fix: modify example notebooks inference

Author: okunator

changelog.d/20230922_082631_oskari.lehtonen.md

@@ -0,0 +1,2 @@
+## Features
+- Add option for mixed precision inference

changelog.d/20231010_143903_oskari.lehtonen.md

@@ -0,0 +1,5 @@
+## Features
+- Add support for `.geojson`, `.feather`, `.parquet` file formats when running inference.
+
+## Style
+- Update the `Ìnferer.infer()` -method api to accept arguments related to saving the model outputs.

examples/lizard_nuclei_segmentation_cellpose.ipynb

@@ -67,7 +67,6 @@
     "    resize=(256, 256), # not actually resizing anything\n",
     "    instance_postproc=\"cellpose\",\n",
     "    normalization=\"minmax\", # minmax normalization was used for training\n",
-    "    save_intermediate=True # this needs to be set to True for latency benchmarks.\n",
     ")"
    ]
   },
@@ -86,7 +85,7 @@
    ],
    "source": [
     "# Run the inference\n",
-    "inferer.infer()"
+    "inferer.infer(save_intermediate=True) # save_intermediate needs to be set to True for latency benchmarks."
    ]
   },
   {
@@ -1330,7 +1329,7 @@
    "name": "python",
    "nbconvert_exporter": "python",
    "pygments_lexer": "ipython3",
-   "version": "3.9.12 (main, Apr  5 2022, 06:56:58) \n[GCC 7.5.0]"
+   "version": "3.9.12"
   },
   "orig_nbformat": 4,
   "vscode": {

examples/pannuke_cellpose_benchmark.ipynb

@@ -899,7 +899,6 @@
     "    out_boundary_weights={\"cellpose\":  False, \"type\": False},\n",
     "    resize=(256, 256), # Not actually resizing anything,\n",
     "    instance_postproc=\"cellpose\",\n",
-    "    save_intermediate=True, # save intermediate soft masks for visualization\n",
     "    batch_size=8,\n",
     "    n_images=50, # Use only the 50 first images of the folder\n",
     "    normalization=\"minmax\",\n",
@@ -955,7 +954,6 @@
     "samples = list(inferer.out_masks.keys())\n",
     "for i, ix in enumerate(ixs):\n",
     "    masks = inferer.out_masks[samples[ix]]\n",
-    "    inter = inferer.soft_masks[samples[ix]]\n",
     "\n",
     "    img_path = save_dir / \"valid\" / f\"images/{samples[ix]}.png\"\n",
     "    img = FileHandler.read_img(img_path)\n",

examples/pannuke_nuclei_segmentation_cellpose.ipynb

@@ -895,7 +895,6 @@
     "    out_boundary_weights={\"cellpose\":  False, \"type\": False},\n",
     "    resize=(252, 252), # resize the input images to a multiple of 14,\n",
     "    instance_postproc=\"cellpose\",\n",
-    "    save_intermediate=True, # save intermediate soft masks for visualization\n",
     "    batch_size=8,\n",
     "    n_images=50, # Use only the 50 first images of the folder\n",
     "    normalization=\"percentile\",\n",
@@ -951,7 +950,6 @@
     "samples = list(inferer.out_masks.keys())\n",
     "for i, ix in enumerate(ixs):\n",
     "    masks = inferer.out_masks[samples[ix]]\n",
-    "    inter = inferer.soft_masks[samples[ix]]\n",
     "\n",
     "    img_path = save_dir / \"valid\" / f\"images/{samples[ix]}.png\"\n",
     "    img = FileHandler.read_img(img_path)\n",

examples/pannuke_nuclei_segmentation_cellpose_dinov2.ipynb

@@ -904,7 +904,6 @@
     "    out_boundary_weights={\"hovernet\":  False, \"inst\": False, \"type\": False},\n",
     "    resize=(256, 256), # Not actually resizing anything,\n",
     "    instance_postproc=\"hovernet\",\n",
-    "    save_intermediate=True, # save intermediate soft masks for visualization\n",
     "    batch_size=8,\n",
     "    n_images=50, # Use only the 50 first images of the folder\n",
     "    normalization=\"minmax\",\n",
@@ -960,7 +959,6 @@
     "samples = list(inferer.out_masks.keys())\n",
     "for i, ix in enumerate(ixs):\n",
     "    masks = inferer.out_masks[samples[ix]]\n",
-    "    inter = inferer.soft_masks[samples[ix]]\n",
     "\n",
     "    img_path = save_dir / \"valid\" / f\"images/{samples[ix]}.png\"\n",
     "    img = FileHandler.read_img(img_path)\n",

examples/pannuke_nuclei_segmentation_cellvit.ipynb

@@ -993,7 +993,6 @@
     "    out_boundary_weights=out_weights,\n",
     "    resize=(256, 256), # Not actually resizing anything,\n",
     "    instance_postproc=\"stardist\",\n",
-    "    save_intermediate=True, # save intermediate soft masks for visualization\n",
     "    batch_size=8,\n",
     "    n_images=50, # Use only the 50 first images of the folder,\n",
     "    normalization=\"percentile\"\n",
@@ -1048,7 +1047,6 @@
     "samples = list(inferer.out_masks.keys())\n",
     "for i, ix in enumerate(ixs):\n",
     "    masks = inferer.out_masks[samples[ix]]\n",
-    "    inter = inferer.soft_masks[samples[ix]]\n",
     "\n",
     "    img_path = save_dir / \"valid\" / f\"images/{samples[ix]}.png\"\n",
     "    img = FileHandler.read_img(img_path)\n",

examples/pannuke_nuclei_segmentation_cppnet.ipynb

@@ -2725,7 +2725,6 @@
     "    out_boundary_weights={\"hovernet\": False, \"inst\": False, \"type\": False},\n",
     "    resize=(256, 256), # Not actually resizing anything,\n",
     "    instance_postproc=\"hovernet\",\n",
-    "    save_intermediate=True, # save intermediate soft masks for visualization\n",
     "    batch_size=8,\n",
     "    n_images=50, # Use only the 50 first images of the folder\n",
     "    normalization=\"minmax\",\n",
@@ -2781,7 +2780,6 @@
     "samples = list(inferer.out_masks.keys())\n",
     "for i, ix in enumerate(ixs):\n",
     "    masks = inferer.out_masks[samples[ix]]\n",
-    "    inter = inferer.soft_masks[samples[ix]]\n",
     "\n",
     "    img_path = save_dir / \"valid\" / f\"images/{samples[ix]}.png\"\n",
     "    img = FileHandler.read_img(img_path)\n",

examples/pannuke_nuclei_segmentation_hovernet.ipynb

@@ -753,7 +753,6 @@
     "    out_boundary_weights={\"omnipose\":  False, \"type\": False},\n",
     "    resize=(256, 256), # Not actually resizing anything,\n",
     "    instance_postproc=\"omnipose\",\n",
-    "    save_intermediate=True, # save intermediate soft masks for visualization\n",
     "    batch_size=8,\n",
     "    n_images=50, # Use only the 50 first images of the folder\n",
     "    normalization=\"percentile\",\n",
@@ -809,7 +808,6 @@
     "samples = list(inferer.out_masks.keys())\n",
     "for i, ix in enumerate(ixs):\n",
     "    masks = inferer.out_masks[samples[ix]]\n",
-    "    inter = inferer.soft_masks[samples[ix]]\n",
     "\n",
     "    img_path = save_dir / \"valid\" / f\"images/{samples[ix]}.png\"\n",
     "    img = FileHandler.read_img(img_path)\n",