Unverified Commit 8aa5ae6b authored by yilian49's avatar yilian49 Committed by GitHub
Browse files

load draft model fix (#7506)

parent 8a323557
......@@ -575,7 +575,13 @@ class DummyModelLoader(BaseModelLoader):
# 2. Post-processing of weights, including assigning specific member variables.
# For `dummy_init`, only the second stage is required.
if hasattr(model, "post_load_weights"):
model.post_load_weights()
if (
model_config.hf_config.architectures[0]
== "DeepseekV3ForCausalLMNextN"
):
model.post_load_weights(is_nextn=True)
else:
model.post_load_weights()
return model.eval()
......
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment