jbilcke-hf HF staff commited on
Commit
4af8a5a
·
1 Parent(s): ada2e97

fix README

Browse files
Files changed (1) hide show
  1. README.md +5 -3
README.md CHANGED
@@ -61,12 +61,14 @@ In practice, a PR (pull request) will be necessary to adapt the UI a bit to acco
61
 
62
  ### LTX-Video
63
 
64
- I have tested training a LoRA model using videos, on a single A100 instance.
 
65
 
66
  ### HunyuanVideo
67
 
68
- I haven't tested it yet, but in theory it should work out of the box.
69
- Please keep in mind that this requires a lot of processing mower.
 
70
 
71
  ### CogVideoX
72
 
 
61
 
62
  ### LTX-Video
63
 
64
+ I have tested training a LTX-Video LoRA model using videos (not images), on a single A100 instance.
65
+ It requires about 18/19 Gb of VRAM, depending on your settings.
66
 
67
  ### HunyuanVideo
68
 
69
+ I have tested training a HunyuanVideo LoRA model using videos (not images),, on a single A100 instance.
70
+
71
+ It requires about 47~49 Gb of VRAM, depending on your settings.
72
 
73
  ### CogVideoX
74