Fix prefill OOM error in the case of large page size (#5081)
This commit is contained in:
@@ -502,6 +502,7 @@ class Scheduler(
|
||||
self.tree_cache = ChunkCache(
|
||||
req_to_token_pool=self.req_to_token_pool,
|
||||
token_to_kv_pool_allocator=self.token_to_kv_pool_allocator,
|
||||
page_size=self.page_size,
|
||||
)
|
||||
else:
|
||||
if self.enable_hierarchical_cache:
|
||||
|
||||
Reference in New Issue
Block a user