Skip to content

Commit

Permalink
Remove not-imported MistralConfig (intel-analytics#10670)
Browse files Browse the repository at this point in the history
  • Loading branch information
xiangyuT authored Apr 7, 2024
1 parent 1a9b820 commit 08018a1
Showing 1 changed file with 0 additions and 9 deletions.
9 changes: 0 additions & 9 deletions python/llm/src/ipex_llm/vllm/config.py
Original file line number Diff line number Diff line change
Expand Up @@ -318,15 +318,6 @@ def _get_and_verify_max_len(
def get_config(model: str,
trust_remote_code: bool,
revision: Optional[str] = None) -> PretrainedConfig:
# NOTE: Because the Mistral model in HF hub does not have
# `configuration_mistral.py`, we cannot use `AutoConfig` to load the
# config. Instead, we use `MistralConfig` directly.
# NOTE: This is a hack. This does not work for local models.
# FIXME: Remove this once the Mistral model is available in the stable
# version of HF transformers.
if "mistral" in model.lower():
return MistralConfig.from_pretrained(model, revision=revision)

try:
config = AutoConfig.from_pretrained(
model, trust_remote_code=trust_remote_code, revision=revision)
Expand Down

0 comments on commit 08018a1

Please sign in to comment.