A high-throughput and memory-efficient inference and serving engine for LLMs
Mirrored from https://github.com/tduncan-lum/Nvidia-Optimization-Stack---VLLM.git
Synced Nov 15, 11:00 UTC
[v1] Rename specialized_manager.py to single_type_kv_cache_manager.py (#17946)
Signed-off-by: Chen Zhang <zhangch99@outlook.com>
Chen Zhang committed
May 11, 2025 at 07:14 UTC
ca66a1674c253b310e1aef94c9766fcf506db187
3 files changed
+3
-2
tests/v1/core/test_specialized_manager.py
+1
-1
@@ -4,7 +4,7 @@ import torch
4
5
from vllm.v1.core.block_pool import BlockPool
6
from vllm.v1.core.kv_cache_utils import BlockHashType, KVCacheBlock
7
-from vllm.v1.core.specialized_manager import SlidingWindowManager
7
+from vllm.v1.core.single_type_kv_cache_manager import SlidingWindowManager
8
from vllm.v1.kv_cache_interface import SlidingWindowSpec
9
10
vllm/v1/core/kv_cache_manager.py
+2
-1
@@ -10,7 +10,8 @@ from vllm.utils import sha256
10
from vllm.v1.core.block_pool import BlockPool
11
from vllm.v1.core.kv_cache_utils import (BlockHashType, KVCacheBlock,
12
hash_request_tokens)
13
-from vllm.v1.core.specialized_manager import get_manager_for_kv_cache_spec
13
+from vllm.v1.core.single_type_kv_cache_manager import (
14
+ get_manager_for_kv_cache_spec)
15
from vllm.v1.kv_cache_interface import KVCacheConfig
16
from vllm.v1.metrics.stats import PrefixCacheStats
17
from vllm.v1.request import Request, RequestStatus