From 9566ae8827f59fd7fd40d042035c7e6eb71aef9b Mon Sep 17 00:00:00 2001 From: ApplEOFDiscord <31272106+ApplEOFDiscord@users.noreply.github.com> Date: Wed, 24 Sep 2025 14:43:46 +0800 Subject: [PATCH] [Bug Fix] disable prefix caching in mm model (#4167) * add http get retry * fix coments * disable prefix caching in mm model * fix unit test --------- Co-authored-by: zhangjunjun04 Co-authored-by: Jiang-Jia-Jun <163579578+Jiang-Jia-Jun@users.noreply.github.com> Co-authored-by: YuBaoku <49938469+EmmonsCurse@users.noreply.github.com> --- fastdeploy/config.py | 2 ++ fastdeploy/engine/args_utils.py | 2 -- 2 files changed, 2 insertions(+), 2 deletions(-) diff --git a/fastdeploy/config.py b/fastdeploy/config.py index c1953577d..6f3e361b2 100644 --- a/fastdeploy/config.py +++ b/fastdeploy/config.py @@ -1540,6 +1540,8 @@ class FDConfig: self.cache_config.postprocess(self.scheduler_config.max_num_batched_tokens, self.scheduler_config.max_num_seqs) self.cache_config.max_block_num_per_seq = int(self.max_model_len // self.cache_config.block_size) + if self.model_config is not None and self.model_config.enable_mm: + self.cache_config.enable_prefix_caching = False if self.guided_decoding_backend == "auto": if current_platform.is_xpu() or self.speculative_config.method is not None: diff --git a/fastdeploy/engine/args_utils.py b/fastdeploy/engine/args_utils.py index f1367b1ae..957932fb2 100644 --- a/fastdeploy/engine/args_utils.py +++ b/fastdeploy/engine/args_utils.py @@ -409,8 +409,6 @@ class EngineArgs: self.enable_prefix_caching = False if self.speculative_config is not None: self.enable_prefix_caching = False - if self.enable_mm: - self.enable_prefix_caching = False if not current_platform.is_cuda(): self.enable_prefix_caching = False if self.dynamic_load_weight: