Clean match_prefix and prepare_for_extend for mem cache V2 (#11200)
This commit is contained in:
@@ -60,7 +60,7 @@ class ChunkCache(BasePrefixCache):
|
||||
]
|
||||
|
||||
# `req.prefix_indices` will be used in `PrefillAdder::add_chunked_req` later
|
||||
req.prefix_indices = kv_indices
|
||||
req.prefix_indices = kv_indices.to(dtype=torch.int64, copy=True)
|
||||
|
||||
def evict(self, num_tokens: int):
|
||||
pass
|
||||
|
||||
Reference in New Issue
Block a user