.

Author: 37nomad

model-with-mlp-1-head-aug-patch-extraction.ipynb

@@ -546,10 +546,8 @@
     "        class_folder = os.path.join(self.image_dir, str(class_label+1))  \n",
     "        img_path = os.path.join(class_folder, img_name)\n",
     "\n",
-    "        # Load the image\n",
     "        image = Image.open(img_path).convert('RGB')\n",
     "\n",
-    "        # Apply transformations\n",
     "        if self.transform:\n",
     "            image = self.transform(image)\n",
     "\n",
@@ -572,7 +570,7 @@
    },
    "outputs": [],
    "source": [
-    "image_directory = r\"/kaggle/input/raf-db-dataset/DATASET/train\"  # Directory containing class subfolders\n",
+    "image_directory = r\"/kaggle/input/raf-db-dataset/DATASET/train\"  \n",
     "csv_file_path = r\"/kaggle/input/raf-db-dataset/train_labels.csv\""
    ]
   },
@@ -668,24 +666,6 @@
     "torch.cuda.empty_cache()"
    ]
   },
-  {
-   "cell_type": "code",
-   "execution_count": 157,
-   "id": "246285eb",
-   "metadata": {
-    "execution": {
-     "iopub.execute_input": "2024-10-27T05:51:15.818480Z",
-     "iopub.status.busy": "2024-10-27T05:51:15.818189Z",
-     "iopub.status.idle": "2024-10-27T05:51:15.822552Z",
-     "shell.execute_reply": "2024-10-27T05:51:15.821714Z",
-     "shell.execute_reply.started": "2024-10-27T05:51:15.818441Z"
-    }
-   },
-   "outputs": [],
-   "source": [
-    "# checkpoint = torch.load(r\"/kaggle/working/models/epoch_19_acc_79.73_PE_aug_1head_unfreeze_tot20epoch_4heads.pth\")"
-   ]
-  },
   {
    "cell_type": "code",
    "execution_count": 158,
@@ -703,9 +683,7 @@
    "outputs": [],
    "source": [
     "criterion = nn.CrossEntropyLoss()\n",
-    "# optimizer = optim.Adam(model.parameters(), lr=0.001)\n",
-    "optimizer = optim.AdamW(model.parameters(), lr=0.001, weight_decay=0.01)\n",
-    "# optimizer.load_state_dict(checkpoint['optimizer_state_dict'])"
+    "optimizer = optim.AdamW(model.parameters(), lr=0.001, weight_decay=0.01)"
    ]
   },
   {
@@ -834,10 +812,8 @@
    "outputs": [],
    "source": [
     "def train(model, train_loader, test_loader, criterion, optimizer, device, num_epochs=20):\n",
-    "    # train_loader = pl.MpDeviceLoader(train_loader, device)\n",
     "    best_accuracy = 75\n",
     "    model.train()\n",
-    "    # os.makedirs('models', exist_ok=True)\n",
     "    for epoch in range(1, num_epochs+1):\n",
     "        running_loss = 0.0\n",
     "\n",
@@ -851,7 +827,6 @@
     "\n",
     "            loss.backward()\n",
     "            optimizer.step()\n",
-    "            # xm.optimizer_step(optimizer)\n",
     "\n",
     "            running_loss += loss.item()\n",
     "        epoch_loss = running_loss / len(train_loader)\n",
@@ -868,7 +843,7 @@
     "                    logits = model(images)\n",
     "                    \n",
     "                    _, predicted = torch.max(logits, 1)\n",
-    "                    total += labels.size(0)  # Add the batch size to total\n",
+    "                    total += labels.size(0)  \n",
     "                    correct += (predicted == labels).sum().item()\n",
     "            accuracy = 100*(correct/total)\n",
     "            print(f\"test_Acc : {accuracy}\")\n",
@@ -2650,7 +2625,7 @@
     "        logits = model(images)\n",
     "        \n",
     "        _, predicted = torch.max(logits, 1)\n",
-    "        total += labels.size(0)  # Add the batch size to total\n",
+    "        total += labels.size(0) \n",
     "        correct += (predicted == labels).sum().item()\n",
     "accuracy = 100*(correct/total)\n",
     "\n",
@@ -2660,25 +2635,6 @@
     "accuracy\n"
    ]
   },
-  {
-   "cell_type": "code",
-   "execution_count": 297,
-   "id": "307848e9",
-   "metadata": {
-    "execution": {
-     "iopub.execute_input": "2024-10-27T08:36:01.208701Z",
-     "iopub.status.busy": "2024-10-27T08:36:01.208262Z",
-     "iopub.status.idle": "2024-10-27T08:36:01.213300Z",
-     "shell.execute_reply": "2024-10-27T08:36:01.212375Z",
-     "shell.execute_reply.started": "2024-10-27T08:36:01.208658Z"
-    },
-    "scrolled": true
-   },
-   "outputs": [],
-   "source": [
-    "# checkpoint = torch.load(r\"/kaggle/working/models/tot20epoch_4heads_epoch_13_loss_0.07219641906097725.pth\")"
-   ]
-  },
   {
    "cell_type": "code",
    "execution_count": 298,