Repository navigation

All pull requests

Labels (14)Milestones (0)New pull request

Open

(66)

Closed

(103)

Fix per-object memory clearing during video corrections
#777Afloat16 opened Oct 15, 2026
Fix mask-only prompt batches in SAM2ImagePredictor
#776Afloat16 opened Oct 9, 2026
Add optional CUDA Graph fast path for the image encoder
#775trueoneplusone opened Sep 19, 2026
docs: add Simplified Chinese translation (README.zh.md)
#774JasonYeYuhe opened Sep 16, 20264 tasks done
1
Support filesystem paths for config_file in build_sam2 / build_sam2_video_predictor
#771Sanjays2402 opened Sep 10, 2026
Perf/reduce redundant inference work
#770SimBuddy opened Sep 5, 2026
2
Fix empty crop_boxes shape in automatic mask generation
#766Dev-Aayush-Kumar opened Aug 22, 2026
SAM2
#764sketchyz88 opened Jul 5, 2026
1
fix: add missing @torch.no_grad() to predict, predict_batch, get_imag…
#763Julianblock opened Jul 3, 2026
fix: remove stale image_format param from set_image docstring
#761Julianblock opened Jul 3, 2026
Bump postcss from 8.4.33 to 8.5.10 in /demo/frontend
#755dependabot[bot] opened May 30, 2026Bot
Bump tmp from 0.2.3 to 0.2.6 in /demo/frontend
#754dependabot[bot] opened May 27, 2026Bot
Bump vite from 5.0.11 to 6.4.2 in /demo/frontend
#752dependabot[bot] opened Apr 7, 2026Bot
Bump picomatch from 2.3.1 to 2.3.2 in /demo/frontend
#750dependabot[bot] opened Apr 7, 2026Bot
Make Hiera input channels and video normalization configurable
#749AryanMoradi opened Apr 7, 2026
3
Bump flatted from 3.2.9 to 3.4.2 in /demo/frontend
#745dependabot[bot] opened Mar 20, 2026Bot
Bump immutable from 4.3.4 to 4.3.8 in /demo/frontend
#744dependabot[bot] opened Mar 5, 2026Bot
Bump rollup from 4.9.5 to 4.59.0 in /demo/frontend
#743dependabot[bot] opened Mar 3, 2026Bot
marge sam2 with old version
#730chen-haijin opened Dec 3, 2025
2
Fix PyAV compatibility issue with DISPLAYMATRIX access
#718paruparu opened Oct 27, 2025
Add section to save segmented frames as video in video_predictor_example.ipynb notebook
#714siromermer opened Oct 9, 2025
1
fix: connected_components.cu has CUDAGuard
#705mheyman opened Aug 29, 2025
2
Add Multi-GPU Parallel Support to Connected Components CUDA Extension
#679moqingyan opened Jul 3, 2025
3
Fix rm ffmpeg not found in docker
#676tornikeo opened Jun 22, 2025
1
Make image predictor thread-safe
#643adamp87 opened May 6, 2025

Navigation Menu

Repository navigation

All pull requests

Fix per-object memory clearing during video corrections

Fix mask-only prompt batches in SAM2ImagePredictor

Add optional CUDA Graph fast path for the image encoder

docs: add Simplified Chinese translation (README.zh.md)

Support filesystem paths for `config_file` in `build_sam2` / `build_sam2_video_predictor`

Perf/reduce redundant inference work

Fix empty crop_boxes shape in automatic mask generation

SAM2

fix: add missing @torch.no_grad() to predict, predict_batch, get_imag…

fix: remove stale image_format param from set_image docstring

Bump postcss from 8.4.33 to 8.5.10 in /demo/frontend

Bump tmp from 0.2.3 to 0.2.6 in /demo/frontend

Bump vite from 5.0.11 to 6.4.2 in /demo/frontend

Bump picomatch from 2.3.1 to 2.3.2 in /demo/frontend

Make Hiera input channels and video normalization configurable

Bump flatted from 3.2.9 to 3.4.2 in /demo/frontend

Bump immutable from 4.3.4 to 4.3.8 in /demo/frontend

Bump rollup from 4.9.5 to 4.59.0 in /demo/frontend

marge sam2 with old version

Fix PyAV compatibility issue with DISPLAYMATRIX access

Add section to save segmented frames as video in video_predictor_example.ipynb notebook

fix: connected_components.cu has CUDAGuard

Add Multi-GPU Parallel Support to Connected Components CUDA Extension

Fix rm ffmpeg not found in docker

Make image predictor thread-safe

Pull requests