A high-throughput and memory-efficient inference and serving engine for LLMs
Mirrored from https://github.com/tduncan-lum/Nvidia-Optimization-Stack---VLLM.git
Synced Nov 11, 17:30 UTC
[BugFix] Fix minimax m2 model rotary_dim (#30384)
Signed-off-by: xuebi <xuebi@minimaxi.com> Co-authored-by: xuebi <xuebi@minimaxi.com>
Roger Young committed
Dec 10, 2025 at 20:58 UTC
d017bceb08eaac7bae2c499124ece737fb4fb22b
1 file changed
+1
-1
vllm/model_executor/models/minimax_m2.py
+1
-1
@@ -201,7 +201,7 @@ class MiniMaxM2Attention(nn.Module):
201
202
self.rotary_emb = get_rope(
203
self.head_dim,
204
- rotary_dim=rotary_dim,
204
+ rotary_dim=self.head_dim,
205
max_position=max_position_embeddings,
206
rope_parameters=rope_parameters,
207
)