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

Fix the chat template for llava-v1.6-34b & format code (#177)

Browse Source
This commit is contained in:
Lianmin Zheng
2024-02-11 05:50:13 -08:00
committed by GitHub
parent 50afed4eaa
commit c51020cf0c
23 changed files with 101 additions and 44 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
python/sglang/utils.py
View File

@@ -22,7 +22,7 @@ def get_available_gpu_memory(gpu_id, distributed=True):
if torch.cuda.current_device() != gpu_id:
print(
f"WARN: current device is not {gpu_id}, but {torch.cuda.current_device()}, ",
f"WARNING: current device is not {gpu_id}, but {torch.cuda.current_device()}, ",
"which may cause useless memory allocation for torch CUDA context.",
)
@@ -95,7 +95,7 @@ def http_request(url, json=None, stream=False, auth_token=None):
return requests.post(url, json=json, stream=True)
headers = {
"Content-Type": "application/json",
"Authentication": f"Bearer {auth_token}"
"Authentication": f"Bearer {auth_token}",
}
return requests.post(url, json=json, stream=True, headers=headers)
else: