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

fix horde worker host and client agent

Browse source
This commit is contained in:
Concedo 2023-07-25 18:18:41 +08:00
parent 48c27a9ce1
commit 66e4b5141e
1 changed files with 6 additions and 3 deletions
Download patch file Download diff file Expand all files Collapse all files

9
koboldcpp.py
View file

@ -1358,11 +1358,14 @@ def show_old_gui():
def run_horde_worker(args, api_key, worker_name): def run_horde_worker(args, api_key, worker_name):
import urllib.request import urllib.request
global friendlymodelname, maxhordectx, maxhordelen, exitcounter, modelbusy global friendlymodelname, maxhordectx, maxhordelen, exitcounter, modelbusy
epurl = f"http://localhost:{args.port}"
if args.host!="":
epurl = f"http://{args.host}:{args.port}"
def make_url_request(url, data, method='POST'): def make_url_request(url, data, method='POST'):
try: try:
request = None request = None
headers = {"apikey": api_key,'User-Agent':'KoboldCpp Embedded Worker v1'} headers = {"apikey": api_key,'User-Agent':'KoboldCpp Embedded Worker v1','Client-Agent':'KoboldCppEmbedWorker:1'}
if method=='POST': if method=='POST':
json_payload = json.dumps(data).encode('utf-8') json_payload = json.dumps(data).encode('utf-8')
request = urllib.request.Request(url, data=json_payload, headers=headers, method=method) request = urllib.request.Request(url, data=json_payload, headers=headers, method=method)
@ -1394,7 +1397,7 @@ def run_horde_worker(args, api_key, worker_name):
cluster = "https://horde.koboldai.net" cluster = "https://horde.koboldai.net"
while exitcounter < 10: while exitcounter < 10:
time.sleep(2) time.sleep(2)
readygo = make_url_request(f'http://localhost:{args.port}/api/v1/info/version', None,'GET') readygo = make_url_request(f'{epurl}/api/v1/info/version', None,'GET')
if readygo: if readygo:
print("Embedded Horde Worker is started.") print("Embedded Horde Worker is started.")
break break
@ -1439,7 +1442,7 @@ def run_horde_worker(args, api_key, worker_name):
#do gen #do gen
while exitcounter < 10: while exitcounter < 10:
if not modelbusy.locked(): if not modelbusy.locked():
current_generation = make_url_request(f'http://localhost:{args.port}/api/v1/generate', current_payload) current_generation = make_url_request(f'{epurl}/api/v1/generate', current_payload)
if current_generation: if current_generation:
break break
else: else: