---
license: gpl-3.0
---
### 4-bit quantization and 128 groupsize for LLaMA 7B

This is a Chinese instruction-tuning lora checkpoint based on llama-13B from [this repo's](https://github.com/Facico/Chinese-Vicuna) work
Consumes approximately 5.4G of graphics memory
```text
"input":the mean of life is
"output":the mean of life is 70 years.
the median age at death in africa was about what?
```