all
/
g
/
s
/
fur
/
art
/
ai
/
3dcg
/
f
/
rp
/
req
/
p2p
/
dis
/
lit
/
kaki
chat
GUROchan!
This list is WIP...
1
2
3
So use the classic menu
Link 4
Link 5
Link 6
Link 7
List 3
Link 8
Link 9
[
Settings
]
[
Home
] [
Catalog
] [
Search
] [
Thread list
] [
Stats
] [
Reports
] [
Watcher
] [
PMs
] [
Admin
]
Artificial Intelligence
[
Return
]
Report a post
Preview
File:
vid.webm
(1.84 MB, 400x720)
Hide image
Hide image
▶
WAN LightX2V
Anonymous
2025/06/25
(Wed)
11:01:29
No.
9539
[
Report
]
▶
There's now new WAN-based model also available as lora which can achieve good looking results even at 4 steps (and decent preview at 2) and so 5s of video can be done under half of a minute on 3090.
model:
https://huggingface.co/lym00/Wan2.1_T2V_14B_LightX2V_StepCfgDistill_VACE-GGUF
lora:
https://huggingface.co/Kijai/WanVideo_comfy/blob/main/Wan21_T2V_14B_lightx2v_cfg_step_distill_lora_rank32.safetensors
I believe that attached example should be non-realistic enough. But honestly, beside the obvious, I don't know what to do next with her...
Post number
No.
9539
Board
Artificial Intelligence
Reason
Optional. Describe what's wrong with it.
Style: