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

bump token limits

Browse source
This commit is contained in:
Concedo 2023-09-27 01:26:00 +08:00
parent 8bf6f7f8b0
commit eb86cd4027
2 changed files with 5 additions and 5 deletions
Download patch file Download diff file Expand all files Collapse all files

4
expose.h
View file

@ -1,7 +1,7 @@
#pragma once #pragma once
const int stop_token_max = 10; const int stop_token_max = 16;
const int ban_token_max = 10; const int ban_token_max = 16;
const int tensor_split_max = 16; const int tensor_split_max = 16;
// match kobold's sampler list and order // match kobold's sampler list and order
enum samplers enum samplers

6
koboldcpp.py
View file

@ -11,9 +11,9 @@ import argparse
import json, sys, http.server, time, asyncio, socket, threading import json, sys, http.server, time, asyncio, socket, threading
from concurrent.futures import ThreadPoolExecutor from concurrent.futures import ThreadPoolExecutor
stop_token_max = 10
sampler_order_max = 7 sampler_order_max = 7
ban_token_max = 10 stop_token_max = 16
ban_token_max = 16
tensor_split_max = 16 tensor_split_max = 16
class load_model_inputs(ctypes.Structure): class load_model_inputs(ctypes.Structure):
@ -749,7 +749,7 @@ def RunServerMultiThreaded(addr, port, embedded_kailite = None):
exitcounter = 999 exitcounter = 999
self.httpd.server_close() self.httpd.server_close()
numThreads = 10 numThreads = 12
threadArr = [] threadArr = []
for i in range(numThreads): for i in range(numThreads):
threadArr.append(Thread(i)) threadArr.append(Thread(i))