minke.yu/sglang
1
0
Fork 0
Code Issues Pull Requests Actions 12 Packages Projects Releases Wiki Activity

[Fix] Fix Qwen3.5 MoE model loading and Mamba cache sharding in PP mode (#21448)

Browse Source 
Co-authored-by: zhangxiaolei123456 <zhangxiaolei.666@bytedance.com>
This commit is contained in:
Feng Su
2026-03-30 11:57:26 +08:00
committed by GitHub
co-authored by zhangxiaolei123456
parent c06ca1526c
commit 9b4dd27478
6 changed files with 78 additions and 8 deletions
Download Patch File Download Diff File Expand all files Collapse all files
test/registered/unit/mem_cache/test_mamba_unittest.py
+2
View File
@@ -99,6 +99,7 @@ class TestMamba(unittest.TestCase):
device=device,
enable_memory_saver=False,
cache_params=mamba2_cache_params,
mamba_layer_ids=mamba_layers,
enable_mamba_extra_buffer=False,
speculative_num_draft_tokens=3,
)
@@ -340,6 +341,7 @@ class TestMamba(unittest.TestCase):
device=device,
enable_memory_saver=False,
cache_params=mamba2_cache_params,
mamba_layer_ids=mamba_layers,
enable_mamba_extra_buffer=False,
speculative_num_draft_tokens=3,
)