03-yolov5_to_onnx.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "e260f146-9844-4854-a29c-7ef08829e4e7",
   "metadata": {},
   "source": [
    "# Export YoloV5 model to ONNX format"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "7b3c563e-535f-431f-b49c-b2008afacf6f",
   "metadata": {},
   "source": [
    "## Install requirements"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "f20ffb88-1685-4448-9af4-a981d83e94a8",
   "metadata": {},
   "outputs": [],
   "source": [
    "!pip install -qr requirements.txt"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f8df2bcd-d5e2-4fa3-8b68-bf8270d021d3",
   "metadata": {},
   "source": [
    "## Download model\n",
    "You can download the model of your choice (different sizes and characteristics) from here: https://github.com/ultralytics/yolov5#pretrained-checkpoints"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "24297e76-2a68-4885-a070-1880eac3f902",
   "metadata": {},
   "source": [
    "## Export the model to ONNX\n",
    "- Eventually modify the image size if you selected a model with the \"6\" suffix, as image size is 1280 (and not 640)\n",
    "- Stay at opset 16 for the moment for OpenVino model server compatibility"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "96665823-af2c-43ed-a828-df19dcdf612f",
   "metadata": {},
   "outputs": [],
   "source": [
    "!python yolov5/export.py --weights yolov5n.pt --include onnx --imgsz 640 --opset 16"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f5b94b3c",
   "metadata": {},
   "source": [
    "Your ONNX model will have the same name as the one you downloaded, but with an `.onnx` extension."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "5faf82da-1a30-4bb6-bdbc-ee641ebbe24b",
   "metadata": {},
   "source": [
    "### Next Step\n",
    "\n",
    "Now you're ready to save the model (`*.onnx`) file to your s3 storage location."
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3.9",
   "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": "3.9.18"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}