Closed JarvisUSTC closed 1 year ago
After preparing the raw data according to https://github.com/guozix/TaI-DPT#datasets, it may take about dozens of minutes to produce the training data structure.
I uploaded a copy of cached filtered captions here, https://drive.google.com/file/d/1RXpaCC2E492GxnPIkyYvxFSIdqf-76wh/view?usp=sharing
Unzip and put all the files under the project root path should work well.
Thanks! I will try it.
Hi, I am so interested in your work and want to reproduce it. However the data directory structure is too complex, could you please provide the training data (filtered captions) for us directly?