Multiscale Object Detection

anirudh · December 24, 2020, 7:00am

https://d2l.ai/chapter_computer-vision/multiscale-object-detection.html

StevenJokess · December 24, 2020, 10:21am

AttributeError: module 'd2l.torch' has no attribute 'multibox_prior'

StevenJokess/d2l-en-read/blob/moreme/Ch13_CV/multiscale_object_detection.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "multiscale-object-detection.ipynb",
      "provenance": []
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"

This file has been truncated. show original

I have !pip install -U git+https://github.com/d2l-ai/d2l-en.git@master
Why did it happen?

It runs well on my computer, but got bugs on colab.

sayakpaul · January 13, 2021, 5:24am

Hi.

Would appreciate it if someone could shed some more light on this statement -

Since the midpoints of anchor boxes anchors overlap with all the units on feature map fmap, the relative spatial positions of the midpoints of the anchors on any image must have a uniform distribution.

Why the uniform distribution condition must be met with?