notebooks/002-openvino-api/002-openvino-api.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "b875b31c",
   "metadata": {},
   "source": [
    "# OpenVINO™ Runtime API Tutorial"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "09abd9f4",
   "metadata": {},
   "source": [
    "This notebook explains the basics of the OpenVINO Runtime API. It covers:\n",
    "\n",
    "- [Loading OpenVINO Runtime and Showing Info](#Loading-OpenVINO-Runtime-and-Showing-Info)\n",
    "- [Loading a Model](#Loading-a-Model)\n",
    "  - [OpenVINO IR Model](#OpenVINO-IR-Model)\n",
    "  - [ONNX Model](#ONNX-Model)\n",
    "  - [PaddlePaddle Model](#PaddlePaddle-Model)\n",
    "  - [TensorFlow Model](#TensorFlow-Model)\n",
    "- [Getting Information about a Model](#Getting-Information-about-a-Model)\n",
    "  - [Model Inputs](#Model-Inputs)\n",
    "  - [Model Outputs](#Model-Outputs)\n",
    "- [Doing Inference on a Model](#Doing-Inference-on-a-Model)\n",
    "- [Reshaping and Resizing](#Reshaping-and-Resizing)\n",
    "  - [Change Image Size](#Change-Image-Size)\n",
    "  - [Change Batch Size](#Change-Batch-Size)\n",
    " - [Caching a Model](#Caching-a-Model)\n",
    "    \n",
    "The notebook is divided into sections with headers. Each section is standalone and does not depend on previous sections. A segmentation and classification OpenVINO IR model and a segmentation ONNX model are provided as examples. These model files can be replaced with your own models. The exact outputs will be different, but the process is the same. "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9ed058f4",
   "metadata": {},
   "source": [
    "## Loading OpenVINO Runtime and Showing Info\n",
    "\n",
    "Initialize OpenVINO Runtime with Core()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "c08b79c4",
   "metadata": {},
   "outputs": [],
   "source": [
    "from openvino.runtime import Core\n",
    "\n",
    "ie = Core()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "68bc4125",
   "metadata": {},
   "source": [
    "OpenVINO Runtime can load a network on a device. A device in this context means a CPU, an Intel GPU, a Neural Compute Stick 2, etc. The `available_devices` property shows the available devices in your system. The \"FULL_DEVICE_NAME\" option to `ie.get_property()` shows the name of the device.\n",
    "\n",
    "In this notebook, the CPU device is used. To use an integrated GPU, use `device_name=\"GPU\"` instead. Be aware that loading a network on GPU will be slower than loading a network on CPU, but inference will likely be faster."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "d0c94f76",
   "metadata": {},
   "outputs": [],
   "source": [
    "devices = ie.available_devices\n",
    "\n",
    "for device in devices:\n",
    "    device_name = ie.get_property(device, \"FULL_DEVICE_NAME\")\n",
    "    print(f\"{device}: {device_name}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "14d62615",
   "metadata": {},
   "source": [
    "## Loading a Model\n",
    "\n",
    "After initializing OpenVINO Runtime, first read the model file with `read_model()`, then compile it to the specified device with the `compile_model()` method. \n",
    "\n",
    "[OpenVINO™ supports several model formats](https://docs.openvino.ai/latest/Supported_Model_Formats.html#doxid-supported-model-formats) and enables developers to convert them to its own OpenVINO IR format using a tool dedicated to this task.\n",
    "\n",
    "### OpenVINO IR Model\n",
    "\n",
    "An OpenVINO IR (Intermediate Representation) model consists of an `.xml` file, containing information about network topology, and a `.bin` file, containing the weights and biases binary data. Models in OpenVINO IR format are obtained by using Model Optimizer tool. The `read_model()` function expects the `.bin` weights file to have the same filename and be located in the same directory as the `.xml` file: `model_weights_file == Path(model_xml).with_suffix(\".bin\")`. If this is the case, specifying the weights file is optional. If the weights file has a different filename, it can be specified using the `weights` parameter in `read_model()`.\n",
    "\n",
    "The OpenVINO [Model Optimizer](https://docs.openvino.ai/latest/openvino_docs_MO_DG_Deep_Learning_Model_Optimizer_DevGuide.html#doxid-openvino-docs-m-o-d-g-deep-learning-model-optimizer-dev-guide) tool is used to convert models to OpenVINO IR format. Model Optimizer reads the original model and creates an OpenVINO IR model (.xml and .bin files) so inference can be performed without delays due to format conversion. Optionally, Model Optimizer can adjust the model to be more suitable for inference, for example, by alternating input shapes, embedding preprocessing and cutting training parts off.\n",
    "For information on how to convert your existing TensorFlow, PyTorch or ONNX model to OpenVINO IR format with Model Optimizer, refer to the [tensorflow-to-openvino](../101-tensorflow-to-openvino/101-tensorflow-to-openvino.ipynb) and [pytorch-onnx-to-openvino](../102-pytorch-onnx-to-openvino/102-pytorch-onnx-to-openvino.ipynb) notebooks. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "523978fe",
   "metadata": {},
   "outputs": [],
   "source": [
    "from openvino.runtime import Core\n",
    "\n",
    "ie = Core()\n",
    "classification_model_xml = \"model/classification.xml\"\n",
    "\n",
    "model = ie.read_model(model=classification_model_xml)\n",
    "compiled_model = ie.compile_model(model=model, device_name=\"CPU\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e5516e87",
   "metadata": {},
   "source": [
    "### ONNX Model\n",
    "\n",
    "[ONNX](https://onnx.ai/) is an open format built to represent machine learning models. ONNX defines a common set of operators - the building blocks of machine learning and deep learning models - and a common file format to enable AI developers to use models with a variety of frameworks, tools, runtimes, and compilers. OpenVINO supports reading models in ONNX format directly,that means they can be used with OpenVINO Runtime without any prior conversion.\n",
    "\n",
    "Reading and loading an ONNX model, which is a single `.onnx` file, works the same way as with an OpenVINO IR model. The `model` argument points to the filename of an ONNX model."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "15833f51",
   "metadata": {},
   "outputs": [],
   "source": [
    "from openvino.runtime import Core\n",
    "\n",
    "ie = Core()\n",
    "onnx_model_path = \"model/segmentation.onnx\"\n",
    "\n",
    "model_onnx = ie.read_model(model=onnx_model_path)\n",
    "compiled_model_onnx = ie.compile_model(model=model_onnx, device_name=\"CPU\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "3d6a41d9-a2de-40a8-b925-fe3e817866d4",
   "metadata": {},
   "source": [
    "The ONNX model can be exported to OpenVINO IR with `serialize()`:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "1610b254-8567-4409-9b26-259b8acd10dd",
   "metadata": {},
   "outputs": [],
   "source": [
    "from openvino.runtime import serialize\n",
    "\n",
    "serialize(model_onnx, xml_path=\"model/exported_onnx_model.xml\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "fb9ba7d2-ef7a-46ed-b8a2-5c47f7b97fec",
   "metadata": {},
   "source": [
    "### PaddlePaddle Model\n",
    "\n",
    "[PaddlePaddle](https://www.paddlepaddle.org.cn/documentation/docs/en/guides/index_en.html) models saved for inference can also be passed to OpenVINO Runtime without any conversion step. Pass the filename with extension to `read_model` and exported an OpenVINO IR with `serialize`\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "d6524ef0-9a76-4254-8614-3a1e24cc9420",
   "metadata": {},
   "outputs": [],
   "source": [
    "from openvino.runtime import Core\n",
    "\n",
    "ie = Core()\n",
    "paddle_model_path = \"model/inference.pdmodel\"\n",
    "\n",
    "model_paddle = ie.read_model(model=paddle_model_path)\n",
    "compiled_model_paddle = ie.compile_model(model=model_paddle, device_name=\"CPU\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "74f8948d-0eed-4b3c-98ae-35870ff82625",
   "metadata": {},
   "outputs": [],
   "source": [
    "from openvino.runtime import serialize\n",
    "\n",
    "serialize(model_paddle, xml_path=\"model/exported_paddle_model.xml\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "1d7fcaff-8ed6-4c96-a46d-2ef813f24eb1",
   "metadata": {},
   "source": [
    "### TensorFlow Model\n",
    "\n",
    "TensorFlow models saved in frozen graph format can also be passed to `read_model` starting in OpenVINO 2022.3.\n",
    "> **NOTE**: Directly loading TensorFlow models is available as a preview feature in the OpenVINO 2022.3 release. Fully functional support will be provided in the upcoming 2023 releases.\n",
    "> Currently support is limited to only frozen graph inference format. Other TensorFlow model formats must be converted to OpenVINO IR using [Model Optimizer](https://docs.openvino.ai/latest/openvino_docs_MO_DG_prepare_model_convert_model_Convert_Model_From_TensorFlow.html)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "85446bed-8787-4fab-b64b-03035ad74bfd",
   "metadata": {},
   "outputs": [],
   "source": [
    "from openvino.runtime import Core\n",
    "\n",
    "ie = Core()\n",
    "tf_model_path = \"model/classification.pb\"\n",
    "\n",
    "model_tf = ie.read_model(model=tf_model_path)\n",
    "compiled_model_tf = ie.compile_model(model=model_tf, device_name=\"CPU\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "48f3cd1c-9343-4fdd-8e8d-9226a4956517",
   "metadata": {},
   "outputs": [],
   "source": [
    "from openvino.runtime import serialize\n",
    "\n",
    "serialize(model_tf, xml_path=\"model/exported_tf_model.xml\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "8ebee450",
   "metadata": {
    "tags": []
   },
   "source": [
    "## Getting Information about a Model\n",
    "\n",
    "The OpenVINO Model instance stores information about the model. Information about the inputs and outputs of the model are in `model.inputs` and `model.outputs`. These are also properties of the CompiledModel instance. While using `model.inputs` and `model.outputs` in the cells below, you can also use `compiled_model.inputs` and `compiled_model.outputs`."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "edc79b32",
   "metadata": {},
   "source": [
    "### Model Inputs\n",
    "\n",
    "Information about all input layers is stored in the `inputs` dictionary."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "5571614d",
   "metadata": {},
   "outputs": [],
   "source": [
    "from openvino.runtime import Core\n",
    "\n",
    "ie = Core()\n",
    "classification_model_xml = \"model/classification.xml\"\n",
    "model = ie.read_model(model=classification_model_xml)\n",
    "model.inputs"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4ebb1299",
   "metadata": {},
   "source": [
    "The cell above shows that the loaded model expects one input with the name _input_. If you loaded a different model, you may see a different input layer name, and you may see more inputs. You may also obtain info about each input layer using `model.input(index)`, where index is a numeric index of the input layers in the model. If a model has only one input, index can be omitted."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "6cf48354",
   "metadata": {},
   "outputs": [],
   "source": [
    "input_layer = model.input(0)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "3f754e26-2223-4b97-9181-46434aa8fbe7",
   "metadata": {},
   "source": [
    "It is often useful to have a reference to the name of the first input layer. For a model with one input, `model.input(0).any_name` gets this name."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "2202aec4-d68f-4cfa-8b8e-3340ec085a8c",
   "metadata": {},
   "outputs": [],
   "source": [
    "input_layer.any_name"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "aea90446",
   "metadata": {},
   "source": [
    "The next cell prints the input layout, precision and shape."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "bd0bc0c1",
   "metadata": {},
   "outputs": [],
   "source": [
    "print(f\"input precision: {input_layer.element_type}\")\n",
    "print(f\"input shape: {input_layer.shape}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d189a73c",
   "metadata": {},
   "source": [
    "This cell shows that the model expects inputs with a shape of [1,3,224,224], and that this is in the `NCHW` layout. This means that the model expects input data with the batch size of 1 (`N`), 3 channels (`C`) , and images with a height (`H`) and width (`W`) equal to 224. The input data is expected to be of `FP32` (floating point) precision."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "155cf48e",
   "metadata": {},
   "source": [
    "### Model Outputs"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "4583eb0e",
   "metadata": {},
   "outputs": [],
   "source": [
    "from openvino.runtime import Core\n",
    "\n",
    "ie = Core()\n",
    "classification_model_xml = \"model/classification.xml\"\n",
    "model = ie.read_model(model=classification_model_xml)\n",
    "model.outputs"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a189c02f",
   "metadata": {},
   "source": [
    "Model output info is stored in `model.outputs`. The cell above shows that the model returns one output, with the `MobilenetV3/Predictions/Softmax` name. Loading a different model will result in different output layer name, and more outputs might be returned. Similar to input, you may also obtain information about each output separately using `model.output(index)`\n",
    "\n",
    "Since this model has one output, follow the same method as for the input layer to get its name."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "88fbbd06",
   "metadata": {},
   "outputs": [],
   "source": [
    "output_layer = model.output(0)\n",
    "output_layer.any_name"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "16ad0240",
   "metadata": {},
   "source": [
    "Getting the output precision and shape is similar to getting the input precision and shape."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "0ee5e14a",
   "metadata": {},
   "outputs": [],
   "source": [
    "print(f\"output precision: {output_layer.element_type}\")\n",
    "print(f\"output shape: {output_layer.shape}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2739f5bb",
   "metadata": {},
   "source": [
    "This cell shows that the model returns outputs with a shape of [1, 1001], where 1 is the batch size (`N`) and 1001 is the number of classes (`C`). The output is returned as 32-bit floating point."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "021708ab",
   "metadata": {},
   "source": [
    "## Doing Inference on a Model\n",
    "\n",
    "> **NOTE** this notebook demonstrates only the basic synchronous inference API. For an async inference example, please refer to [Async API notebook](../115-async-api/115-async-api.ipynb)\n",
    "\n",
    "The diagram below shows a typical inference pipeline with OpenVINO\n",
    "\n",
    "![image.png](https://docs.openvino.ai/latest/_images/IMPLEMENT_PIPELINE_with_API_C.svg)\n",
    "\n",
    "Creating OpenVINO Core and model compilation is covered in the previous steps. The next step is preparing an inference request. To do inference on a model, first create an inference request by calling the `create_infer_request()` method of `CompiledModel`, `compiled_model` that was loaded with `compile_model()`. Then, call the `infer()` method of `InferRequest`. It expects one argument: `inputs`. This is a dictionary that maps input layer names to input data or list of input data in np.ndarray format, where the position of the input tensor corresponds to input index. If a model has a single input, wrapping to a dictionary or list can be omitted. "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "3ec2eac8",
   "metadata": {},
   "source": [
    "**Load the network**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "298c80b5",
   "metadata": {},
   "outputs": [],
   "source": [
    "from openvino.runtime import Core\n",
    "\n",
    "ie = Core()\n",
    "classification_model_xml = \"model/classification.xml\"\n",
    "model = ie.read_model(model=classification_model_xml)\n",
    "compiled_model = ie.compile_model(model=model, device_name=\"CPU\")\n",
    "input_layer = compiled_model.input(0)\n",
    "output_layer = compiled_model.output(0)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "173cd1c9",
   "metadata": {},
   "source": [
    "**Load an image and convert to the input shape**\n",
    "\n",
    "To propagate an image through the network, it needs to be loaded into an array, resized to the shape that the network expects, and converted to the input layout of the network."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "1f23c43a",
   "metadata": {},
   "outputs": [],
   "source": [
    "import cv2\n",
    "\n",
    "image_filename = \"../data/image/coco_hollywood.jpg\"\n",
    "image = cv2.imread(image_filename)\n",
    "image.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6bf541d8",
   "metadata": {},
   "source": [
    "The image has a shape of (663,994,3). It is 663 pixels in height, 994 pixels in width, and has 3 color channels. A reference to the height and width expected by the network is obtained and the image is resized to these dimensions."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "c9f97da2",
   "metadata": {},
   "outputs": [],
   "source": [
    "# N,C,H,W = batch size, number of channels, height, width.\n",
    "N, C, H, W = input_layer.shape\n",
    "# OpenCV resize expects the destination size as (width, height).\n",
    "resized_image = cv2.resize(src=image, dsize=(W, H))\n",
    "resized_image.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "436a6d74",
   "metadata": {},
   "source": [
    "Now, the image has the width and height that the network expects. This is still in `HWC` format and must be changed to `NCHW` format. First, call the `np.transpose()` method to change to `CHW` and then add the `N` dimension (where `N`= 1) by calling the `np.expand_dims()` method. Next, convert the data to `FP32` with `np.astype()` method."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "d09b7275",
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "\n",
    "input_data = np.expand_dims(np.transpose(resized_image, (2, 0, 1)), 0).astype(np.float32)\n",
    "input_data.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "af110efb",
   "metadata": {},
   "source": [
    "**Do inference**\n",
    "\n",
    "Now that the input data is in the right shape, run inference. The CompiledModel inference result is a dictionary where keys are the Output class instances (the same keys in `compiled_model.outputs` that can also be obtained with `compiled_model.output(index)`) and values - predicted result in np.array format."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "098c8cb2",
   "metadata": {
    "pycharm": {
     "name": "#%%\n"
    }
   },
   "outputs": [],
   "source": [
    "# for single input models only\n",
    "result = compiled_model(input_data)[output_layer]\n",
    "\n",
    "# for multiple inputs in a list\n",
    "result = compiled_model([input_data])[output_layer]\n",
    "\n",
    "# or using a dictionary, where the key is input tensor name or index\n",
    "result = compiled_model({input_layer.any_name: input_data})[output_layer]"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "978a131e",
   "metadata": {
    "pycharm": {
     "name": "#%% md\n"
    }
   },
   "source": [
    "You can also create `InferRequest` and run `infer` method on request."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "bf94022c",
   "metadata": {},
   "outputs": [],
   "source": [
    "request = compiled_model.create_infer_request()\n",
    "request.infer(inputs={input_layer.any_name: input_data})\n",
    "result = request.get_output_tensor(output_layer.index).data"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "53bf7c61",
   "metadata": {},
   "source": [
    "The `.infer()` function sets output tensor, that can be reached, using `get_output_tensor()`. Since this network returns one output, and the reference to the output layer is in the `output_layer.index` parameter, you can get the data with `request.get_output_tensor(output_layer.index)`. To get a numpy array from the output, use the `.data` parameter."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "4a0f63b5",
   "metadata": {},
   "outputs": [],
   "source": [
    "result.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "7e5834ee",
   "metadata": {},
   "source": [
    "The output shape is (1,1001), which is the expected output shape. This shape indicates that the network returns probabilities for 1001 classes. To learn more about this notion, refer to the [hello world notebook](../001-hello-world/001-hello-world.ipynb)."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ec6a9be1",
   "metadata": {},
   "source": [
    "## Reshaping and Resizing\n",
    "\n",
    "### Change Image Size"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4239b10c",
   "metadata": {},
   "source": [
    "Instead of reshaping the image to fit the model, it is also possible to reshape the model to fit the image. Be aware that not all models support reshaping, and models that do, may not support all input shapes. The model accuracy may also suffer if you reshape the model input shape.\n",
    "\n",
    "First check the input shape of the model, then reshape it to the new input shape."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "bc1a69f8",
   "metadata": {},
   "outputs": [],
   "source": [
    "from openvino.runtime import Core, PartialShape\n",
    "\n",
    "ie = Core()\n",
    "segmentation_model_xml = \"model/segmentation.xml\"\n",
    "segmentation_model = ie.read_model(model=segmentation_model_xml)\n",
    "segmentation_input_layer = segmentation_model.input(0)\n",
    "segmentation_output_layer = segmentation_model.output(0)\n",
    "\n",
    "print(\"~~~~ ORIGINAL MODEL ~~~~\")\n",
    "print(f\"input shape: {segmentation_input_layer.shape}\")\n",
    "print(f\"output shape: {segmentation_output_layer.shape}\")\n",
    "\n",
    "new_shape = PartialShape([1, 3, 544, 544])\n",
    "segmentation_model.reshape({segmentation_input_layer.any_name: new_shape})\n",
    "segmentation_compiled_model = ie.compile_model(model=segmentation_model, device_name=\"CPU\")\n",
    "# help(segmentation_compiled_model)\n",
    "print(\"~~~~ RESHAPED MODEL ~~~~\")\n",
    "print(f\"model input shape: {segmentation_input_layer.shape}\")\n",
    "print(\n",
    "    f\"compiled_model input shape: \"\n",
    "    f\"{segmentation_compiled_model.input(index=0).shape}\"\n",
    ")\n",
    "print(f\"compiled_model output shape: {segmentation_output_layer.shape}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2104cef6",
   "metadata": {},
   "source": [
    "The input shape for the segmentation network is [1,3,512,512], with the `NCHW` layout: the network expects 3-channel images with a width and height of 512 and a batch size of 1. Reshape the network with the `.reshape()` method of `IENetwork` to make it accept input images with a width and height of 544. This segmentation network always returns arrays with the input width and height of equal value. Therefore, setting the input dimensions to 544x544 also modifies the output dimensions. After reshaping, compile the network once again."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "249d697a",
   "metadata": {},
   "source": [
    "### Change Batch Size"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ded79c8f",
   "metadata": {},
   "source": [
    "Use the `.reshape()` method to set the batch size, by increasing the first element of `new_shape`. For example, to set a batch size of two, set `new_shape = (2,3,544,544)` in the cell above. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "a49d65c6",
   "metadata": {},
   "outputs": [],
   "source": [
    "from openvino.runtime import Core, PartialShape\n",
    "\n",
    "ie = Core()\n",
    "segmentation_model_xml = \"model/segmentation.xml\"\n",
    "segmentation_model = ie.read_model(model=segmentation_model_xml)\n",
    "segmentation_input_layer = segmentation_model.input(0)\n",
    "segmentation_output_layer = segmentation_model.output(0)\n",
    "new_shape = PartialShape([2, 3, 544, 544])\n",
    "segmentation_model.reshape({segmentation_input_layer.any_name: new_shape})\n",
    "segmentation_compiled_model = ie.compile_model(model=segmentation_model, device_name=\"CPU\")\n",
    "\n",
    "print(f\"input shape: {segmentation_input_layer.shape}\")\n",
    "print(f\"output shape: {segmentation_output_layer.shape}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "40c2f3d2",
   "metadata": {},
   "source": [
    "The output shows that by setting the batch size to 2, the first element (`N`) of the input and output shape has a value of 2. Propagate the input image through the network to see the result:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "0eb487fa",
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "from openvino.runtime import Core, PartialShape\n",
    "\n",
    "ie = Core()\n",
    "segmentation_model_xml = \"model/segmentation.xml\"\n",
    "segmentation_model = ie.read_model(model=segmentation_model_xml)\n",
    "segmentation_input_layer = segmentation_model.input(0)\n",
    "segmentation_output_layer = segmentation_model.output(0)\n",
    "new_shape = PartialShape([2, 3, 544, 544])\n",
    "segmentation_model.reshape({segmentation_input_layer.any_name: new_shape})\n",
    "segmentation_compiled_model = ie.compile_model(model=segmentation_model, device_name=\"CPU\")\n",
    "input_data = np.random.rand(2, 3, 544, 544)\n",
    "\n",
    "output = segmentation_compiled_model([input_data])\n",
    "\n",
    "print(f\"input data shape: {input_data.shape}\")\n",
    "print(f\"result data data shape: {segmentation_output_layer.shape}\")"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "a9657dc5-9713-4d2b-a324-c8cd6195e79a",
   "metadata": {},
   "source": [
    "## Caching a Model\n",
    "\n",
    "For some devices, like GPU, loading a model can take some time. Model Caching solves this issue by caching the model in a cache directory. If `ie.compile_model(model=net, device_name=device_name, config=config_dict)` is set, caching will be used. This option checks if a model exists in the cache. If so, it loads it from the cache. If not, it loads the model regularly, and stores it in the cache, so that the next time the model is loaded when this option is set, the model will be loaded from the cache.\n",
    "\n",
    "In the cell below, we create a *model_cache* directory as a subdirectory of *model*, where the model will be cached for the specified device. The model will be loaded to the GPU. After running this cell once, the model will be cached, so subsequent runs of this cell will load the model from the cache.\n",
    "\n",
    "*Note: Model Caching is also available on CPU devices*"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "1d235185-18f7-4cf0-8cb2-1ecba279318a",
   "metadata": {},
   "outputs": [],
   "source": [
    "import time\n",
    "from pathlib import Path\n",
    "\n",
    "from openvino.runtime import Core\n",
    "\n",
    "ie = Core()\n",
    "\n",
    "device_name = \"GPU\" \n",
    "\n",
    "if device_name in ie.available_devices:\n",
    "    cache_path = Path(\"model/model_cache\")\n",
    "    cache_path.mkdir(exist_ok=True)\n",
    "    # Enable caching for OpenVINO Runtime. To disable caching set enable_caching = False\n",
    "    enable_caching = True\n",
    "    config_dict = {\"CACHE_DIR\": str(cache_path)} if enable_caching else {}\n",
    "\n",
    "    classification_model_xml = \"model/classification.xml\"\n",
    "    model = ie.read_model(model=classification_model_xml)\n",
    "\n",
    "    start_time = time.perf_counter()\n",
    "    compiled_model = ie.compile_model(model=model, device_name=device_name, config=config_dict)\n",
    "    end_time = time.perf_counter()\n",
    "    print(f\"Loading the network to the {device_name} device took {end_time-start_time:.2f} seconds.\")"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "48e0a860-c93c-4b93-a684-f53cd66ec2e3",
   "metadata": {},
   "source": [
    "After running the previous cell, we know the model exists in the cache directory. Then, we delete the compiled model and load it again. Now, we measure the time it takes now."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "b5a7686b-b9e0-44a6-8b6e-5b299d085eac",
   "metadata": {},
   "outputs": [],
   "source": [
    "if device_name in ie.available_devices:\n",
    "    del compiled_model\n",
    "    start_time = time.perf_counter()\n",
    "    compiled_model = ie.compile_model(model=model, device_name=device_name, config=config_dict)\n",
    "    end_time = time.perf_counter()\n",
    "    print(f\"Loading the network to the {device_name} device took {end_time-start_time:.2f} seconds.\")"
   ]
  }
 ],
 "metadata": {
  "interpreter": {
   "hash": "ae617ccb002f72b3ab6d0069d721eac67ac2a969e83c083c4321cfcab0437cd1"
  },
  "kernelspec": {
   "display_name": "Python 3",
   "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": "3.8.10"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}