Lightrun Answers was designed to reduce the constant googling that comes with debugging 3rd party libraries. It collects links to all the places you might be looking at while hunting down a tough bug.
And, if you’re still stuck at the end, we’re happy to hop on a call to see how we can help out.
st.cache fails for TF2 SavedModels (HuggingFace, Tensorflow Hub, etc.)
See original GitHub issueThere is a particular TensorFlow type that st.cache fails to cache: tensorflow.python.util.object_identity._ObjectIdentityWrapper. This is apparently a very common type and used in TF2 SavedModels, including HuggingFace tokenizers and TensorFlow Hub models.
I tried overriding hash_funcs with id and lamba _: None, but both resulted in the program hanging. Without overriding hash_funcs, I get a Streamlit error saying that Streamlit doesn’t know how to cache the type.
Issue Analytics
- State:
- Created 3 years ago
- Comments:7 (1 by maintainers)
Top Results From Across the Web
Top Related Medium Post
Top Related StackOverflow Question
Troubleshoot Live Code
Top Related Reddit Thread
Top Related Hackernoon Post
Top Related Tweet
Top Related Dev.to Post
Top Related Hashnode Post
@sai-krishna-msk I might but can’t recall how. We eventually moved to expose an HTTP API that the Streamlit app calls. FastAPI was our choice on the server side and httpx the choice on the client side.
But, Streamlit recently release .89 which has two new cache decorators that might interest you. See: https://docs.streamlit.io/en/stable/changelog.html#version-0-89-0
Closing since this is resolved by using our newer caching methods.