top | item 46046475

(no title)

enoch2090 | 3 months ago

Surprisingly, SAM3 works bad on engineering drawings while SAM2 kinda works, and VLMs like Qwen3-VL works as well

discuss

order

zubiaur|3 months ago

Had good luck with Gemini 2.5, SAM3 failed miserably with PIDs.

retinaros|3 months ago

yeah I tried too. Im trying a fine tuning on PIDs.

enoch2090|3 months ago

Looking forward to your progress! Just checked the paper and it says the underlying backbone is still DETR. My guess would be that SAM3 uses more video frames during the training process and caused the dilution of sparse engineering-paper-like data.