类库 › kandinsky-5
kandinskylab

kandinskylab/kandinsky-5

Kandinsky 5.0 是一系列用于视频和图像生成的扩散模型。它能够根据文本提示和/或输入图像来生成相应的视频或图像,提供了Pro和Lite等不同版本。

kandinskylab/kandinsky-5

技术栈

测试

pytest

网络

Requests
查看全部依赖 (26)

依赖

NumPy Pydantic accelerate aiohttp anyio av beautifulsoup4 bitsandbytes diffusers einops imageio imageio-ffmpeg ipykernel ipywidgets ninja omegaconf opencv-python packaging peft pillow sentencepiece torch torchvision tqdm transformers websocket-client

截图

https://user-images.githubusercontent.com/25423296/163456779-a8556205-d0a5-45e2-ac17-42d089e3c3f8.png
assets/sbs/kandinsky_5_video_lite_vs_sora.jpg
assets/sbs/kandinsky_5_video_lite_vs_wan_2.1_14B.jpg
assets/sbs/kandinsky_5_video_lite_vs_wan_2.2_5B.jpg
assets/sbs/kandinsky_5_video_lite_vs_wan_2.2_A14B.jpg
assets/sbs/kandinsky_5_video_lite_vs_wan_2.1_1.3B.jpg
assets/sbs/kandinsky_5_video_lite_10s_vs_kandinsky_5_video_lite_distill_10s.jpg

评论

Home - Wiki
Copyright © 2011-2026 iteam. Current version is 2.155.2. UTC+08:00, 2026-05-17 06:52
浙ICP备14020137号-1 $Map of visitor$