Pre-release quality checks

[chmjkb]

This release introduces a lot of new, user-facing changes. Therefore, we should perform some manual-quality checks. For each change one should:

• Read the documentation and verify if it's correct and clear
• Build basic applications with a given feature
• Try to break it!

Ideally, the change you're testing should NOT be the one you were developing.

List of changes to test:

• Instance Segmentation (Instance Segmentation API #825) — new API, needs end-to-end testing @msluszniak
• Selfie Segmentation (Selfie segmentation #720) — new model, test segmentation output quality @barhanc
• RF-DETR object detection (Add RF-Detr support #797) — new model via the generic OD interface @barhanc
• LFM2.5-VL-1.6B VLM (Add support for LFM2.5-VL-1.6B #813) — vision language model, test image+prompt inputs @barhanc
• LFM2.5-1.2B-Instruct LLM (Add support for LiquidAI/LFM2.5-1.2B-Instruct #812) — text-only liquid model - @NorbertKlockiewicz
• Semantic segmentation (torchvision models) (Export quicker CV models for torchvision #799) @msluszniak
• Bare RN (feat!: add modular resource fetcher adapters for Expo and bare React Native #759) - @chmjkb
• Factory flow (refactor!: add static factories for remaining models #937) - @NorbertKlockiewicz

[msluszniak]

🚀