xinyun/vllm
1
0
Fork 0
mirror of https://git.datalinker.icu/vllm-project/vllm.git synced 2026-05-06 13:24:35 +08:00
Code Issues Packages Projects Releases Wiki Activity

minor

Browse Source 
Signed-off-by: Woosuk Kwon <woosuk@thinkingmachines.ai>
This commit is contained in:
Woosuk Kwon 2025-09-15 19:53:58 +00:00
parent 69b17891a3
commit f1981db101
1 changed files with 3 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

3
vllm/v1/worker/gpu/model_runner.py
View File

@ -48,6 +48,7 @@ class GPUModelRunner:
if self.cache_config.cache_dtype == "auto": if self.cache_config.cache_dtype == "auto":
self.kv_cache_dtype = self.dtype self.kv_cache_dtype = self.dtype
else: else:
# Quantized KV cache.
self.kv_cache_dtype = STR_DTYPE_TO_TORCH_DTYPE[ self.kv_cache_dtype = STR_DTYPE_TO_TORCH_DTYPE[
self.cache_config.cache_dtype] self.cache_config.cache_dtype]
@ -217,6 +218,8 @@ class GPUModelRunner:
self.kv_cache_config.kv_cache_groups): self.kv_cache_config.kv_cache_groups):
block_table = block_tables[i] block_table = block_tables[i]
slot_mapping = slot_mappings[i] slot_mapping = slot_mappings[i]
num_common_prefix_blocks = 0
return InputBatch( return InputBatch(
req_ids=req_ids, req_ids=req_ids,