yumoqing/llmage
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

bugfix

Browse Source
This commit is contained in:
yumoqing 2026-02-11 17:24:58 +08:00
parent e88b4f61c6
commit 86030862f0
2 changed files with 16 additions and 7 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
llmage/init.py
View File

@ -6,6 +6,7 @@ from llmage.keling import keling_token
from llmage.llmclient import (
b64media2url,
get_llm,
inference_generator,
inference,
get_llmproviders,
get_llms_sort_by_provider,
@ -21,6 +22,7 @@ def load_llmage():
env.b64media2url = b64media2url
env.hex2base64 = hex2base64
env.inference = inference
env.inference_generator = inference_generator
env.get_llms_by_catelog = get_llms_by_catelog
env.get_llmcatelogs = get_llmcatelogs
env.checkCustomerBalance = checkCustomerBalance

15
llmage/llmclient.py
View File

@ -380,7 +380,7 @@ def b64media2url(request, mediafile):
url = entire_url('/idfile?path=') + env.quote(mediafile)
return url
async def inference(request, *args, params_kw=None, **kw):
async def inference_generator(request, *args, params_kw=None, **kw):
env = request._run_ns.copy()
if not params_kw:
params_kw = env.params_kw
@ -390,6 +390,7 @@ async def inference(request, *args, params_kw=None, **kw):
dbname = env.get_module_dbname('llmage')
db = env.DBPools()
async with db.sqlorContext(dbname) as sor:
f == None
llm = await get_llm(llmid)
if not params_kw.model:
params_kw.model = llm.model
@ -397,11 +398,17 @@ async def inference(request, *args, params_kw=None, **kw):
llm.stream = 'sync'
if llm.stream == 'async':
f = partial(async_uapi_request, request, llm, sor, params_kw=params_kw)
return await env.stream_response(request, f)
if llm.stream == 'sync':
elif llm.stream == 'sync':
f = partial(sync_uapi_request, request, llm, sor, params_kw=params_kw)
return await env.stream_response(request, f)
# env.update(llm)
else:
uapi = UAPI(request, sor=sor)
f = partial(uapi_request, request, llm, sor, params_kw=params_kw)
async for d in f():
yield d
async def inference(request, *args, params_kw=None, **kw):
env = request._run_ns.copy()
f = partial(inference_generator, *args, params_kw=params_kw, **kw)
return await env.stream_response(request, f)