docs: update example notebooks

Author: okunator

examples/pannuke_cellpose_benchmark.ipynb

@@ -36,7 +36,7 @@
    "outputs": [],
    "source": [
     "from cellseg_models_pytorch.models import get_model\n",
-    "from cellseg_models_pytorch.inference import SlidingWindowInferer, ResizeInferer\n",
+    "from cellseg_models_pytorch.inference import ResizeInferer\n",
     "\n",
     "# the pannuke classes\n",
     "type_classes = {\n",
@@ -50,12 +50,12 @@
     "\n",
     "# Initializee the model and input and ckpt paths\n",
     "model = get_model(\"cellpose\", type=\"base\", ntypes=len(type_classes))\n",
-    "input_images = \"/path/to/pannuke/fold1/images/\" # modify this path\n",
-    "ckpt_path = \"/path/to/fold23_weights.ckpt\" # modify this path\n",
+    "input_images = \"/path/to/pannuke/fold1/images/\"  # modify this path\n",
+    "ckpt_path = \"/path/to/fold23_weights.ckpt\"  # modify this path\n",
     "\n",
     "# Initialize the output head args\n",
-    "out_activations={\"type\": \"softmax\", \"cellpose\": None}\n",
-    "out_boundary_weights={\"type\": False, \"cellpose\": False}\n",
+    "out_activations = {\"type\": \"softmax\", \"cellpose\": None}\n",
+    "out_boundary_weights = {\"type\": False, \"cellpose\": False}\n",
     "\n",
     "# Initialize the inferer\n",
     "inferer = ResizeInferer(\n",
@@ -64,9 +64,9 @@
     "    checkpoint_path=ckpt_path,\n",
     "    out_activations=out_activations,\n",
     "    out_boundary_weights=out_boundary_weights,\n",
-    "    resize=(256, 256), # not actually resizing anything\n",
+    "    resize=(256, 256),  # not actually resizing anything\n",
     "    instance_postproc=\"cellpose\",\n",
-    "    normalization=\"minmax\", # minmax normalization was used for training\n",
+    "    normalization=\"minmax\",  # minmax normalization was used for training\n",
     ")"
    ]
   },
@@ -85,7 +85,9 @@
    ],
    "source": [
     "# Run the inference\n",
-    "inferer.infer(save_intermediate=True) # save_intermediate needs to be set to True for latency benchmarks."
+    "inferer.infer(\n",
+    "    save_intermediate=True\n",
+    ")  # save_intermediate needs to be set to True for latency benchmarks."
    ]
   },
   {
@@ -184,7 +186,7 @@
    "source": [
     "from cellseg_models_pytorch.utils.seg_benchmark import SegBenchmarker\n",
     "\n",
-    "in_dir = \"/path/to/pannuke/fold1/labels\" # modify this path\n",
+    "in_dir = \"/path/to/pannuke/fold1/labels\"  # modify this path\n",
     "bm = SegBenchmarker(\n",
     "    true_path=in_dir,\n",
     "    inferer=inferer,\n",
@@ -424,7 +426,7 @@
     }
    ],
    "source": [
-    "# If we filter out also the examples that got \n",
+    "# If we filter out also the examples that got\n",
     "res_df[res_df[[\"pq\", \"dq\", \"sq\"]] > 0].mean(axis=0)"
    ]
   },
@@ -524,7 +526,7 @@
     }
    ],
    "source": [
-    "res_df.loc[res_df['pq'] >= 0].groupby(\"type\").mean(numeric_only=True)"
+    "res_df.loc[res_df[\"pq\"] >= 0].groupby(\"type\").mean(numeric_only=True)"
    ]
   },
   {