pre-final notebook

Author: giocopp

.pre-commit-config.yaml

@@ -59,7 +59,7 @@ repos:
         #Automatic linting fixes
         types_or: [ python, pyi, jupyter ]
         args: [ --fix ]
-        #stages: [manual]
+        stages: [manual]
       - id: ruff-format
         types_or: [ python, pyi, jupyter ]
-        #stages: [manual]
+        stages: [manual]

notebooks/exploration.ipynb

@@ -431,7 +431,7 @@
     "import numpy as np\n",
     "\n",
     "\n",
-    "def compute_basic_stats(split=\"train\", category=\"all\", max_images=100):\n",
+    "def compute_basic_stats(split=\"train\", category=\"all\"):\n",
     "    img_dir = dataset_root / split / \"images\" / category\n",
     "    label_dir = dataset_root / split / \"labels\" / category\n",
     "\n",

notebooks/try_few_shot.ipynb

@@ -934,60 +934,6 @@
     "By varying `K` we can study how performance improves as we provide more labelled examples to the model.\n"
    ]
   },
-  {
-   "cell_type": "code",
-   "execution_count": 18,
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "# ============================================================\n",
-    "# 9a. K-shot inference on test images\n",
-    "# ============================================================\n",
-    "\n",
-    "import numpy as np\n",
-    "import torch\n",
-    "\n",
-    "\n",
-    "def evaluate_kshot_iou(encoder, train_dataset, test_dataset, K=5, num_samples=None):\n",
-    "    \"\"\"\n",
-    "    Evaluate K-shot IoU on 'num_samples' random test images.\n",
-    "    For each test image, randomly sample K *distinct* support images from the train set.\n",
-    "    \"\"\"\n",
-    "    encoder.eval()\n",
-    "    rng = np.random.default_rng(42)\n",
-    "\n",
-    "    # If num_samples is None, evaluate on all test samples\n",
-    "    if num_samples is None:\n",
-    "        num_samples = len(test_dataset)\n",
-    "\n",
-    "    ious = []\n",
-    "    for _ in range(num_samples):\n",
-    "        # pick random test index\n",
-    "        ti = rng.integers(0, len(test_dataset))\n",
-    "        img_q, mask_q = test_dataset[ti]\n",
-    "\n",
-    "        # pick K distinct support indices\n",
-    "        support_indices = rng.choice(len(train_dataset), size=K, replace=False)\n",
-    "        support_imgs = []\n",
-    "        support_masks = []\n",
-    "        for si in support_indices:\n",
-    "            img_s, mask_s = train_dataset[si]\n",
-    "            support_imgs.append(img_s)\n",
-    "            support_masks.append(mask_s)\n",
-    "        support_imgs = torch.stack(support_imgs, dim=0)  # [K,3,H,W]\n",
-    "        support_masks = torch.stack(support_masks, dim=0)  # [K,1,H,W]\n",
-    "\n",
-    "        # run K-shot prediction\n",
-    "        logits = k_shot_predict(encoder, support_imgs, support_masks, img_q)  # [1,2,H,W]\n",
-    "\n",
-    "        iou = iou_from_logits(logits, mask_q.unsqueeze(0))\n",
-    "        ious.append(iou)\n",
-    "\n",
-    "    ious = np.array(ious)\n",
-    "    print(f\"{K}-shot mean IoU over {num_samples} test samples: {ious.mean():.3f}\")\n",
-    "    return ious"
-   ]
-  },
   {
    "cell_type": "markdown",
    "metadata": {},