sambit mohapatra
Published © GPL3+

Hardware Accelerated Real-time Perception in 3D (HARP-3D)

End-to-end demonstration of 3D object detection in LiDAR point clouds using a deep neural network running on the ULTRA96V2.

IntermediateFull instructions provided15 hours2,242

Things used in this project

Hardware components

Ultra96-V2
Avnet Ultra96-V2
×1
nvidia rtx2080
×1

Software apps and online services

Vivado Design Suite
Xilinx Vivado Design Suite
Xilinx Vitis-AI
VS Code
Microsoft VS Code
TensorFlow
TensorFlow
Balena Etcher
Mayavi

Story

Read more

Custom parts and enclosures

Ultra96v2 dcf and arch zip

For use in Vitis-AI

Trained model - elf

Trained and compiled model for key point detection from range images

test data

extract and copy to ultra96.
Refer the inference code about updating the inference path

Code

inference application

Python
copy it to Ultra96
{
 "metadata": {
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.6-final"
  },
  "orig_nbformat": 2,
  "kernelspec": {
   "name": "Python 3.7.6 64-bit",
   "display_name": "Python 3.7.6 64-bit",
   "metadata": {
    "interpreter": {
     "hash": "31f2aee4e71d21fbe5cf8b01ff0e069b9275f58929596ceb00d14d90e3e16cd6"
    }
   }
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2,
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "from ctypes import *\n",
    "import cv2\n",
    "import numpy as np\n",
    "import runner\n",
    "import os\n",
    "import xir.graph\n",
    "import pathlib\n",
    "import xir.subgraph\n",
    "import threading\n",
    "import time\n",
    "import sys\n",
    "import argparse\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "from scipy.special import softmax\n",
    "\n",
    "%matplotlib inline\n",
    "\n",
    "show_count = 0\n",
    "\n",
    "def preprocess_fn(image_path):\n",
    "    '''\n",
    "    Image pre-processing.\n",
    "    Rearranges from BGR to RGB then normalizes to range 0:1\n",
    "    input arg: path of image file\n",
    "    return: numpy array\n",
    "    '''\n",
    "    # image = cv2.imread(image_path)\n",
    "    image = np.load(image_path)\n",
    "    print(image.shape)\n",
    "\n",
    "    test_x = image[0:64, 0:256, 3]  # get only depth channel\n",
    "    # plt.imshow(test_x)\n",
    "    # plt.show()\n",
    "    # plt.imsave(\"depth.png\", test_x)\n",
    "    test_x /= np.max(test_x)\n",
    "    \n",
    "    return test_x\n",
    "\n",
    "def get_subgraph (g):\n",
    "    sub = []\n",
    "    root = g.get_root_subgraph()\n",
    "    sub = [ s for s in root.children\n",
    "            if s.metadata.get_attr_str (\"device\") == \"DPU\"]\n",
    "    return sub \n",
    "\n",
    "\n",
    "def runDPU(id,start,dpu,img):\n",
    "\n",
    "    global show_count\n",
    "\n",
    "    '''get tensor'''\n",
    "    inputTensors = dpu.get_input_tensors()\n",
    "    outputTensors = dpu.get_output_tensors()\n",
    "    outputHeight = outputTensors[0].dims[1]\n",
    "    outputWidth = outputTensors[0].dims[2]\n",
    "    outputChannel = outputTensors[0].dims[3]\n",
    "    outputSize = outputHeight*outputWidth*outputChannel\n",
    "\n",
    "    batchSize = inputTensors[0].dims[0]\n",
    "    n_of_images = len(img)\n",
    "    count = 0\n",
    "    write_index = start\n",
    "    while count < n_of_images:\n",
    "        if (count+batchSize<=n_of_images):\n",
    "            runSize = batchSize\n",
    "        else:\n",
    "            runSize=n_of_images-count\n",
    "        shapeIn = (runSize,) + tuple([inputTensors[0].dims[i] for i in range(inputTensors[0].ndim)][1:])\n",
    "\n",
    "        '''prepare batch input/output '''\n",
    "        outputData = []\n",
    "        inputData = []\n",
    "        outputData.append(np.empty((runSize,outputHeight,outputWidth,outputChannel), dtype = np.float32, order = 'C'))\n",
    "        inputData.append(np.empty((shapeIn), dtype = np.float32, order = 'C'))\n",
    "\n",
    "        '''init input image to input buffer '''\n",
    "        for j in range(runSize):\n",
    "            imageRun = inputData[0]\n",
    "            imageRun[j,...] = img[(count+j)% n_of_images].reshape(inputTensors[0].dims[1],inputTensors[0].dims[2],inputTensors[0].dims[3])\n",
    "\n",
    "        '''run with batch '''\n",
    "        job_id = dpu.execute_async(inputData,outputData)\n",
    "        dpu.wait(job_id)\n",
    "\n",
    "        # get predictions direclty here\n",
    "        predictions = outputData[0][0]\n",
    "        predictions = softmax(predictions, -1)  # softmax along channels\n",
    "        print(\"predictions shape: \", predictions.shape)\n",
    "        mask = np.argmax(predictions, -1)  # along channels\n",
    "        print(\"mask shape: \", mask.shape)\n",
    "\n",
    "        img_test = np.squeeze(imageRun[0], -1)\n",
    "        \n",
    "        # if show_count < 4:\n",
    "        #     plt.imshow(img_test)\n",
    "        #     plt.show()\n",
    "        #     plt.imsave(\"depth_\"+str(count)+'.png', img_test)\n",
    "        # # mask = mask[0]\n",
    "        #     plt.imshow(mask)\n",
    "        #     plt.show()\n",
    "        #     show_count += 1\n",
    "        #     plt.imsave(\"keypoints_\"+str(count)+\".png\", mask)\n",
    "\n",
    "\n",
    "        # for j in range(len(outputData)):\n",
    "        #     outputData[j] = outputData[j].reshape(runSize, outputSize)\n",
    "\n",
    "        '''store output vectors '''\n",
    "        # for j in range(runSize):\n",
    "        #     out_q[write_index] = outputData[0][j]\n",
    "        #     write_index += 1\n",
    "        \n",
    "        count = count + runSize\n",
    "        break\n",
    "        \n",
    "\n",
    "\n",
    "def app(image_dir,threads,model):\n",
    "\n",
    "    listimage=os.listdir(image_dir)\n",
    "    runTotal = len(listimage)\n",
    "\n",
    "    global out_q\n",
    "    out_q = [None] * runTotal\n",
    "\n",
    "    g = xir.graph.Graph.deserialize(pathlib.Path(model))\n",
    "    subgraphs = get_subgraph (g)\n",
    "    assert len(subgraphs) == 1 # only one DPU kernel\n",
    "    all_dpu_runners = []\n",
    "    for i in range(threads):\n",
    "        all_dpu_runners.append(runner.Runner(subgraphs[0], \"run\"))\n",
    "\n",
    "    ''' preprocess images '''\n",
    "    print('Pre-processing',runTotal,'images...')\n",
    "    img = []\n",
    "    for i in range(runTotal):\n",
    "        path = os.path.join(image_dir,listimage[i])\n",
    "        img.append(preprocess_fn(path))\n",
    "\n",
    "    '''run threads '''\n",
    "    print('Starting',threads,'threads...')\n",
    "    threadAll = []\n",
    "    start=0\n",
    "    for i in range(threads):\n",
    "        if (i==threads-1):\n",
    "            end = len(img)\n",
    "        else:\n",
    "            end = start+(len(img)//threads)\n",
    "        in_q = img[start:end]\n",
    "        t1 = threading.Thread(target=runDPU, args=(i,start,all_dpu_runners[i], in_q))\n",
    "        threadAll.append(t1)\n",
    "        start=end\n",
    "\n",
    "    time1 = time.time()\n",
    "    for x in threadAll:\n",
    "        x.start()\n",
    "    for x in threadAll:\n",
    "        x.join()\n",
    "    time2 = time.time()\n",
    "    timetotal = time2 - time1\n",
    "\n",
    "    fps = float(runTotal / timetotal)\n",
    "    print(\"FPS=%.2f, total frames = %.0f , time=%.4f seconds\" %(fps,runTotal, timetotal))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "output_type": "stream",
     "name": "stdout",
     "text": [
      "hello to Xilinx app development using DPU\n",
      "Pre-processing 50 images...\n",
      "(66, 259, 6)\n",
      "(66, 259, 6)\n",
      "(66, 259, 6)\n",
      "(66, 259, 6)\n",
      "(66, 259, 6)\n",
      "(66, 259, 6)\n",
      "(66, 259, 6)\n",
      "(66, 259, 6)\n",
      "(66, 259, 6)\n",
      "(66, 259, 6)\n",
      "(66, 259, 6)\n",
      "(66, 259, 6)\n",
      "(66, 259, 6)\n",
      "(66, 259, 6)\n",
      "(66, 259, 6)\n",
      "(66, 259, 6)\n",
      "(66, 259, 6)\n",
      "(66, 259, 6)\n",
      "(66, 259, 6)\n",
      "(66, 259, 6)\n",
      "(66, 259, 6)\n",
      "(66, 259, 6)\n",
      "(66, 259, 6)\n",
      "(66, 259, 6)\n",
      "(66, 259, 6)\n",
      "(66, 259, 6)\n",
      "(66, 259, 6)\n",
      "(66, 259, 6)\n",
      "(66, 259, 6)\n",
      "(66, 259, 6)\n",
      "(66, 259, 6)\n",
      "(66, 259, 6)\n",
      "(66, 259, 6)\n",
      "(66, 259, 6)\n",
      "(66, 259, 6)\n",
      "(66, 259, 6)\n",
      "(66, 259, 6)\n",
      "(66, 259, 6)\n",
      "(66, 259, 6)\n",
      "(66, 259, 6)\n",
      "(66, 259, 6)\n",
      "(66, 259, 6)\n",
      "(66, 259, 6)\n",
      "(66, 259, 6)\n",
      "(66, 259, 6)\n",
      "(66, 259, 6)\n",
      "(66, 259, 6)\n",
      "(66, 259, 6)\n",
      "(66, 259, 6)\n",
      "(66, 259, 6)\n",
      "Starting 1 threads...\n",
      "predictions shape:  (64, 256, 2)\n",
      "mask shape:  (64, 256)\n"
     ]
    },
    {
     "output_type": "display_data",
     "data": {
      "text/plain": "<Figure size 432x288 with 1 Axes>",
      "image/svg+xml": "<?xml version=\"1.0\" encoding=\"utf-8\" standalone=\"no\"?>\n<!DOCTYPE svg PUBLIC \"-//W3C//DTD SVG 1.1//EN\"\n  \"http://www.w3.org/Graphics/SVG/1.1/DTD/svg11.dtd\">\n<!-- Created with matplotlib (https://matplotlib.org/) -->\n<svg height=\"117.923437pt\" version=\"1.1\" viewBox=\"0 0 371.275781 117.923437\" width=\"371.275781pt\" xmlns=\"http://www.w3.org/2000/svg\" xmlns:xlink=\"http://www.w3.org/1999/xlink\">\n <metadata>\n  <rdf:RDF xmlns:cc=\"http://creativecommons.org/ns#\" xmlns:dc=\"http://purl.org/dc/elements/1.1/\" xmlns:rdf=\"http://www.w3.org/1999/02/22-rdf-syntax-ns#\">\n   <cc:Work>\n    <dc:type rdf:resource=\"http://purl.org/dc/dcmitype/StillImage\"></dc:type>\n    <dc:date>2020-11-08T22:34:13.683312</dc:date>\n    <dc:format>image/svg+xml</dc:format>\n    <dc:creator>\n     <cc:Agent>\n      <dc:title>Matplotlib v3.3.2, https://matplotlib.org/</dc:title>\n     </cc:Agent>\n    </dc:creator>\n   </cc:Work>\n  </rdf:RDF>\n </metadata>\n <defs>\n  <style type=\"text/css\">*{stroke-linecap:butt;stroke-linejoin:round;}</style>\n </defs>\n <g id=\"figure_1\">\n  <g id=\"patch_1\">\n   <path d=\"M -0 117.923437 \nL 371.275781 117.923437 \nL 371.275781 0 \nL -0 0 \nz\n\" style=\"fill:none;\"></path>\n  </g>\n  <g id=\"axes_1\">\n   <g id=\"patch_2\">\n    <path d=\"M 26.925 94.045312 \nL 361.725 94.045312 \nL 361.725 10.345312 \nL 26.925 10.345312 \nz\n\" style=\"fill:#ffffff;\"></path>\n   </g>\n   <g clip-path=\"url(#p1d2369f4b1)\">\n    <image height=\"84\" id=\"imagebcb865c6d3\" transform=\"scale(1 -1)translate(0 -84)\" width=\"335\" x=\"26.925\" xlink:href=\"data:image/png;base64,\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\" y=\"-10.045312\"></image>\n   </g>\n   <g id=\"matplotlib.axis_1\">\n    <g id=\"xtick_1\">\n     <g id=\"line2d_1\">\n      <defs>\n       <path d=\"M 0 0 \nL 0 3.5 \n\" id=\"m952b463b8c\" style=\"stroke:#000000;stroke-width:0.8;\"></path>\n      </defs>\n      <g>\n       <use style=\"stroke:#000000;stroke-width:0.8;\" x=\"27.578906\" xlink:href=\"#m952b463b8c\" y=\"94.045312\"></use>\n      </g>\n     </g>\n     <g id=\"text_1\">\n      <!-- 0 -->\n      <g transform=\"translate(24.397656 108.64375)scale(0.1 -0.1)\">\n       <defs>\n        <path d=\"M 31.78125 66.40625 \nQ 24.171875 66.40625 20.328125 58.90625 \nQ 16.5 51.421875 16.5 36.375 \nQ 16.5 21.390625 20.328125 13.890625 \nQ 24.171875 6.390625 31.78125 6.390625 \nQ 39.453125 6.390625 43.28125 13.890625 \nQ 47.125 21.390625 47.125 36.375 \nQ 47.125 51.421875 43.28125 58.90625 \nQ 39.453125 66.40625 31.78125 66.40625 \nz\nM 31.78125 74.21875 \nQ 44.046875 74.21875 50.515625 64.515625 \nQ 56.984375 54.828125 56.984375 36.375 \nQ 56.984375 17.96875 50.515625 8.265625 \nQ 44.046875 -1.421875 31.78125 -1.421875 \nQ 19.53125 -1.421875 13.0625 8.265625 \nQ 6.59375 17.96875 6.59375 36.375 \nQ 6.59375 54.828125 13.0625 64.515625 \nQ 19.53125 74.21875 31.78125 74.21875 \nz\n\" id=\"DejaVuSans-48\"></path>\n       </defs>\n       <use xlink:href=\"#DejaVuSans-48\"></use>\n      </g>\n     </g>\n    </g>\n    <g id=\"xtick_2\">\n     <g id=\"line2d_2\">\n      <g>\n       <use style=\"stroke:#000000;stroke-width:0.8;\" x=\"92.969531\" xlink:href=\"#m952b463b8c\" y=\"94.045312\"></use>\n      </g>\n     </g>\n     <g id=\"text_2\">\n      <!-- 50 -->\n      <g transform=\"translate(86.607031 108.64375)scale(0.1 -0.1)\">\n       <defs>\n        <path d=\"M 10.796875 72.90625 \nL 49.515625 72.90625 \nL 49.515625 64.59375 \nL 19.828125 64.59375 \nL 19.828125 46.734375 \nQ 21.96875 47.46875 24.109375 47.828125 \nQ 26.265625 48.1875 28.421875 48.1875 \nQ 40.625 48.1875 47.75 41.5 \nQ 54.890625 34.8125 54.890625 23.390625 \nQ 54.890625 11.625 47.5625 5.09375 \nQ 40.234375 -1.421875 26.90625 -1.421875 \nQ 22.3125 -1.421875 17.546875 -0.640625 \nQ 12.796875 0.140625 7.71875 1.703125 \nL 7.71875 11.625 \nQ 12.109375 9.234375 16.796875 8.0625 \nQ 21.484375 6.890625 26.703125 6.890625 \nQ 35.15625 6.890625 40.078125 11.328125 \nQ 45.015625 15.765625 45.015625 23.390625 \nQ 45.015625 31 40.078125 35.4375 \nQ 35.15625 39.890625 26.703125 39.890625 \nQ 22.75 39.890625 18.8125 39.015625 \nQ 14.890625 38.140625 10.796875 36.28125 \nz\n\" id=\"DejaVuSans-53\"></path>\n       </defs>\n       <use xlink:href=\"#DejaVuSans-53\"></use>\n       <use x=\"63.623047\" xlink:href=\"#DejaVuSans-48\"></use>\n      </g>\n     </g>\n    </g>\n    <g id=\"xtick_3\">\n     <g id=\"line2d_3\">\n      <g>\n       <use style=\"stroke:#000000;stroke-width:0.8;\" x=\"158.360156\" xlink:href=\"#m952b463b8c\" y=\"94.045312\"></use>\n      </g>\n     </g>\n     <g id=\"text_3\">\n      <!-- 100 -->\n      <g transform=\"translate(148.816406 108.64375)scale(0.1 -0.1)\">\n       <defs>\n        <path d=\"M 12.40625 8.296875 \nL 28.515625 8.296875 \nL 28.515625 63.921875 \nL 10.984375 60.40625 \nL 10.984375 69.390625 \nL 28.421875 72.90625 \nL 38.28125 72.90625 \nL 38.28125 8.296875 \nL 54.390625 8.296875 \nL 54.390625 0 \nL 12.40625 0 \nz\n\" id=\"DejaVuSans-49\"></path>\n       </defs>\n       <use xlink:href=\"#DejaVuSans-49\"></use>\n       <use x=\"63.623047\" xlink:href=\"#DejaVuSans-48\"></use>\n       <use x=\"127.246094\" xlink:href=\"#DejaVuSans-48\"></use>\n      </g>\n     </g>\n    </g>\n    <g id=\"xtick_4\">\n     <g id=\"line2d_4\">\n      <g>\n       <use style=\"stroke:#000000;stroke-width:0.8;\" x=\"223.750781\" xlink:href=\"#m952b463b8c\" y=\"94.045312\"></use>\n      </g>\n     </g>\n     <g id=\"text_4\">\n      <!-- 150 -->\n      <g transform=\"translate(214.207031 108.64375)scale(0.1 -0.1)\">\n       <use xlink:href=\"#DejaVuSans-49\"></use>\n       <use x=\"63.623047\" xlink:href=\"#DejaVuSans-53\"></use>\n       <use x=\"127.246094\" xlink:href=\"#DejaVuSans-48\"></use>\n      </g>\n     </g>\n    </g>\n    <g id=\"xtick_5\">\n     <g id=\"line2d_5\">\n      <g>\n       <use style=\"stroke:#000000;stroke-width:0.8;\" x=\"289.141406\" xlink:href=\"#m952b463b8c\" y=\"94.045312\"></use>\n      </g>\n     </g>\n     <g id=\"text_5\">\n      <!-- 200 -->\n      <g transform=\"translate(279.597656 108.64375)scale(0.1 -0.1)\">\n       <defs>\n        <path d=\"M 19.1875 8.296875 \nL 53.609375 8.296875 \nL 53.609375 0 \nL 7.328125 0 \nL 7.328125 8.296875 \nQ 12.9375 14.109375 22.625 23.890625 \nQ 32.328125 33.6875 34.8125 36.53125 \nQ 39.546875 41.84375 41.421875 45.53125 \nQ 43.3125 49.21875 43.3125 52.78125 \nQ 43.3125 58.59375 39.234375 62.25 \nQ 35.15625 65.921875 28.609375 65.921875 \nQ 23.96875 65.921875 18.8125 64.3125 \nQ 13.671875 62.703125 7.8125 59.421875 \nL 7.8125 69.390625 \nQ 13.765625 71.78125 18.9375 73 \nQ 24.125 74.21875 28.421875 74.21875 \nQ 39.75 74.21875 46.484375 68.546875 \nQ 53.21875 62.890625 53.21875 53.421875 \nQ 53.21875 48.921875 51.53125 44.890625 \nQ 49.859375 40.875 45.40625 35.40625 \nQ 44.1875 33.984375 37.640625 27.21875 \nQ 31.109375 20.453125 19.1875 8.296875 \nz\n\" id=\"DejaVuSans-50\"></path>\n       </defs>\n       <use xlink:href=\"#DejaVuSans-50\"></use>\n       <use x=\"63.623047\" xlink:href=\"#DejaVuSans-48\"></use>\n       <use x=\"127.246094\" xlink:href=\"#DejaVuSans-48\"></use>\n      </g>\n     </g>\n    </g>\n    <g id=\"xtick_6\">\n     <g id=\"line2d_6\">\n      <g>\n       <use style=\"stroke:#000000;stroke-width:0.8;\" x=\"354.532031\" xlink:href=\"#m952b463b8c\" y=\"94.045312\"></use>\n      </g>\n     </g>\n     <g id=\"text_6\">\n      <!-- 250 -->\n      <g transform=\"translate(344.988281 108.64375)scale(0.1 -0.1)\">\n       <use xlink:href=\"#DejaVuSans-50\"></use>\n       <use x=\"63.623047\" xlink:href=\"#DejaVuSans-53\"></use>\n       <use x=\"127.246094\" xlink:href=\"#DejaVuSans-48\"></use>\n      </g>\n     </g>\n    </g>\n   </g>\n   <g id=\"matplotlib.axis_2\">\n    <g id=\"ytick_1\">\n     <g id=\"line2d_7\">\n      <defs>\n       <path d=\"M 0 0 \nL -3.5 0 \n\" id=\"m28f6c51fc7\" style=\"stroke:#000000;stroke-width:0.8;\"></path>\n      </defs>\n      <g>\n       <use style=\"stroke:#000000;stroke-width:0.8;\" x=\"26.925\" xlink:href=\"#m28f6c51fc7\" y=\"10.999219\"></use>\n      </g>\n     </g>\n     <g id=\"text_7\">\n      <!-- 0 -->\n      <g transform=\"translate(13.5625 14.798437)scale(0.1 -0.1)\">\n       <use xlink:href=\"#DejaVuSans-48\"></use>\n      </g>\n     </g>\n    </g>\n    <g id=\"ytick_2\">\n     <g id=\"line2d_8\">\n      <g>\n       <use style=\"stroke:#000000;stroke-width:0.8;\" x=\"26.925\" xlink:href=\"#m28f6c51fc7\" y=\"37.155469\"></use>\n      </g>\n     </g>\n     <g id=\"text_8\">\n      <!-- 20 -->\n      <g transform=\"translate(7.2 40.954687)scale(0.1 -0.1)\">\n       <use xlink:href=\"#DejaVuSans-50\"></use>\n       <use x=\"63.623047\" xlink:href=\"#DejaVuSans-48\"></use>\n      </g>\n     </g>\n    </g>\n    <g id=\"ytick_3\">\n     <g id=\"line2d_9\">\n      <g>\n       <use style=\"stroke:#000000;stroke-width:0.8;\" x=\"26.925\" xlink:href=\"#m28f6c51fc7\" y=\"63.311719\"></use>\n      </g>\n     </g>\n     <g id=\"text_9\">\n      <!-- 40 -->\n      <g transform=\"translate(7.2 67.110937)scale(0.1 -0.1)\">\n       <defs>\n        <path d=\"M 37.796875 64.3125 \nL 12.890625 25.390625 \nL 37.796875 25.390625 \nz\nM 35.203125 72.90625 \nL 47.609375 72.90625 \nL 47.609375 25.390625 \nL 58.015625 25.390625 \nL 58.015625 17.1875 \nL 47.609375 17.1875 \nL 47.609375 0 \nL 37.796875 0 \nL 37.796875 17.1875 \nL 4.890625 17.1875 \nL 4.890625 26.703125 \nz\n\" id=\"DejaVuSans-52\"></path>\n       </defs>\n       <use xlink:href=\"#DejaVuSans-52\"></use>\n       <use x=\"63.623047\" xlink:href=\"#DejaVuSans-48\"></use>\n      </g>\n     </g>\n    </g>\n    <g id=\"ytick_4\">\n     <g id=\"line2d_10\">\n      <g>\n       <use style=\"stroke:#000000;stroke-width:0.8;\" x=\"26.925\" xlink:href=\"#m28f6c51fc7\" y=\"89.467969\"></use>\n      </g>\n     </g>\n     <g id=\"text_10\">\n      <!-- 60 -->\n      <g transform=\"translate(7.2 93.267187)scale(0.1 -0.1)\">\n       <defs>\n        <path d=\"M 33.015625 40.375 \nQ 26.375 40.375 22.484375 35.828125 \nQ 18.609375 31.296875 18.609375 23.390625 \nQ 18.609375 15.53125 22.484375 10.953125 \nQ 26.375 6.390625 33.015625 6.390625 \nQ 39.65625 6.390625 43.53125 10.953125 \nQ 47.40625 15.53125 47.40625 23.390625 \nQ 47.40625 31.296875 43.53125 35.828125 \nQ 39.65625 40.375 33.015625 40.375 \nz\nM 52.59375 71.296875 \nL 52.59375 62.3125 \nQ 48.875 64.0625 45.09375 64.984375 \nQ 41.3125 65.921875 37.59375 65.921875 \nQ 27.828125 65.921875 22.671875 59.328125 \nQ 17.53125 52.734375 16.796875 39.40625 \nQ 19.671875 43.65625 24.015625 45.921875 \nQ 28.375 48.1875 33.59375 48.1875 \nQ 44.578125 48.1875 50.953125 41.515625 \nQ 57.328125 34.859375 57.328125 23.390625 \nQ 57.328125 12.15625 50.6875 5.359375 \nQ 44.046875 -1.421875 33.015625 -1.421875 \nQ 20.359375 -1.421875 13.671875 8.265625 \nQ 6.984375 17.96875 6.984375 36.375 \nQ 6.984375 53.65625 15.1875 63.9375 \nQ 23.390625 74.21875 37.203125 74.21875 \nQ 40.921875 74.21875 44.703125 73.484375 \nQ 48.484375 72.75 52.59375 71.296875 \nz\n\" id=\"DejaVuSans-54\"></path>\n       </defs>\n       <use xlink:href=\"#DejaVuSans-54\"></use>\n       <use x=\"63.623047\" xlink:href=\"#DejaVuSans-48\"></use>\n      </g>\n     </g>\n    </g>\n   </g>\n   <g id=\"patch_3\">\n    <path d=\"M 26.925 94.045312 \nL 26.925 10.345312 \n\" style=\"fill:none;stroke:#000000;stroke-linecap:square;stroke-linejoin:miter;stroke-width:0.8;\"></path>\n   </g>\n   <g id=\"patch_4\">\n    <path d=\"M 361.725 94.045312 \nL 361.725 10.345312 \n\" style=\"fill:none;stroke:#000000;stroke-linecap:square;stroke-linejoin:miter;stroke-width:0.8;\"></path>\n   </g>\n   <g id=\"patch_5\">\n    <path d=\"M 26.925 94.045312 \nL 361.725 94.045312 \n\" style=\"fill:none;stroke:#000000;stroke-linecap:square;stroke-linejoin:miter;stroke-width:0.8;\"></path>\n   </g>\n   <g id=\"patch_6\">\n    <path d=\"M 26.925 10.345312 \nL 361.725 10.345312 \n\" style=\"fill:none;stroke:#000000;stroke-linecap:square;stroke-linejoin:miter;stroke-width:0.8;\"></path>\n   </g>\n  </g>\n </g>\n <defs>\n  <clipPath id=\"p1d2369f4b1\">\n   <rect height=\"83.7\" width=\"334.8\" x=\"26.925\" y=\"10.345312\"></rect>\n  </clipPath>\n </defs>\n</svg>\n",
      "image/png": "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\n"
     },
     "metadata": {
      "needs_background": "light"
     }
    },
    {
     "output_type": "display_data",
     "data": {
      "text/plain": "<Figure size 432x288 with 1 Axes>",
      "image/svg+xml": "<?xml version=\"1.0\" encoding=\"utf-8\" standalone=\"no\"?>\n<!DOCTYPE svg PUBLIC \"-//W3C//DTD SVG 1.1//EN\"\n  \"http://www.w3.org/Graphics/SVG/1.1/DTD/svg11.dtd\">\n<!-- Created with matplotlib (https://matplotlib.org/) -->\n<svg height=\"117.923437pt\" version=\"1.1\" viewBox=\"0 0 371.275781 117.923437\" width=\"371.275781pt\" xmlns=\"http://www.w3.org/2000/svg\" xmlns:xlink=\"http://www.w3.org/1999/xlink\">\n <metadata>\n  <rdf:RDF xmlns:cc=\"http://creativecommons.org/ns#\" xmlns:dc=\"http://purl.org/dc/elements/1.1/\" xmlns:rdf=\"http://www.w3.org/1999/02/22-rdf-syntax-ns#\">\n   <cc:Work>\n    <dc:type rdf:resource=\"http://purl.org/dc/dcmitype/StillImage\"></dc:type>\n    <dc:date>2020-11-08T22:34:15.205949</dc:date>\n    <dc:format>image/svg+xml</dc:format>\n    <dc:creator>\n     <cc:Agent>\n      <dc:title>Matplotlib v3.3.2, https://matplotlib.org/</dc:title>\n     </cc:Agent>\n    </dc:creator>\n   </cc:Work>\n  </rdf:RDF>\n </metadata>\n <defs>\n  <style type=\"text/css\">*{stroke-linecap:butt;stroke-linejoin:round;}</style>\n </defs>\n <g id=\"figure_1\">\n  <g id=\"patch_1\">\n   <path d=\"M -0 117.923437 \nL 371.275781 117.923437 \nL 371.275781 0 \nL -0 0 \nz\n\" style=\"fill:none;\"></path>\n  </g>\n  <g id=\"axes_1\">\n   <g id=\"patch_2\">\n    <path d=\"M 26.925 94.045312 \nL 361.725 94.045312 \nL 361.725 10.345312 \nL 26.925 10.345312 \nz\n\" style=\"fill:#ffffff;\"></path>\n   </g>\n   <g clip-path=\"url(#pdc8d6ed2d1)\">\n    <image height=\"84\" id=\"imagef8546f5fac\" transform=\"scale(1 -1)translate(0 -84)\" width=\"335\" x=\"26.925\" xlink:href=\"data:image/png;base64,\niVBORw0KGgoAAAANSUhEUgAAAU8AAABUCAYAAAACq7UeAAABaUlEQVR4nO3WsW0CUQAFwf8NiR1cRozowgUQOaYR+nA7LodOkI4GSLzJgTRTwYtWb57nZR0A/MvH1gMA3pF4AgTiCRCIJ0AgngCBeAIE4gkQiCdAIJ4AgXgCBOIJEIgnQCCeAIF4AgTiCRCIJ0AgngCBeAIE4gkQiCdAIJ4AgXgCBOIJEIgnQCCeAIF4AgTiCRCIJ0AgngCBeAIE4gkQiCdAIJ4AgXgCBOIJEIgnQCCeAIF4AgTiCRCIJ0AgngCBeAIE4gkQiCdAIJ4AgXgCBOIJEIgnQCCeAIF4AgTiCRCIJ0AgngCBeAIE4gkQiCdAIJ4AgXgCBOIJEIgnQCCeAIF4AgTiCRCIJ0AgngCBeAIE4gkQiCdAIJ4AgXgCBOIJEIgnQCCePLU/Hcft93vslmXrKfCSxJOn7odlXH/+xvz63HoKvKR5npd16xEA78bzBAjEEyAQT4BAPAEC8QQIxBMgEE+AQDwBggf8jgh+cS6wHwAAAABJRU5ErkJggg==\" y=\"-10.045312\"></image>\n   </g>\n   <g id=\"matplotlib.axis_1\">\n    <g id=\"xtick_1\">\n     <g id=\"line2d_1\">\n      <defs>\n       <path d=\"M 0 0 \nL 0 3.5 \n\" id=\"m3df349d0fa\" style=\"stroke:#000000;stroke-width:0.8;\"></path>\n      </defs>\n      <g>\n       <use style=\"stroke:#000000;stroke-width:0.8;\" x=\"27.578906\" xlink:href=\"#m3df349d0fa\" y=\"94.045312\"></use>\n      </g>\n     </g>\n     <g id=\"text_1\">\n      <!-- 0 -->\n      <g transform=\"translate(24.397656 108.64375)scale(0.1 -0.1)\">\n       <defs>\n        <path d=\"M 31.78125 66.40625 \nQ 24.171875 66.40625 20.328125 58.90625 \nQ 16.5 51.421875 16.5 36.375 \nQ 16.5 21.390625 20.328125 13.890625 \nQ 24.171875 6.390625 31.78125 6.390625 \nQ 39.453125 6.390625 43.28125 13.890625 \nQ 47.125 21.390625 47.125 36.375 \nQ 47.125 51.421875 43.28125 58.90625 \nQ 39.453125 66.40625 31.78125 66.40625 \nz\nM 31.78125 74.21875 \nQ 44.046875 74.21875 50.515625 64.515625 \nQ 56.984375 54.828125 56.984375 36.375 \nQ 56.984375 17.96875 50.515625 8.265625 \nQ 44.046875 -1.421875 31.78125 -1.421875 \nQ 19.53125 -1.421875 13.0625 8.265625 \nQ 6.59375 17.96875 6.59375 36.375 \nQ 6.59375 54.828125 13.0625 64.515625 \nQ 19.53125 74.21875 31.78125 74.21875 \nz\n\" id=\"DejaVuSans-48\"></path>\n       </defs>\n       <use xlink:href=\"#DejaVuSans-48\"></use>\n      </g>\n     </g>\n    </g>\n    <g id=\"xtick_2\">\n     <g id=\"line2d_2\">\n      <g>\n       <use style=\"stroke:#000000;stroke-width:0.8;\" x=\"92.969531\" xlink:href=\"#m3df349d0fa\" y=\"94.045312\"></use>\n      </g>\n     </g>\n     <g id=\"text_2\">\n      <!-- 50 -->\n      <g transform=\"translate(86.607031 108.64375)scale(0.1 -0.1)\">\n       <defs>\n        <path d=\"M 10.796875 72.90625 \nL 49.515625 72.90625 \nL 49.515625 64.59375 \nL 19.828125 64.59375 \nL 19.828125 46.734375 \nQ 21.96875 47.46875 24.109375 47.828125 \nQ 26.265625 48.1875 28.421875 48.1875 \nQ 40.625 48.1875 47.75 41.5 \nQ 54.890625 34.8125 54.890625 23.390625 \nQ 54.890625 11.625 47.5625 5.09375 \nQ 40.234375 -1.421875 26.90625 -1.421875 \nQ 22.3125 -1.421875 17.546875 -0.640625 \nQ 12.796875 0.140625 7.71875 1.703125 \nL 7.71875 11.625 \nQ 12.109375 9.234375 16.796875 8.0625 \nQ 21.484375 6.890625 26.703125 6.890625 \nQ 35.15625 6.890625 40.078125 11.328125 \nQ 45.015625 15.765625 45.015625 23.390625 \nQ 45.015625 31 40.078125 35.4375 \nQ 35.15625 39.890625 26.703125 39.890625 \nQ 22.75 39.890625 18.8125 39.015625 \nQ 14.890625 38.140625 10.796875 36.28125 \nz\n\" id=\"DejaVuSans-53\"></path>\n       </defs>\n       <use xlink:href=\"#DejaVuSans-53\"></use>\n       <use x=\"63.623047\" xlink:href=\"#DejaVuSans-48\"></use>\n      </g>\n     </g>\n    </g>\n    <g id=\"xtick_3\">\n     <g id=\"line2d_3\">\n      <g>\n       <use style=\"stroke:#000000;stroke-width:0.8;\" x=\"158.360156\" xlink:href=\"#m3df349d0fa\" y=\"94.045312\"></use>\n      </g>\n     </g>\n     <g id=\"text_3\">\n      <!-- 100 -->\n      <g transform=\"translate(148.816406 108.64375)scale(0.1 -0.1)\">\n       <defs>\n        <path d=\"M 12.40625 8.296875 \nL 28.515625 8.296875 \nL 28.515625 63.921875 \nL 10.984375 60.40625 \nL 10.984375 69.390625 \nL 28.421875 72.90625 \nL 38.28125 72.90625 \nL 38.28125 8.296875 \nL 54.390625 8.296875 \nL 54.390625 0 \nL 12.40625 0 \nz\n\" id=\"DejaVuSans-49\"></path>\n       </defs>\n       <use xlink:href=\"#DejaVuSans-49\"></use>\n       <use x=\"63.623047\" xlink:href=\"#DejaVuSans-48\"></use>\n       <use x=\"127.246094\" xlink:href=\"#DejaVuSans-48\"></use>\n      </g>\n     </g>\n    </g>\n    <g id=\"xtick_4\">\n     <g id=\"line2d_4\">\n      <g>\n       <use style=\"stroke:#000000;stroke-width:0.8;\" x=\"223.750781\" xlink:href=\"#m3df349d0fa\" y=\"94.045312\"></use>\n      </g>\n     </g>\n     <g id=\"text_4\">\n      <!-- 150 -->\n      <g transform=\"translate(214.207031 108.64375)scale(0.1 -0.1)\">\n       <use xlink:href=\"#DejaVuSans-49\"></use>\n       <use x=\"63.623047\" xlink:href=\"#DejaVuSans-53\"></use>\n       <use x=\"127.246094\" xlink:href=\"#DejaVuSans-48\"></use>\n      </g>\n     </g>\n    </g>\n    <g id=\"xtick_5\">\n     <g id=\"line2d_5\">\n      <g>\n       <use style=\"stroke:#000000;stroke-width:0.8;\" x=\"289.141406\" xlink:href=\"#m3df349d0fa\" y=\"94.045312\"></use>\n      </g>\n     </g>\n     <g id=\"text_5\">\n      <!-- 200 -->\n      <g transform=\"translate(279.597656 108.64375)scale(0.1 -0.1)\">\n       <defs>\n        <path d=\"M 19.1875 8.296875 \nL 53.609375 8.296875 \nL 53.609375 0 \nL 7.328125 0 \nL 7.328125 8.296875 \nQ 12.9375 14.109375 22.625 23.890625 \nQ 32.328125 33.6875 34.8125 36.53125 \nQ 39.546875 41.84375 41.421875 45.53125 \nQ 43.3125 49.21875 43.3125 52.78125 \nQ 43.3125 58.59375 39.234375 62.25 \nQ 35.15625 65.921875 28.609375 65.921875 \nQ 23.96875 65.921875 18.8125 64.3125 \nQ 13.671875 62.703125 7.8125 59.421875 \nL 7.8125 69.390625 \nQ 13.765625 71.78125 18.9375 73 \nQ 24.125 74.21875 28.421875 74.21875 \nQ 39.75 74.21875 46.484375 68.546875 \nQ 53.21875 62.890625 53.21875 53.421875 \nQ 53.21875 48.921875 51.53125 44.890625 \nQ 49.859375 40.875 45.40625 35.40625 \nQ 44.1875 33.984375 37.640625 27.21875 \nQ 31.109375 20.453125 19.1875 8.296875 \nz\n\" id=\"DejaVuSans-50\"></path>\n       </defs>\n       <use xlink:href=\"#DejaVuSans-50\"></use>\n       <use x=\"63.623047\" xlink:href=\"#DejaVuSans-48\"></use>\n       <use x=\"127.246094\" xlink:href=\"#DejaVuSans-48\"></use>\n      </g>\n     </g>\n    </g>\n    <g id=\"xtick_6\">\n     <g id=\"line2d_6\">\n      <g>\n       <use style=\"stroke:#000000;stroke-width:0.8;\" x=\"354.532031\" xlink:href=\"#m3df349d0fa\" y=\"94.045312\"></use>\n      </g>\n     </g>\n     <g id=\"text_6\">\n      <!-- 250 -->\n      <g transform=\"translate(344.988281 108.64375)scale(0.1 -0.1)\">\n       <use xlink:href=\"#DejaVuSans-50\"></use>\n       <use x=\"63.623047\" xlink:href=\"#DejaVuSans-53\"></use>\n       <use x=\"127.246094\" xlink:href=\"#DejaVuSans-48\"></use>\n      </g>\n     </g>\n    </g>\n   </g>\n   <g id=\"matplotlib.axis_2\">\n    <g id=\"ytick_1\">\n     <g id=\"line2d_7\">\n      <defs>\n       <path d=\"M 0 0 \nL -3.5 0 \n\" id=\"m20b333a647\" style=\"stroke:#000000;stroke-width:0.8;\"></path>\n      </defs>\n      <g>\n       <use style=\"stroke:#000000;stroke-width:0.8;\" x=\"26.925\" xlink:href=\"#m20b333a647\" y=\"10.999219\"></use>\n      </g>\n     </g>\n     <g id=\"text_7\">\n      <!-- 0 -->\n      <g transform=\"translate(13.5625 14.798437)scale(0.1 -0.1)\">\n       <use xlink:href=\"#DejaVuSans-48\"></use>\n      </g>\n     </g>\n    </g>\n    <g id=\"ytick_2\">\n     <g id=\"line2d_8\">\n      <g>\n       <use style=\"stroke:#000000;stroke-width:0.8;\" x=\"26.925\" xlink:href=\"#m20b333a647\" y=\"37.155469\"></use>\n      </g>\n     </g>\n     <g id=\"text_8\">\n      <!-- 20 -->\n      <g transform=\"translate(7.2 40.954687)scale(0.1 -0.1)\">\n       <use xlink:href=\"#DejaVuSans-50\"></use>\n       <use x=\"63.623047\" xlink:href=\"#DejaVuSans-48\"></use>\n      </g>\n     </g>\n    </g>\n    <g id=\"ytick_3\">\n     <g id=\"line2d_9\">\n      <g>\n       <use style=\"stroke:#000000;stroke-width:0.8;\" x=\"26.925\" xlink:href=\"#m20b333a647\" y=\"63.311719\"></use>\n      </g>\n     </g>\n     <g id=\"text_9\">\n      <!-- 40 -->\n      <g transform=\"translate(7.2 67.110937)scale(0.1 -0.1)\">\n       <defs>\n        <path d=\"M 37.796875 64.3125 \nL 12.890625 25.390625 \nL 37.796875 25.390625 \nz\nM 35.203125 72.90625 \nL 47.609375 72.90625 \nL 47.609375 25.390625 \nL 58.015625 25.390625 \nL 58.015625 17.1875 \nL 47.609375 17.1875 \nL 47.609375 0 \nL 37.796875 0 \nL 37.796875 17.1875 \nL 4.890625 17.1875 \nL 4.890625 26.703125 \nz\n\" id=\"DejaVuSans-52\"></path>\n       </defs>\n       <use xlink:href=\"#DejaVuSans-52\"></use>\n       <use x=\"63.623047\" xlink:href=\"#DejaVuSans-48\"></use>\n      </g>\n     </g>\n    </g>\n    <g id=\"ytick_4\">\n     <g id=\"line2d_10\">\n      <g>\n       <use style=\"stroke:#000000;stroke-width:0.8;\" x=\"26.925\" xlink:href=\"#m20b333a647\" y=\"89.467969\"></use>\n      </g>\n     </g>\n     <g id=\"text_10\">\n      <!-- 60 -->\n      <g transform=\"translate(7.2 93.267187)scale(0.1 -0.1)\">\n       <defs>\n        <path d=\"M 33.015625 40.375 \nQ 26.375 40.375 22.484375 35.828125 \nQ 18.609375 31.296875 18.609375 23.390625 \nQ 18.609375 15.53125 22.484375 10.953125 \nQ 26.375 6.390625 33.015625 6.390625 \nQ 39.65625 6.390625 43.53125 10.953125 \nQ 47.40625 15.53125 47.40625 23.390625 \nQ 47.40625 31.296875 43.53125 35.828125 \nQ 39.65625 40.375 33.015625 40.375 \nz\nM 52.59375 71.296875 \nL 52.59375 62.3125 \nQ 48.875 64.0625 45.09375 64.984375 \nQ 41.3125 65.921875 37.59375 65.921875 \nQ 27.828125 65.921875 22.671875 59.328125 \nQ 17.53125 52.734375 16.796875 39.40625 \nQ 19.671875 43.65625 24.015625 45.921875 \nQ 28.375 48.1875 33.59375 48.1875 \nQ 44.578125 48.1875 50.953125 41.515625 \nQ 57.328125 34.859375 57.328125 23.390625 \nQ 57.328125 12.15625 50.6875 5.359375 \nQ 44.046875 -1.421875 33.015625 -1.421875 \nQ 20.359375 -1.421875 13.671875 8.265625 \nQ 6.984375 17.96875 6.984375 36.375 \nQ 6.984375 53.65625 15.1875 63.9375 \nQ 23.390625 74.21875 37.203125 74.21875 \nQ 40.921875 74.21875 44.703125 73.484375 \nQ 48.484375 72.75 52.59375 71.296875 \nz\n\" id=\"DejaVuSans-54\"></path>\n       </defs>\n       <use xlink:href=\"#DejaVuSans-54\"></use>\n       <use x=\"63.623047\" xlink:href=\"#DejaVuSans-48\"></use>\n      </g>\n     </g>\n    </g>\n   </g>\n   <g id=\"patch_3\">\n    <path d=\"M 26.925 94.045312 \nL 26.925 10.345312 \n\" style=\"fill:none;stroke:#000000;stroke-linecap:square;stroke-linejoin:miter;stroke-width:0.8;\"></path>\n   </g>\n   <g id=\"patch_4\">\n    <path d=\"M 361.725 94.045312 \nL 361.725 10.345312 \n\" style=\"fill:none;stroke:#000000;stroke-linecap:square;stroke-linejoin:miter;stroke-width:0.8;\"></path>\n   </g>\n   <g id=\"patch_5\">\n    <path d=\"M 26.925 94.045312 \nL 361.725 94.045312 \n\" style=\"fill:none;stroke:#000000;stroke-linecap:square;stroke-linejoin:miter;stroke-width:0.8;\"></path>\n   </g>\n   <g id=\"patch_6\">\n    <path d=\"M 26.925 10.345312 \nL 361.725 10.345312 \n\" style=\"fill:none;stroke:#000000;stroke-linecap:square;stroke-linejoin:miter;stroke-width:0.8;\"></path>\n   </g>\n  </g>\n </g>\n <defs>\n  <clipPath id=\"pdc8d6ed2d1\">\n   <rect height=\"83.7\" width=\"334.8\" x=\"26.925\" y=\"10.345312\"></rect>\n  </clipPath>\n </defs>\n</svg>\n",
      "image/png": "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\n"
     },
     "metadata": {
      "needs_background": "light"
     }
    },
    {
     "output_type": "stream",
     "name": "stdout",
     "text": [
      "FPS=16.49, total frames = 50 , time=3.0328 seconds\n"
     ]
    }
   ],
   "source": [
    "if __name__ == \"__main__\":\n",
    "    print(\"hello to Xilinx app development using DPU\")\n",
    "    os.system('dexplorer -w')\n",
    "\n",
    "    app(\"/home/root/Vitis-AI/dex_read\", 1, \"/home/root/Vitis-AI/Custom_models/dpu_u3d_kp.elf\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ]
}

datagen.py

Python
calibration data generator
import os
import numpy as np


calib_images_path = './calib_images'
calib_batch_size = 100

def get_calib_data(iter):
    """
    Function provides calibration images to the quantizer from the training set
    """

    frames = os.listdir(calib_images_path)
    # np.random.shuffle(frames)
    num_frames = len(frames)
    print("number of calibration images : ", num_frames)

    out_train_x_normalized = np.zeros((calib_batch_size, 64, 256, 1))

    frame_indices = list(range(iter*calib_batch_size, calib_batch_size + (iter * calib_batch_size)))

    for i, frame in enumerate(frame_indices):
        f_path = calib_images_path + '/' + frames[frame]
        print(f_path)
        file = np.load(f_path)
        file = file[0: 64, 0: 256, :]
        out_train_x_normalized[i] = np.expand_dims(file[:, :, 3], -1)  # depth channel
        out_train_x_normalized /= np.max(out_train_x_normalized)  # normalize
    
    return {"input_1": out_train_x_normalized}
    

if __name__ == "__main__":
    # keras_convert(keras_json=None, tf_ckpt=None, keras_hdf5='/home/sambit/Xilinx_Works/Vitis-AI-Tutorials-DenseNetX_DPUv2/files/from_docker_tf/trained_unet_1ch_input.h5')
    for i in range(10):
        get_calib_data(i)

inference application - .py file

Python
copy to Ultra96 and change paths to .elf and test data
from ctypes import *
import cv2
import numpy as np
import runner
import os
import xir.graph
import pathlib
import xir.subgraph
import threading
import time
import sys
import argparse
import numpy as np
import matplotlib.pyplot as plt
from scipy.special import softmax

def preprocess_fn(image_path):
    '''
    Image pre-processing.
    Rearranges from BGR to RGB then normalizes to range 0:1
    input arg: path of image file
    return: numpy array
    '''
    # image = cv2.imread(image_path)
    image = np.load(image_path)
    print(image.shape)

    test_x = image[0:64, 0:256, 3]  # get only depth channel
    plt.imshow(test_x)
    plt.imsave("depth.png", test_x)
    test_x /= np.max(test_x)
    
    return test_x

def get_subgraph (g):
    sub = []
    root = g.get_root_subgraph()
    sub = [ s for s in root.children
            if s.metadata.get_attr_str ("device") == "DPU"]
    return sub 


def runDPU(id,start,dpu,img):

    '''get tensor'''
    inputTensors = dpu.get_input_tensors()
    outputTensors = dpu.get_output_tensors()
    outputHeight = outputTensors[0].dims[1]
    outputWidth = outputTensors[0].dims[2]
    outputChannel = outputTensors[0].dims[3]
    outputSize = outputHeight*outputWidth*outputChannel

    batchSize = inputTensors[0].dims[0]
    n_of_images = len(img)
    count = 0
    write_index = start
    while count < n_of_images:
        if (count+batchSize<=n_of_images):
            runSize = batchSize
        else:
            runSize=n_of_images-count
        shapeIn = (runSize,) + tuple([inputTensors[0].dims[i] for i in range(inputTensors[0].ndim)][1:])

        '''prepare batch input/output '''
        outputData = []
        inputData = []
        outputData.append(np.empty((runSize,outputHeight,outputWidth,outputChannel), dtype = np.float32, order = 'C'))
        inputData.append(np.empty((shapeIn), dtype = np.float32, order = 'C'))

        '''init input image to input buffer '''
        for j in range(runSize):
            imageRun = inputData[0]
            imageRun[j,...] = img[(count+j)% n_of_images].reshape(inputTensors[0].dims[1],inputTensors[0].dims[2],inputTensors[0].dims[3])

        '''run with batch '''
        job_id = dpu.execute_async(inputData,outputData)
        dpu.wait(job_id)

        # get predictions direclty here
        predictions = outputData[0][0]
        predictions = softmax(predictions, -1)  # softmax along channels
        print("predictions shape: ", predictions.shape)
        mask = np.argmax(predictions, -1)  # along channels
        print("mask shape: ", mask.shape)

        img_test = np.squeeze(imageRun[0], -1)
        plt.imshow(img_test)
        plt.imsave("depth_"+str(count)+'.png', img_test)
        # mask = mask[0]
        plt.imshow(mask)
        plt.imsave("keypoints_"+str(count)+".png", mask)


        # for j in range(len(outputData)):
        #     outputData[j] = outputData[j].reshape(runSize, outputSize)

        '''store output vectors '''
        # for j in range(runSize):
        #     out_q[write_index] = outputData[0][j]
        #     write_index += 1
        
        count = count + runSize
        


def app(image_dir,threads,model):

    listimage=os.listdir(image_dir)
    runTotal = len(listimage)

    global out_q
    out_q = [None] * runTotal

    g = xir.graph.Graph.deserialize(pathlib.Path(model))
    subgraphs = get_subgraph (g)
    assert len(subgraphs) == 1 # only one DPU kernel
    all_dpu_runners = []
    for i in range(threads):
        all_dpu_runners.append(runner.Runner(subgraphs[0], "run"))

    ''' preprocess images '''
    print('Pre-processing',runTotal,'images...')
    img = []
    for i in range(runTotal):
        path = os.path.join(image_dir,listimage[i])
        img.append(preprocess_fn(path))

    '''run threads '''
    print('Starting',threads,'threads...')
    threadAll = []
    start=0
    for i in range(threads):
        if (i==threads-1):
            end = len(img)
        else:
            end = start+(len(img)//threads)
        in_q = img[start:end]
        t1 = threading.Thread(target=runDPU, args=(i,start,all_dpu_runners[i], in_q))
        threadAll.append(t1)
        start=end

    time1 = time.time()
    for x in threadAll:
        x.start()
    for x in threadAll:
        x.join()
    time2 = time.time()
    timetotal = time2 - time1

    fps = float(runTotal / timetotal)
    print("FPS=%.2f, total frames = %.0f , time=%.4f seconds" %(fps,runTotal, timetotal))


    ''' post-processing '''
    # classes = ['airplane','automobile','bird','cat','deer','dog','frog','horse','ship','truck']  
    # correct = 0
    # wrong = 0
    # print('output buffer length:',len(out_q))
    # for i in range(len(out_q)):
    #     argmax = np.argmax((out_q[i]))
    #     prediction = classes[argmax]
    #     ground_truth, _ = listimage[i].split('_')
    #     if (ground_truth==prediction):
    #         correct += 1
    #     else:
    #         wrong += 1
    # accuracy = correct/len(out_q)
    # print('Correct:',correct,'Wrong:',wrong,'Accuracy:', accuracy)


if __name__ == "__main__":
    print("hello to Xilinx app development using DPU")
    os.system('dexplorer -w')

    app("/home/root/Vitis-AI/dex_read", 1, "/home/root/Vitis-AI/Custom_models/dpu_u3d_kp.elf")

Credits

sambit mohapatra

sambit mohapatra

1 project • 5 followers
System engineer and researcher in the field of deep learning for perception systems with a leading multinational automotive supplier.

Comments