Fix set_epoch bug in Jupyter notebook

Add hasattr check before calling model.set_epoch() to prevent
AttributeError when training baseline BERT models that don't
have this method. Only BP-SOM models have set_epoch().

Bug found during Colab testing.

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

Author: antalvdb

BPSOM_SST2_Experiment.ipynb

@@ -605,84 +605,7 @@
    "execution_count": null,
    "metadata": {},
    "outputs": [],
-   "source": [
-    "def train_epoch(model, dataloader, optimizer, scheduler, device, epoch, max_epochs):\n",
-    "    \"\"\"Train for one epoch.\"\"\"\n",
-    "    model.train()\n",
-    "    model.set_epoch(epoch, max_epochs)\n",
-    "    \n",
-    "    total_loss = 0\n",
-    "    correct = 0\n",
-    "    total = 0\n",
-    "    \n",
-    "    progress_bar = tqdm(dataloader, desc=f\"Epoch {epoch+1}/{max_epochs} [Train]\")\n",
-    "    \n",
-    "    for batch in progress_bar:\n",
-    "        input_ids = batch['input_ids'].to(device)\n",
-    "        attention_mask = batch['attention_mask'].to(device)\n",
-    "        labels = batch['labels'].to(device)\n",
-    "        \n",
-    "        optimizer.zero_grad()\n",
-    "        \n",
-    "        outputs = model(\n",
-    "            input_ids=input_ids,\n",
-    "            attention_mask=attention_mask,\n",
-    "            labels=labels\n",
-    "        )\n",
-    "        \n",
-    "        loss = outputs.loss\n",
-    "        loss.backward()\n",
-    "        optimizer.step()\n",
-    "        scheduler.step()\n",
-    "        \n",
-    "        total_loss += loss.item()\n",
-    "        predictions = torch.argmax(outputs.logits, dim=-1)\n",
-    "        correct += (predictions == labels).sum().item()\n",
-    "        total += labels.size(0)\n",
-    "        \n",
-    "        progress_bar.set_postfix({\n",
-    "            'loss': f\"{loss.item():.4f}\",\n",
-    "            'acc': f\"{correct/total*100:.2f}%\"\n",
-    "        })\n",
-    "    \n",
-    "    return {\n",
-    "        'loss': total_loss / len(dataloader),\n",
-    "        'accuracy': correct / total * 100\n",
-    "    }\n",
-    "\n",
-    "\n",
-    "def eval_epoch(model, dataloader, device, desc=\"Eval\"):\n",
-    "    \"\"\"Evaluate model.\"\"\"\n",
-    "    model.eval()\n",
-    "    \n",
-    "    total_loss = 0\n",
-    "    correct = 0\n",
-    "    total = 0\n",
-    "    \n",
-    "    with torch.no_grad():\n",
-    "        for batch in tqdm(dataloader, desc=desc):\n",
-    "            input_ids = batch['input_ids'].to(device)\n",
-    "            attention_mask = batch['attention_mask'].to(device)\n",
-    "            labels = batch['labels'].to(device)\n",
-    "            \n",
-    "            outputs = model(\n",
-    "                input_ids=input_ids,\n",
-    "                attention_mask=attention_mask,\n",
-    "                labels=labels\n",
-    "            )\n",
-    "            \n",
-    "            total_loss += outputs.loss.item()\n",
-    "            predictions = torch.argmax(outputs.logits, dim=-1)\n",
-    "            correct += (predictions == labels).sum().item()\n",
-    "            total += labels.size(0)\n",
-    "    \n",
-    "    return {\n",
-    "        'loss': total_loss / len(dataloader),\n",
-    "        'accuracy': correct / total * 100\n",
-    "    }\n",
-    "\n",
-    "print(\"✓ Training functions defined\")"
-   ]
+   "source": "def train_epoch(model, dataloader, optimizer, scheduler, device, epoch, max_epochs):\n    \"\"\"Train for one epoch.\"\"\"\n    model.train()\n    \n    # Set epoch for BP-SOM models (baseline models don't have this method)\n    if hasattr(model, 'set_epoch'):\n        model.set_epoch(epoch, max_epochs)\n    \n    total_loss = 0\n    correct = 0\n    total = 0\n    \n    progress_bar = tqdm(dataloader, desc=f\"Epoch {epoch+1}/{max_epochs} [Train]\")\n    \n    for batch in progress_bar:\n        input_ids = batch['input_ids'].to(device)\n        attention_mask = batch['attention_mask'].to(device)\n        labels = batch['labels'].to(device)\n        \n        optimizer.zero_grad()\n        \n        outputs = model(\n            input_ids=input_ids,\n            attention_mask=attention_mask,\n            labels=labels\n        )\n        \n        loss = outputs.loss\n        loss.backward()\n        optimizer.step()\n        scheduler.step()\n        \n        total_loss += loss.item()\n        predictions = torch.argmax(outputs.logits, dim=-1)\n        correct += (predictions == labels).sum().item()\n        total += labels.size(0)\n        \n        progress_bar.set_postfix({\n            'loss': f\"{loss.item():.4f}\",\n            'acc': f\"{correct/total*100:.2f}%\"\n        })\n    \n    return {\n        'loss': total_loss / len(dataloader),\n        'accuracy': correct / total * 100\n}\n\n\ndef eval_epoch(model, dataloader, device, desc=\"Eval\"):\n    \"\"\"Evaluate model.\"\"\"\n    model.eval()\n    \n    total_loss = 0\n    correct = 0\n    total = 0\n    \n    with torch.no_grad():\n        for batch in tqdm(dataloader, desc=desc):\n            input_ids = batch['input_ids'].to(device)\n            attention_mask = batch['attention_mask'].to(device)\n            labels = batch['labels'].to(device)\n            \n            outputs = model(\n                input_ids=input_ids,\n                attention_mask=attention_mask,\n                labels=labels\n            )\n            \n            total_loss += outputs.loss.item()\n            predictions = torch.argmax(outputs.logits, dim=-1)\n            correct += (predictions == labels).sum().item()\n            total += labels.size(0)\n    \n    return {\n        'loss': total_loss / len(dataloader),\n        'accuracy': correct / total * 100\n    }\n\nprint(\"✓ Training functions defined\")"
   },
   {
    "cell_type": "markdown",
@@ -1170,4 +1093,4 @@
  },
  "nbformat": 4,
  "nbformat_minor": 4
-}
+}