ayo/ipex-llm
1
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions 1

[NPU] fix transpose_value = False for NPU optimize_model=True (#12525)

Browse source
This commit is contained in:
Ruonan Wang 2024-12-10 23:51:39 -08:00 committed by GitHub
parent 588bfa24dc
commit 41ef4974ab
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
1 changed files with 1 additions and 0 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

1
python/llm/src/ipex_llm/transformers/npu_models/mp_models_base.py
View file

@ -225,6 +225,7 @@ class LLMBaseNNFactory(NNFactory):
head_dim=head_dim, head_dim=head_dim,
) )
new_key_states = key_states new_key_states = key_states
new_value_states = value_states
if mode == "decode": if mode == "decode":
key_states = self.concat(past_key, key_states, axis=-2) key_states = self.concat(past_key, key_states, axis=-2)