vbatts/llama.cpp
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

tighten timings

Browse source
This commit is contained in:
Concedo 2023-10-25 20:44:20 +08:00
parent c9983a72d6
commit 98d1dba256
1 changed files with 1 additions and 1 deletions
Download patch file Download diff file Expand all files Collapse all files

2
koboldcpp.py
View file

@ -527,7 +527,7 @@ class ServerRequestHandler(http.server.SimpleHTTPRequestHandler):
current_token = 0 current_token = 0
incomplete_token_buffer = bytearray() incomplete_token_buffer = bytearray()
await asyncio.sleep(0.1) #anti race condition, prevent check from overtaking generate await asyncio.sleep(0.05) #anti race condition, prevent check from overtaking generate
while True: while True:
streamDone = handle.has_finished() #exit next loop on done streamDone = handle.has_finished() #exit next loop on done
tokenStr = "" tokenStr = ""