REDCODE/NemesisAI
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Added rop of pending updates on bot start, reset command, AnswerChat method, GPU offload, limit to response lenght, context reduced to 2048, flash attention, 4 parallel decode queues, --keep of the original 810 tokens (which is the starting prompt)

Browse Source
This commit is contained in:
Samuele Lorefice
2024-12-26 03:24:56 +01:00
parent 296d150282
commit 4167c75279
5 changed files with 49 additions and 20 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
TelegramBot/.env
View File

@@ -1,4 +1,4 @@
TELEGRAM_BOT_TOKEN=yourTokenHere
OPENAI_BASE_URL=http://llm-server/
OPENAI_MODEL=Qwen2.5-7B-Instruct-Q8_0.gguf
OPENAI_MODEL=Qwen2.5-7B-Instruct-Q8.gguf
OPENAI_API_KEY=MyApiKey