I'm testing multiple small VLMs-as-judges, all parts of pipeline are different model families
let me know below if you want me to test any other models, these are very convenient
I'm working on a bit of a something, here's a spoiler
I learned so much from the process about VLM labelling & judging currently adding instance segmentation and adding more infra options
Some users find the developer-built orchestrated VLM pipeline for dataset labeling intriguing and enjoyable, offering help with their own GPU and credits.
I have a GPU at home, Hugging Face credits to spend on Jobs and Inference Providers, help this girl with what you'd like to tweak
enjoying my time a lot lol
I'm testing multiple small VLMs-as-judges, all parts of pipeline are different model families
let me know below if you want me to test any other models, these are very convenient
I think I'll try Molmo somewhere here and moondream3 for a more general problem
I have a GPU at home, Hugging Face credits to spend on Jobs and Inference Providers, help this girl with what you'd like to tweak
enjoying my time a lot lol
@mervenoyann 12B as a judge is intriguing
@mervenoyann Maybe you could help out with pushing out smaller version of the whale. https://huggingface.co/kshitijthakkar/deepseek-v4-mini-1B-init/discussions/1
@mervenoyann seeking smallest but quite capable models that can run on a 3090 well