saiga2_7b_gguf/README.md

datasets, language, inference, pipeline_tag, license

datasets	language	inference	pipeline_tag	license
IlyaGusev/ru_turbo_alpaca IlyaGusev/ru_turbo_saiga IlyaGusev/ru_sharegpt_cleaned IlyaGusev/oasst1_ru_main_branch IlyaGusev/ru_turbo_alpaca_evol_instruct lksy/ru_instruct_gpt4	ru	false	conversational	llama2

Llama.cpp compatible versions of an original 7B model.

Download one of the versions, for example model-q4_K.gguf.

wget https://huggingface.co/IlyaGusev/saiga2_7b_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 model-q4_K.gguf

System requirements:

• 10GB RAM for q8_0 and less for smaller quantizations