Search results

parse_action_history_for_propagation()感觉这个函数中的传播类型判定不太对

Status: Open.

#394 In facebookresearch/sam3;

· come105 opened on Jan 5, 2026

High confidence for negative text prompt ("person without safety helmet") even when all persons are wearing helmets

Status: Open.

#393 In facebookresearch/sam3;

· likebean opened on Jan 3, 2026

HELP for Reproduce of the GMOT40 results

Status: Open.

#392 In facebookresearch/sam3;

· JasonShen-SH opened on Jan 3, 2026

Can SAM3 be deployed using Triton Inference Server? How should the configuration file be written?

Status: Open.

#391 In facebookresearch/sam3;

· myself-wss opened on Jan 2, 2026

Are there any plans to release tiny or base variant of Sam 3?

Status: Open.

#389 In facebookresearch/sam3;

· Anshull-py opened on Jan 1, 2026

The TensorRT result of ViT PatchEmbed module is incorrect

Status: Open.

#388 In facebookresearch/sam3;

· icandoits opened on Dec 31, 2025

How to fine-tune SAM 3 with SAM3InteractiveImagePredictor (for SAM 1 task)?

Status: Open.

#386 In facebookresearch/sam3;

· MaxTeselkin opened on Dec 29, 2025

Training on a aleady trained weight

Status: Open.

#384 In facebookresearch/sam3;

· aniket-professional2025 opened on Dec 27, 2025

Questions about SAM3 efficiency and improvements over SAM1/SAM2

Status: Open.

#382 In facebookresearch/sam3;

· broucemonkey-stack opened on Dec 26, 2025

Produce incomplete mask when an object split into 2 parts over time

Status: Open.

#381 In facebookresearch/sam3;

· hengrui0516 opened on Dec 26, 2025

why is the metric bbox_AP used

Status: Open.

#380 In facebookresearch/sam3;

· Troy0630 opened on Dec 26, 2025

Example script/demo for Video PCS using Image Exemplars

Status: Open.

#379 In facebookresearch/sam3;

· Daria8976 opened on Dec 25, 2025