Tucano: Advancing Neural Text Generation for Portuguese
Paper
•
2411.07854
•
Published
•
7
ViTucano-v1 (Portuguese)
ViTucano is our first attempt at creating a vision assistant natively pretrained in Portuguese. ViTucano is built on top of the Tucano series.
-
TucanoBR/ViTucano-2b8-v1
Image-Text-to-Text • 3B • Updated • 85 • 7Note 2.8 billion-parameter version of the ViTucano (vision-Tucano).
TucanoBR/ViTucano-1b5-v1
Image-Text-to-Text • 2B • Updated • 16 • 3Note 1.5 billion-parameter version of the ViTucano (vision-Tucano).
TucanoBR/ViTucano-Pretrain
Updated • 112 • 1Note ViTucano-Pretrain is a translation of the original liuhaotian/LLaVA-Pretrain, obtained via Google's translation API.
TucanoBR/ViTucano-SFT
Viewer • Updated • 517k • 60 • 2Note ViTucano-SFT is a dataset for visual instruction tuning.