Spaces:

rinong
/

StyleGAN-NADA

Runtime error

App Files Files Community

StyleGAN-NADA

Commit History

May 18, 2022

16:06
Increased default value of StyleCLIP step

rinong

ac190d6
13:55
Removed default from checkboxgroup

rinong

040da24
13:48
Set default style, UI fix

rinong

ec0bccf
13:41
UI update

rinong

f515965
13:36
UI update

rinong

04b671c
13:19
Added more models

rinong

b92def5
12:18
Documentation fixes.

rinong

4fdf50b
06:39
Fixed misnamed model

rinong

98ef928
06:36
Added missing model, re-arranged options

rinong

d48a8f6
06:30
Added some credits, cleaned up comments

rinong

4f361bc
06:27
Added a bunch of models

rinong

cb4b687

May 17, 2022

17:14
Increased StyleCLIP video length. Added more descriptions.

rinong

951f6cb
16:39
Squeeze un-edited latents

rinong

cf27c3e
16:35
Unifited latent np-torch conversion

rinong

16c6665
16:18
Added styleclip support to video generation

rinong

bbbf71e
15:56
StyleSpace list now uses correct order.

rinong

f9cb70e
12:55
Added tensor dim expansion for edit directions.

rinong

9a29e97
12:42
Modified s_code generation

rinong

6f52ac4
12:31
Modified s_dict generation

rinong

4663a72
11:53
Fixed typing bugs in styleclip projection

rinong

f2ea589
11:40
Added clip to requirements

rinong

4e5af99
11:35
Added StyleCLIP support

rinong

fcf0449
08:43
Bug fixes

rinong

6104a4e
08:32
Fixed np latent when not editing

rinong

65a078b
08:26
Fixed predict args, minor UI change

rinong

83dc64b
08:09
Overhauled editing UI, output to gallery

rinong

210c702
05:47
Updated to new blocks SDK, testing visibility

rinong

7a331ca

May 16, 2022

19:36
Fixed loop arg to ffmpeg

rinong

a3a7b2b
19:29
Videos now loop.

rinong

9cdf7e4
18:38
Cast paths to str

rinong

abcc9dc
18:29
Layout changes, fixed inversion input

rinong

cbac29e
17:09
Changed blocks format to tabs

rinong

1edbcae
16:58
Updated visibility change callbacks

rinong

96b8eee
16:37
Updated visibility change callbacks

rinong

62fe8a5
16:31
Updated visibility change callbacks

rinong

bb723df
16:16
Updated visibility change callbacks

rinong

a39b130
16:13
Updated visibility change callbacks

rinong

cc5a813
16:08
Changed visibility settings. Removed some models for quicker debugging

rinong

4ec364e
16:02
Changed to blocks API

rinong

8e301b6
08:56
Boundaries now load on cpu

rinong

2771b0d
08:43
Fixed editing file paths

rinong

7abf923
08:05
Enabled queue

rinong

89e8a1e
08:00
Added editing direction boundaries

rinong

2cc5eae
07:56
Models now loaded on appropriate device

rinong

a85cdf5
07:50
Added joker, anime models.

rinong

1408f30
07:01
Fixed missing op import

rinong

a390b6f
06:58
Input image now uses filepath

rinong

558a786
06:53
Cuda uses now check for device

rinong

720df82
06:49
Added explicit device to e4e construction

rinong

ea00796
06:47
Updated to use arbitrary model paths

rinong

5479d05

Commit History

May 18, 2022

Increased default value of StyleCLIP step

Removed default from checkboxgroup

Set default style, UI fix

UI update

UI update

Added more models

Documentation fixes.

Fixed misnamed model

Added missing model, re-arranged options

Added some credits, cleaned up comments

Added a bunch of models

May 17, 2022

Increased StyleCLIP video length. Added more descriptions.

Squeeze un-edited latents

Unifited latent np-torch conversion

Added styleclip support to video generation

StyleSpace list now uses correct order.

Added tensor dim expansion for edit directions.

Modified s_code generation

Modified s_dict generation

Fixed typing bugs in styleclip projection

Added clip to requirements

Added StyleCLIP support

Bug fixes

Fixed np latent when not editing

Fixed predict args, minor UI change

Overhauled editing UI, output to gallery

Updated to new blocks SDK, testing visibility

May 16, 2022

Fixed loop arg to ffmpeg

Videos now loop.

Cast paths to str

Layout changes, fixed inversion input

Changed blocks format to tabs

Updated visibility change callbacks

Updated visibility change callbacks

Updated visibility change callbacks

Updated visibility change callbacks

Updated visibility change callbacks

Changed visibility settings. Removed some models for quicker debugging

Changed to blocks API

Boundaries now load on cpu

Fixed editing file paths

Enabled queue

Added editing direction boundaries

Models now loaded on appropriate device

Added joker, anime models.

Fixed missing op import

Input image now uses filepath

Cuda uses now check for device

Added explicit device to e4e construction

Updated to use arbitrary model paths