Files
TinyLlama_v1.1_math_code-wo…/README.md

5 lines
362 B
Markdown
Raw Normal View History

---
license: mit
---
This model is trained off of Josephgflowers/TinyLlama-v1.1-Agent-Rag-Nerd-v1 which is trained off of the tinyllama v1.1. This model was trained using the tinyllama chat format, zephyr.
I trained this model using a vast amount of textbooks are all you need style synthetic datasets as well as some common chat datasets and the Cinder dataset.