Closed kiahsa2002 closed 4 years ago
Hi, you can have a look at the way we store the data in npz files and convert your data accordingly. Here is a code snippet that allows you to do this:
import numpy as np
loader = dict(np.load('data/stack_overflow.npz', allow_pickle=True))
print(loader.keys())
print(loader['arrival_times'])
print(loader['marks'])
Both arrival_times
and marks
are lists of lists. If you save your data to an npz
file using the same format (using the function np.savez
), you should be able to load it with our function dpp.data.load_dataset
.
Thank you for your response with details.
Regards, Kiahsa
From: Oleksandr Shchur notifications@github.com Sent: Monday, June 8, 2020 2:10:19 PM To: shchur/ifl-tpp ifl-tpp@noreply.github.com Cc: kiahsa2002 kiahsa2002@outlook.com; Author author@noreply.github.com Subject: Re: [shchur/ifl-tpp] ATM dataset testing (#7)
Hi, you can have a look at the way we store the data in npz files and convert your data accordingly. Here is a code snippet that allows you to do this:
import numpy as np
loader = dict(np.load('data/stack_overflow.npz', allow_pickle=True)) print(loader.keys()) print(loader['arrival_times']) print(loader['marks'])
Both arrival_times and marks are lists of lists. If you save your data to an npz file using the same format (using the function np.savezhttps://numpy.org/doc/stable/reference/generated/numpy.savez.html), you should be able to load it with our function dpp.data.load_dataset.
— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHubhttps://github.com/shchur/ifl-tpp/issues/7#issuecomment-640563864, or unsubscribehttps://github.com/notifications/unsubscribe-auth/ALYAXNDQQKRDXRREJF23UGDRVTIKXANCNFSM4NRDODKQ.
Hi,
I'm new to TPP.
I was trying to run your interactive ipython file with the ATM dataset as CSV file. I see all datasets you used are .npz. Do I have to preprocess ATM csv file to .npz? How you set the sequence length? Since the ATM dataset I have contains only 4 months of data.
I request you to clarify.
Regards, kiahsa