google / google/torchax

How torchax handle tied_embedding?

Open
#68 1 comment 0 reactions 0 assignees View on GitHub
Dominant language
Python
Stars
243
Forks
38
Avg merge
15h 32m
Merged PRs (30d)
2

Description

This is just my Observation: loading weights from safetensors (or maybe getting the weights from state_dict()) results in two weights even when the model uses tied_embedding, model.embed_tokens and lm_head, If we use these two weights in the forward call, we'll chagne both weights instead of just one, no?

Contributor guide

Open the contributing guide

Assessment

This issue has not been assessed yet.

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.