---
license: mit
---
The model is trained based on LLAMA 2. Please adhere to license. 
We use a subset of TULU v2 instruction tuning corpus to train the model. 
We use a subset of TULU v2 instruction tuning corpus to train the model.