Llama.cpp compatible versions of an original 13B model.
Download one of the versions, for example model-q4_K.gguf.
wget https://huggingface.co/IlyaGusev/saiga2_13b_gguf/resolve/main/model-q4_K.gguf
Download interact_llamacpp.py
wget https://raw.githubusercontent.com/IlyaGusev/rulm/master/self_instruct/src/interact_llamacpp.py
How to run:
pip install llama-cpp-python fire
python3 interact_llamacpp.py ggml-model-q4_K.gguf
System requirements:
- 18GB RAM for q8_K
- 10GB RAM for q4_K
- Downloads last month
- 0