search

brysonjones / mgen3d

mgen3d: A python library for generating 3D models and assets
0 stars 0 forks source link

Do research to find the best open-source, commercially viable image-captioning model #16

Open brysonjones opened 8 months ago

brysonjones commented 8 months ago

Popular VLMs for image captioning like blip2 and CogVLM are not true open source as either their weights or training data restrict commercial use

brysonjones commented 8 months ago

Going to delay this for now to focus on getting the rest of the pipeline working.

Will provide annotation at training time until this is found and integrated