Activation_stats.plot_layer_stats always shows 400 activations

bgraysea · January 5, 2021, 9:13am

fastai/fastbook/blob/3916b71bdf2f9e587ac82f3c2ef4aabd05b8f51c/13_convolutions.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "#hide\n",
    "!pip install -Uqq fastbook\n",
    "import fastbook\n",
    "fastbook.setup_book()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [

This file has been truncated. show original

Why is the x axis always 400? I thought from summary that the number of activations per layer is going down?

Sequential (Input shape: 64) ============================================================================ Layer (type) Output Shape Param # Trainable ============================================================================ 64 x 4 x 14 x 14 Conv2d 40 True ReLU ____________________________________________________________________________ 64 x 8 x 7 x 7 Conv2d 296 True ReLU ____________________________________________________________________________ 64 x 16 x 4 x 4 Conv2d 1168 True ReLU ____________________________________________________________________________ 64 x 32 x 2 x 2 Conv2d 4640 True ReLU ____________________________________________________________________________ 64 x 2 x 1 x 1 Conv2d 578 True ____________________________________________________________________________

bgraysea · January 8, 2021, 2:59pm

Any pointers on the above? I must be mis-understanding something; as I’m struggling to figure out why the count on the x axis of the graphs doesn’t mach the Summary: Output

bgraysea · January 11, 2021, 10:33am

4×14×14 = 784
8x7x7=392
16x4x4=256

bgraysea · January 13, 2021, 11:49am

Ah; breakthrough

The x axis is batch number; not activation
So the y axis is showing the statistic of all activations in that layer for each batch; so there where 400 batches