Example Fine-Tuning Script

#6
by onurulu17 - opened

Hi OmLab team,

Thank you for sharing the omlab/omdet-turbo-swin-tiny-hf model — it's an exciting contribution to the open-vocabulary object detection space.

I'm currently working on fine-tuning this model on a domain-specific dataset using attribute-rich captions. I’d like to train model using LoRA for efficient adaptation. However, since this model has a custom architecture, it would be very helpful to have an official or minimal working example of a fine-tuning script.

Could you please provide an example or notebook that demonstrates how to fine-tune this model, particularly with:

A custom dataset in COCO-style format (or your preferred format),

LoRA integration (e.g., via peft),

Focused fine-tuning on model,

Evaluation or inference pipeline after training.

Such a script would be a great starting point for the community and help others quickly adapt the model to different domains.

Thanks in advance for your support!

Best regards,
Computer Engineer Onur ULU

Sign up or log in to comment