zigma
zigma copied to clipboard
Any code and how to change the dataset to webdataset format?
I want to make the webdataset format of MSCOCO dataset, could you tell me how to make that in your code?
You can follow the READMe in this repo, they also use webdataset.
https://github.com/bytedance/1d-tokenizer?tab=readme-ov-file