Blockly?
#1
by
Bit101
- opened
have that bug, to fix try to disable gpu offload
it fixed in future llama versions
read here about
https://huggingface.co/Qwen/Qwen2-7B-Instruct-GGUF/discussions/1
If that's the case it should be fixed in the future KCPP versions.
Or use this fork:
https://github.com/Nexesenex/kobold.cpp
This should have the latest commits merged already.