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

LLM：Fix moe indexs error on cpu (#10791)

Browse source
This commit is contained in:
Wang, Jian4 2024-04-18 15:56:52 +08:00 committed by GitHub
parent cbe7b5753f
commit 14ca42a048
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
2 changed files with 2 additions and 2 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

2
python/llm/src/ipex_llm/transformers/models/mixtral.py
View file

@ -102,7 +102,7 @@ def mixtral_moeblock_forward(self,
final_hidden_states = expert_layer(hidden_states, weight) final_hidden_states = expert_layer(hidden_states, weight)
else: else:
final_hidden_states = final_hidden_states + expert_layer(hidden_states, weight) final_hidden_states = final_hidden_states + expert_layer(hidden_states, weight)
elif bs < 256: elif bs < 256 and hidden_states.device.type == 'xpu':
final_hidden_states = torch.zeros((batch_size * sequence_length, hidden_dim), final_hidden_states = torch.zeros((batch_size * sequence_length, hidden_dim),
dtype=hidden_states.dtype, device=hidden_states.device) dtype=hidden_states.dtype, device=hidden_states.device)
import linear_q4_0 import linear_q4_0

2
python/llm/src/ipex_llm/transformers/models/qwen2_moe.py
View file

@ -567,7 +567,7 @@ def qwen2moe_moeblock_forward(self, hidden_states: torch.Tensor):
final_hidden_states = expert_layer(hidden_states) * weight final_hidden_states = expert_layer(hidden_states) * weight
else: else:
final_hidden_states = final_hidden_states + expert_layer(hidden_states) * weight final_hidden_states = final_hidden_states + expert_layer(hidden_states) * weight
elif bs < 256: elif bs < 256 and hidden_states.device.type == 'xpu':
final_hidden_states = torch.zeros((batch_size * sequence_length, hidden_dim), final_hidden_states = torch.zeros((batch_size * sequence_length, hidden_dim),
dtype=hidden_states.dtype, device=hidden_states.device) dtype=hidden_states.dtype, device=hidden_states.device)
import linear_q4_0 import linear_q4_0