Skip to content

Commit

Permalink
Revert "Fix bug when loading 4bit checkpoint quantized in INC (huggin…
Browse files Browse the repository at this point in the history
…gface#1447)"

This reverts commit 4bdf434.
  • Loading branch information
xin3he authored Nov 5, 2024
1 parent 309e0c4 commit 38deac1
Showing 1 changed file with 0 additions and 3 deletions.
3 changes: 0 additions & 3 deletions examples/text-generation/utils.py
Original file line number Diff line number Diff line change
Expand Up @@ -269,9 +269,6 @@ def setup_model(args, model_dtype, model_kwargs, logger):
original_model=org_model,
**model_kwargs,
)
# TODO: This will be removed in v1.19 Synapse release
# the loaded model should have the same dtype as original_model
model = model.to(model_kwargs["torch_dtype"])
else:
if args.assistant_model is not None:
assistant_model = AutoModelForCausalLM.from_pretrained(
Expand Down

0 comments on commit 38deac1

Please sign in to comment.