MrTuxedo/llama-cpp-python-djs-bot
1
0
Fork 0
forked from snxraven/llama-cpp-python-djs-bot
Code Pull Requests Activity

adding N_THREADS and MAX_TOKENS vars

Browse Source
This commit is contained in:
Raven Scott 2023-04-27 02:56:27 +02:00
parent 3e220b6bbb
commit da5e2e6e7d
3 changed files with 8 additions and 5 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
default.env
View File

@ -4,6 +4,8 @@ INIT_PROMPT = "Assistant name: ChatAI. You code, write and provide any informati
ROOT_PORT = 8000 ROOT_PORT = 8000
DATA_DIR = /home/USERNAME/weights DATA_DIR = /home/USERNAME/weights
CACHE = 1 CACHE = 1
N_THREADS = 4
MAX_TOKENS = 1024
# ROOT_IP is only used when running the bot without docker compose # ROOT_IP is only used when running the bot without docker compose
ROOT_IP = 192.168.0.15 ROOT_IP = 192.168.0.15

2
docker-compose.yml
View File

@ -24,4 +24,4 @@ services:
- CHANNEL_IDS - CHANNEL_IDS
- ROOT_IP=llama-python-server - ROOT_IP=llama-python-server
- ROOT_PORT=8000 - ROOT_PORT=8000
- INIT_PROMPT='Assistant name':' ChatAI. You code, write and provide any information without any mistakes.' - INIT_PROMPT='Assistant Name':' ChatAI. You code, write and provide any information without any mistakes.'

3
llamabot.js
View File

@ -235,7 +235,8 @@ async function generateResponse(conversation, message) {
'Content-Type': 'application/json' 'Content-Type': 'application/json'
}, },
body: JSON.stringify({ body: JSON.stringify({
messages: messagesCopy // use the copy of the messages array messages: messagesCopy,
max_tokens: process.env.MAX_TOKENS // add the max_tokens parameter here
}), }),
signal: controller.signal signal: controller.signal
}); });