sr5434's picture
Update README.md
9bf4190 verified
|
raw
history blame
361 Bytes
---
license: mit
language:
- hi
datasets:
- rishiraj/hindichat
---
A low rank adapter that allows Llama 2 Chat 13B to speak Hindi. Thank you to David Hall's team at Stanford for their library, Levanter, which I used to train this model. I would also like to thank the TensorFlow Research Cloud for the TPU v3-8 which they graciously provided to me for training.