ReShot — copy the shot, not the actors. Video → depth map, OpenPose skeleton or canny lines for Seedance 2.0/2.5 reference video, MiniMax H3 Fun ControlNet and Wan VACE. Apache-2.0. Open-sourced by Maosika 猫斯卡 (www.maosika.com), the AI short-drama production system.
Python ai-short-dramaai-videocannycomfyuicontrolnetdepth-anythingdepth-estimationdwposemaosikaminimaxopenposepose-controlnetpose-estimationseedanceshort-dramavideo-generationvideo-to-videowan
Commits on main
docs: README (en/zh) names the only official channels and states there is no exe/Windows installer
770d722 wangfeng committed 1 month ago
scripts: mirror the three default weights to ModelScope (sha256-verified, token from env); model card documents weights/ and the modelscope download path
36d6d79 wangfeng committed 2 months ago
docs: ModelScope model card (docs/MODELSCOPE.md, mirrored to modelscope.cn/models/maosika/reshot); README nav links to it
066b936 wangfeng committed 2 months ago
landing: HF Space page leads with the skeleton demo; install line, what-it-does, presets and links updated for 0.5
1ba6b14 wangfeng committed 2 months ago
docs: social preview card shows reference → skeleton → three takes; scripts/make_social_preview.py renders it
b8bebfe wangfeng committed 2 months ago
docs: skeleton demo — the three takes re-made on MiniMax H3 from the pose video (demo-pose.gif/mp4), prompts with seeds; replaces the three-panel GIF
8f2a2e8 wangfeng committed 2 months ago
docs: MiniMax H3 verification of skeleton reference videos — three demo takes re-made with pose as <Video 1>, prompts with seeds, comparison image
52987fd wangfeng committed 2 months ago
0.5.0: --control pose (DWPose skeletons, tracked + smoothed, keypoints JSON) and --control canny; web page picks depth / skeleton / lines; --pose-detect-every; docs, Colab, CI, demo assets
8ebc883 wangfeng committed 2 months ago
docs: link the ReShot article on maosika.com (en/zh) from the About sections and the landing page
09dc9cc wangfeng committed 2 months ago
colab: measured time (58 s for the 12 s demo on a T4 at fast)
22e9c6d wangfeng committed 2 months ago
colab: call the reshot console script (PyPI 0.4.0 has no __main__)
210a1cc wangfeng committed 2 months ago
Colab notebook (examples/reshot_colab.ipynb) + python -m reshot; links point at it instead of a paid Space
fcf1159 wangfeng committed 2 months ago
0.4.1: ZeroGPU Space (demo/space), run(backend=) for resident processes, landing page links to it
062c585 wangfeng committed 2 months ago
community: Discussions, issue chooser, depth-quality template, Community section in both READMEs
33b02f3 wangfeng committed 2 months ago
docs: Apple Silicon row measured at the fast default (46 ms/frame on M2 Max), the old 500 ms figure was the full-size run
329d45d wangfeng committed 2 months ago
docs: web page screenshots (corridor demo clip) in README, HF Space page and PyPI text; scripts/shoot_web_ui.py drives headless Chrome over CDP
d60101e wangfeng committed 2 months ago
0.4.0: local web page — reshot with no arguments opens a Studio Air UI (drop clip → target → quality → depth map, side-by-side preview, download, prompt); stdlib only
1a14bbf wangfeng committed 2 months ago
0.3.8: fix MPS low-watermark error on torch >= 2.14; AGENT_INSTALL.md for AI coding agents
9665da5 wangfeng committed 2 months ago
docs: AGENT_INSTALL.md — step-by-step guide for AI coding agents, plus a copy-paste prompt in the READMEs
10a1efb wangfeng committed 2 months ago
0.3.7: on PyPI — pip install reshot / uvx reshot; install instructions updated everywhere
b7cf345 wangfeng committed 2 months ago
0.3.6: --quality fast is the default (model sees 644x364 for 16:9); resolutions written out in full everywhere; run prints the exact model resolution
7f1a472 wangfeng committed 2 months ago
0.3.5: --quality auto|full|fast is the user's choice; measured full-vs-fast difference in the docs
e525be4 wangfeng committed 2 months ago
0.3.4: 8 GB GPUs get input-size 364 automatically (measured on a 3080 Ti); GPU peak in metrics; 'What you need' table
1ddb763 wangfeng committed 2 months ago
metrics: GPU peak (allocated/reserved) per run; hidden --cuda-memory-fraction to emulate a smaller card
6c9b9d2 wangfeng committed 2 months ago
0.3.3: memory estimate fitted on six RTX 4090 runs (2.0 GiB base + 20 B/pixel-frame)
008dc52 wangfeng committed 2 months ago
docs: the reference clip was a Seedance 2.0 render, not MiniMax H3
0842e1d wangfeng committed 2 months ago