Need hash_func for type torch._C._TensorBase

Summary

A community member noticed a “cannot hash of type” issue. More information can be found in this forum discussion.

Full error message

This is the info provided by hertzmann:

“UnhashableType: Cannot hash object of type torch._C._TensorBase”.

"I’m trying to cache initialization of a StyleGAN2 class datastructure (which includes loading model weights from a file or online). I don’t want the model weights to be reloaded (and other data structures re-initialized) every time a new image is created.

I tried refactoring my code so that @st.cache is in front of a function that only does one thing: torch.load(filename) and the same error occurs."

[hertzmann]

Here’s a code snippet that gives the originally-reported error:

import torch
import streamlit as st

@st.cache(allow_output_mutation=True)
def multby2(data):
	return data * 2

st.write(multby2(torch.zeros( (1) )))

which gives:

UnhashableTypeError: Cannot hash object of type torch._C._TensorBase, found in the arguments of multby2().

Note that this variant fails even with allow_output_mutation=True

[tc87]

Hey @hertzmann,

Thanks for providing the above 🙏. Going to circle back with the team with the new info and then will get back to you, but might not be able to respond before next week. If not, hope you have a nice weekend and will reach back out soon.