-
Notifications
You must be signed in to change notification settings - Fork 0
/
lpw_stable_diffusion.py
1164 lines (1065 loc) · 56.4 KB
/
lpw_stable_diffusion.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
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
import inspect
import re
from typing import Callable, List, Optional, Union
import numpy as np
import torch
import diffusers
import PIL
from diffusers import SchedulerMixin, StableDiffusionPipeline
from diffusers.models import AutoencoderKL, UNet2DConditionModel
from diffusers.pipelines.stable_diffusion import StableDiffusionPipelineOutput, StableDiffusionSafetyChecker
from diffusers.utils import deprecate, logging
from packaging import version
from transformers import CLIPFeatureExtractor, CLIPTextModel, CLIPTokenizer
try:
from diffusers.utils import PIL_INTERPOLATION
except ImportError:
if version.parse(version.parse(PIL.__version__).base_version) >= version.parse("9.1.0"):
PIL_INTERPOLATION = {
"linear": PIL.Image.Resampling.BILINEAR,
"bilinear": PIL.Image.Resampling.BILINEAR,
"bicubic": PIL.Image.Resampling.BICUBIC,
"lanczos": PIL.Image.Resampling.LANCZOS,
"nearest": PIL.Image.Resampling.NEAREST,
}
else:
PIL_INTERPOLATION = {
"linear": PIL.Image.LINEAR,
"bilinear": PIL.Image.BILINEAR,
"bicubic": PIL.Image.BICUBIC,
"lanczos": PIL.Image.LANCZOS,
"nearest": PIL.Image.NEAREST,
}
# ------------------------------------------------------------------------------
logger = logging.get_logger(__name__) # pylint: disable=invalid-name
re_attention = re.compile(
r"""
\\\(|
\\\)|
\\\[|
\\]|
\\\\|
\\|
\(|
\[|
:([+-]?[.\d]+)\)|
\)|
]|
[^\\()\[\]:]+|
:
""",
re.X,
)
def parse_prompt_attention(text):
"""
Parses a string with attention tokens and returns a list of pairs: text and its associated weight.
Accepted tokens are:
(abc) - increases attention to abc by a multiplier of 1.1
(abc:3.12) - increases attention to abc by a multiplier of 3.12
[abc] - decreases attention to abc by a multiplier of 1.1
\( - literal character '('
\[ - literal character '['
\) - literal character ')'
\] - literal character ']'
\\ - literal character '\'
anything else - just text
>>> parse_prompt_attention('normal text')
[['normal text', 1.0]]
>>> parse_prompt_attention('an (important) word')
[['an ', 1.0], ['important', 1.1], [' word', 1.0]]
>>> parse_prompt_attention('(unbalanced')
[['unbalanced', 1.1]]
>>> parse_prompt_attention('\(literal\]')
[['(literal]', 1.0]]
>>> parse_prompt_attention('(unnecessary)(parens)')
[['unnecessaryparens', 1.1]]
>>> parse_prompt_attention('a (((house:1.3)) [on] a (hill:0.5), sun, (((sky))).')
[['a ', 1.0],
['house', 1.5730000000000004],
[' ', 1.1],
['on', 1.0],
[' a ', 1.1],
['hill', 0.55],
[', sun, ', 1.1],
['sky', 1.4641000000000006],
['.', 1.1]]
"""
res = []
round_brackets = []
square_brackets = []
round_bracket_multiplier = 1.1
square_bracket_multiplier = 1 / 1.1
def multiply_range(start_position, multiplier):
for p in range(start_position, len(res)):
res[p][1] *= multiplier
for m in re_attention.finditer(text):
text = m.group(0)
weight = m.group(1)
if text.startswith("\\"):
res.append([text[1:], 1.0])
elif text == "(":
round_brackets.append(len(res))
elif text == "[":
square_brackets.append(len(res))
elif weight is not None and len(round_brackets) > 0:
multiply_range(round_brackets.pop(), float(weight))
elif text == ")" and len(round_brackets) > 0:
multiply_range(round_brackets.pop(), round_bracket_multiplier)
elif text == "]" and len(square_brackets) > 0:
multiply_range(square_brackets.pop(), square_bracket_multiplier)
else:
res.append([text, 1.0])
for pos in round_brackets:
multiply_range(pos, round_bracket_multiplier)
for pos in square_brackets:
multiply_range(pos, square_bracket_multiplier)
if len(res) == 0:
res = [["", 1.0]]
# merge runs of identical weights
i = 0
while i + 1 < len(res):
if res[i][1] == res[i + 1][1]:
res[i][0] += res[i + 1][0]
res.pop(i + 1)
else:
i += 1
return res
def get_prompts_with_weights(pipe: StableDiffusionPipeline, prompt: List[str], max_length: int):
r"""
Tokenize a list of prompts and return its tokens with weights of each token.
No padding, starting or ending token is included.
"""
tokens = []
weights = []
truncated = False
for text in prompt:
texts_and_weights = parse_prompt_attention(text)
text_token = []
text_weight = []
for word, weight in texts_and_weights:
# tokenize and discard the starting and the ending token
token = pipe.tokenizer(word).input_ids[1:-1]
text_token += token
# copy the weight by length of token
text_weight += [weight] * len(token)
# stop if the text is too long (longer than truncation limit)
if len(text_token) > max_length:
truncated = True
break
# truncate
if len(text_token) > max_length:
truncated = True
text_token = text_token[:max_length]
text_weight = text_weight[:max_length]
tokens.append(text_token)
weights.append(text_weight)
if truncated:
logger.warning("Prompt was truncated. Try to shorten the prompt or increase max_embeddings_multiples")
return tokens, weights
def pad_tokens_and_weights(tokens, weights, max_length, bos, eos, no_boseos_middle=True, chunk_length=77):
r"""
Pad the tokens (with starting and ending tokens) and weights (with 1.0) to max_length.
"""
max_embeddings_multiples = (max_length - 2) // (chunk_length - 2)
weights_length = max_length if no_boseos_middle else max_embeddings_multiples * chunk_length
for i in range(len(tokens)):
tokens[i] = [bos] + tokens[i] + [eos] * (max_length - 1 - len(tokens[i]))
if no_boseos_middle:
weights[i] = [1.0] + weights[i] + [1.0] * (max_length - 1 - len(weights[i]))
else:
w = []
if len(weights[i]) == 0:
w = [1.0] * weights_length
else:
for j in range(max_embeddings_multiples):
w.append(1.0) # weight for starting token in this chunk
w += weights[i][j * (chunk_length - 2) : min(len(weights[i]), (j + 1) * (chunk_length - 2))]
w.append(1.0) # weight for ending token in this chunk
w += [1.0] * (weights_length - len(w))
weights[i] = w[:]
return tokens, weights
def get_unweighted_text_embeddings(
pipe: StableDiffusionPipeline,
text_input: torch.Tensor,
chunk_length: int,
no_boseos_middle: Optional[bool] = True,
):
"""
When the length of tokens is a multiple of the capacity of the text encoder,
it should be split into chunks and sent to the text encoder individually.
"""
max_embeddings_multiples = (text_input.shape[1] - 2) // (chunk_length - 2)
if max_embeddings_multiples > 1:
text_embeddings = []
for i in range(max_embeddings_multiples):
# extract the i-th chunk
text_input_chunk = text_input[:, i * (chunk_length - 2) : (i + 1) * (chunk_length - 2) + 2].clone()
# cover the head and the tail by the starting and the ending tokens
text_input_chunk[:, 0] = text_input[0, 0]
text_input_chunk[:, -1] = text_input[0, -1]
text_embedding = pipe.text_encoder(text_input_chunk)[0]
if no_boseos_middle:
if i == 0:
# discard the ending token
text_embedding = text_embedding[:, :-1]
elif i == max_embeddings_multiples - 1:
# discard the starting token
text_embedding = text_embedding[:, 1:]
else:
# discard both starting and ending tokens
text_embedding = text_embedding[:, 1:-1]
text_embeddings.append(text_embedding)
text_embeddings = torch.concat(text_embeddings, axis=1)
else:
text_embeddings = pipe.text_encoder(text_input)[0]
return text_embeddings
def get_weighted_text_embeddings(
pipe: StableDiffusionPipeline,
prompt: Union[str, List[str]],
uncond_prompt: Optional[Union[str, List[str]]] = None,
max_embeddings_multiples: Optional[int] = 3,
no_boseos_middle: Optional[bool] = False,
skip_parsing: Optional[bool] = False,
skip_weighting: Optional[bool] = False,
**kwargs,
):
r"""
Prompts can be assigned with local weights using brackets. For example,
prompt 'A (very beautiful) masterpiece' highlights the words 'very beautiful',
and the embedding tokens corresponding to the words get multiplied by a constant, 1.1.
Also, to regularize of the embedding, the weighted embedding would be scaled to preserve the original mean.
Args:
pipe (`StableDiffusionPipeline`):
Pipe to provide access to the tokenizer and the text encoder.
prompt (`str` or `List[str]`):
The prompt or prompts to guide the image generation.
uncond_prompt (`str` or `List[str]`):
The unconditional prompt or prompts for guide the image generation. If unconditional prompt
is provided, the embeddings of prompt and uncond_prompt are concatenated.
max_embeddings_multiples (`int`, *optional*, defaults to `3`):
The max multiple length of prompt embeddings compared to the max output length of text encoder.
no_boseos_middle (`bool`, *optional*, defaults to `False`):
If the length of text token is multiples of the capacity of text encoder, whether reserve the starting and
ending token in each of the chunk in the middle.
skip_parsing (`bool`, *optional*, defaults to `False`):
Skip the parsing of brackets.
skip_weighting (`bool`, *optional*, defaults to `False`):
Skip the weighting. When the parsing is skipped, it is forced True.
"""
max_length = (pipe.tokenizer.model_max_length - 2) * max_embeddings_multiples + 2
if isinstance(prompt, str):
prompt = [prompt]
if not skip_parsing:
prompt_tokens, prompt_weights = get_prompts_with_weights(pipe, prompt, max_length - 2)
if uncond_prompt is not None:
if isinstance(uncond_prompt, str):
uncond_prompt = [uncond_prompt]
uncond_tokens, uncond_weights = get_prompts_with_weights(pipe, uncond_prompt, max_length - 2)
else:
prompt_tokens = [
token[1:-1] for token in pipe.tokenizer(prompt, max_length=max_length, truncation=True).input_ids
]
prompt_weights = [[1.0] * len(token) for token in prompt_tokens]
if uncond_prompt is not None:
if isinstance(uncond_prompt, str):
uncond_prompt = [uncond_prompt]
uncond_tokens = [
token[1:-1]
for token in pipe.tokenizer(uncond_prompt, max_length=max_length, truncation=True).input_ids
]
uncond_weights = [[1.0] * len(token) for token in uncond_tokens]
# round up the longest length of tokens to a multiple of (model_max_length - 2)
max_length = max([len(token) for token in prompt_tokens])
if uncond_prompt is not None:
max_length = max(max_length, max([len(token) for token in uncond_tokens]))
max_embeddings_multiples = min(
max_embeddings_multiples,
(max_length - 1) // (pipe.tokenizer.model_max_length - 2) + 1,
)
max_embeddings_multiples = max(1, max_embeddings_multiples)
max_length = (pipe.tokenizer.model_max_length - 2) * max_embeddings_multiples + 2
# pad the length of tokens and weights
bos = pipe.tokenizer.bos_token_id
eos = pipe.tokenizer.eos_token_id
prompt_tokens, prompt_weights = pad_tokens_and_weights(
prompt_tokens,
prompt_weights,
max_length,
bos,
eos,
no_boseos_middle=no_boseos_middle,
chunk_length=pipe.tokenizer.model_max_length,
)
prompt_tokens = torch.tensor(prompt_tokens, dtype=torch.long, device=pipe.device)
if uncond_prompt is not None:
uncond_tokens, uncond_weights = pad_tokens_and_weights(
uncond_tokens,
uncond_weights,
max_length,
bos,
eos,
no_boseos_middle=no_boseos_middle,
chunk_length=pipe.tokenizer.model_max_length,
)
uncond_tokens = torch.tensor(uncond_tokens, dtype=torch.long, device=pipe.device)
# get the embeddings
text_embeddings = get_unweighted_text_embeddings(
pipe,
prompt_tokens,
pipe.tokenizer.model_max_length,
no_boseos_middle=no_boseos_middle,
)
prompt_weights = torch.tensor(prompt_weights, dtype=text_embeddings.dtype, device=pipe.device)
if uncond_prompt is not None:
uncond_embeddings = get_unweighted_text_embeddings(
pipe,
uncond_tokens,
pipe.tokenizer.model_max_length,
no_boseos_middle=no_boseos_middle,
)
uncond_weights = torch.tensor(uncond_weights, dtype=uncond_embeddings.dtype, device=pipe.device)
# assign weights to the prompts and normalize in the sense of mean
# TODO: should we normalize by chunk or in a whole (current implementation)?
if (not skip_parsing) and (not skip_weighting):
previous_mean = text_embeddings.float().mean(axis=[-2, -1]).to(text_embeddings.dtype)
text_embeddings *= prompt_weights.unsqueeze(-1)
current_mean = text_embeddings.float().mean(axis=[-2, -1]).to(text_embeddings.dtype)
text_embeddings *= (previous_mean / current_mean).unsqueeze(-1).unsqueeze(-1)
if uncond_prompt is not None:
previous_mean = uncond_embeddings.float().mean(axis=[-2, -1]).to(uncond_embeddings.dtype)
uncond_embeddings *= uncond_weights.unsqueeze(-1)
current_mean = uncond_embeddings.float().mean(axis=[-2, -1]).to(uncond_embeddings.dtype)
uncond_embeddings *= (previous_mean / current_mean).unsqueeze(-1).unsqueeze(-1)
if uncond_prompt is not None:
return text_embeddings, uncond_embeddings
return text_embeddings, None
def preprocess_image(image):
w, h = image.size
w, h = map(lambda x: x - x % 32, (w, h)) # resize to integer multiple of 32
image = image.resize((w, h), resample=PIL_INTERPOLATION["lanczos"])
image = np.array(image).astype(np.float32) / 255.0
image = image[None].transpose(0, 3, 1, 2)
image = torch.from_numpy(image)
return 2.0 * image - 1.0
def preprocess_mask(mask, scale_factor=8):
mask = mask.convert("L")
w, h = mask.size
w, h = map(lambda x: x - x % 32, (w, h)) # resize to integer multiple of 32
mask = mask.resize((w // scale_factor, h // scale_factor), resample=PIL_INTERPOLATION["nearest"])
mask = np.array(mask).astype(np.float32) / 255.0
mask = np.tile(mask, (4, 1, 1))
mask = mask[None].transpose(0, 1, 2, 3) # what does this step do?
mask = 1 - mask # repaint white, keep black
mask = torch.from_numpy(mask)
return mask
class StableDiffusionLongPromptWeightingPipeline(StableDiffusionPipeline):
r"""
Pipeline for text-to-image generation using Stable Diffusion without tokens length limit, and support parsing
weighting in prompt.
This model inherits from [`DiffusionPipeline`]. Check the superclass documentation for the generic methods the
library implements for all the pipelines (such as downloading or saving, running on a particular device, etc.)
Args:
vae ([`AutoencoderKL`]):
Variational Auto-Encoder (VAE) Model to encode and decode images to and from latent representations.
text_encoder ([`CLIPTextModel`]):
Frozen text-encoder. Stable Diffusion uses the text portion of
[CLIP](https://huggingface.co/docs/transformers/model_doc/clip#transformers.CLIPTextModel), specifically
the [clip-vit-large-patch14](https://huggingface.co/openai/clip-vit-large-patch14) variant.
tokenizer (`CLIPTokenizer`):
Tokenizer of class
[CLIPTokenizer](https://huggingface.co/docs/transformers/v4.21.0/en/model_doc/clip#transformers.CLIPTokenizer).
unet ([`UNet2DConditionModel`]): Conditional U-Net architecture to denoise the encoded image latents.
scheduler ([`SchedulerMixin`]):
A scheduler to be used in combination with `unet` to denoise the encoded image latents. Can be one of
[`DDIMScheduler`], [`LMSDiscreteScheduler`], or [`PNDMScheduler`].
safety_checker ([`StableDiffusionSafetyChecker`]):
Classification module that estimates whether generated images could be considered offensive or harmful.
Please, refer to the [model card](https://huggingface.co/CompVis/stable-diffusion-v1-4) for details.
feature_extractor ([`CLIPFeatureExtractor`]):
Model that extracts features from generated images to be used as inputs for the `safety_checker`.
"""
if version.parse(version.parse(diffusers.__version__).base_version) >= version.parse("0.9.0"):
def __init__(
self,
vae: AutoencoderKL,
text_encoder: CLIPTextModel,
tokenizer: CLIPTokenizer,
unet: UNet2DConditionModel,
scheduler: SchedulerMixin,
safety_checker: StableDiffusionSafetyChecker,
feature_extractor: CLIPFeatureExtractor,
requires_safety_checker: bool = True,
):
super().__init__(
vae=vae,
text_encoder=text_encoder,
tokenizer=tokenizer,
unet=unet,
scheduler=scheduler,
safety_checker=safety_checker,
feature_extractor=feature_extractor,
requires_safety_checker=requires_safety_checker,
)
self.__init__additional__()
else:
def __init__(
self,
vae: AutoencoderKL,
text_encoder: CLIPTextModel,
tokenizer: CLIPTokenizer,
unet: UNet2DConditionModel,
scheduler: SchedulerMixin,
safety_checker: StableDiffusionSafetyChecker,
feature_extractor: CLIPFeatureExtractor,
):
super().__init__(
vae=vae,
text_encoder=text_encoder,
tokenizer=tokenizer,
unet=unet,
scheduler=scheduler,
safety_checker=safety_checker,
feature_extractor=feature_extractor,
)
self.__init__additional__()
def __init__additional__(self):
if not hasattr(self, "vae_scale_factor"):
setattr(self, "vae_scale_factor", 2 ** (len(self.vae.config.block_out_channels) - 1))
@property
def _execution_device(self):
r"""
Returns the device on which the pipeline's models will be executed. After calling
`pipeline.enable_sequential_cpu_offload()` the execution device can only be inferred from Accelerate's module
hooks.
"""
if self.device != torch.device("meta") or not hasattr(self.unet, "_hf_hook"):
return self.device
for module in self.unet.modules():
if (
hasattr(module, "_hf_hook")
and hasattr(module._hf_hook, "execution_device")
and module._hf_hook.execution_device is not None
):
return torch.device(module._hf_hook.execution_device)
return self.device
def _encode_prompt(
self,
prompt,
device,
num_images_per_prompt,
do_classifier_free_guidance,
negative_prompt,
max_embeddings_multiples,
):
r"""
Encodes the prompt into text encoder hidden states.
Args:
prompt (`str` or `list(int)`):
prompt to be encoded
device: (`torch.device`):
torch device
num_images_per_prompt (`int`):
number of images that should be generated per prompt
do_classifier_free_guidance (`bool`):
whether to use classifier free guidance or not
negative_prompt (`str` or `List[str]`):
The prompt or prompts not to guide the image generation. Ignored when not using guidance (i.e., ignored
if `guidance_scale` is less than `1`).
max_embeddings_multiples (`int`, *optional*, defaults to `3`):
The max multiple length of prompt embeddings compared to the max output length of text encoder.
"""
batch_size = len(prompt) if isinstance(prompt, list) else 1
if negative_prompt is None:
negative_prompt = [""] * batch_size
elif isinstance(negative_prompt, str):
negative_prompt = [negative_prompt] * batch_size
if batch_size != len(negative_prompt):
raise ValueError(
f"`negative_prompt`: {negative_prompt} has batch size {len(negative_prompt)}, but `prompt`:"
f" {prompt} has batch size {batch_size}. Please make sure that passed `negative_prompt` matches"
" the batch size of `prompt`."
)
text_embeddings, uncond_embeddings = get_weighted_text_embeddings(
pipe=self,
prompt=prompt,
uncond_prompt=negative_prompt if do_classifier_free_guidance else None,
max_embeddings_multiples=max_embeddings_multiples,
)
bs_embed, seq_len, _ = text_embeddings.shape
text_embeddings = text_embeddings.repeat(1, num_images_per_prompt, 1)
text_embeddings = text_embeddings.view(bs_embed * num_images_per_prompt, seq_len, -1)
if do_classifier_free_guidance:
bs_embed, seq_len, _ = uncond_embeddings.shape
uncond_embeddings = uncond_embeddings.repeat(1, num_images_per_prompt, 1)
uncond_embeddings = uncond_embeddings.view(bs_embed * num_images_per_prompt, seq_len, -1)
text_embeddings = torch.cat([uncond_embeddings, text_embeddings])
return text_embeddings
def check_inputs(self, prompt, height, width, strength, callback_steps):
if not isinstance(prompt, str) and not isinstance(prompt, list):
raise ValueError(f"`prompt` has to be of type `str` or `list` but is {type(prompt)}")
if strength < 0 or strength > 1:
raise ValueError(f"The value of strength should in [0.0, 1.0] but is {strength}")
if height % 8 != 0 or width % 8 != 0:
raise ValueError(f"`height` and `width` have to be divisible by 8 but are {height} and {width}.")
if (callback_steps is None) or (
callback_steps is not None and (not isinstance(callback_steps, int) or callback_steps <= 0)
):
raise ValueError(
f"`callback_steps` has to be a positive integer but is {callback_steps} of type"
f" {type(callback_steps)}."
)
def get_timesteps(self, num_inference_steps, strength, device, is_text2img):
if is_text2img:
return self.scheduler.timesteps.to(device), num_inference_steps
else:
# get the original timestep using init_timestep
offset = self.scheduler.config.get("steps_offset", 0)
init_timestep = int(num_inference_steps * strength) + offset
init_timestep = min(init_timestep, num_inference_steps)
t_start = max(num_inference_steps - init_timestep + offset, 0)
timesteps = self.scheduler.timesteps[t_start:].to(device)
return timesteps, num_inference_steps - t_start
def run_safety_checker(self, image, device, dtype):
if self.safety_checker is not None:
safety_checker_input = self.feature_extractor(self.numpy_to_pil(image), return_tensors="pt").to(device)
image, has_nsfw_concept = self.safety_checker(
images=image, clip_input=safety_checker_input.pixel_values.to(dtype)
)
else:
has_nsfw_concept = None
return image, has_nsfw_concept
def decode_latents(self, latents):
latents = 1 / 0.18215 * latents
image = self.vae.decode(latents).sample
image = (image / 2 + 0.5).clamp(0, 1)
# we always cast to float32 as this does not cause significant overhead and is compatible with bfloa16
image = image.cpu().permute(0, 2, 3, 1).float().numpy()
return image
def prepare_extra_step_kwargs(self, generator, eta):
# prepare extra kwargs for the scheduler step, since not all schedulers have the same signature
# eta (η) is only used with the DDIMScheduler, it will be ignored for other schedulers.
# eta corresponds to η in DDIM paper: https://arxiv.org/abs/2010.02502
# and should be between [0, 1]
accepts_eta = "eta" in set(inspect.signature(self.scheduler.step).parameters.keys())
extra_step_kwargs = {}
if accepts_eta:
extra_step_kwargs["eta"] = eta
# check if the scheduler accepts generator
accepts_generator = "generator" in set(inspect.signature(self.scheduler.step).parameters.keys())
if accepts_generator:
extra_step_kwargs["generator"] = generator
return extra_step_kwargs
def prepare_latents(self, image, timestep, batch_size, height, width, dtype, device, generator, latents=None):
if image is None:
shape = (
batch_size,
self.unet.in_channels,
height // self.vae_scale_factor,
width // self.vae_scale_factor,
)
if latents is None:
# randn does not work reproducibly on mps
rand_device = "cpu" if device.type == "mps" else device
if isinstance(generator, list):
shape = (1,) + shape[1:]
latents = [
torch.randn(shape, generator=generator[i], device=rand_device, dtype=dtype)
for i in range(batch_size)
]
latents = torch.cat(latents, dim=0).to(device)
else:
latents = torch.randn(shape, generator=generator, device=rand_device, dtype=dtype).to(device)
else:
if latents.shape != shape:
raise ValueError(f"Unexpected latents shape, got {latents.shape}, expected {shape}")
latents = latents.to(device)
# scale the initial noise by the standard deviation required by the scheduler
latents = latents * self.scheduler.init_noise_sigma
return latents, None, None
else:
init_latent_dist = self.vae.encode(image).latent_dist
init_latents = init_latent_dist.sample(generator=generator)
init_latents = 0.18215 * init_latents
init_latents = torch.cat([init_latents] * batch_size, dim=0)
init_latents_orig = init_latents
shape = init_latents.shape
# add noise to latents using the timesteps
if device.type == "mps":
noise = torch.randn(shape, generator=generator, device="cpu", dtype=dtype).to(device)
else:
noise = torch.randn(shape, generator=generator, device=device, dtype=dtype)
latents = self.scheduler.add_noise(init_latents, noise, timestep)
return latents, init_latents_orig, noise
@torch.no_grad()
def __call__(
self,
prompt: Union[str, List[str]],
negative_prompt: Optional[Union[str, List[str]]] = None,
image: Union[torch.FloatTensor, PIL.Image.Image] = None,
mask_image: Union[torch.FloatTensor, PIL.Image.Image] = None,
height: int = 512,
width: int = 512,
num_inference_steps: int = 50,
guidance_scale: float = 7.5,
strength: float = 0.8,
num_images_per_prompt: Optional[int] = 1,
eta: float = 0.0,
generator: Optional[torch.Generator] = None,
latents: Optional[torch.FloatTensor] = None,
max_embeddings_multiples: Optional[int] = 3,
output_type: Optional[str] = "pil",
return_dict: bool = True,
callback: Optional[Callable[[int, int, torch.FloatTensor], None]] = None,
is_cancelled_callback: Optional[Callable[[], bool]] = None,
callback_steps: Optional[int] = 1,
**kwargs,
):
r"""
Function invoked when calling the pipeline for generation.
Args:
prompt (`str` or `List[str]`):
The prompt or prompts to guide the image generation.
negative_prompt (`str` or `List[str]`, *optional*):
The prompt or prompts not to guide the image generation. Ignored when not using guidance (i.e., ignored
if `guidance_scale` is less than `1`).
image (`torch.FloatTensor` or `PIL.Image.Image`):
`Image`, or tensor representing an image batch, that will be used as the starting point for the
process.
mask_image (`torch.FloatTensor` or `PIL.Image.Image`):
`Image`, or tensor representing an image batch, to mask `image`. White pixels in the mask will be
replaced by noise and therefore repainted, while black pixels will be preserved. If `mask_image` is a
PIL image, it will be converted to a single channel (luminance) before use. If it's a tensor, it should
contain one color channel (L) instead of 3, so the expected shape would be `(B, H, W, 1)`.
height (`int`, *optional*, defaults to 512):
The height in pixels of the generated image.
width (`int`, *optional*, defaults to 512):
The width in pixels of the generated image.
num_inference_steps (`int`, *optional*, defaults to 50):
The number of denoising steps. More denoising steps usually lead to a higher quality image at the
expense of slower inference.
guidance_scale (`float`, *optional*, defaults to 7.5):
Guidance scale as defined in [Classifier-Free Diffusion Guidance](https://arxiv.org/abs/2207.12598).
`guidance_scale` is defined as `w` of equation 2. of [Imagen
Paper](https://arxiv.org/pdf/2205.11487.pdf). Guidance scale is enabled by setting `guidance_scale >
1`. Higher guidance scale encourages to generate images that are closely linked to the text `prompt`,
usually at the expense of lower image quality.
strength (`float`, *optional*, defaults to 0.8):
Conceptually, indicates how much to transform the reference `image`. Must be between 0 and 1.
`image` will be used as a starting point, adding more noise to it the larger the `strength`. The
number of denoising steps depends on the amount of noise initially added. When `strength` is 1, added
noise will be maximum and the denoising process will run for the full number of iterations specified in
`num_inference_steps`. A value of 1, therefore, essentially ignores `image`.
num_images_per_prompt (`int`, *optional*, defaults to 1):
The number of images to generate per prompt.
eta (`float`, *optional*, defaults to 0.0):
Corresponds to parameter eta (η) in the DDIM paper: https://arxiv.org/abs/2010.02502. Only applies to
[`schedulers.DDIMScheduler`], will be ignored for others.
generator (`torch.Generator`, *optional*):
A [torch generator](https://pytorch.org/docs/stable/generated/torch.Generator.html) to make generation
deterministic.
latents (`torch.FloatTensor`, *optional*):
Pre-generated noisy latents, sampled from a Gaussian distribution, to be used as inputs for image
generation. Can be used to tweak the same generation with different prompts. If not provided, a latents
tensor will ge generated by sampling using the supplied random `generator`.
max_embeddings_multiples (`int`, *optional*, defaults to `3`):
The max multiple length of prompt embeddings compared to the max output length of text encoder.
output_type (`str`, *optional*, defaults to `"pil"`):
The output format of the generate image. Choose between
[PIL](https://pillow.readthedocs.io/en/stable/): `PIL.Image.Image` or `np.array`.
return_dict (`bool`, *optional*, defaults to `True`):
Whether or not to return a [`~pipelines.stable_diffusion.StableDiffusionPipelineOutput`] instead of a
plain tuple.
callback (`Callable`, *optional*):
A function that will be called every `callback_steps` steps during inference. The function will be
called with the following arguments: `callback(step: int, timestep: int, latents: torch.FloatTensor)`.
is_cancelled_callback (`Callable`, *optional*):
A function that will be called every `callback_steps` steps during inference. If the function returns
`True`, the inference will be cancelled.
callback_steps (`int`, *optional*, defaults to 1):
The frequency at which the `callback` function will be called. If not specified, the callback will be
called at every step.
Returns:
`None` if cancelled by `is_cancelled_callback`,
[`~pipelines.stable_diffusion.StableDiffusionPipelineOutput`] or `tuple`:
[`~pipelines.stable_diffusion.StableDiffusionPipelineOutput`] if `return_dict` is True, otherwise a `tuple.
When returning a tuple, the first element is a list with the generated images, and the second element is a
list of `bool`s denoting whether the corresponding generated image likely represents "not-safe-for-work"
(nsfw) content, according to the `safety_checker`.
"""
# 0. Default height and width to unet
height = height or self.unet.config.sample_size * self.vae_scale_factor
width = width or self.unet.config.sample_size * self.vae_scale_factor
# 1. Check inputs. Raise error if not correct
self.check_inputs(prompt, height, width, strength, callback_steps)
# 2. Define call parameters
batch_size = 1 if isinstance(prompt, str) else len(prompt)
device = self._execution_device
# here `guidance_scale` is defined analog to the guidance weight `w` of equation (2)
# of the Imagen paper: https://arxiv.org/pdf/2205.11487.pdf . `guidance_scale = 1`
# corresponds to doing no classifier free guidance.
do_classifier_free_guidance = guidance_scale > 1.0
# 3. Encode input prompt
text_embeddings = self._encode_prompt(
prompt,
device,
num_images_per_prompt,
do_classifier_free_guidance,
negative_prompt,
max_embeddings_multiples,
)
dtype = text_embeddings.dtype
# 4. Preprocess image and mask
if isinstance(image, PIL.Image.Image):
image = preprocess_image(image)
if image is not None:
image = image.to(device=self.device, dtype=dtype)
if isinstance(mask_image, PIL.Image.Image):
mask_image = preprocess_mask(mask_image, self.vae_scale_factor)
if mask_image is not None:
mask = mask_image.to(device=self.device, dtype=dtype)
mask = torch.cat([mask] * batch_size * num_images_per_prompt)
else:
mask = None
# 5. set timesteps
self.scheduler.set_timesteps(num_inference_steps, device=device)
timesteps, num_inference_steps = self.get_timesteps(num_inference_steps, strength, device, image is None)
latent_timestep = timesteps[:1].repeat(batch_size * num_images_per_prompt)
# 6. Prepare latent variables
latents, init_latents_orig, noise = self.prepare_latents(
image,
latent_timestep,
batch_size * num_images_per_prompt,
height,
width,
dtype,
device,
generator,
latents,
)
# 7. Prepare extra step kwargs. TODO: Logic should ideally just be moved out of the pipeline
extra_step_kwargs = self.prepare_extra_step_kwargs(generator, eta)
# 8. Denoising loop
for i, t in enumerate(self.progress_bar(timesteps)):
# expand the latents if we are doing classifier free guidance
latent_model_input = torch.cat([latents] * 2) if do_classifier_free_guidance else latents
latent_model_input = self.scheduler.scale_model_input(latent_model_input, t)
# predict the noise residual
noise_pred = self.unet(latent_model_input, t, encoder_hidden_states=text_embeddings).sample
# perform guidance
if do_classifier_free_guidance:
noise_pred_uncond, noise_pred_text = noise_pred.chunk(2)
noise_pred = noise_pred_uncond + guidance_scale * (noise_pred_text - noise_pred_uncond)
# compute the previous noisy sample x_t -> x_t-1
latents = self.scheduler.step(noise_pred, t, latents, **extra_step_kwargs).prev_sample
if mask is not None:
# masking
init_latents_proper = self.scheduler.add_noise(init_latents_orig, noise, torch.tensor([t]))
latents = (init_latents_proper * mask) + (latents * (1 - mask))
# call the callback, if provided
if i % callback_steps == 0:
if callback is not None:
callback(i, t, latents)
if is_cancelled_callback is not None and is_cancelled_callback():
return None
# 9. Post-processing
image = self.decode_latents(latents)
# 10. Run safety checker
image, has_nsfw_concept = self.run_safety_checker(image, device, text_embeddings.dtype)
# 11. Convert to PIL
if output_type == "pil":
image = self.numpy_to_pil(image)
if not return_dict:
return image, has_nsfw_concept
return StableDiffusionPipelineOutput(images=image, nsfw_content_detected=has_nsfw_concept)
def text2img(
self,
prompt: Union[str, List[str]],
negative_prompt: Optional[Union[str, List[str]]] = None,
height: int = 512,
width: int = 512,
num_inference_steps: int = 50,
guidance_scale: float = 7.5,
num_images_per_prompt: Optional[int] = 1,
eta: float = 0.0,
generator: Optional[torch.Generator] = None,
latents: Optional[torch.FloatTensor] = None,
max_embeddings_multiples: Optional[int] = 3,
output_type: Optional[str] = "pil",
return_dict: bool = True,
callback: Optional[Callable[[int, int, torch.FloatTensor], None]] = None,
is_cancelled_callback: Optional[Callable[[], bool]] = None,
callback_steps: Optional[int] = 1,
**kwargs,
):
r"""
Function for text-to-image generation.
Args:
prompt (`str` or `List[str]`):
The prompt or prompts to guide the image generation.
negative_prompt (`str` or `List[str]`, *optional*):
The prompt or prompts not to guide the image generation. Ignored when not using guidance (i.e., ignored
if `guidance_scale` is less than `1`).
height (`int`, *optional*, defaults to 512):
The height in pixels of the generated image.
width (`int`, *optional*, defaults to 512):
The width in pixels of the generated image.
num_inference_steps (`int`, *optional*, defaults to 50):
The number of denoising steps. More denoising steps usually lead to a higher quality image at the
expense of slower inference.
guidance_scale (`float`, *optional*, defaults to 7.5):
Guidance scale as defined in [Classifier-Free Diffusion Guidance](https://arxiv.org/abs/2207.12598).
`guidance_scale` is defined as `w` of equation 2. of [Imagen
Paper](https://arxiv.org/pdf/2205.11487.pdf). Guidance scale is enabled by setting `guidance_scale >
1`. Higher guidance scale encourages to generate images that are closely linked to the text `prompt`,
usually at the expense of lower image quality.
num_images_per_prompt (`int`, *optional*, defaults to 1):
The number of images to generate per prompt.
eta (`float`, *optional*, defaults to 0.0):
Corresponds to parameter eta (η) in the DDIM paper: https://arxiv.org/abs/2010.02502. Only applies to
[`schedulers.DDIMScheduler`], will be ignored for others.
generator (`torch.Generator`, *optional*):
A [torch generator](https://pytorch.org/docs/stable/generated/torch.Generator.html) to make generation
deterministic.
latents (`torch.FloatTensor`, *optional*):
Pre-generated noisy latents, sampled from a Gaussian distribution, to be used as inputs for image
generation. Can be used to tweak the same generation with different prompts. If not provided, a latents
tensor will ge generated by sampling using the supplied random `generator`.
max_embeddings_multiples (`int`, *optional*, defaults to `3`):
The max multiple length of prompt embeddings compared to the max output length of text encoder.
output_type (`str`, *optional*, defaults to `"pil"`):
The output format of the generate image. Choose between
[PIL](https://pillow.readthedocs.io/en/stable/): `PIL.Image.Image` or `np.array`.
return_dict (`bool`, *optional*, defaults to `True`):
Whether or not to return a [`~pipelines.stable_diffusion.StableDiffusionPipelineOutput`] instead of a
plain tuple.
callback (`Callable`, *optional*):
A function that will be called every `callback_steps` steps during inference. The function will be
called with the following arguments: `callback(step: int, timestep: int, latents: torch.FloatTensor)`.
is_cancelled_callback (`Callable`, *optional*):
A function that will be called every `callback_steps` steps during inference. If the function returns
`True`, the inference will be cancelled.
callback_steps (`int`, *optional*, defaults to 1):
The frequency at which the `callback` function will be called. If not specified, the callback will be
called at every step.
Returns:
[`~pipelines.stable_diffusion.StableDiffusionPipelineOutput`] or `tuple`:
[`~pipelines.stable_diffusion.StableDiffusionPipelineOutput`] if `return_dict` is True, otherwise a `tuple.
When returning a tuple, the first element is a list with the generated images, and the second element is a
list of `bool`s denoting whether the corresponding generated image likely represents "not-safe-for-work"
(nsfw) content, according to the `safety_checker`.
"""
return self.__call__(
prompt=prompt,
negative_prompt=negative_prompt,
height=height,
width=width,
num_inference_steps=num_inference_steps,
guidance_scale=guidance_scale,
num_images_per_prompt=num_images_per_prompt,
eta=eta,
generator=generator,
latents=latents,
max_embeddings_multiples=max_embeddings_multiples,
output_type=output_type,
return_dict=return_dict,
callback=callback,
is_cancelled_callback=is_cancelled_callback,
callback_steps=callback_steps,
**kwargs,
)
def img2img(
self,
image: Union[torch.FloatTensor, PIL.Image.Image],
prompt: Union[str, List[str]],
negative_prompt: Optional[Union[str, List[str]]] = None,
strength: float = 0.8,
num_inference_steps: Optional[int] = 50,
guidance_scale: Optional[float] = 7.5,
num_images_per_prompt: Optional[int] = 1,
eta: Optional[float] = 0.0,
generator: Optional[torch.Generator] = None,
max_embeddings_multiples: Optional[int] = 3,
output_type: Optional[str] = "pil",
return_dict: bool = True,
callback: Optional[Callable[[int, int, torch.FloatTensor], None]] = None,
is_cancelled_callback: Optional[Callable[[], bool]] = None,
callback_steps: Optional[int] = 1,
**kwargs,
):
r"""
Function for image-to-image generation.
Args:
image (`torch.FloatTensor` or `PIL.Image.Image`):
`Image`, or tensor representing an image batch, that will be used as the starting point for the
process.
prompt (`str` or `List[str]`):
The prompt or prompts to guide the image generation.
negative_prompt (`str` or `List[str]`, *optional*):
The prompt or prompts not to guide the image generation. Ignored when not using guidance (i.e., ignored
if `guidance_scale` is less than `1`).
strength (`float`, *optional*, defaults to 0.8):
Conceptually, indicates how much to transform the reference `image`. Must be between 0 and 1.
`image` will be used as a starting point, adding more noise to it the larger the `strength`. The