+- Python Forum (https://python-forum.io)
+-- Forum: Python Coding (https://python-forum.io/forum-7.html)
+--- Forum: Data Science (https://python-forum.io/forum-44.html)
+--- Thread: Analysis (/thread-34802.html)
Analysis - tech93 - Sep-02-2021
Hi,
[1.1] “Appendix J: Multiple (Maximum and Range) Pooling Statistics in a Convolution Network” in the book, Understanding
Deep Learning. Download from https://www.understandingdeeplearning.com
Apply multiple pooling as shown in [1.1] on any image data set.
How can we apply multiple pooling as shown in [1.1] on any image data set.
Thanks..
RE: Analysis - jefsummers - Sep-02-2021
And your code so far....?
RE: Analysis - tech93 - Sep-03-2021
The process diagram in chapter 6.11 and the recipe in Appendix J provides a reasonable outline for working with Keras to solve the problem but not sure how to progress for this.
Thanks..
RE: Analysis - jefsummers - Sep-05-2021
Most likely you want to start in a notebook style app for your program - Jupyter notebooks, Google's colab, whatever.
After a markdown cell describing what you want to do, first coding cell should be your imports.
Once the imports work (you may need to pip install tensorflow depeneding on your system) next cell would start to read your data
Next cell display that data to be sure it is in the format you expected
Next cell start your analysis
And show your work so far.
And please keep this publicly posted. There is no reason for private messaging here.
RE: Analysis - tech93 - Sep-05-2021
is there any need to do pip install tensorflow if whole of this exercise is tried to be done on Google's colab as there we need not to do such pip installations as it has all the required packages inbuilt in itself?
Thanks..
RE: Analysis - jefsummers - Sep-05-2021
That's why I said depending on your system. You are correct, Google's colab does not require you to install tensorflow. If instead you were using Watson, Jupyter lab, Jupyter notebook, or VSCode you would.
The instances where you would need to do that appears to outnumber those where you don't.