File size: 801 Bytes
3e1f7b5 bb60e90 3e1f7b5 bb60e90 3e1f7b5 bb60e90 3e1f7b5 bb60e90 3e1f7b5 bb60e90 3e1f7b5 bb60e90 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 |
---
library_name: transformers
tags:
- latex
- image-to-text
datasets:
- lamm-mit/OleehyO-latex-formulas
- OleehyO/latex-formulas
license: apache-2.0
---
## Model Summary
Cephalo is a series of multimodal materials science focused vision large language models (V-LLMs) designed to integrate visual and linguistic data for advanced understanding and interaction in human-AI or multi-agent AI frameworks.
![image/png](https://cdn-uploads.huggingface.co/production/uploads/623ce1c6b66fedf374859fe7/kl5GWBP9WS0D4uwd1t3S7.png)
## Model Capabilities
This version of Cephalo, lamm-mit/Cephalo-LaTeX-Phi-3-vision-128k-4b-beta, is trained to convert images of equations to LaTeX code. This version is trained on a larger dataset and for more epochs than lamm-mit/Cephalo-LaTeX-Phi-3-vision-128k-4b-alpha. |