Chapter 7 - Further research Implementing TTA

kofi · September 4, 2020, 3:16pm

I ended up implementing the crops as transforms and modifying the batch transforms of the validation dataloader. You can find it here.

github.com

asiedubrempong/fastai-experiments/blob/master/tta.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# TTA\n",
    "---\n",
    "This notebook contains the solution to this research question in chapter 7 of the fastai book:\n",
    "\n",
    "*Use the fastai documentation to build a function that crops an image to a square in each of the four corners, then implement a TTA method that averages the predictions on a center crop and those four crops. Did it help? Is it better than the TTA method of fastai?*"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "from fastai2.vision.all import *"

This file has been truncated. show original

Hope it helps.