",
       "Epoch 6/100\n",
-      "\u001b[1m16/17\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m━━\u001b[0m \u001b[1m0s\u001b[0m 154ms/step - accuracy: 0.4741 - loss: 1.5593\n",
-      "Epoch 6: ReduceLROnPlateau reducing learning rate to 0.0005000000237487257.\n",
-      "\u001b[1m17/17\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 169ms/step - accuracy: 0.4767 - loss: 1.5513 - val_accuracy: 0.0253 - val_loss: 7.2059 - learning_rate: 0.0010\n",
+      "\u001b[1m17/17\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 169ms/step - accuracy: 0.7942 - loss: 0.6128 - val_accuracy: 0.0517 - val_loss: 3.4742 - learning_rate: 5.0000e-04\n",
       "Epoch 7/100\n",
-      "\u001b[1m17/17\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 164ms/step - accuracy: 0.5727 - loss: 1.2536 - val_accuracy: 0.0507 - val_loss: 5.7393 - learning_rate: 5.0000e-04\n",
+      "\u001b[1m17/17\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 175ms/step - accuracy: 0.8509 - loss: 0.4914 - val_accuracy: 0.1033 - val_loss: 3.5149 - learning_rate: 5.0000e-04\n",
       "Epoch 8/100\n",
-      "\u001b[1m17/17\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 161ms/step - accuracy: 0.6373 - loss: 1.0414 - val_accuracy: 0.0439 - val_loss: 5.9624 - learning_rate: 5.0000e-04\n",
+      "\u001b[1m17/17\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 166ms/step - accuracy: 0.8810 - loss: 0.4040 - val_accuracy: 0.0887 - val_loss: 3.4355 - learning_rate: 5.0000e-04\n",
       "Epoch 9/100\n",
-      "\u001b[1m17/17\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 161ms/step - accuracy: 0.7017 - loss: 0.8525 - val_accuracy: 0.0653 - val_loss: 3.8481 - learning_rate: 5.0000e-04\n",
+      "\u001b[1m17/17\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 160ms/step - accuracy: 0.8782 - loss: 0.3804 - val_accuracy: 0.0838 - val_loss: 3.5252 - learning_rate: 5.0000e-04\n",
       "Epoch 10/100\n",
-      "\u001b[1m17/17\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 160ms/step - accuracy: 0.7394 - loss: 0.7493 - val_accuracy: 0.0429 - val_loss: 4.9525 - learning_rate: 5.0000e-04\n",
+      "\u001b[1m17/17\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 166ms/step - accuracy: 0.9003 - loss: 0.3083 - val_accuracy: 0.0409 - val_loss: 3.9652 - learning_rate: 5.0000e-04\n",
       "Epoch 11/100\n",
-      "\u001b[1m17/17\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 170ms/step - accuracy: 0.7771 - loss: 0.6399 - val_accuracy: 0.0916 - val_loss: 3.6721 - learning_rate: 5.0000e-04\n",
+      "\u001b[1m16/17\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m━━\u001b[0m \u001b[1m0s\u001b[0m 146ms/step - accuracy: 0.9139 - loss: 0.2681\n",
+      "Epoch 11: ReduceLROnPlateau reducing learning rate to 0.0002500000118743628.\n",
+      "\u001b[1m17/17\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 160ms/step - accuracy: 0.9135 - loss: 0.2698 - val_accuracy: 0.0409 - val_loss: 3.6442 - learning_rate: 5.0000e-04\n",
       "Epoch 12/100\n",
-      "\u001b[1m17/17\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 175ms/step - accuracy: 0.8078 - loss: 0.5768 - val_accuracy: 0.0419 - val_loss: 4.3079 - learning_rate: 5.0000e-04\n",
+      "\u001b[1m17/17\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 164ms/step - accuracy: 0.9171 - loss: 0.2677 - val_accuracy: 0.0897 - val_loss: 3.7882 - learning_rate: 2.5000e-04\n",
       "Epoch 13/100\n",
-      "\u001b[1m17/17\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 165ms/step - accuracy: 0.8301 - loss: 0.5190 - val_accuracy: 0.0848 - val_loss: 3.3119 - learning_rate: 5.0000e-04\n",
+      "\u001b[1m17/17\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 166ms/step - accuracy: 0.9362 - loss: 0.2024 - val_accuracy: 0.6209 - val_loss: 1.2906 - learning_rate: 2.5000e-04\n",
       "Epoch 14/100\n",
-      "\u001b[1m17/17\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 172ms/step - accuracy: 0.8531 - loss: 0.4740 - val_accuracy: 0.0789 - val_loss: 3.5706 - learning_rate: 5.0000e-04\n",
+      "\u001b[1m17/17\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 172ms/step - accuracy: 0.9436 - loss: 0.1950 - val_accuracy: 0.3587 - val_loss: 2.3459 - learning_rate: 2.5000e-04\n",
       "Epoch 15/100\n",
-      "\u001b[1m17/17\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 158ms/step - accuracy: 0.8778 - loss: 0.3741 - val_accuracy: 0.0760 - val_loss: 3.4350 - learning_rate: 5.0000e-04\n",
+      "\u001b[1m17/17\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 172ms/step - accuracy: 0.9492 - loss: 0.1696 - val_accuracy: 0.1472 - val_loss: 3.4519 - learning_rate: 2.5000e-04\n",
       "Epoch 16/100\n",
-      "\u001b[1m17/17\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 164ms/step - accuracy: 0.8828 - loss: 0.3522 - val_accuracy: 0.1530 - val_loss: 3.2019 - learning_rate: 5.0000e-04\n",
+      "\u001b[1m16/17\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m━━\u001b[0m \u001b[1m0s\u001b[0m 151ms/step - accuracy: 0.9516 - loss: 0.1625\n",
+      "Epoch 16: ReduceLROnPlateau reducing learning rate to 0.0001250000059371814.\n",
+      "\u001b[1m17/17\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 166ms/step - accuracy: 0.9514 - loss: 0.1623 - val_accuracy: 0.5439 - val_loss: 1.6761 - learning_rate: 2.5000e-04\n",
       "Epoch 17/100\n",
-      "\u001b[1m17/17\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 167ms/step - accuracy: 0.8956 - loss: 0.3078 - val_accuracy: 0.0624 - val_loss: 3.5924 - learning_rate: 5.0000e-04\n",
+      "\u001b[1m17/17\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 171ms/step - accuracy: 0.9592 - loss: 0.1414 - val_accuracy: 0.2778 - val_loss: 3.0580 - learning_rate: 1.2500e-04\n",
       "Epoch 18/100\n",
-      "\u001b[1m17/17\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 158ms/step - accuracy: 0.8968 - loss: 0.3075 - val_accuracy: 0.0906 - val_loss: 3.4831 - learning_rate: 5.0000e-04\n",
+      "\u001b[1m17/17\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 177ms/step - accuracy: 0.9646 - loss: 0.1259 - val_accuracy: 0.8889 - val_loss: 0.4246 - learning_rate: 1.2500e-04\n",
       "Epoch 19/100\n",
-      "\u001b[1m16/17\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m━━\u001b[0m \u001b[1m0s\u001b[0m 152ms/step - accuracy: 0.9154 - loss: 0.2567\n",
-      "Epoch 19: ReduceLROnPlateau reducing learning rate to 0.0002500000118743628.\n",
-      "\u001b[1m17/17\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 166ms/step - accuracy: 0.9153 - loss: 0.2573 - val_accuracy: 0.0634 - val_loss: 3.5743 - learning_rate: 5.0000e-04\n",
+      "\u001b[1m17/17\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 166ms/step - accuracy: 0.9648 - loss: 0.1228 - val_accuracy: 0.4152 - val_loss: 2.2136 - learning_rate: 1.2500e-04\n",
       "Epoch 20/100\n",
-      "\u001b[1m17/17\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 172ms/step - accuracy: 0.9301 - loss: 0.2228 - val_accuracy: 0.0887 - val_loss: 3.9743 - learning_rate: 2.5000e-04\n",
+      "\u001b[1m17/17\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 168ms/step - accuracy: 0.9677 - loss: 0.1173 - val_accuracy: 0.6511 - val_loss: 1.2789 - learning_rate: 1.2500e-04\n",
       "Epoch 21/100\n",
-      "\u001b[1m17/17\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 171ms/step - accuracy: 0.9278 - loss: 0.2197 - val_accuracy: 0.8109 - val_loss: 0.6371 - learning_rate: 2.5000e-04\n",
+      "\u001b[1m16/17\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m━━\u001b[0m \u001b[1m0s\u001b[0m 146ms/step - accuracy: 0.9638 - loss: 0.1178\n",
+      "Epoch 21: ReduceLROnPlateau reducing learning rate to 6.25000029685907e-05.\n",
+      "\u001b[1m17/17\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 161ms/step - accuracy: 0.9637 - loss: 0.1186 - val_accuracy: 0.7222 - val_loss: 1.0544 - learning_rate: 1.2500e-04\n",
       "Epoch 22/100\n",
-      "\u001b[1m17/17\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 167ms/step - accuracy: 0.9439 - loss: 0.1691 - val_accuracy: 0.1140 - val_loss: 3.0610 - learning_rate: 2.5000e-04\n",
+      "\u001b[1m17/17\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 173ms/step - accuracy: 0.9685 - loss: 0.1068 - val_accuracy: 0.8694 - val_loss: 0.4784 - learning_rate: 6.2500e-05\n",
       "Epoch 23/100\n",
-      "\u001b[1m17/17\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 170ms/step - accuracy: 0.9528 - loss: 0.1569 - val_accuracy: 0.3928 - val_loss: 2.1225 - learning_rate: 2.5000e-04\n",
+      "\u001b[1m17/17\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 170ms/step - accuracy: 0.9716 - loss: 0.0956 - val_accuracy: 0.7953 - val_loss: 0.7334 - learning_rate: 6.2500e-05\n",
       "Epoch 24/100\n",
-      "\u001b[1m16/17\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m━━\u001b[0m \u001b[1m0s\u001b[0m 151ms/step - accuracy: 0.9551 - loss: 0.1495\n",
-      "Epoch 24: ReduceLROnPlateau reducing learning rate to 0.0001250000059371814.\n",
-      "\u001b[1m17/17\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 167ms/step - accuracy: 0.9547 - loss: 0.1500 - val_accuracy: 0.0916 - val_loss: 3.7034 - learning_rate: 2.5000e-04\n",
+      "\u001b[1m16/17\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m━━\u001b[0m \u001b[1m0s\u001b[0m 148ms/step - accuracy: 0.9712 - loss: 0.0965\n",
+      "Epoch 24: ReduceLROnPlateau reducing learning rate to 3.125000148429535e-05.\n",
+      "\u001b[1m17/17\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 164ms/step - accuracy: 0.9711 - loss: 0.0968 - val_accuracy: 0.7719 - val_loss: 0.8042 - learning_rate: 6.2500e-05\n",
       "Epoch 25/100\n",
-      "\u001b[1m17/17\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 158ms/step - accuracy: 0.9599 - loss: 0.1338 - val_accuracy: 0.3168 - val_loss: 2.7638 - learning_rate: 1.2500e-04\n",
+      "\u001b[1m17/17\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 163ms/step - accuracy: 0.9692 - loss: 0.0952 - val_accuracy: 0.8811 - val_loss: 0.4304 - learning_rate: 3.1250e-05\n",
       "Epoch 26/100\n",
-      "\u001b[1m17/17\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 167ms/step - accuracy: 0.9622 - loss: 0.1254 - val_accuracy: 0.1979 - val_loss: 3.0109 - learning_rate: 1.2500e-04\n",
+      "\u001b[1m17/17\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 168ms/step - accuracy: 0.9727 - loss: 0.0918 - val_accuracy: 0.8713 - val_loss: 0.4808 - learning_rate: 3.1250e-05\n",
       "Epoch 27/100\n",
-      "\u001b[1m16/17\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m━━\u001b[0m \u001b[1m0s\u001b[0m 151ms/step - accuracy: 0.9652 - loss: 0.1105\n",
-      "Epoch 27: ReduceLROnPlateau reducing learning rate to 6.25000029685907e-05.\n",
-      "\u001b[1m17/17\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 168ms/step - accuracy: 0.9651 - loss: 0.1112 - val_accuracy: 0.7768 - val_loss: 0.8127 - learning_rate: 1.2500e-04\n",
+      "\u001b[1m16/17\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m━━\u001b[0m \u001b[1m0s\u001b[0m 153ms/step - accuracy: 0.9752 - loss: 0.0891\n",
+      "Epoch 27: ReduceLROnPlateau reducing learning rate to 1.5625000742147677e-05.\n",
+      "\u001b[1m17/17\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 168ms/step - accuracy: 0.9752 - loss: 0.0892 - val_accuracy: 0.8519 - val_loss: 0.5557 - learning_rate: 3.1250e-05\n",
       "Epoch 28/100\n",
-      "\u001b[1m17/17\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 166ms/step - accuracy: 0.9574 - loss: 0.1175 - val_accuracy: 0.7573 - val_loss: 0.9513 - learning_rate: 6.2500e-05\n",
-      "Epoch 29/100\n",
-      "\u001b[1m17/17\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 163ms/step - accuracy: 0.9597 - loss: 0.1185 - val_accuracy: 0.8791 - val_loss: 0.4387 - learning_rate: 6.2500e-05\n",
-      "Epoch 30/100\n",
-      "\u001b[1m17/17\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 161ms/step - accuracy: 0.9656 - loss: 0.1038 - val_accuracy: 0.8743 - val_loss: 0.4810 - learning_rate: 6.2500e-05\n",
-      "Epoch 31/100\n",
-      "\u001b[1m17/17\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 166ms/step - accuracy: 0.9637 - loss: 0.1088 - val_accuracy: 0.7895 - val_loss: 0.7801 - learning_rate: 6.2500e-05\n",
-      "Epoch 32/100\n",
-      "\u001b[1m16/17\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m━━\u001b[0m \u001b[1m0s\u001b[0m 151ms/step - accuracy: 0.9670 - loss: 0.1002\n",
-      "Epoch 32: ReduceLROnPlateau reducing learning rate to 3.125000148429535e-05.\n",
-      "\u001b[1m17/17\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 165ms/step - accuracy: 0.9668 - loss: 0.1011 - val_accuracy: 0.7934 - val_loss: 0.7480 - learning_rate: 6.2500e-05\n",
-      "Epoch 33/100\n",
-      "\u001b[1m17/17\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 167ms/step - accuracy: 0.9672 - loss: 0.1050 - val_accuracy: 0.8840 - val_loss: 0.4324 - learning_rate: 3.1250e-05\n",
-      "Epoch 34/100\n",
-      "\u001b[1m17/17\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 161ms/step - accuracy: 0.9754 - loss: 0.0859 - val_accuracy: 0.8977 - val_loss: 0.4095 - learning_rate: 3.1250e-05\n",
-      "Epoch 35/100\n",
-      "\u001b[1m17/17\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 172ms/step - accuracy: 0.9718 - loss: 0.0974 - val_accuracy: 0.8889 - val_loss: 0.4155 - learning_rate: 3.1250e-05\n",
-      "Epoch 36/100\n",
-      "\u001b[1m17/17\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 172ms/step - accuracy: 0.9708 - loss: 0.0970 - val_accuracy: 0.8986 - val_loss: 0.3981 - learning_rate: 3.1250e-05\n",
-      "Epoch 37/100\n",
-      "\u001b[1m17/17\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 158ms/step - accuracy: 0.9722 - loss: 0.0954 - val_accuracy: 0.8733 - val_loss: 0.4616 - learning_rate: 3.1250e-05\n",
-      "Epoch 38/100\n",
-      "\u001b[1m17/17\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 163ms/step - accuracy: 0.9702 - loss: 0.0894 - val_accuracy: 0.8343 - val_loss: 0.6653 - learning_rate: 3.1250e-05\n",
-      "Epoch 39/100\n",
-      "\u001b[1m16/17\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m━━\u001b[0m \u001b[1m0s\u001b[0m 147ms/step - accuracy: 0.9748 - loss: 0.0883\n",
-      "Epoch 39: ReduceLROnPlateau reducing learning rate to 1.5625000742147677e-05.\n",
-      "\u001b[1m17/17\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 163ms/step - accuracy: 0.9748 - loss: 0.0880 - val_accuracy: 0.8402 - val_loss: 0.6305 - learning_rate: 3.1250e-05\n",
-      "Epoch 40/100\n",
-      "\u001b[1m17/17\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 164ms/step - accuracy: 0.9773 - loss: 0.0844 - val_accuracy: 0.8694 - val_loss: 0.5243 - learning_rate: 1.5625e-05\n",
-      "Epoch 41/100\n",
-      "\u001b[1m17/17\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 166ms/step - accuracy: 0.9761 - loss: 0.0825 - val_accuracy: 0.8830 - val_loss: 0.4835 - learning_rate: 1.5625e-05\n",
-      "Epoch 42/100\n",
-      "\u001b[1m16/17\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m━━\u001b[0m \u001b[1m0s\u001b[0m 154ms/step - accuracy: 0.9745 - loss: 0.0862\n",
-      "Epoch 42: ReduceLROnPlateau reducing learning rate to 7.812500371073838e-06.\n",
-      "\u001b[1m17/17\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 169ms/step - accuracy: 0.9746 - loss: 0.0861 - val_accuracy: 0.8869 - val_loss: 0.4429 - learning_rate: 1.5625e-05\n",
-      "Epoch 43/100\n",
-      "\u001b[1m17/17\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 168ms/step - accuracy: 0.9765 - loss: 0.0854 - val_accuracy: 0.8908 - val_loss: 0.4340 - learning_rate: 7.8125e-06\n",
-      "Epoch 44/100\n",
-      "\u001b[1m17/17\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 163ms/step - accuracy: 0.9786 - loss: 0.0839 - val_accuracy: 0.8947 - val_loss: 0.4642 - learning_rate: 7.8125e-06\n",
-      "Epoch 45/100\n",
-      "\u001b[1m16/17\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m━━\u001b[0m \u001b[1m0s\u001b[0m 150ms/step - accuracy: 0.9722 - loss: 0.0899\n",
-      "Epoch 45: ReduceLROnPlateau reducing learning rate to 3.906250185536919e-06.\n",
-      "\u001b[1m17/17\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 165ms/step - accuracy: 0.9722 - loss: 0.0896 - val_accuracy: 0.8879 - val_loss: 0.4508 - learning_rate: 7.8125e-06\n",
-      "Epoch 46/100\n",
-      "\u001b[1m17/17\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 163ms/step - accuracy: 0.9758 - loss: 0.0813 - val_accuracy: 0.8918 - val_loss: 0.4233 - learning_rate: 3.9063e-06\n",
-      "Epoch 46: early stopping\n",
-      "Restoring model weights from the end of the best epoch: 36.\n"
+      "\u001b[1m17/17\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 171ms/step - accuracy: 0.9800 - loss: 0.0760 - val_accuracy: 0.8626 - val_loss: 0.5024 - learning_rate: 1.5625e-05\n",
+      "Epoch 28: early stopping\n",
+      "Restoring model weights from the end of the best epoch: 18.\n"
      ]
     }
    ],
@@ -1155,7 +1115,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 22,
+   "execution_count": 28,
    "id": "ce766ea6-c00e-49c3-b759-ae277ed3089f",
    "metadata": {},
    "outputs": [
@@ -1163,17 +1123,17 @@
      "name": "stdout",
      "output_type": "stream",
      "text": [
-      "\u001b[1m5/5\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 39ms/step - accuracy: 0.8803 - loss: 0.4273 \n",
-      "Test loss: 0.3981, Test accuracy: 0.8986\n"
+      "\u001b[1m5/5\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 40ms/step - accuracy: 0.8663 - loss: 0.4895 \n",
+      "Test loss: 0.4246, Test accuracy: 0.8889\n"
      ]
     },
     {
      "data": {
       "text/plain": [
-       "(0.3981378972530365, 0.8986355066299438)"
+       "(0.4246201515197754, 0.8888888955116272)"
       ]
      },
-     "execution_count": 22,
+     "execution_count": 28,
      "metadata": {},
      "output_type": "execute_result"
     }
@@ -1279,7 +1239,7 @@
    "source": [
     "# Train the model\n",
     "model_cbam_aug = build_efficientnet_cbam(img_height, img_width, NUM_CLASSES)\n",
-    "history_cbam_aug = train_model(model_cbam_aug, aug_ds['train'], aug_ds['test'], '/workspace/models/efficientdetv2_cbam/noaug_model', epochs=EPOCHS)"
+    "history_cbam_aug = train_model(model_cbam_aug, aug_ds['train'], aug_ds['test'], '/workspace/models/efficientdetv2_cbam/aug_model', epochs=EPOCHS)"
    ]
   },
   {

efficientdetv2_cbam/aug_model/model.keras

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:73bd30b0f31458985e8da4ae9c3b39a04e3e5cad18301375498aa7b0279d5869
+size 92027091

efficientdetv2_cbam/noaug_model/model.keras

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:73bd30b0f31458985e8da4ae9c3b39a04e3e5cad18301375498aa7b0279d5869
-size 92027091
+oid sha256:4f22575fe10322aa4bcc4ee71a7ce8b0731225887b98bc19d5561e3e5e76d782
+size 92027118