apps/FastDeploy
1
0
Fork 0
mirror of https://github.com/PaddlePaddle/FastDeploy.git synced 2026-04-23 00:17:25 +08:00
Code Issues Actions 19 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/spec_decode/base.py
+1 -1
View File
@@ -71,7 +71,7 @@ class Proposer(ABC):
self.max_ngram_size = self.speculative_config.max_ngram_size
self.min_ngram_size = self.speculative_config.min_ngram_size
self.enable_mm = self.model_config.enable_mm
self.enable_mm = self.fd_config.enable_mm_runtime
spec_logger.info(f"Speculate config: {self.speculative_config}")
fastdeploy/spec_decode/mtp.py
+1 -1
View File
@@ -103,7 +103,7 @@ class MTPProposer(Proposer):
self.num_main_model_layers = self.model_config.num_hidden_layers
self.local_rank = local_rank
self.device_id = device_id
self.use_attn_mask_offset = self.enable_mm and self.fd_config.deploy_modality != "text"
self.use_attn_mask_offset = self.enable_mm
self._update_mtp_config(main_model)
self._load_model()