Update config.json with the correct state
#1
by
dsikka
- opened
The model is compressed however was incorrectly saved using the "frozen" quantization_status, which results in incorrect loading through transformers.
mgoin
changed pull request status to
merged