570b0f5d94
Add batch inference feature to WD Tagger
...
as same as standalone version
2024-05-21 00:46:15 +09:00
8feb29de40
fix #97 ( #98 )
...
and a bug: caption file extension not work on save
2024-05-20 19:07:40 +09:00
2f99cc54d8
Add Z3D threshold to image interrogator function ( #96 )
...
* Update tab_edit_caption_of_selected_image.py
* Update tab_edit_caption_of_selected_image.py
fix interrogate_selected_image() missing threshold_z3d
* Update tab_edit_caption_of_selected_image.py
remove print
2024-05-18 09:53:59 +09:00
6d929765eb
Add missing Z3D threshold ( #95 )
...
Currently attempting to use the Z3D interrogator doesn't work because the threshold value isn't getting passed into the load_dataset function. This fixes that.
Co-authored-by: The Divine Heir <the.divine.heir.kuro@gmail.com>
2024-05-17 14:06:48 +09:00
f726ec3f04
revert model path to system default
...
TRANSFORMERS_CACHE
HUGGINGFACE_HUB_CACHE
2024-05-08 04:47:51 +09:00
4621d76286
move model path
2024-05-08 04:39:03 +09:00
d1a8edae13
blip2 model path
2024-05-08 04:24:33 +09:00
c3252d8325
Merge changes in standalone version ( #93 )
...
* Merge changes in standalone version
- New Taggers and Custom Tagger
- a little bit stable UI
2024-05-08 03:58:58 +09:00
7a2f4c53fb
rename package ( #80 )
2023-06-05 19:12:00 +09:00
e883ae6063
Update block_tag_select.py
...
fix for #77
2023-05-02 20:08:23 +09:00
c1c78f5976
Add max image resolution option ( #70 )
...
* Add max image resolution option
* disable webUI's custom feature when max_res is specified
---------
Co-authored-by: toshiaki1729 <116595002+toshiaki1729@users.noreply.github.com>
2023-04-02 20:15:32 +09:00
2781e4b6b8
Hot fix for #69
2023-04-01 22:41:54 +09:00
a9b30a94ea
Initial support for gradio 3.23.0 ( #64 )
...
This commit will make it incompatible with old webUI.
2023-03-26 18:40:20 +09:00
2b3bd4a724
first aid for #62 again
2023-03-26 03:41:16 +09:00
af4508f462
Revert "first aid for #62"
...
This reverts commit 12e326e8f3
2023-03-26 03:35:10 +09:00
12e326e8f3
first aid for #62
...
disables type hint for List[…]
to avoid error on gradio v3.23.0
2023-03-26 02:31:51 +09:00
b8730ad53b
Feature to replace new-line with comma on loading
2023-03-23 09:55:06 +09:00
1644f630ab
move unload button
2023-03-23 09:07:27 +09:00
518406286d
first aid for #60
2023-03-22 23:24:22 +09:00
cb8f3a72f4
enable unloading dataset with single click ( #59 )
2023-03-20 23:02:23 +09:00
6f8abe148a
fix 57
2023-03-10 23:49:37 +09:00
ca95abb441
oops
2023-03-10 12:32:27 +09:00
b70eb4b18b
just move variables
2023-03-10 12:28:50 +09:00
dfe589a71c
Apply change in caption to all displayed images ( #54 )
2023-03-08 23:12:32 +09:00
43d8d8f65f
implement some features to manage token count ( #52 )
...
Add:
- Count and truncate by tokens amount (#40 )
- Sort tags by token count
Change:
- Move "use raw clip token..." setting to "Settings" tab
- Show tag frequency, length or token count depending on the "Sort by"
2023-03-07 22:23:48 +09:00
09272f2840
implement token count
2023-03-07 00:57:19 +09:00
25e51b2de8
Add type hint for developer (need python>=3.7) ( #49 )
...
Change nothing in behavior at all
2023-03-06 11:15:59 +09:00
2406fc8139
move from wrong pos
2023-03-05 21:15:41 +09:00
d94664e5ae
version 0.1.1
...
implement #46
minor fix
2023-03-05 21:10:47 +09:00
79f10c8365
fix #46
...
and some minor fix in behavior
2023-03-05 21:06:55 +09:00
262128e1ce
Huge refactoring of UI related codes ( #48 )
2023-03-05 19:03:55 +09:00
8f1f045e59
fix #45
2023-02-28 19:27:40 +09:00
7191240d59
fix always using json metadata even if it is disabled
2023-02-26 14:32:13 +09:00
d93c84a6f9
refactoring
2023-02-26 14:29:18 +09:00
a646ebd065
Accept kohya-ss's metadata json as both input and output
2023-02-26 14:14:15 +09:00
9db35c0a45
use DirectML
2023-02-26 12:33:31 +09:00
c0aa72e100
enable wd-v14-tagger for AMD GPU users ( #43 )
2023-02-26 12:24:27 +09:00
77a4c42ab0
use pathlib instead of os.path
...
I'm not sure but may fix #42
2023-02-25 05:28:49 +09:00
c01f7c54f7
first aid for #42
...
Add option to force image gallery to use temporary files
2023-02-25 04:15:31 +09:00
09cdc65b30
use S=R threshold values for wd-v1-4-taggers
2023-02-25 00:40:49 +09:00
4b2a4a49e8
Fix: doesn't sort caption when save from dialog
2023-02-24 23:26:27 +09:00
952bd8f119
Added tag sorting to UI
...
This implementation utilizes the pre-existing DatasetTagEditor.sort_tags method.
Currently sorting is hardcoded as ascending alphabetical ("123abc"), but
it should be trivial to provide a UI for the various other options
provided by the method if anyone particularly wants alternative sorting.
Tags may be sorted as a batch for all filtered tags via a new button
- Found under: Batch Edit Captions > Extras > Sort selected tags
- The "Extras" menu was created as part of this commit
There is also an option to automatically sort tags upon clicking the
"Apply changes to selected image" button when editing tags via the
"Edit Caption of Selected Image" menu
- Disabled by default
- Enabled via: "Sort caption on save" checkbox (same menu)
2023-02-23 14:51:52 -05:00
8190f910e9
fix change in behavior ( #39 )
2023-02-18 16:29:34 +09:00
ddde7b756e
fix #39
2023-02-18 16:24:32 +09:00
7123254545
fix holding images on RAM when move or delete ( #39 )
2023-02-18 16:19:36 +09:00
6b1c76849a
Squashed commit of the following:
...
commit d445c1e15ded70cb511dfbf46ad414063d877294
Author: toshiaki1729 <116595002+toshiaki1729@users.noreply.github.com>
Date: Sat Feb 18 01:28:26 2023 +0900
Update dataset_tag_editor.py
commit 08342e87eb0501b2bd793ed91df0c2bc9e3ad9c9
Author: toshiaki1729 <116595002+toshiaki1729@users.noreply.github.com>
Date: Sat Feb 18 01:27:30 2023 +0900
Add file path info to image not for gallery to make temp file (#38 )
2023-02-18 01:36:13 +09:00
0eff08c170
split tag if comma in text when search and replace
2023-02-15 00:47:07 +09:00
4e00a69570
implement #37 : prefix/suffix and regex on tag search
2023-02-09 20:36:27 +09:00
dc77b5a1b6
add sort by "Length" option in TagSelectUI
2023-02-09 19:41:30 +09:00
7e2558fb75
implement #36
2023-02-07 22:49:03 +09:00
toshiaki1729
Zak
Ririan
ArrowM
Chris Mekarski
Christopher Crockett