初始化项目，由ModelHub XC社区提供模型

Model: IlyaGusev/saiga2_7b_gguf
Source: Original Platform

README.md

@@ -0,0 +1,36 @@
+---
+datasets:
+- 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
+language:
+- ru
+inference: false
+pipeline_tag: conversational
+license: llama2
+---
+
+Llama.cpp compatible versions of an original [7B model](https://huggingface.co/IlyaGusev/saiga2_7b_lora).
+
+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](https://raw.githubusercontent.com/IlyaGusev/rulm/master/self_instruct/src/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