Running into NotImplementedError while trying to implement DenseNet

maxmatical · August 22, 2018, 3:10pm

The error comes from the Transition blocks. TransitionBlock is what I wrote up myself, and TransitionBlock_test is using code form another implementation of the same network. I even modified the variables to be the same between both, but I still get an error in my code while the other one runs fine.

github.com

maxmatical/pytorch-projects/blob/master/TransitionBlock.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import torch\n",
    "import torchvision\n",
    "import torchvision.transforms as transforms\n",
    "import numpy as np\n",
    "import math\n",
    "\n",
    "import torch.nn as nn\n",
    "import torch.nn.functional as F\n",
    "\n",
    "from torchsummary import summary\n"
   ]
  },

This file has been truncated. show original

morpheusPrime · August 22, 2018, 3:34pm

You have typo in the first Transition Block definition. You wrote foward instead of forward. So it can’t find the forward function

Source: Personal experience

maxmatical · August 22, 2018, 3:36pm

oh my god thank you. i thought i was losing it