Spaces:
Paused
Paused
Apply for community grant: Personal project (gpu and storage)
#1
by
aletrn
- opened
I'm trying to port on HuggingFace the LISA project:
https://github.com/dvlab-research/LISA
This project is an attempt to perform "reasoning" image segmentation, combining both LLAVA (https://github.com/haotian-liu/LLaVA) and Segment Anything models.
Right now the demo in this space works but I plant to add python tests and refactor the project to help me migrate from pytorch to GGML.
Can you help me? Right now the demo works loading the LISA model in 4bit and fp16 bit precision.
Thank you!
Best regards,
Alessandro
I make this space public again, please don't restart the project, I'm trying to minimize my bill.