Public/koboldcpp
1
0
Fork 0
mirror of https://github.com/LostRuins/koboldcpp.git synced 2026-09-17 08:19:49 +02:00
Code Issues Packages Projects Releases Wiki Activity

Merge branch 'concedo_experimental' of https://github.com/LostRuins/llamacpp-for-kobold into concedo_experimental

Browse Source
This commit is contained in:
Concedo
2024-01-31 14:43:03 +08:00
parent 13dcf4b556 e00e17b3f9
commit 6bb4be0e7f
1 changed files with 7 additions and 3 deletions
Download Patch File Download Diff File Expand all files Collapse all files
koboldcpp.py
+7 -3
View File
@@ -2433,7 +2433,9 @@ def main(launch_args,start_server=True):
modelname = os.path.abspath(args.model_param)
print(args)
print(f"==========\nLoading model: {modelname} \n[Threads: {args.threads}, BlasThreads: {args.blasthreads}, SmartContext: {args.smartcontext}, ContextShift: {not (args.noshift)}]")
# Flush stdout for win32 issue with regards to piping in terminals,
# especially before handing over to C++ context.
print(f"==========\nLoading model: {modelname} \n[Threads: {args.threads}, BlasThreads: {args.blasthreads}, SmartContext: {args.smartcontext}, ContextShift: {not (args.noshift)}]", flush=True)
loadok = load_model(modelname)
print("Load Model OK: " + str(loadok))
@@ -2507,10 +2509,12 @@ def main(launch_args,start_server=True):
if start_server:
if args.remotetunnel:
setuptunnel()
print(f"======\nPlease connect to custom endpoint at {epurl}")
# Flush stdout for previous win32 issue so the client can see output.
print(f"======\nPlease connect to custom endpoint at {epurl}", flush=True)
asyncio.run(RunServerMultiThreaded(args.host, args.port, embedded_kailite, embedded_kcpp_docs))
else:
print(f"Server was not started, main function complete. Idling.")
# Flush stdout for previous win32 issue so the client can see output.
print(f"Server was not started, main function complete. Idling.", flush=True)
def run_in_queue(launch_args, input_queue, output_queue):
main(launch_args, start_server=False)