How to cache my mixture

[renmengjie7]

I noticed the annotation

# If you're using Seqio, we suggest caching your mixture as they take a while to generate.

But I don't know how to do it

[shayne-longpre]

Here are some resources, and there should be more info in the documentation: /~https://github.com/google/seqio#optional-offline-caching.

This caching is for if you are using the same vocabulary as T5. If you want to train a different model, stream the output and save their pretokenized text format, as we do in run_example.py.

[sunyi06200]

So where can add a seqio.CacheDatasetPlaceholder(required=False)?Or is there a more detailed steps in how to use the five original submix you provided, when I've downloaded them.