biondizzle/vllm
1
0
Fork 0
Code Issues Pull Requests Actions 2 Packages Projects Releases Wiki Activity

[Core][Model runner refactoring 1/N] Refactor attn metadata term (#4518)

Browse Source
This commit is contained in:
SangBin Cho
2024-05-04 02:20:12 +09:00
committed by GitHub
parent 2d7bce9cd5
commit 3521ba4f25
27 changed files with 554 additions and 525 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
tests/spec_decode/e2e/conftest.py
View File

@@ -45,7 +45,7 @@ class AsyncLLM:
gpu_memory_utilization: float = 0.9,
swap_space: int = 4,
enforce_eager: bool = False,
max_context_len_to_capture: int = 8192,
max_seq_len_to_capture: int = 8192,
disable_custom_all_reduce: bool = False,
**kwargs,
) -> None:
@@ -66,7 +66,7 @@ class AsyncLLM:
gpu_memory_utilization=gpu_memory_utilization,
swap_space=swap_space,
enforce_eager=enforce_eager,
max_context_len_to_capture=max_context_len_to_capture,
max_seq_len_to_capture=max_seq_len_to_capture,
engine_use_ray=True,
disable_custom_all_reduce=disable_custom_all_reduce,
**kwargs,