feat(vision): 7-node suite — RF-DETR, Mask2Former, DA-V2, BlazeFace + more by ryan-t-christensen · Pull Request #1081 · rocketride-org/rocketride-server

Commits on Jun 5, 2026

feat(vision): add vision node suite and detection models

authored and
asclearuc
committed
feat(vision): add dense_resize image util for memory-bounded inference

authored and
asclearuc
committed
feat(vision): add detection model loaders (RF-DETR, MM-G-DINO, Mask2Former)

authored and
asclearuc
committed
refactor(detect): swap YOLO-World for RF-DETR + MM-Grounding-DINO

authored and
asclearuc
committed
refactor(detect_segment): swap SAM3 for Mask2Former instance + semantic

authored and
asclearuc
committed
refactor(depth_estimate): lock to Depth-Anything V2 Small, image-only

authored and
asclearuc
committed
refactor(vision): specialize describe into caption-only Florence-2 node

authored and
asclearuc
committed
feat(vision): add face_detection node (MediaPipe BlazeFace)

authored and
asclearuc
committed
feat(vision): add pose_estimation node (RTMPose via rtmlib)

authored and
asclearuc
committed
feat(vision): add background_removal node (BiRefNet)

authored and
asclearuc
committed
fix(ui): textarea helperText + sticky outlined label

authored and
asclearuc
committed
chore(deps): pin transformers==4.53.3 for vision suite compatibility

authored and
asclearuc
committed
chore(video_composer): add missing icon

authored and
asclearuc
committed
fix(vision): address CodeRabbit review on PR #1081

authored and
asclearuc
committed
refactor(face_detection): remove non-functional biometric-chain guard

authored and
asclearuc
committed
refactor(face_detection): drop unused torch device-probe; add real face test
asclearuc
committed
feat(depth_estimate): serve via model server
asclearuc
committed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

feat(vision): 7-node suite — RF-DETR, Mask2Former, DA-V2, BlazeFace + more#1081

feat(vision): 7-node suite — RF-DETR, Mask2Former, DA-V2, BlazeFace + more#1081
ryan-t-christensen wants to merge 29 commits into
developfrom
feat/vision

Commits on Jun 5, 2026

feat(vision): add vision node suite and detection models

feat(vision): add dense_resize image util for memory-bounded inference

feat(vision): add detection model loaders (RF-DETR, MM-G-DINO, Mask2Former)

refactor(detect): swap YOLO-World for RF-DETR + MM-Grounding-DINO

refactor(detect_segment): swap SAM3 for Mask2Former instance + semantic

refactor(depth_estimate): lock to Depth-Anything V2 Small, image-only

refactor(vision): specialize describe into caption-only Florence-2 node

feat(vision): add face_detection node (MediaPipe BlazeFace)

feat(vision): add pose_estimation node (RTMPose via rtmlib)

feat(vision): add background_removal node (BiRefNet)

fix(ui): textarea helperText + sticky outlined label

chore(deps): pin transformers==4.53.3 for vision suite compatibility

chore(video_composer): add missing icon

fix(vision): address CodeRabbit review on PR #1081

refactor(face_detection): remove non-functional biometric-chain guard

refactor(face_detection): drop unused torch device-probe; add real face test

feat(depth_estimate): serve via model server

Commits on Jun 7, 2026

@

feat(vision): serve depth/detect/detect_segment via model server

Commits on Jun 9, 2026

feat(vision): serve caption + background_removal via model server

refactor(vision): harden converted nodes + complete fulltest coverage

Commits on Jun 10, 2026

feat(pose_estimation): serve via model server; fix caption GPU dtype

Merge branch 'develop' into feat/vision

fix(depends): use engine_cache_dir for excludes.txt path

fix(face_detection): prevent engine abort, handle missing libGLESv2

Commits on Jun 11, 2026

fix(face_detection): require libEGL.so.1; skip mediapipe in contract-check

fix(nodes): address CodeRabbit review findings

Commits on Jun 12, 2026

fix(nodes): address CodeRabbit review findings

fix(nodes): address CodeRabbit review findings