ai-models/WanVideo_comfy_fp8_scaled
2
0
Fork 0
mirror of https://www.modelscope.cn/Kijai/WanVideo_comfy_fp8_scaled.git synced 2026-04-02 22:12:53 +08:00
Code Issues Packages Projects Releases Wiki Activity

Upload SCAIL/Wan21-14B-SCAIL-preview_fp8_scaled_mixed.safetensors with ModelScope SDK (batch 1/1)

Browse Source
This commit is contained in:
systemd
2026-02-27 11:24:59 +00:00
parent ffd6279832
commit 1ffd509551
7 changed files with 34 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

13
SkyReelsV3/README.md Normal file
View File

@ -0,0 +1,13 @@
SkyReels V3 models in ComfyUI compatible fp8 scaled format:
Source:
https://huggingface.co/collections/Skywork/skyreels-v3
A2V (TalkingAvatar) is Multitalk variant, works in WanVideoWrapper
R2V (Reference to video) is Phantom variant, should work in any Phantom workflow
V2V is video extension and works same way as Pusa with per token timesteps, so runs in Pusa workflows in WanVideoWrapper
V2V_shot is variant of above that generates new scenes from video, this one still lacks proper workflow but does run same as above.