forked from portu-sim/sd-webui-bmab
-
Notifications
You must be signed in to change notification settings - Fork 0
/
ui.py
773 lines (733 loc) · 42.1 KB
/
ui.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
import random
import gradio as gr
from modules import sd_models
from modules import sd_vae
from modules import ui_components
from modules import shared
from modules import extras
from modules import images
from sd_bmab import constants
from sd_bmab import util
from sd_bmab import detectors
from sd_bmab import parameters
from sd_bmab.base import context
from sd_bmab.base import filter
from sd_bmab.base import installer
from sd_bmab import pipeline
from sd_bmab import masking
from sd_bmab.util import debug_print
from sd_bmab.processors.controlnet import Openpose
from sd_bmab.processors.postprocess import Watermark
bmab_version = 'v24.04.26.0'
final_images = []
last_process = None
bmab_script = None
gallery_select_index = 0
def create_ui(bscript, is_img2img):
class ListOv(list):
def __iadd__(self, x):
self.append(x)
return self
ui_checkpoints = [constants.checkpoint_default]
ui_checkpoints.extend([str(x) for x in sd_models.checkpoints_list.keys()])
ui_vaes = [constants.vae_default]
ui_vaes.extend([str(x) for x in sd_vae.vae_dict.keys()])
elem = ListOv()
with gr.Group():
with gr.Row():
with gr.Column():
elem += gr.Checkbox(label=f'Enable BMAB', value=False)
with gr.Column():
btn_stop = ui_components.ToolButton('⏹️', visible=True, interactive=True, tooltip='stop generation', elem_id='bmab_stop_generation')
with gr.Accordion(f'BMAB Preprocessor', open=False):
with gr.Row():
with gr.Tab('Context', id='bmab_context', elem_id='bmab_context_tabs'):
with gr.Tab('Generic'):
with gr.Row():
with gr.Column():
with gr.Row():
checkpoint_models = gr.Dropdown(label='CheckPoint', visible=True, value=ui_checkpoints[0], choices=ui_checkpoints)
elem += checkpoint_models
with gr.Column():
with gr.Row():
vaes_models = gr.Dropdown(label='SD VAE', visible=True, value=ui_vaes[0], choices=ui_vaes)
elem += vaes_models
with gr.Row():
gr.Markdown(constants.checkpoint_description)
with gr.Row():
elem += gr.Slider(minimum=0, maximum=1.5, value=1, step=0.001, label='txt2img noise multiplier for hires.fix (EXPERIMENTAL)', elem_id='bmab_txt2img_noise_multiplier')
with gr.Row():
elem += gr.Slider(minimum=0, maximum=1, value=0, step=0.01, label='txt2img extra noise multiplier for hires.fix (EXPERIMENTAL)', elem_id='bmab_txt2img_extra_noise_multiplier')
with gr.Row():
with gr.Column():
with gr.Row():
dd_hiresfix_filter1 = gr.Dropdown(label='Hires.fix filter before upscale', visible=True, value=filter.filters[0], choices=filter.filters)
elem += dd_hiresfix_filter1
with gr.Column():
with gr.Row():
dd_hiresfix_filter2 = gr.Dropdown(label='Hires.fix filter after upscale', visible=True, value=filter.filters[0], choices=filter.filters)
elem += dd_hiresfix_filter2
with gr.Tab('Kohya Hires.fix'):
with gr.Row():
with gr.Column():
elem += gr.Checkbox(label='Enable Kohya hires.fix', value=False)
with gr.Row():
gr.HTML(constants.kohya_hiresfix_description)
with gr.Row():
elem += gr.Slider(minimum=0, maximum=0.5, step=0.01, label="Stop at, first", value=0.15)
elem += gr.Slider(minimum=1, maximum=10, step=1, label="Depth, first", value=3)
with gr.Row():
elem += gr.Slider(minimum=0, maximum=0.5, step=0.01, label="Stop at, second", value=0.4)
elem += gr.Slider(minimum=1, maximum=10, step=1, label="Depth, second", value=4)
with gr.Row():
elem += gr.Dropdown(['bicubic', 'bilinear', 'nearest', 'nearest-exact'], label='Layer scaler', value='bicubic')
elem += gr.Slider(minimum=0.1, maximum=1.0, step=0.05, label="Downsampling scale", value=0.5)
elem += gr.Slider(minimum=1.0, maximum=4.0, step=0.1, label="Upsampling scale", value=2.0)
with gr.Row():
elem += gr.Checkbox(label="Smooth scaling", value=True)
elem += gr.Checkbox(label="Early upsampling", value=False)
elem += gr.Checkbox(label='Disable for additional passes', value=True)
with gr.Tab('Resample', id='bmab_resample', elem_id='bmab_resample_tabs'):
with gr.Row():
with gr.Column():
elem += gr.Checkbox(label='Enable self resample', value=False)
with gr.Column():
elem += gr.Checkbox(label='Save image before processing', value=False)
with gr.Row():
elem += gr.Checkbox(label='Enable resample before hires.fix', value=False)
with gr.Row():
with gr.Column():
with gr.Row():
resample_models = gr.Dropdown(label='CheckPoint', visible=True, value=ui_checkpoints[0], choices=ui_checkpoints)
elem += resample_models
with gr.Column():
with gr.Row():
resample_vaes = gr.Dropdown(label='SD VAE', visible=True, value=ui_vaes[0], choices=ui_vaes)
elem += resample_vaes
with gr.Row():
with gr.Column(min_width=100):
methods = ['txt2img-1pass', 'txt2img-2pass', 'img2img-1pass']
elem += gr.Dropdown(label='Resample method', visible=True, value=methods[0], choices=methods)
with gr.Column():
dd_resample_filter = gr.Dropdown(label='Resample filter', visible=True, value=filter.filters[0], choices=filter.filters)
elem += dd_resample_filter
with gr.Row():
elem += gr.Textbox(placeholder='prompt. if empty, use main prompt', lines=3, visible=True, value='', label='Resample prompt')
with gr.Row():
elem += gr.Textbox(placeholder='negative prompt. if empty, use main negative prompt', lines=3, visible=True, value='', label='Resample negative prompt')
with gr.Row():
with gr.Column(min_width=100):
with gr.Row():
with gr.Column(min_width=50):
asamplers = [constants.sampler_default]
asamplers.extend([x.name for x in shared.list_samplers()])
elem += gr.Dropdown(label='Sampling method', visible=True, value=asamplers[0], choices=asamplers)
with gr.Column(min_width=50):
ascheduler = util.get_scueduler_list()
elem += gr.Dropdown(label='Scheduler', elem_id="resample_scheduler", choices=ascheduler, value=ascheduler[0])
with gr.Column(min_width=100):
upscalers = [constants.fast_upscaler]
upscalers.extend([x.name for x in shared.sd_upscalers])
elem += gr.Dropdown(label='Upscaler', visible=True, value=upscalers[0], choices=upscalers)
with gr.Row():
with gr.Column(min_width=100):
elem += gr.Slider(minimum=1, maximum=150, value=20, step=1, label='Resample Sampling Steps', elem_id='bmab_resample_steps')
elem += gr.Slider(minimum=1, maximum=30, value=7, step=0.5, label='Resample CFG Scale', elem_id='bmab_resample_cfg_scale')
elem += gr.Slider(minimum=0, maximum=1, value=0.75, step=0.01, label='Resample Denoising Strength', elem_id='bmab_resample_denoising')
elem += gr.Slider(minimum=0.0, maximum=2, value=0.5, step=0.05, label='Resample strength', elem_id='bmab_resample_cn_strength')
elem += gr.Slider(minimum=0.0, maximum=1.0, value=0.1, step=0.01, label='Resample begin', elem_id='bmab_resample_cn_begin')
elem += gr.Slider(minimum=0.0, maximum=1.0, value=0.9, step=0.01, label='Resample end', elem_id='bmab_resample_cn_end')
with gr.Tab('Pretraining', id='bmab_pretraining', elem_id='bmab_pretraining_tabs'):
with gr.Row():
elem += gr.Checkbox(label='Enable pretraining detailer', value=False)
with gr.Row():
elem += gr.Checkbox(label='Enable pretraining before hires.fix', value=False)
with gr.Row():
with gr.Column():
with gr.Row():
pretraining_checkpoint_models = gr.Dropdown(label='CheckPoint', visible=True, value=ui_checkpoints[0], choices=ui_checkpoints)
elem += pretraining_checkpoint_models
with gr.Column():
with gr.Row():
pretraining_vaes_models = gr.Dropdown(label='SD VAE', visible=True, value=ui_vaes[0], choices=ui_vaes)
elem += pretraining_vaes_models
with gr.Row():
with gr.Column(min_width=100):
with gr.Row():
models = ['Select Model']
models.extend(util.list_pretraining_models())
pretraining_models = gr.Dropdown(label='Pretraining Model', visible=True, value=models[0], choices=models, elem_id='bmab_pretraining_models')
elem += pretraining_models
with gr.Column(min_width=100):
dd_pretraining_filter = gr.Dropdown(label='Pretraining filter', visible=True, value=filter.filters[0], choices=filter.filters)
elem += dd_pretraining_filter
with gr.Row():
elem += gr.Textbox(placeholder='prompt. if empty, use main prompt', lines=3, visible=True, value='', label='Pretraining prompt')
with gr.Row():
elem += gr.Textbox(placeholder='negative prompt. if empty, use main negative prompt', lines=3, visible=True, value='', label='Pretraining negative prompt')
with gr.Row():
with gr.Column(min_width=100):
with gr.Row():
with gr.Column(min_width=50):
asamplers = [constants.sampler_default]
asamplers.extend([x.name for x in shared.list_samplers()])
elem += gr.Dropdown(label='Sampling method', visible=True, value=asamplers[0], choices=asamplers)
with gr.Column(min_width=50):
ascheduler = util.get_scueduler_list()
elem += gr.Dropdown(label='Scheduler', elem_id="pretraining_scheduler", choices=ascheduler, value=ascheduler[0])
with gr.Row():
with gr.Column(min_width=100):
elem += gr.Slider(minimum=1, maximum=150, value=20, step=1, label='Pretraining sampling steps', elem_id='bmab_pretraining_steps')
elem += gr.Slider(minimum=1, maximum=30, value=7, step=0.5, label='Pretraining CFG scale', elem_id='bmab_pretraining_cfg_scale')
elem += gr.Slider(minimum=0, maximum=1, value=0.75, step=0.01, label='Pretraining denoising Strength', elem_id='bmab_pretraining_denoising')
elem += gr.Slider(minimum=0, maximum=128, value=4, step=1, label='Pretraining dilation', elem_id='bmab_pretraining_dilation')
elem += gr.Slider(minimum=0.1, maximum=1, value=0.35, step=0.01, label='Pretraining box threshold', elem_id='bmab_pretraining_box_threshold')
with gr.Tab('Edge', elem_id='bmab_edge_tabs'):
with gr.Row():
elem += gr.Checkbox(label='Enable edge enhancement', value=False)
with gr.Row():
elem += gr.Slider(minimum=1, maximum=255, value=50, step=1, label='Edge low threshold')
elem += gr.Slider(minimum=1, maximum=255, value=200, step=1, label='Edge high threshold')
with gr.Row():
elem += gr.Slider(minimum=0, maximum=1, value=0.5, step=0.05, label='Edge strength')
gr.Markdown('')
with gr.Tab('Resize', elem_id='bmab_preprocess_resize_tab'):
with gr.Row():
elem += gr.Checkbox(label='Enable resize (intermediate)', value=False)
with gr.Row():
elem += gr.Checkbox(label='Resized by person', value=True)
with gr.Row():
gr.HTML(constants.resize_description)
with gr.Row():
with gr.Column():
methods = ['stretching', 'inpaint', 'inpaint+lama', 'inpaint_only', 'inpaint_only+lama']
elem += gr.Dropdown(label='Method', visible=True, value=methods[0], choices=methods)
with gr.Column():
align = [x for x in util.alignment.keys()]
elem += gr.Dropdown(label='Alignment', visible=True, value=align[4], choices=align)
with gr.Row():
with gr.Column():
dd_resize_filter = gr.Dropdown(label='Resize filter', visible=True, value=filter.filters[0], choices=filter.filters)
elem += dd_resize_filter
with gr.Column():
gr.Markdown('')
with gr.Row():
elem += gr.Slider(minimum=0.50, maximum=0.95, value=0.85, step=0.01, label='Resize by person intermediate')
with gr.Row():
elem += gr.Slider(minimum=0, maximum=1, value=0.75, step=0.01, label='Denoising Strength for inpaint and inpaint+lama', elem_id='bmab_resize_intermediate_denoising')
with gr.Tab('Refiner', id='bmab_refiner', elem_id='bmab_refiner_tabs'):
with gr.Row():
elem += gr.Checkbox(label='Enable refiner', value=False)
with gr.Row():
with gr.Column():
with gr.Row():
refiner_models = gr.Dropdown(label='CheckPoint for refiner', visible=True, value=ui_checkpoints[0], choices=ui_checkpoints)
elem += refiner_models
with gr.Column():
with gr.Row():
vaes = [constants.vae_default]
vaes.extend([str(x) for x in sd_vae.vae_dict.keys()])
refiner_vaes = gr.Dropdown(label='SD VAE', visible=True, value=ui_vaes[0], choices=ui_vaes)
elem += refiner_vaes
with gr.Row():
elem += gr.Checkbox(label='Use this checkpoint for detailing(Face, Person, Hand)', value=True)
with gr.Row():
elem += gr.Textbox(placeholder='prompt. if empty, use main prompt', lines=3, visible=True, value='', label='Prompt')
with gr.Row():
elem += gr.Textbox(placeholder='negative prompt. if empty, use main negative prompt', lines=3, visible=True, value='', label='Negative Prompt')
with gr.Row():
with gr.Column(min_width=100):
with gr.Row():
with gr.Column(min_width=50):
asamplers = [constants.sampler_default]
asamplers.extend([x.name for x in shared.list_samplers()])
elem += gr.Dropdown(label='Sampling method', visible=True, value=asamplers[0], choices=asamplers)
with gr.Column(min_width=50):
ascheduler = util.get_scueduler_list()
elem += gr.Dropdown(label='Scheduler', elem_id="refiner_scheduler", choices=ascheduler, value=ascheduler[0])
with gr.Column(min_width=100):
upscalers = [constants.fast_upscaler]
upscalers.extend([x.name for x in shared.sd_upscalers])
elem += gr.Dropdown(label='Upscaler', visible=True, value=upscalers[0], choices=upscalers)
with gr.Row():
with gr.Column(min_width=100):
elem += gr.Slider(minimum=1, maximum=150, value=20, step=1, label='Refiner Sampling Steps', elem_id='bmab_refiner_steps')
elem += gr.Slider(minimum=1, maximum=30, value=7, step=0.5, label='Refiner CFG Scale', elem_id='bmab_refiner_cfg_scale')
elem += gr.Slider(minimum=0, maximum=1, value=0.75, step=0.01, label='Refiner Denoising Strength', elem_id='bmab_refiner_denoising')
with gr.Row():
with gr.Column(min_width=100):
elem += gr.Slider(minimum=0, maximum=4, value=1, step=0.1, label='Refiner Scale', elem_id='bmab_refiner_scale')
elem += gr.Slider(minimum=0, maximum=2048, value=0, step=1, label='Refiner Width', elem_id='bmab_refiner_width')
elem += gr.Slider(minimum=0, maximum=2048, value=0, step=1, label='Refiner Height', elem_id='bmab_refiner_height')
with gr.Accordion(f'BMAB', open=False):
with gr.Row():
with gr.Tabs(elem_id='bmab_tabs'):
with gr.Tab('Basic', elem_id='bmab_basic_tabs'):
with gr.Row():
with gr.Column():
elem += gr.Slider(minimum=0, maximum=2, value=1, step=0.05, label='Contrast')
elem += gr.Slider(minimum=0, maximum=2, value=1, step=0.05, label='Brightness')
elem += gr.Slider(minimum=-5, maximum=5, value=1, step=0.1, label='Sharpeness')
elem += gr.Slider(minimum=0, maximum=2, value=1, step=0.01, label='Color')
with gr.Column():
elem += gr.Slider(minimum=-2000, maximum=+2000, value=0, step=1, label='Color temperature')
elem += gr.Slider(minimum=0, maximum=1, value=0, step=0.05, label='Noise alpha')
elem += gr.Slider(minimum=0, maximum=1, value=0, step=0.05, label='Noise alpha at final stage')
with gr.Tab('Imaging', elem_id='bmab_imaging_tabs'):
with gr.Row():
elem += gr.Image(source='upload', type='pil')
with gr.Row():
elem += gr.Checkbox(label='Blend enabled', value=False)
with gr.Row():
with gr.Column():
elem += gr.Slider(minimum=0, maximum=1, value=1, step=0.05, label='Blend alpha')
with gr.Column():
gr.Markdown('')
with gr.Row():
elem += gr.Checkbox(label='Enable detect', value=False)
with gr.Row():
elem += gr.Textbox(placeholder='1girl', visible=True, value='', label='Prompt')
with gr.Tab('Person', elem_id='bmab_person_tabs'):
with gr.Row():
elem += gr.Checkbox(label='Enable person detailing for landscape', value=False)
with gr.Row():
elem += gr.Checkbox(label='Enable best quality (EXPERIMENTAL, Use more GPU)', value=False)
elem += gr.Checkbox(label='Force upscale ratio 1:1 without area limit', value=False)
with gr.Row():
elem += gr.Checkbox(label='Block over-scaled image', value=True)
elem += gr.Checkbox(label='Auto Upscale if Block over-scaled image enabled', value=True)
with gr.Row():
with gr.Column(min_width=100):
elem += gr.Slider(minimum=1, maximum=8, value=4, step=0.01, label='Upscale Ratio')
elem += gr.Slider(minimum=0, maximum=20, value=3, step=1, label='Dilation mask')
elem += gr.Slider(minimum=0.01, maximum=1, value=0.1, step=0.01, label='Large person area limit')
elem += gr.Slider(minimum=0, maximum=20, value=1, step=1, label='Limit')
elem += gr.Slider(minimum=0, maximum=2, value=1, step=0.01, visible=shared.opts.data.get('bmab_test_function', False), label='Background color (HIDDEN)')
elem += gr.Slider(minimum=0, maximum=30, value=0, step=1, visible=shared.opts.data.get('bmab_test_function', False), label='Background blur (HIDDEN)')
with gr.Column(min_width=100):
elem += gr.Slider(minimum=0, maximum=1, value=0.4, step=0.01, label='Denoising Strength')
elem += gr.Slider(minimum=1, maximum=30, value=7, step=0.5, label='CFG Scale')
gr.Markdown('')
with gr.Tab('Face', elem_id='bmab_face_tabs'):
with gr.Row():
elem += gr.Checkbox(label='Enable face detailing', value=False)
with gr.Row():
elem += gr.Checkbox(label='Enable face detailing before upscale', value=False)
with gr.Row():
elem += gr.Checkbox(label='Disable extra networks in prompt (LORA, Hypernetwork, ...)', value=False)
with gr.Row():
with gr.Column(min_width=100):
elem += gr.Dropdown(label='Face detailing sort by', choices=['Score', 'Size', 'Left', 'Right', 'Center'], type='value', value='Score')
with gr.Column(min_width=100):
elem += gr.Slider(minimum=0, maximum=20, value=1, step=1, label='Limit')
with gr.Tab('Face1', elem_id='bmab_face1_tabs'):
with gr.Row():
elem += gr.Textbox(placeholder='prompt. if empty, use main prompt', lines=3, visible=True, value='', label='Prompt')
with gr.Row():
elem += gr.Textbox(placeholder='negative prompt. if empty, use main negative prompt', lines=3, visible=True, value='', label='Negative Prompt')
with gr.Tab('Face2', elem_id='bmab_face2_tabs'):
with gr.Row():
elem += gr.Textbox(placeholder='prompt. if empty, use main prompt', lines=3, visible=True, value='', label='Prompt')
with gr.Row():
elem += gr.Textbox(placeholder='negative prompt. if empty, use main negative prompt', lines=3, visible=True, value='', label='Negative Prompt')
with gr.Tab('Face3', elem_id='bmab_face3_tabs'):
with gr.Row():
elem += gr.Textbox(placeholder='prompt. if empty, use main prompt', lines=3, visible=True, value='', label='Prompt')
with gr.Row():
elem += gr.Textbox(placeholder='negative prompt. if empty, use main negative prompt', lines=3, visible=True, value='', label='Negative Prompt')
with gr.Tab('Face4', elem_id='bmab_face4_tabs'):
with gr.Row():
elem += gr.Textbox(placeholder='prompt. if empty, use main prompt', lines=3, visible=True, value='', label='Prompt')
with gr.Row():
elem += gr.Textbox(placeholder='negative prompt. if empty, use main negative prompt', lines=3, visible=True, value='', label='Negative Prompt')
with gr.Tab('Face5', elem_id='bmab_face5_tabs'):
with gr.Row():
elem += gr.Textbox(placeholder='prompt. if empty, use main prompt', lines=3, visible=True, value='', label='Prompt')
with gr.Row():
elem += gr.Textbox(placeholder='negative prompt. if empty, use main negative prompt', lines=3, visible=True, value='', label='Negative Prompt')
with gr.Row():
with gr.Tab('Parameters', elem_id='bmab_parameter_tabs'):
with gr.Row():
elem += gr.Checkbox(label='Overide Parameters', value=False)
with gr.Row():
with gr.Column(min_width=100):
elem += gr.Slider(minimum=64, maximum=2048, value=512, step=8, label='Width')
elem += gr.Slider(minimum=64, maximum=2048, value=512, step=8, label='Height')
with gr.Column(min_width=100):
elem += gr.Slider(minimum=1, maximum=30, value=7, step=0.5, label='CFG Scale')
elem += gr.Slider(minimum=1, maximum=150, value=20, step=1, label='Steps')
elem += gr.Slider(minimum=0, maximum=64, value=4, step=1, label='Mask Blur')
with gr.Row():
with gr.Column(min_width=100):
with gr.Row():
with gr.Column(min_width=50):
face_models = gr.Dropdown(label='CheckPoint for face', visible=True, value=ui_checkpoints[0], choices=ui_checkpoints)
elem += face_models
with gr.Column(min_width=50):
face_vaes = gr.Dropdown(label='SD VAE for face', visible=True, value=ui_vaes[0], choices=ui_vaes)
elem += face_vaes
with gr.Row():
with gr.Column(min_width=50):
asamplers = [constants.sampler_default]
asamplers.extend([x.name for x in shared.list_samplers()])
elem += gr.Dropdown(label='Sampler', elem_id="face_sampler", visible=True, value=asamplers[0], choices=asamplers)
with gr.Column(min_width=50):
ascheduler = util.get_scueduler_list()
elem += gr.Dropdown(label='Scheduler', elem_id="face_scheduler", choices=ascheduler, value=ascheduler[0])
with gr.Row():
inpaint_area = gr.Radio(label='Inpaint area', choices=['Whole picture', 'Only masked'], type='value', value='Only masked')
elem += inpaint_area
with gr.Row():
elem += gr.Slider(label='Only masked padding, pixels', minimum=0, maximum=256, step=4, value=32)
with gr.Row():
choices = detectors.list_face_detectors()
elem += gr.Dropdown(label='Detection Model', choices=choices, type='value', value=choices[0])
with gr.Column():
elem += gr.Slider(minimum=0, maximum=1, value=0.4, step=0.01, label='Face Denoising Strength', elem_id='bmab_face_denoising_strength')
elem += gr.Slider(minimum=0, maximum=64, value=4, step=1, label='Face Dilation', elem_id='bmab_face_dilation')
elem += gr.Slider(minimum=0.1, maximum=1, value=0.35, step=0.01, label='Face Box threshold')
elem += gr.Checkbox(label='Skip face detailing by area', value=False)
elem += gr.Slider(minimum=0.0, maximum=3.0, value=0.26, step=0.01, label='Face area (MegaPixel)')
with gr.Tab('Hand', elem_id='bmab_hand_tabs'):
with gr.Row():
elem += gr.Checkbox(label='Enable hand detailing (EXPERIMENTAL)', value=False)
elem += gr.Checkbox(label='Block over-scaled image', value=True)
with gr.Row():
elem += gr.Checkbox(label='Enable best quality (EXPERIMENTAL, Use more GPU)', value=False)
with gr.Row():
elem += gr.Dropdown(label='Method', visible=True, interactive=True, value='subframe', choices=['subframe', 'each hand', 'inpaint each hand', 'at once', 'depth hand refiner'])
with gr.Row():
elem += gr.Textbox(placeholder='prompt. if empty, use main prompt', lines=3, visible=True, value='', label='Prompt')
with gr.Row():
elem += gr.Textbox(placeholder='negative prompt. if empty, use main negative prompt', lines=3, visible=True, value='', label='Negative Prompt')
with gr.Row():
with gr.Column():
elem += gr.Slider(minimum=0, maximum=1, value=0.4, step=0.01, label='Denoising Strength')
elem += gr.Slider(minimum=1, maximum=30, value=7, step=0.5, label='CFG Scale')
elem += gr.Checkbox(label='Auto Upscale if Block over-scaled image enabled', value=True)
with gr.Column():
elem += gr.Slider(minimum=1, maximum=4, value=2, step=0.01, label='Upscale Ratio')
elem += gr.Slider(minimum=0, maximum=1, value=0.3, step=0.01, label='Box Threshold')
elem += gr.Slider(minimum=0, maximum=0.3, value=0.1, step=0.01, label='Box Dilation')
with gr.Row():
inpaint_area = gr.Radio(label='Inpaint area', choices=['Whole picture', 'Only masked'], type='value', value='Whole picture')
elem += inpaint_area
with gr.Row():
with gr.Column():
elem += gr.Slider(label='Only masked padding, pixels', minimum=0, maximum=256, step=4, value=32)
with gr.Column():
gr.Markdown('')
with gr.Row():
elem += gr.Textbox(placeholder='Additional parameter for advanced user', visible=True, value='', label='Additional Parameter')
with gr.Tab('ControlNet', elem_id='bmab_controlnet_tabs'):
with gr.Row():
elem += gr.Checkbox(label='Enable ControlNet access', value=False)
with gr.Row():
elem += gr.Checkbox(label='Process with BMAB refiner', value=False)
with gr.Row():
with gr.Tab('Noise', elem_id='bmab_cn_noise_tabs'):
with gr.Row():
elem += gr.Checkbox(label='Enable noise', value=False)
with gr.Row():
with gr.Column():
elem += gr.Slider(minimum=0.0, maximum=2, value=0.4, step=0.05, elem_id='bmab_cn_noise', label='Noise strength')
elem += gr.Slider(minimum=0.0, maximum=1.0, value=0.1, step=0.01, elem_id='bmab_cn_noise_begin', label='Noise begin')
elem += gr.Slider(minimum=0.0, maximum=1.0, value=0.9, step=0.01, elem_id='bmab_cn_noise_end', label='Noise end')
elem += gr.Radio(label='Hire-fix option for noise', choices=['Both', 'Low res only', 'High res only'], type='value', value='Both')
with gr.Column():
gr.Markdown('')
with gr.Tab('Pose', elem_id='bmab_cn_pose_tabs'):
with gr.Row():
elem += gr.Checkbox(label='Enable pose', value=False)
with gr.Row():
with gr.Column():
elem += gr.Slider(minimum=0.0, maximum=2, value=0.3, step=0.05, elem_id='bmab_cn_pose', label='Pose strength')
elem += gr.Slider(minimum=0.0, maximum=1.0, value=0.0, step=0.01, elem_id='bmab_cn_pose_begin', label='Pose begin')
elem += gr.Slider(minimum=0.0, maximum=1.0, value=0.1, step=0.01, elem_id='bmab_cn_pose_end', label='Pose end')
elem += gr.Checkbox(label='Face only', value=False)
poses = ['Random']
poses.extend(Openpose.list_pose())
dd_pose = gr.Dropdown(label='Pose Selection', interactive=True, visible=True, value=poses[0], choices=poses)
elem += dd_pose
with gr.Column():
pose_image = gr.Image(elem_id='bmab_pose_image')
with gr.Accordion(f'BMAB Postprocessor', open=False):
with gr.Row():
with gr.Tab('Resize by person', elem_id='bmab_postprocess_resize_tab'):
with gr.Row():
elem += gr.Checkbox(label='Enable resize by person', value=False)
mode = ['Inpaint', 'ControlNet inpaint+lama']
elem += gr.Dropdown(label='Mode', visible=True, value=mode[0], choices=mode)
with gr.Row():
with gr.Column():
elem += gr.Slider(minimum=0.70, maximum=0.95, value=0.85, step=0.01, label='Resize by person')
with gr.Column():
elem += gr.Slider(minimum=0, maximum=1, value=0.6, step=0.01, label='Denoising Strength for Inpaint, ControlNet')
with gr.Row():
with gr.Column():
gr.Markdown('')
with gr.Column():
elem += gr.Slider(minimum=4, maximum=128, value=30, step=1, label='Mask Dilation')
with gr.Tab('Upscale', elem_id='bmab_postprocess_upscale_tab'):
with gr.Row():
with gr.Column(min_width=100):
elem += gr.Checkbox(label='Enable upscale at final stage', value=False)
elem += gr.Checkbox(label='Detailing after upscale', value=True)
with gr.Column(min_width=100):
gr.Markdown('')
with gr.Row():
with gr.Column(min_width=100):
upscalers = [x.name for x in shared.sd_upscalers]
elem += gr.Dropdown(label='Upscaler', visible=True, value=upscalers[0], choices=upscalers)
elem += gr.Slider(minimum=1, maximum=4, value=1.5, step=0.1, label='Upscale ratio')
with gr.Tab('Filter', id='bmab_final_filter', elem_id='bmab_final_filter_tab'):
with gr.Row():
dd_final_filter = gr.Dropdown(label='Final filter', visible=True, value=filter.filters[0], choices=filter.filters)
elem += dd_final_filter
with gr.Tab('Watermark', id='bmab_watermark', elem_id='bmab_watermark'):
elem += gr.Checkbox(label='Watermark enabled', value=False)
with gr.Row():
with gr.Column(min_width=100):
fonts = Watermark.list_fonts()
elem += gr.Dropdown(label='Watermark Font', visible=True, value=fonts[0], choices=fonts)
align = [x for x in Watermark.alignment.keys()]
elem += gr.Dropdown(label='Watermark Alignment', visible=True, value=align[5], choices=align)
elem += gr.Dropdown(label='Watermark Text Alignment', visible=True, value='left', choices=['left', 'right', 'center'])
elem += gr.Dropdown(label='Watermark Text Rotate', visible=True, value='0', choices=['0', '90', '180', '270'])
elem += gr.Textbox(label='Watermark Text Color', visible=True, value='#000000')
elem += gr.Textbox(label='Watermark Background Color', visible=True, value='#000000')
with gr.Column(min_width=100):
elem += gr.Slider(minimum=4, maximum=128, value=12, step=1, label='Font Size')
elem += gr.Slider(minimum=0, maximum=100, value=100, step=1, label='Transparency')
elem += gr.Slider(minimum=0, maximum=100, value=0, step=1, label='Background Transparency')
elem += gr.Slider(minimum=0, maximum=100, value=5, step=1, label='Margin')
with gr.Row():
elem += gr.Textbox(placeholder='watermark text here', lines=1, max_lines=10, visible=True, value='', label='Watermark or Image path')
with gr.Accordion(f'BMAB Refresh, Config, Preset, Installer', open=False):
with gr.Row():
configs = parameters.Parameters().list_config()
config = '' if not configs else configs[0]
with gr.Tab('Configuration', elem_id='bmab_configuration_tabs'):
with gr.Row():
with gr.Column(scale=2):
with gr.Row():
config_dd = gr.Dropdown(label='Configuration', visible=True, interactive=True, allow_custom_value=True, value=config, choices=configs)
elem += config_dd
load_btn = ui_components.ToolButton('⬇️', visible=True, interactive=True, tooltip='load configuration', elem_id='bmab_load_configuration')
save_btn = ui_components.ToolButton('⬆️', visible=True, interactive=True, tooltip='save configuration', elem_id='bmab_save_configuration')
reset_btn = ui_components.ToolButton('🔃', visible=True, interactive=True, tooltip='reset to default', elem_id='bmab_reset_configuration')
with gr.Column(scale=1):
gr.Markdown('')
with gr.Row():
with gr.Column(scale=1):
btn_refresh_all = gr.Button('Refresh ALL', visible=True, interactive=True, elem_id='bmab_refresh_all')
with gr.Column(scale=1):
gr.Markdown('')
with gr.Column(scale=1):
gr.Markdown('')
with gr.Column(scale=1):
gr.Markdown('')
with gr.Tab('Preset', elem_id='bmab_configuration_tabs'):
with gr.Row():
with gr.Column(min_width=100):
gr.Markdown('Preset Loader : preset override UI configuration.')
with gr.Row():
presets = parameters.Parameters().list_preset()
with gr.Column(min_width=100):
with gr.Row():
preset_dd = gr.Dropdown(label='Preset', visible=True, interactive=True, allow_custom_value=True, value=presets[0], choices=presets)
elem += preset_dd
refresh_btn = ui_components.ToolButton('🔄', visible=True, interactive=True, tooltip='refresh preset', elem_id='bmab_preset_refresh')
with gr.Tab('Toy', elem_id='bmab_toy_tabs'):
with gr.Row():
merge_result = gr.Markdown('Result here')
with gr.Row():
random_checkpoint = gr.Button('Merge Random Checkpoint', visible=True, interactive=True, elem_id='bmab_merge_random_checkpoint')
with gr.Tab('Installer', elem_id='bmab_install_tabs'):
with gr.Row():
pkgs = ['GroundingDINO']
dd_pkg = gr.Dropdown(label='Package', visible=True, value=pkgs[0], choices=pkgs)
btn_install = ui_components.ToolButton('🔄', visible=True, interactive=True, tooltip='Install package', elem_id='bmab_btn_install')
with gr.Row():
markdown_install = gr.Markdown('')
with gr.Accordion(f'BMAB Testroom', open=False, visible=shared.opts.data.get('bmab_for_developer', False)):
with gr.Row():
gallery = gr.Gallery(label='Images', value=[], elem_id='bmab_testroom_gallery')
result_image = gr.Image(elem_id='bmab_result_image')
with gr.Row():
btn_fetch_images = ui_components.ToolButton('🔄', visible=True, interactive=True, tooltip='fetch images', elem_id='bmab_fetch_images')
btn_process_pipeline = ui_components.ToolButton('▶️', visible=True, interactive=True, tooltip='fetch images', elem_id='bmab_fetch_images')
gr.Markdown(f'<div style="text-align: right; vertical-align: bottom"><span style="color: green">{bmab_version}</span></div>')
def load_config(*args):
name = args[0]
ret = parameters.Parameters().load_config(name)
return ret
def save_config(*args):
name = parameters.Parameters().get_save_config_name(args)
parameters.Parameters().save_config(args)
return {
config_dd: {
'choices': parameters.Parameters().list_config(),
'value': name,
'__type__': 'update'
}
}
def reset_config(*args):
return parameters.Parameters().get_default()
def refresh_preset(*args):
return {
preset_dd: {
'choices': parameters.Parameters().list_preset(),
'value': 'None',
'__type__': 'update'
}
}
def merge_random_checkpoint(*args):
def find_random(k, f):
for v in k:
if v.startswith(f):
return v
result = ''
checkpoints = [str(x) for x in sd_models.checkpoints_list.keys()]
target = random.choices(checkpoints, k=3)
multiplier = random.randrange(10, 90, 1) / 100
index = random.randrange(0x10000000, 0xFFFFFFFF, 1)
output = f'bmab_random_{format(index, "08X")}'
extras.run_modelmerger(None, target[0], target[1], target[2], 'Weighted sum', multiplier, False, output, 'safetensors', 0, None, '', True, True, True, '{}')
result += f'{output}.safetensors generated<br>'
for x in range(1, random.randrange(0, 5, 1)):
checkpoints = [str(x) for x in sd_models.checkpoints_list.keys()]
br = find_random(checkpoints, f'{output}.safetensors')
if br is None:
return
index = random.randrange(0x10000000, 0xFFFFFFFF, 1)
output = f'bmab_random_{format(index, "08X")}'
target = random.choices(checkpoints, k=2)
multiplier = random.randrange(10, 90, 1) / 100
extras.run_modelmerger(None, br, target[0], target[1], 'Weighted sum', multiplier, False, output, 'safetensors', 0, None, '', True, True, True, '{}')
result += f'{output}.safetensors generated<br>'
debug_print('done')
return {
merge_result: {
'value': result,
'__type__': 'update'
}
}
def fetch_images(*args):
global gallery_select_index
gallery_select_index = 0
return {
gallery: {
'value': final_images,
'__type__': 'update'
}
}
def process_pipeline(*args):
config, a = parameters.parse_args(args)
preview = final_images[gallery_select_index]
p = last_process
ctx = context.Context.newContext(bmab_script, p, a, gallery_select_index)
preview = pipeline.process(ctx, preview)
images.save_image(
preview, p.outpath_samples, '',
p.all_seeds[gallery_select_index], p.all_prompts[gallery_select_index],
shared.opts.samples_format, p=p, suffix="-testroom")
return {
result_image: {
'value': preview,
'__type__': 'update'
}
}
refresh_targets = [dd_hiresfix_filter1, dd_hiresfix_filter2, dd_resample_filter, dd_resize_filter, dd_final_filter, dd_pretraining_filter]
refresh_targets.extend([checkpoint_models, vaes_models, refiner_models, refiner_vaes, face_models, face_vaes, resample_models, resample_vaes])
refresh_targets.extend([pretraining_checkpoint_models, pretraining_vaes_models, pretraining_models, dd_pose])
def reload_filter(*args):
filter.reload_filters()
inputs = list(args)
_checkpoints = [constants.checkpoint_default]
_checkpoints.extend([str(x) for x in sd_models.checkpoints_list.keys()])
_vaes = [constants.vae_default]
_vaes.extend([str(x) for x in sd_vae.vae_dict.keys()])
_pretraining_models = ['Select Model']
_pretraining_models.extend(util.list_pretraining_models())
_poses = ['Random']
_poses.extend(Openpose.list_pose())
values = [
filter.filters, filter.filters, filter.filters, filter.filters, filter.filters, filter.filters,
_checkpoints, _vaes, _checkpoints, _vaes, _checkpoints, _vaes, _checkpoints, _vaes,
_checkpoints, _vaes, _pretraining_models, _poses
]
ret = {
t: {
'choices': v,
'value': v[0] if i not in v else i,
'__type__': 'update'
}
for t, i, v in zip(refresh_targets, inputs, values)
}
return ret
def image_selected(data: gr.SelectData, *args):
debug_print(data.index)
global gallery_select_index
gallery_select_index = data.index
def hit_install(*args):
pkg_name = args[0]
if pkg_name == 'GroundingDINO':
installer.install_groudingdino()
msg = f'{pkg_name} installed'
else:
msg = 'Nothing installed.'
return {
markdown_install: {
'value': msg,
'__type__': 'update'
}
}
def stop_process(*args):
bscript.stop_generation = True
gr.Info('Waiting for processing done.')
def pose_selected(*args):
print(args)
return Openpose.get_pose(args[0])
load_btn.click(load_config, inputs=[config_dd], outputs=elem)
save_btn.click(save_config, inputs=elem, outputs=[config_dd])
reset_btn.click(reset_config, outputs=elem)
refresh_btn.click(refresh_preset, outputs=elem)
random_checkpoint.click(merge_random_checkpoint, outputs=[merge_result])
btn_fetch_images.click(fetch_images, outputs=[gallery])
btn_refresh_all.click(
reload_filter,
inputs=refresh_targets,
outputs=refresh_targets,
)
btn_process_pipeline.click(process_pipeline, inputs=elem, outputs=[result_image])
gallery.select(image_selected, inputs=[gallery])
btn_install.click(hit_install, inputs=[dd_pkg], outputs=[markdown_install])
btn_stop.click(stop_process)
dd_pose.select(pose_selected, inputs=[dd_pose], outputs=[pose_image])
return elem
def on_ui_settings():
shared.opts.add_option('bmab_debug_print', shared.OptionInfo(False, 'Print debug message.', section=('bmab', 'BMAB')))
shared.opts.add_option('bmab_debug_logging', shared.OptionInfo(False, 'Enable developer logging.', section=('bmab', 'BMAB')))
shared.opts.add_option('bmab_show_extends', shared.OptionInfo(False, 'Show before processing image. (DO NOT ENABLE IN CLOUD)', section=('bmab', 'BMAB')))
shared.opts.add_option('bmab_test_function', shared.OptionInfo(False, 'Show Test Function', section=('bmab', 'BMAB')))
shared.opts.add_option('bmab_keep_original_setting', shared.OptionInfo(False, 'Keep original setting', section=('bmab', 'BMAB')))
shared.opts.add_option('bmab_save_image_before_process', shared.OptionInfo(False, 'Save image that before processing', section=('bmab', 'BMAB')))
shared.opts.add_option('bmab_save_image_after_process', shared.OptionInfo(False, 'Save image that after processing (some bugs)', section=('bmab', 'BMAB')))
shared.opts.add_option('bmab_for_developer', shared.OptionInfo(False, 'Show developer hidden function.', section=('bmab', 'BMAB')))
shared.opts.add_option('bmab_use_dino_predict', shared.OptionInfo(False, 'Use GroudingDINO for detecting hand. GroudingDINO should be installed manually.', section=('bmab', 'BMAB')))
shared.opts.add_option('bmab_max_detailing_element', shared.OptionInfo(
default=0, label='Max Detailing Element', component=gr.Slider, component_args={'minimum': 0, 'maximum': 10, 'step': 1}, section=('bmab', 'BMAB')))
shared.opts.add_option('bmab_detail_full', shared.OptionInfo(True, 'Allways use FULL, VAE type for encode when detail anything. (v1.6.0)', section=('bmab', 'BMAB')))
shared.opts.add_option('bmab_optimize_vram', shared.OptionInfo(default='None', label='Checkpoint for Person, Face, Hand', component=gr.Radio, component_args={'choices': ['None', 'low vram', 'med vram']}, section=('bmab', 'BMAB')))
mask_names = masking.list_mask_names()
shared.opts.add_option('bmab_mask_model', shared.OptionInfo(default=mask_names[0], label='Masking model', component=gr.Radio, component_args={'choices': mask_names}, section=('bmab', 'BMAB')))
shared.opts.add_option('bmab_use_specific_model', shared.OptionInfo(False, 'Use specific model', section=('bmab', 'BMAB')))
shared.opts.add_option('bmab_model', shared.OptionInfo(default='', label='Checkpoint for Person, Face, Hand', component=gr.Textbox, component_args='', section=('bmab', 'BMAB')))
shared.opts.add_option('bmab_cn_openpose', shared.OptionInfo(default='control_v11p_sd15_openpose_fp16 [73c2b67d]', label='ControlNet openpose model', component=gr.Textbox, component_args='', section=('bmab', 'BMAB')))
shared.opts.add_option('bmab_cn_lineart', shared.OptionInfo(default='control_v11p_sd15_lineart_fp16 [5c23b17d]', label='ControlNet lineart model', component=gr.Textbox, component_args='', section=('bmab', 'BMAB')))
shared.opts.add_option('bmab_cn_inpaint', shared.OptionInfo(default='control_v11p_sd15_inpaint_fp16 [be8bc0ed]', label='ControlNet inpaint model', component=gr.Textbox, component_args='', section=('bmab', 'BMAB')))
shared.opts.add_option('bmab_cn_tile_resample', shared.OptionInfo(default='control_v11f1e_sd15_tile_fp16 [3b860298]', label='ControlNet tile model', component=gr.Textbox, component_args='', section=('bmab', 'BMAB')))
shared.opts.add_option('bmab_cn_inpaint_depth_hand', shared.OptionInfo(default='control_sd15_inpaint_depth_hand_fp16 [09456e54]', label='ControlNet tile model', component=gr.Textbox, component_args='', section=('bmab', 'BMAB')))
shared.opts.add_option('img2img_extra_noise',shared.OptionInfo(default=0.0, label="Extra noise multiplier for img2img and hires fix", component=gr.Slider, component_args={"minimum": 0.0, "maximum": 1.0, "step": 0.01}, section=('bmab', 'BMAB')))
shared.opts.add_option('randn_source',shared.OptionInfo(default="NV", label="RNG", component=gr.Radio, component_args={"choices": ["GPU", "CPU", "NV"]}, section=('bmab', 'BMAB')))