Closed adelightday closed 5 months ago
Hi, I didn't quite understand your question. Are you having trouble configuring the environment for DEKM_dense.py? If so, you can use this Colab notebook, where I successfully run DEKM_dense.py on the 20NEWS dataset.
Thank you for attaching the Colab notebook! Just wondering, how do I lunch an experiment for DEKM_dense.py? Is it the same with DEKM.py one? python DEKM.py MNIST for MNIST dataset?
how do I lunch an experiment for DEKM_dense.py
First, select the dataset you want to use for the experiment by revising this line: ds_name = ['REUTERS', '20NEWS', 'RCV1'][1] # for 20NEWS
. Then, run all the notebook cells.
Is it the same with DEKM.py one?
A little different. In the Colab notebook for DEKM_dense.py, you select the dataset by modifying this line. In contrast, for DEKM.py, you specify the dataset through command-line arguments python DEKM.py MNIST
or python DEKM.py USPS
.
Tysm for the answer!!
Hi, I want to ask if there are alternative dependencies to clone the code locally. I have tried using Python 3.8 and venv but there is still an error for DEKM_dense.py
It stated "No matching distribution found for tensorflow==2.4.1"
Many thanks for considering my request.