apps/FastDeploy
1
0
Fork 0
mirror of https://github.com/PaddlePaddle/FastDeploy.git synced 2026-04-25 09:57:51 +08:00
Code Issues Actions 9 Packages Projects Releases Wiki Activity

Split enable_mm (#7183)

Browse Source 
Co-authored-by: liuruian <liuruian@MacBook-Pro.local>
This commit is contained in:
K11OntheBoat
2026-04-08 11:25:41 +08:00
committed by GitHub
parent 8496ec71a6
commit bb48bcbaa2
33 changed files with 109 additions and 69 deletions
Download Patch File Download Diff File Expand all files Collapse all files
fastdeploy/model_executor/layers/attention/mla_attention_backend.py
+1 -1
View File
@@ -263,7 +263,7 @@ class MLAAttentionBackend(AttentionBackend):
self.rope_theta: float = (
10000.0 if fd_config.model_config.rope_theta is None else fd_config.model_config.rope_theta
)
self.rope_3d: bool = getattr(fd_config.model_config, "rope_3d", False)
self.rope_3d: bool = fd_config.enable_rope_3d_runtime
self.causal: bool = getattr(fd_config.model_config, "causal", True)
self.speculative_method = fd_config.speculative_config.method
self.use_speculate: bool = self.speculative_method is not None