Spaces:

---

SAVSNET

/

README

like 0

Running

Dear SAVSNET Team,

I hope you're doing well. I’m currently building an AI assistant for pet owners that helps detect early symptoms of illness based on syndromic classification. I've been exploring your excellent models on Hugging Face and wanted to ask for some clarification regarding:

What is the key difference between the PetHarbor and PetBERT models in terms of intended use (e.g., token classification, masked LM, syndrome classification)?

Which model is currently best suited for multi-label syndrome classification across the full set of 20 syndromic labels (e.g., including Dental, Hepatic, Cardiovascular, etc.)?

Is there a publicly available fine-tuned version of either model (or another you recommend) that outputs 20-class multi-label classification, suitable for real-world symptom prediction from short owner-reported text?

I’ve successfully deployed PetHarbor in a FastAPI wrapper, but it only appears to return 11 prediction scores, which makes me wonder if I might be using the wrong architecture or model head. I'd be very grateful if you could help clarify the best model setup and use case here.

I also attached a result from model.

Thank you so much for your amazing research and resources. Looking forward to your guidance!

Best regards,
Saad