Bug about the version of Pytorch during preparing datasets.

Labbeti / conette-audio-captioning

CoNeTTE: An efficient Audio Captioning system leveraging multiple datasets with Task Embedding

https://arxiv.org/pdf/2309.00454.pdf

11 stars 0 forks source link

Bug about the version of Pytorch during preparing datasets. #4

Closed LittleFlyingSheep closed 5 months ago

LittleFlyingSheep commented 5 months ago

Hi, I find that here is a bug when preparing the datasets following Readme.

Specifically, the error information is shown as below:

ERROR! Intel® Extension for PyTorch* needs to work with PyTorch 2.2.*, but PyTorch 1.13.1+cu117 is found. Please switch to the matching version and run `again.`

However, Readme highlights that the version of Pytorch should be <2.0. This is contradictory.

Could you please help me to address this bug for the use of DCASE 2024 Task 6 baseline?

Labbeti commented 5 months ago

Hi! Thank you very much for reporting this error.

This repository is not the code used to train the DCASE baseline, it is available at https://github.com/Labbeti/dcase2024-task6-baseline. The requirements are different, and the baseline is not the CoNeTTE model but a CNext-trans model.

Anyway, I am really sorry for this error, I will try to correct it. This is probably due to the current requirements, which do not contain a specific version.

LittleFlyingSheep commented 5 months ago

收到