apps/FastDeploy
1
0
Fork 0
mirror of https://github.com/PaddlePaddle/FastDeploy.git synced 2026-04-23 17:11:21 +08:00
Code Issues Actions 23 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/engine/sched/resource_manager_v1.py
+3 -3
View File
@@ -205,11 +205,11 @@ class ResourceManagerV1(ResourceManager):
self.need_block_num_map = dict()
self.encoder_cache = None
if config.model_config.enable_mm and config.cache_config.max_encoder_cache > 0:
if config.enable_mm_runtime and config.cache_config.max_encoder_cache > 0:
self.encoder_cache = EncoderCacheManager(config.cache_config.max_encoder_cache)
self.processor_cache = None
if config.model_config.enable_mm and config.cache_config.max_processor_cache > 0:
if config.enable_mm_runtime and config.cache_config.max_processor_cache > 0:
max_processor_cache_in_bytes = int(config.cache_config.max_processor_cache * 1024 * 1024 * 1024)
self.processor_cache = ProcessorCacheManager(max_processor_cache_in_bytes)
@@ -550,7 +550,7 @@ class ResourceManagerV1(ResourceManager):
num_new_tokens = token_budget // self.config.cache_config.block_size * self.config.cache_config.block_size
request.with_image = False
if not self.config.model_config.enable_mm:
if not self.config.enable_mm_runtime:
return num_new_tokens
inputs = request.multimodal_inputs