Add resnet (#246)

* Resnet support added

* Tests fixed
Shapes matching for Resnet50-onnx
Example of Resnet50 to onnx conversion (basic)

* Removed optional conversion from PIL to np.ndarray and now it it's made default
Fixed test accordingly

* Refactoring of pil2ndarray

* Partial support of convnext preprocessing
Resize logic

* normalize canonical value

* Style changes for review

* new: update resnet repo

---------

Co-authored-by: d.rudenko <dimitriyrudenk@gmail.com>
Co-authored-by: George Panchuk <george.panchuk@qdrant.tech>

experiments/Example. Convert Resnet50 to ONNX.ipynb

@@ -0,0 +1,122 @@
+{
+ "cells": [
+  {
+   "cell_type": "code",
+   "execution_count": 1,
+   "id": "4bdb2a91-fa2a-4cee-ad5a-176cc957394d",
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2024-05-23T12:15:28.171586Z",
+     "start_time": "2024-05-23T12:15:28.076314Z"
+    }
+   },
+   "outputs": [
+    {
+     "ename": "ModuleNotFoundError",
+     "evalue": "No module named 'torch'",
+     "output_type": "error",
+     "traceback": [
+      "\u001B[0;31m---------------------------------------------------------------------------\u001B[0m",
+      "\u001B[0;31mModuleNotFoundError\u001B[0m                       Traceback (most recent call last)",
+      "Cell \u001B[0;32mIn[1], line 1\u001B[0m\n\u001B[0;32m----> 1\u001B[0m \u001B[38;5;28;01mimport\u001B[39;00m \u001B[38;5;21;01mtorch\u001B[39;00m\n\u001B[1;32m      2\u001B[0m \u001B[38;5;28;01mimport\u001B[39;00m \u001B[38;5;21;01mtorch\u001B[39;00m\u001B[38;5;21;01m.\u001B[39;00m\u001B[38;5;21;01monnx\u001B[39;00m\n\u001B[1;32m      3\u001B[0m \u001B[38;5;28;01mimport\u001B[39;00m \u001B[38;5;21;01mtorchvision\u001B[39;00m\u001B[38;5;21;01m.\u001B[39;00m\u001B[38;5;21;01mmodels\u001B[39;00m \u001B[38;5;28;01mas\u001B[39;00m \u001B[38;5;21;01mmodels\u001B[39;00m\n",
+      "\u001B[0;31mModuleNotFoundError\u001B[0m: No module named 'torch'"
+     ]
+    }
+   ],
+   "source": [
+    "import torch\n",
+    "import torch.onnx\n",
+    "import torchvision.models as models\n",
+    "import torchvision.transforms as transforms\n",
+    "from PIL import Image\n",
+    "import numpy as np\n",
+    "from tests.config import TEST_MISC_DIR\n",
+    "\n",
+    "# Load pre-trained ResNet-50 model\n",
+    "resnet = models.resnet50(pretrained=True)\n",
+    "resnet = torch.nn.Sequential(*(list(resnet.children())[:-1]))  # Remove the last fully connected layer\n",
+    "resnet.eval()\n",
+    "\n",
+    "# Define preprocessing transform\n",
+    "preprocess = transforms.Compose([\n",
+    "    transforms.Resize(256),\n",
+    "    transforms.CenterCrop(224),\n",
+    "    transforms.ToTensor(),\n",
+    "    transforms.Normalize(mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225]),\n",
+    "])\n",
+    "\n",
+    "# Load and preprocess the image\n",
+    "def preprocess_image(image_path):\n",
+    "    input_image = Image.open(image_path)\n",
+    "    input_tensor = preprocess(input_image)\n",
+    "    input_batch = input_tensor.unsqueeze(0)  # Add batch dimension\n",
+    "    return input_batch\n",
+    "\n",
+    "# Example input for exporting\n",
+    "input_image = preprocess_image('example.jpg')\n",
+    "\n",
+    "# Export the model to ONNX with dynamic axes\n",
+    "torch.onnx.export(\n",
+    "    resnet, \n",
+    "    input_image, \n",
+    "    \"model.onnx\", \n",
+    "    export_params=True, \n",
+    "    opset_version=9, \n",
+    "    input_names=['input'], \n",
+    "    output_names=['output'],\n",
+    "    dynamic_axes={'input': {0: 'batch_size'}, 'output': {0: 'batch_size'}}\n",
+    ")\n",
+    "\n",
+    "# Load ONNX model\n",
+    "import onnx\n",
+    "import onnxruntime as ort\n",
+    "\n",
+    "onnx_model = onnx.load(\"model.onnx\")\n",
+    "ort_session = ort.InferenceSession(\"model.onnx\")\n",
+    "\n",
+    "# Run inference and extract feature vectors\n",
+    "def extract_feature_vectors(image_paths):\n",
+    "    input_images = [preprocess_image(image_path) for image_path in image_paths]\n",
+    "    input_batch = torch.cat(input_images, dim=0)  # Combine images into a single batch\n",
+    "    ort_inputs = {ort_session.get_inputs()[0].name: input_batch.numpy()}\n",
+    "    ort_outs = ort_session.run(None, ort_inputs)\n",
+    "    return ort_outs[0]\n",
+    "\n",
+    "# Example usage\n",
+    "images = [TEST_MISC_DIR / \"image.jpeg\", str(TEST_MISC_DIR / \"small_image.jpeg\")]  # Replace with your image paths\n",
+    "feature_vectors = extract_feature_vectors(images)\n",
+    "print(\"Feature vector shape:\", feature_vectors.shape)\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "outputs": [],
+   "source": [],
+   "metadata": {
+    "collapsed": false
+   },
+   "id": "baa650c4cb3e0e6d"
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "Python 3 (ipykernel)",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.12.2"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}

fastembed/image/onnx_embedding.py

@@ -18,7 +18,17 @@
             "hf": "Qdrant/clip-ViT-B-32-vision",
         },
         "model_file": "model.onnx",
-    }
+    },
+    {
+        "model": "Qdrant/resnet50-onnx",
+        "dim": 2048,
+        "description": "ResNet-50 from `Deep Residual Learning for Image Recognition <https://arxiv.org/abs/1512.03385>`__.",
+        "size_in_GB": 0.1,
+        "sources": {
+            "hf": "Qdrant/resnet50-onnx",
+        },
+        "model_file": "model.onnx",
+    },
 ]
 
 

fastembed/image/onnx_image_model.py

@@ -46,15 +46,19 @@ def load_onnx_model(
         )
         self.processor = load_preprocessor(model_dir=model_dir)
 
+    def _build_onnx_input(self, encoded: np.ndarray) -> Dict[str, np.ndarray]:
+        return {node.name: encoded for node in self.model.get_inputs()}
+
     def onnx_embed(self, images: List[PathInput]) -> OnnxOutputContext:
         with contextlib.ExitStack():
             image_files = [Image.open(image) for image in images]
             encoded = self.processor(image_files)
-        onnx_input = {"pixel_values": encoded}
+        onnx_input = self._build_onnx_input(encoded)
         onnx_input = self._preprocess_onnx_input(onnx_input)
-
         model_output = self.model.run(None, onnx_input)
-        embeddings = model_output[0]
+
+        embeddings = model_output[0].reshape(len(images), -1)
+
         return OnnxOutputContext(
             model_output=embeddings
         )
@@ -82,7 +86,6 @@ def _embed_images(
 
         if parallel is None or is_small:
             for batch in iter_batch(images, batch_size):
-                # open and preprocess images
                 yield from self._post_process_onnx_output(self.onnx_embed(batch))
         else:
             start_method = "forkserver" if "forkserver" in get_all_start_methods() else "spawn"

fastembed/image/transform/functional.py

@@ -14,14 +14,17 @@ def convert_to_rgb(image: Image.Image) -> Image.Image:
 
 
 def center_crop(
-    image: Image.Image,
+    image: Union[Image.Image, np.ndarray],
     size: Tuple[int, int],
 ) -> np.ndarray:
-    orig_height, orig_width = image.height, image.width
-    crop_height, crop_width = size
+    if isinstance(image, np.ndarray):
+        _, orig_height, orig_width = image.shape
+    else:
+        orig_height, orig_width = image.height, image.width
+        # (H, W, C) -> (C, H, W)
+        image = np.array(image).transpose((2, 0, 1))
 
-    # (H, W, C) -> (C, H, W)
-    image = np.array(image).transpose((2, 0, 1))
+    crop_height, crop_width = size
 
     # left upper corner (0, 0)
     top = (orig_height - crop_height) // 2
@@ -96,7 +99,7 @@ def normalize(
 def resize(
     image: Image,
     size: Union[int, Tuple[int, int]],
-    resample: Image.Resampling = Image.Resampling.BICUBIC,
+    resample: Image.Resampling = Image.Resampling.BILINEAR,
 ) -> Image:
     if isinstance(size, tuple):
         return image.resize(size, resample)
@@ -109,9 +112,14 @@ def resize(
         new_size = (new_short, new_long)
     else:
         new_size = (new_long, new_short)
-
-    return image.resize(new_size, Image.Resampling.BICUBIC)
+    return image.resize(new_size, resample)
 
 
 def rescale(image: np.ndarray, scale: float, dtype=np.float32) -> np.ndarray:
     return (image * scale).astype(dtype)
+
+
+def pil2ndarray(image: Union[Image.Image, np.ndarray]):
+    if isinstance(image, Image.Image):
+        return np.asarray(image).transpose((2, 0, 1))
+    return image

fastembed/image/transform/operators.py

@@ -2,13 +2,13 @@
 
 import numpy as np
 from PIL import Image
-
 from fastembed.image.transform.functional import (
     center_crop,
     normalize,
     resize,
     convert_to_rgb,
     rescale,
+    pil2ndarray
 )
 
 
@@ -59,68 +59,110 @@ def __init__(self, scale: float = 1 / 255):
     def __call__(self, images: List[np.ndarray]) -> List[np.ndarray]:
         return [rescale(image, scale=self.scale) for image in images]
 
+class PILtoNDarray(Transform):
+    def __call__(self, images: List[Union[Image.Image, np.ndarray]]) -> List[np.ndarray]:
+        return [pil2ndarray(image) for image in images]
 
 class Compose:
     def __init__(self, transforms: List[Transform]):
         self.transforms = transforms
 
-    def __call__(
-        self, images: Union[List[Image.Image], List[np.ndarray]]
-    ) -> Union[List[np.ndarray], List[Image.Image]]:
+    def __call__(self, images: Union[List[Image.Image], List[np.ndarray]]) -> Union[List[np.ndarray], List[Image.Image]]:
         for transform in self.transforms:
             images = transform(images)
         return images
 
     @classmethod
     def from_config(cls, config: Dict[str, Any]) -> "Compose":
         """Creates processor from a config dict.
-
-        Args:
-            config (Dict[str, Any]): Configuration dictionary.
-
-                Valid keys:
-                    - do_resize
-                    - size
-                    - do_center_crop
-                    - crop_size
-                    - do_rescale
-                    - rescale_factor
-                    - do_normalize
-                    - image_mean
-                    - image_std
-                Valid size keys (nested):
-                    - {"height", "width"}
-                    - {"shortest_edge"}
-
-        Returns:
-            Compose: Image processor.
+            Args:
+                config (Dict[str, Any]): Configuration dictionary.
+
+                    Valid keys:
+                        - do_resize
+                        - size
+                        - do_center_crop
+                        - crop_size
+                        - do_rescale
+                        - rescale_factor
+                        - do_normalize
+                        - image_mean
+                        - image_std
+                    Valid size keys (nested):
+                        - {"height", "width"}
+                        - {"shortest_edge"}
+
+            Returns:
+                Compose: Image processor.
         """
-        transforms = [ConvertToRGB()]
-        if config.get("do_resize", False):
-            size = config["size"]
-            if "shortest_edge" in size:
-                size = size["shortest_edge"]
-            elif "height" in size and "width" in size:
-                size = (size["height"], size["width"])
-            else:
-                raise ValueError(
-                    "Size must contain either 'shortest_edge' or 'height' and 'width'."
-                )
-            transforms.append(
-                Resize(size=size, resample=config.get("resample", Image.Resampling.BICUBIC))
-            )
-        if config.get("do_center_crop", False):
-            crop_size = config["crop_size"]
-            if isinstance(crop_size, int):
-                crop_size = (crop_size, crop_size)
-            elif isinstance(crop_size, dict):
-                crop_size = (crop_size["height"], crop_size["width"])
+        transforms = []
+        cls._get_convert_to_rgb(transforms, config)
+        cls._get_resize(transforms, config)
+        cls._get_center_crop(transforms, config)
+        cls._get_pil2ndarray(transforms, config)
+        cls._get_rescale(transforms, config)
+        cls._get_normalize(transforms, config)
+        return cls(transforms=transforms)
+
+    @staticmethod
+    def _get_convert_to_rgb(transforms: List[Transform], config: Dict[str, Any]):
+        transforms.append(ConvertToRGB())
+
+    @staticmethod
+    def _get_resize(transforms: List[Transform], config: Dict[str, Any]):
+        mode = config.get('image_processor_type', 'CLIPImageProcessor')
+        if mode == 'CLIPImageProcessor':
+            if config.get("do_resize", False):
+                size = config["size"]
+                if "shortest_edge" in size:
+                    size = size["shortest_edge"]
+                elif "height" in size and "width" in size:
+                    size = (size["height"], size["width"])
+                else:
+                    raise ValueError("Size must contain either 'shortest_edge' or 'height' and 'width'.")
+                transforms.append(Resize(size=size, resample=config.get("resample", Image.Resampling.BICUBIC)))
+        elif mode == 'ConvNextFeatureExtractor':
+            if 'size' in config and "shortest_edge" not in config['size']:
+                raise ValueError(f"Size dictionary must contain 'shortest_edge' key. Got {config['size'].keys()}")
+            shortest_edge = config['size']["shortest_edge"]
+            crop_pct = config.get("crop_pct", 0.875)
+            if shortest_edge < 384:
+                # maintain same ratio, resizing shortest edge to shortest_edge/crop_pct
+                resize_shortest_edge = int(shortest_edge / crop_pct)
+                transforms.append(Resize(size=resize_shortest_edge, resample=config.get("resample", Image.Resampling.BICUBIC)))
+                transforms.append(CenterCrop(size=(shortest_edge, shortest_edge)))
             else:
-                raise ValueError(f"Invalid crop size: {crop_size}")
-            transforms.append(CenterCrop(size=crop_size))
+                transforms.append(Resize(size=(shortest_edge, shortest_edge), resample=config.get("resample", Image.Resampling.BICUBIC)))
+
+    @staticmethod
+    def _get_center_crop(transforms: List[Transform], config: Dict[str, Any]):
+        mode = config.get('image_processor_type', 'CLIPImageProcessor')
+        if mode == 'CLIPImageProcessor':
+            if config.get("do_center_crop", False):
+                crop_size = config["crop_size"]
+                if isinstance(crop_size, int):
+                    crop_size = (crop_size, crop_size)
+                elif isinstance(crop_size, dict):
+                    crop_size = (crop_size["height"], crop_size["width"])
+                else:
+                    raise ValueError(f"Invalid crop size: {crop_size}")
+                transforms.append(CenterCrop(size=crop_size))
+        elif mode == 'ConvNextFeatureExtractor':
+            pass
+        else:
+            raise ValueError(f"Preprocessor {mode} is not supported")
+
+    @staticmethod
+    def _get_pil2ndarray(transforms: List[Transform], config: Dict[str, Any]):
+        transforms.append(PILtoNDarray())
+
+    @staticmethod
+    def _get_rescale(transforms: List[Transform], config: Dict[str, Any]):
         if config.get("do_rescale", True):
             rescale_factor = config.get("rescale_factor", 1 / 255)
             transforms.append(Rescale(scale=rescale_factor))
+
+    @staticmethod
+    def _get_normalize(transforms: List[Transform], config: Dict[str, Any]):
         if config.get("do_normalize", False):
             transforms.append(Normalize(mean=config["image_mean"], std=config["image_std"]))
-        return cls(transforms=transforms)