EngineX-Hygon/sglang
5
0
Fork 0
Code Issues Pull Requests Actions 7 Projects Releases Wiki Activity

[CI] Fix the problem of hf runner too slow (#1202)

Browse Source
This commit is contained in:
Ying Sheng
2024-08-24 18:35:55 -07:00
committed by GitHub
parent b20daf982a
commit e61d13acdf
1 changed files with 0 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
python/sglang/test/runners.py
View File

@@ -86,7 +86,6 @@ class HFRunner:
self.tokenizer = AutoTokenizer.from_pretrained( self.tokenizer = AutoTokenizer.from_pretrained(
model_path, model_path,
torch_dtype=torch_dtype, torch_dtype=torch_dtype,
trust_remote_code=True,
) )
self.is_generation_model = is_generation_model self.is_generation_model = is_generation_model
@@ -96,7 +95,6 @@ class HFRunner:
model_path, model_path,
torch_dtype=torch_dtype, torch_dtype=torch_dtype,
low_cpu_mem_usage=True, low_cpu_mem_usage=True,
trust_remote_code=True,
).cuda() ).cuda()
else: else:
from sentence_transformers import SentenceTransformer from sentence_transformers import SentenceTransformer