Spaces:
Running
on
Zero
Running
on
Zero
File size: 79,654 Bytes
0324143 |
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 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262 1263 1264 1265 1266 1267 1268 1269 1270 1271 1272 1273 1274 1275 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285 1286 1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337 1338 1339 1340 1341 1342 1343 1344 1345 1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365 1366 1367 1368 1369 1370 1371 1372 1373 1374 1375 1376 1377 1378 1379 1380 1381 1382 1383 1384 1385 1386 1387 1388 1389 1390 1391 1392 1393 1394 1395 1396 1397 1398 1399 1400 1401 1402 1403 1404 1405 1406 1407 1408 1409 1410 1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421 1422 1423 1424 1425 1426 1427 1428 1429 1430 1431 1432 1433 1434 1435 1436 1437 1438 1439 1440 1441 1442 1443 1444 1445 1446 1447 1448 1449 1450 1451 1452 1453 1454 1455 1456 1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467 1468 1469 1470 1471 1472 1473 1474 1475 1476 1477 1478 1479 1480 1481 1482 1483 1484 1485 1486 1487 1488 1489 1490 1491 1492 1493 1494 1495 1496 1497 1498 1499 1500 1501 1502 1503 1504 1505 1506 1507 1508 1509 1510 1511 1512 1513 1514 1515 1516 1517 1518 1519 1520 1521 1522 1523 1524 1525 1526 1527 1528 1529 1530 1531 1532 1533 |
# Copyright 2024 DiffEdit Authors and Pix2Pix Zero Authors and The HuggingFace Team. All rights reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
import inspect
from dataclasses import dataclass
from typing import Any, Callable, Dict, List, Optional, Union
import numpy as np
import PIL.Image
import torch
from packaging import version
from transformers import CLIPImageProcessor, CLIPTextModel, CLIPTokenizer
from ...configuration_utils import FrozenDict
from ...image_processor import VaeImageProcessor
from ...loaders import LoraLoaderMixin, TextualInversionLoaderMixin
from ...models import AutoencoderKL, UNet2DConditionModel
from ...models.lora import adjust_lora_scale_text_encoder
from ...schedulers import DDIMInverseScheduler, KarrasDiffusionSchedulers
from ...utils import (
PIL_INTERPOLATION,
USE_PEFT_BACKEND,
BaseOutput,
deprecate,
logging,
replace_example_docstring,
scale_lora_layers,
unscale_lora_layers,
)
from ...utils.torch_utils import randn_tensor
from ..pipeline_utils import DiffusionPipeline, StableDiffusionMixin
from ..stable_diffusion import StableDiffusionPipelineOutput
from ..stable_diffusion.safety_checker import StableDiffusionSafetyChecker
logger = logging.get_logger(__name__) # pylint: disable=invalid-name
@dataclass
class DiffEditInversionPipelineOutput(BaseOutput):
"""
Output class for Stable Diffusion pipelines.
Args:
latents (`torch.FloatTensor`)
inverted latents tensor
images (`List[PIL.Image.Image]` or `np.ndarray`)
List of denoised PIL images of length `num_timesteps * batch_size` or numpy array of shape `(num_timesteps,
batch_size, height, width, num_channels)`. PIL images or numpy array present the denoised images of the
diffusion pipeline.
"""
latents: torch.FloatTensor
images: Union[List[PIL.Image.Image], np.ndarray]
EXAMPLE_DOC_STRING = """
```py
>>> import PIL
>>> import requests
>>> import torch
>>> from io import BytesIO
>>> from diffusers import StableDiffusionDiffEditPipeline
>>> def download_image(url):
... response = requests.get(url)
... return PIL.Image.open(BytesIO(response.content)).convert("RGB")
>>> img_url = "https://github.com/Xiang-cd/DiffEdit-stable-diffusion/raw/main/assets/origin.png"
>>> init_image = download_image(img_url).resize((768, 768))
>>> pipe = StableDiffusionDiffEditPipeline.from_pretrained(
... "stabilityai/stable-diffusion-2-1", torch_dtype=torch.float16
... )
>>> pipe = pipe.to("cuda")
>>> pipeline.scheduler = DDIMScheduler.from_config(pipeline.scheduler.config)
>>> pipeline.inverse_scheduler = DDIMInverseScheduler.from_config(pipeline.scheduler.config)
>>> pipeline.enable_model_cpu_offload()
>>> mask_prompt = "A bowl of fruits"
>>> prompt = "A bowl of pears"
>>> mask_image = pipe.generate_mask(image=init_image, source_prompt=prompt, target_prompt=mask_prompt)
>>> image_latents = pipe.invert(image=init_image, prompt=mask_prompt).latents
>>> image = pipe(prompt=prompt, mask_image=mask_image, image_latents=image_latents).images[0]
```
"""
EXAMPLE_INVERT_DOC_STRING = """
```py
>>> import PIL
>>> import requests
>>> import torch
>>> from io import BytesIO
>>> from diffusers import StableDiffusionDiffEditPipeline
>>> def download_image(url):
... response = requests.get(url)
... return PIL.Image.open(BytesIO(response.content)).convert("RGB")
>>> img_url = "https://github.com/Xiang-cd/DiffEdit-stable-diffusion/raw/main/assets/origin.png"
>>> init_image = download_image(img_url).resize((768, 768))
>>> pipe = StableDiffusionDiffEditPipeline.from_pretrained(
... "stabilityai/stable-diffusion-2-1", torch_dtype=torch.float16
... )
>>> pipe = pipe.to("cuda")
>>> pipeline.scheduler = DDIMScheduler.from_config(pipeline.scheduler.config)
>>> pipeline.inverse_scheduler = DDIMInverseScheduler.from_config(pipeline.scheduler.config)
>>> pipeline.enable_model_cpu_offload()
>>> prompt = "A bowl of fruits"
>>> inverted_latents = pipe.invert(image=init_image, prompt=prompt).latents
```
"""
def auto_corr_loss(hidden_states, generator=None):
reg_loss = 0.0
for i in range(hidden_states.shape[0]):
for j in range(hidden_states.shape[1]):
noise = hidden_states[i : i + 1, j : j + 1, :, :]
while True:
roll_amount = torch.randint(noise.shape[2] // 2, (1,), generator=generator).item()
reg_loss += (noise * torch.roll(noise, shifts=roll_amount, dims=2)).mean() ** 2
reg_loss += (noise * torch.roll(noise, shifts=roll_amount, dims=3)).mean() ** 2
if noise.shape[2] <= 8:
break
noise = torch.nn.functional.avg_pool2d(noise, kernel_size=2)
return reg_loss
def kl_divergence(hidden_states):
return hidden_states.var() + hidden_states.mean() ** 2 - 1 - torch.log(hidden_states.var() + 1e-7)
# Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion_img2img.preprocess
def preprocess(image):
deprecation_message = "The preprocess method is deprecated and will be removed in diffusers 1.0.0. Please use VaeImageProcessor.preprocess(...) instead"
deprecate("preprocess", "1.0.0", deprecation_message, standard_warn=False)
if isinstance(image, torch.Tensor):
return image
elif isinstance(image, PIL.Image.Image):
image = [image]
if isinstance(image[0], PIL.Image.Image):
w, h = image[0].size
w, h = (x - x % 8 for x in (w, h)) # resize to integer multiple of 8
image = [np.array(i.resize((w, h), resample=PIL_INTERPOLATION["lanczos"]))[None, :] for i in image]
image = np.concatenate(image, axis=0)
image = np.array(image).astype(np.float32) / 255.0
image = image.transpose(0, 3, 1, 2)
image = 2.0 * image - 1.0
image = torch.from_numpy(image)
elif isinstance(image[0], torch.Tensor):
image = torch.cat(image, dim=0)
return image
def preprocess_mask(mask, batch_size: int = 1):
if not isinstance(mask, torch.Tensor):
# preprocess mask
if isinstance(mask, PIL.Image.Image) or isinstance(mask, np.ndarray):
mask = [mask]
if isinstance(mask, list):
if isinstance(mask[0], PIL.Image.Image):
mask = [np.array(m.convert("L")).astype(np.float32) / 255.0 for m in mask]
if isinstance(mask[0], np.ndarray):
mask = np.stack(mask, axis=0) if mask[0].ndim < 3 else np.concatenate(mask, axis=0)
mask = torch.from_numpy(mask)
elif isinstance(mask[0], torch.Tensor):
mask = torch.stack(mask, dim=0) if mask[0].ndim < 3 else torch.cat(mask, dim=0)
# Batch and add channel dim for single mask
if mask.ndim == 2:
mask = mask.unsqueeze(0).unsqueeze(0)
# Batch single mask or add channel dim
if mask.ndim == 3:
# Single batched mask, no channel dim or single mask not batched but channel dim
if mask.shape[0] == 1:
mask = mask.unsqueeze(0)
# Batched masks no channel dim
else:
mask = mask.unsqueeze(1)
# Check mask shape
if batch_size > 1:
if mask.shape[0] == 1:
mask = torch.cat([mask] * batch_size)
elif mask.shape[0] > 1 and mask.shape[0] != batch_size:
raise ValueError(
f"`mask_image` with batch size {mask.shape[0]} cannot be broadcasted to batch size {batch_size} "
f"inferred by prompt inputs"
)
if mask.shape[1] != 1:
raise ValueError(f"`mask_image` must have 1 channel, but has {mask.shape[1]} channels")
# Check mask is in [0, 1]
if mask.min() < 0 or mask.max() > 1:
raise ValueError("`mask_image` should be in [0, 1] range")
# Binarize mask
mask[mask < 0.5] = 0
mask[mask >= 0.5] = 1
return mask
class StableDiffusionDiffEditPipeline(
DiffusionPipeline, StableDiffusionMixin, TextualInversionLoaderMixin, LoraLoaderMixin
):
r"""
<Tip warning={true}>
This is an experimental feature!
</Tip>
Pipeline for text-guided image inpainting using Stable Diffusion and DiffEdit.
This model inherits from [`DiffusionPipeline`]. Check the superclass documentation for the generic methods
implemented for all pipelines (downloading, saving, running on a particular device, etc.).
The pipeline also inherits the following loading and saving methods:
- [`~loaders.TextualInversionLoaderMixin.load_textual_inversion`] for loading textual inversion embeddings
- [`~loaders.LoraLoaderMixin.load_lora_weights`] for loading LoRA weights
- [`~loaders.LoraLoaderMixin.save_lora_weights`] for saving LoRA weights
Args:
vae ([`AutoencoderKL`]):
Variational Auto-Encoder (VAE) model to encode and decode images to and from latent representations.
text_encoder ([`~transformers.CLIPTextModel`]):
Frozen text-encoder ([clip-vit-large-patch14](https://huggingface.co./openai/clip-vit-large-patch14)).
tokenizer ([`~transformers.CLIPTokenizer`]):
A `CLIPTokenizer` to tokenize text.
unet ([`UNet2DConditionModel`]):
A `UNet2DConditionModel` to denoise the encoded image latents.
scheduler ([`SchedulerMixin`]):
A scheduler to be used in combination with `unet` to denoise the encoded image latents.
inverse_scheduler ([`DDIMInverseScheduler`]):
A scheduler to be used in combination with `unet` to fill in the unmasked part of the input latents.
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./runwayml/stable-diffusion-v1-5) for more details
about a model's potential harms.
feature_extractor ([`~transformers.CLIPImageProcessor`]):
A `CLIPImageProcessor` to extract features from generated images; used as inputs to the `safety_checker`.
"""
model_cpu_offload_seq = "text_encoder->unet->vae"
_optional_components = ["safety_checker", "feature_extractor", "inverse_scheduler"]
_exclude_from_cpu_offload = ["safety_checker"]
def __init__(
self,
vae: AutoencoderKL,
text_encoder: CLIPTextModel,
tokenizer: CLIPTokenizer,
unet: UNet2DConditionModel,
scheduler: KarrasDiffusionSchedulers,
safety_checker: StableDiffusionSafetyChecker,
feature_extractor: CLIPImageProcessor,
inverse_scheduler: DDIMInverseScheduler,
requires_safety_checker: bool = True,
):
super().__init__()
if hasattr(scheduler.config, "steps_offset") and scheduler.config.steps_offset != 1:
deprecation_message = (
f"The configuration file of this scheduler: {scheduler} is outdated. `steps_offset`"
f" should be set to 1 instead of {scheduler.config.steps_offset}. Please make sure "
"to update the config accordingly as leaving `steps_offset` might led to incorrect results"
" in future versions. If you have downloaded this checkpoint from the Hugging Face Hub,"
" it would be very nice if you could open a Pull request for the `scheduler/scheduler_config.json`"
" file"
)
deprecate("steps_offset!=1", "1.0.0", deprecation_message, standard_warn=False)
new_config = dict(scheduler.config)
new_config["steps_offset"] = 1
scheduler._internal_dict = FrozenDict(new_config)
if hasattr(scheduler.config, "skip_prk_steps") and scheduler.config.skip_prk_steps is False:
deprecation_message = (
f"The configuration file of this scheduler: {scheduler} has not set the configuration"
" `skip_prk_steps`. `skip_prk_steps` should be set to True in the configuration file. Please make"
" sure to update the config accordingly as not setting `skip_prk_steps` in the config might lead to"
" incorrect results in future versions. If you have downloaded this checkpoint from the Hugging Face"
" Hub, it would be very nice if you could open a Pull request for the"
" `scheduler/scheduler_config.json` file"
)
deprecate("skip_prk_steps not set", "1.0.0", deprecation_message, standard_warn=False)
new_config = dict(scheduler.config)
new_config["skip_prk_steps"] = True
scheduler._internal_dict = FrozenDict(new_config)
if safety_checker is None and requires_safety_checker:
logger.warning(
f"You have disabled the safety checker for {self.__class__} by passing `safety_checker=None`. Ensure"
" that you abide to the conditions of the Stable Diffusion license and do not expose unfiltered"
" results in services or applications open to the public. Both the diffusers team and Hugging Face"
" strongly recommend to keep the safety filter enabled in all public facing circumstances, disabling"
" it only for use-cases that involve analyzing network behavior or auditing its results. For more"
" information, please have a look at https://github.com/huggingface/diffusers/pull/254 ."
)
if safety_checker is not None and feature_extractor is None:
raise ValueError(
"Make sure to define a feature extractor when loading {self.__class__} if you want to use the safety"
" checker. If you do not want to use the safety checker, you can pass `'safety_checker=None'` instead."
)
is_unet_version_less_0_9_0 = hasattr(unet.config, "_diffusers_version") and version.parse(
version.parse(unet.config._diffusers_version).base_version
) < version.parse("0.9.0.dev0")
is_unet_sample_size_less_64 = hasattr(unet.config, "sample_size") and unet.config.sample_size < 64
if is_unet_version_less_0_9_0 and is_unet_sample_size_less_64:
deprecation_message = (
"The configuration file of the unet has set the default `sample_size` to smaller than"
" 64 which seems highly unlikely .If you're checkpoint is a fine-tuned version of any of the"
" following: \n- CompVis/stable-diffusion-v1-4 \n- CompVis/stable-diffusion-v1-3 \n-"
" CompVis/stable-diffusion-v1-2 \n- CompVis/stable-diffusion-v1-1 \n- runwayml/stable-diffusion-v1-5"
" \n- runwayml/stable-diffusion-inpainting \n you should change 'sample_size' to 64 in the"
" configuration file. Please make sure to update the config accordingly as leaving `sample_size=32`"
" in the config might lead to incorrect results in future versions. If you have downloaded this"
" checkpoint from the Hugging Face Hub, it would be very nice if you could open a Pull request for"
" the `unet/config.json` file"
)
deprecate("sample_size<64", "1.0.0", deprecation_message, standard_warn=False)
new_config = dict(unet.config)
new_config["sample_size"] = 64
unet._internal_dict = FrozenDict(new_config)
self.register_modules(
vae=vae,
text_encoder=text_encoder,
tokenizer=tokenizer,
unet=unet,
scheduler=scheduler,
safety_checker=safety_checker,
feature_extractor=feature_extractor,
inverse_scheduler=inverse_scheduler,
)
self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1)
self.image_processor = VaeImageProcessor(vae_scale_factor=self.vae_scale_factor)
self.register_to_config(requires_safety_checker=requires_safety_checker)
# Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion.StableDiffusionPipeline._encode_prompt
def _encode_prompt(
self,
prompt,
device,
num_images_per_prompt,
do_classifier_free_guidance,
negative_prompt=None,
prompt_embeds: Optional[torch.FloatTensor] = None,
negative_prompt_embeds: Optional[torch.FloatTensor] = None,
lora_scale: Optional[float] = None,
**kwargs,
):
deprecation_message = "`_encode_prompt()` is deprecated and it will be removed in a future version. Use `encode_prompt()` instead. Also, be aware that the output format changed from a concatenated tensor to a tuple."
deprecate("_encode_prompt()", "1.0.0", deprecation_message, standard_warn=False)
prompt_embeds_tuple = self.encode_prompt(
prompt=prompt,
device=device,
num_images_per_prompt=num_images_per_prompt,
do_classifier_free_guidance=do_classifier_free_guidance,
negative_prompt=negative_prompt,
prompt_embeds=prompt_embeds,
negative_prompt_embeds=negative_prompt_embeds,
lora_scale=lora_scale,
**kwargs,
)
# concatenate for backwards comp
prompt_embeds = torch.cat([prompt_embeds_tuple[1], prompt_embeds_tuple[0]])
return prompt_embeds
# Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion.StableDiffusionPipeline.encode_prompt
def encode_prompt(
self,
prompt,
device,
num_images_per_prompt,
do_classifier_free_guidance,
negative_prompt=None,
prompt_embeds: Optional[torch.FloatTensor] = None,
negative_prompt_embeds: Optional[torch.FloatTensor] = None,
lora_scale: Optional[float] = None,
clip_skip: Optional[int] = None,
):
r"""
Encodes the prompt into text encoder hidden states.
Args:
prompt (`str` or `List[str]`, *optional*):
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]`, *optional*):
The prompt or prompts not to guide the image generation. If not defined, one has to pass
`negative_prompt_embeds` instead. Ignored when not using guidance (i.e., ignored if `guidance_scale` is
less than `1`).
prompt_embeds (`torch.FloatTensor`, *optional*):
Pre-generated text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt weighting. If not
provided, text embeddings will be generated from `prompt` input argument.
negative_prompt_embeds (`torch.FloatTensor`, *optional*):
Pre-generated negative text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt
weighting. If not provided, negative_prompt_embeds will be generated from `negative_prompt` input
argument.
lora_scale (`float`, *optional*):
A LoRA scale that will be applied to all LoRA layers of the text encoder if LoRA layers are loaded.
clip_skip (`int`, *optional*):
Number of layers to be skipped from CLIP while computing the prompt embeddings. A value of 1 means that
the output of the pre-final layer will be used for computing the prompt embeddings.
"""
# set lora scale so that monkey patched LoRA
# function of text encoder can correctly access it
if lora_scale is not None and isinstance(self, LoraLoaderMixin):
self._lora_scale = lora_scale
# dynamically adjust the LoRA scale
if not USE_PEFT_BACKEND:
adjust_lora_scale_text_encoder(self.text_encoder, lora_scale)
else:
scale_lora_layers(self.text_encoder, lora_scale)
if prompt is not None and isinstance(prompt, str):
batch_size = 1
elif prompt is not None and isinstance(prompt, list):
batch_size = len(prompt)
else:
batch_size = prompt_embeds.shape[0]
if prompt_embeds is None:
# textual inversion: process multi-vector tokens if necessary
if isinstance(self, TextualInversionLoaderMixin):
prompt = self.maybe_convert_prompt(prompt, self.tokenizer)
text_inputs = self.tokenizer(
prompt,
padding="max_length",
max_length=self.tokenizer.model_max_length,
truncation=True,
return_tensors="pt",
)
text_input_ids = text_inputs.input_ids
untruncated_ids = self.tokenizer(prompt, padding="longest", return_tensors="pt").input_ids
if untruncated_ids.shape[-1] >= text_input_ids.shape[-1] and not torch.equal(
text_input_ids, untruncated_ids
):
removed_text = self.tokenizer.batch_decode(
untruncated_ids[:, self.tokenizer.model_max_length - 1 : -1]
)
logger.warning(
"The following part of your input was truncated because CLIP can only handle sequences up to"
f" {self.tokenizer.model_max_length} tokens: {removed_text}"
)
if hasattr(self.text_encoder.config, "use_attention_mask") and self.text_encoder.config.use_attention_mask:
attention_mask = text_inputs.attention_mask.to(device)
else:
attention_mask = None
if clip_skip is None:
prompt_embeds = self.text_encoder(text_input_ids.to(device), attention_mask=attention_mask)
prompt_embeds = prompt_embeds[0]
else:
prompt_embeds = self.text_encoder(
text_input_ids.to(device), attention_mask=attention_mask, output_hidden_states=True
)
# Access the `hidden_states` first, that contains a tuple of
# all the hidden states from the encoder layers. Then index into
# the tuple to access the hidden states from the desired layer.
prompt_embeds = prompt_embeds[-1][-(clip_skip + 1)]
# We also need to apply the final LayerNorm here to not mess with the
# representations. The `last_hidden_states` that we typically use for
# obtaining the final prompt representations passes through the LayerNorm
# layer.
prompt_embeds = self.text_encoder.text_model.final_layer_norm(prompt_embeds)
if self.text_encoder is not None:
prompt_embeds_dtype = self.text_encoder.dtype
elif self.unet is not None:
prompt_embeds_dtype = self.unet.dtype
else:
prompt_embeds_dtype = prompt_embeds.dtype
prompt_embeds = prompt_embeds.to(dtype=prompt_embeds_dtype, device=device)
bs_embed, seq_len, _ = prompt_embeds.shape
# duplicate text embeddings for each generation per prompt, using mps friendly method
prompt_embeds = prompt_embeds.repeat(1, num_images_per_prompt, 1)
prompt_embeds = prompt_embeds.view(bs_embed * num_images_per_prompt, seq_len, -1)
# get unconditional embeddings for classifier free guidance
if do_classifier_free_guidance and negative_prompt_embeds is None:
uncond_tokens: List[str]
if negative_prompt is None:
uncond_tokens = [""] * batch_size
elif prompt is not None and type(prompt) is not type(negative_prompt):
raise TypeError(
f"`negative_prompt` should be the same type to `prompt`, but got {type(negative_prompt)} !="
f" {type(prompt)}."
)
elif isinstance(negative_prompt, str):
uncond_tokens = [negative_prompt]
elif 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`."
)
else:
uncond_tokens = negative_prompt
# textual inversion: process multi-vector tokens if necessary
if isinstance(self, TextualInversionLoaderMixin):
uncond_tokens = self.maybe_convert_prompt(uncond_tokens, self.tokenizer)
max_length = prompt_embeds.shape[1]
uncond_input = self.tokenizer(
uncond_tokens,
padding="max_length",
max_length=max_length,
truncation=True,
return_tensors="pt",
)
if hasattr(self.text_encoder.config, "use_attention_mask") and self.text_encoder.config.use_attention_mask:
attention_mask = uncond_input.attention_mask.to(device)
else:
attention_mask = None
negative_prompt_embeds = self.text_encoder(
uncond_input.input_ids.to(device),
attention_mask=attention_mask,
)
negative_prompt_embeds = negative_prompt_embeds[0]
if do_classifier_free_guidance:
# duplicate unconditional embeddings for each generation per prompt, using mps friendly method
seq_len = negative_prompt_embeds.shape[1]
negative_prompt_embeds = negative_prompt_embeds.to(dtype=prompt_embeds_dtype, device=device)
negative_prompt_embeds = negative_prompt_embeds.repeat(1, num_images_per_prompt, 1)
negative_prompt_embeds = negative_prompt_embeds.view(batch_size * num_images_per_prompt, seq_len, -1)
if isinstance(self, LoraLoaderMixin) and USE_PEFT_BACKEND:
# Retrieve the original scale by scaling back the LoRA layers
unscale_lora_layers(self.text_encoder, lora_scale)
return prompt_embeds, negative_prompt_embeds
# Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion.StableDiffusionPipeline.run_safety_checker
def run_safety_checker(self, image, device, dtype):
if self.safety_checker is None:
has_nsfw_concept = None
else:
if torch.is_tensor(image):
feature_extractor_input = self.image_processor.postprocess(image, output_type="pil")
else:
feature_extractor_input = self.image_processor.numpy_to_pil(image)
safety_checker_input = self.feature_extractor(feature_extractor_input, return_tensors="pt").to(device)
image, has_nsfw_concept = self.safety_checker(
images=image, clip_input=safety_checker_input.pixel_values.to(dtype)
)
return image, has_nsfw_concept
# Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion.StableDiffusionPipeline.prepare_extra_step_kwargs
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
# Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion.StableDiffusionPipeline.decode_latents
def decode_latents(self, latents):
deprecation_message = "The decode_latents method is deprecated and will be removed in 1.0.0. Please use VaeImageProcessor.postprocess(...) instead"
deprecate("decode_latents", "1.0.0", deprecation_message, standard_warn=False)
latents = 1 / self.vae.config.scaling_factor * latents
image = self.vae.decode(latents, return_dict=False)[0]
image = (image / 2 + 0.5).clamp(0, 1)
# we always cast to float32 as this does not cause significant overhead and is compatible with bfloat16
image = image.cpu().permute(0, 2, 3, 1).float().numpy()
return image
def check_inputs(
self,
prompt,
strength,
callback_steps,
negative_prompt=None,
prompt_embeds=None,
negative_prompt_embeds=None,
):
if (strength is None) or (strength is not None and (strength < 0 or strength > 1)):
raise ValueError(
f"The value of `strength` should in [0.0, 1.0] but is, but is {strength} of type {type(strength)}."
)
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)}."
)
if prompt is not None and prompt_embeds is not None:
raise ValueError(
f"Cannot forward both `prompt`: {prompt} and `prompt_embeds`: {prompt_embeds}. Please make sure to"
" only forward one of the two."
)
elif prompt is None and prompt_embeds is None:
raise ValueError(
"Provide either `prompt` or `prompt_embeds`. Cannot leave both `prompt` and `prompt_embeds` undefined."
)
elif prompt is not None and (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 negative_prompt is not None and negative_prompt_embeds is not None:
raise ValueError(
f"Cannot forward both `negative_prompt`: {negative_prompt} and `negative_prompt_embeds`:"
f" {negative_prompt_embeds}. Please make sure to only forward one of the two."
)
if prompt_embeds is not None and negative_prompt_embeds is not None:
if prompt_embeds.shape != negative_prompt_embeds.shape:
raise ValueError(
"`prompt_embeds` and `negative_prompt_embeds` must have the same shape when passed directly, but"
f" got: `prompt_embeds` {prompt_embeds.shape} != `negative_prompt_embeds`"
f" {negative_prompt_embeds.shape}."
)
def check_source_inputs(
self,
source_prompt=None,
source_negative_prompt=None,
source_prompt_embeds=None,
source_negative_prompt_embeds=None,
):
if source_prompt is not None and source_prompt_embeds is not None:
raise ValueError(
f"Cannot forward both `source_prompt`: {source_prompt} and `source_prompt_embeds`: {source_prompt_embeds}."
" Please make sure to only forward one of the two."
)
elif source_prompt is None and source_prompt_embeds is None:
raise ValueError(
"Provide either `source_image` or `source_prompt_embeds`. Cannot leave all both of the arguments undefined."
)
elif source_prompt is not None and (
not isinstance(source_prompt, str) and not isinstance(source_prompt, list)
):
raise ValueError(f"`source_prompt` has to be of type `str` or `list` but is {type(source_prompt)}")
if source_negative_prompt is not None and source_negative_prompt_embeds is not None:
raise ValueError(
f"Cannot forward both `source_negative_prompt`: {source_negative_prompt} and `source_negative_prompt_embeds`:"
f" {source_negative_prompt_embeds}. Please make sure to only forward one of the two."
)
if source_prompt_embeds is not None and source_negative_prompt_embeds is not None:
if source_prompt_embeds.shape != source_negative_prompt_embeds.shape:
raise ValueError(
"`source_prompt_embeds` and `source_negative_prompt_embeds` must have the same shape when passed"
f" directly, but got: `source_prompt_embeds` {source_prompt_embeds.shape} !="
f" `source_negative_prompt_embeds` {source_negative_prompt_embeds.shape}."
)
def get_timesteps(self, num_inference_steps, strength, device):
# get the original timestep using init_timestep
init_timestep = min(int(num_inference_steps * strength), num_inference_steps)
t_start = max(num_inference_steps - init_timestep, 0)
timesteps = self.scheduler.timesteps[t_start * self.scheduler.order :]
return timesteps, num_inference_steps - t_start
def get_inverse_timesteps(self, num_inference_steps, strength, device):
# get the original timestep using init_timestep
init_timestep = min(int(num_inference_steps * strength), num_inference_steps)
t_start = max(num_inference_steps - init_timestep, 0)
# safety for t_start overflow to prevent empty timsteps slice
if t_start == 0:
return self.inverse_scheduler.timesteps, num_inference_steps
timesteps = self.inverse_scheduler.timesteps[:-t_start]
return timesteps, num_inference_steps - t_start
# Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion.StableDiffusionPipeline.prepare_latents
def prepare_latents(self, batch_size, num_channels_latents, height, width, dtype, device, generator, latents=None):
shape = (
batch_size,
num_channels_latents,
int(height) // self.vae_scale_factor,
int(width) // self.vae_scale_factor,
)
if isinstance(generator, list) and len(generator) != batch_size:
raise ValueError(
f"You have passed a list of generators of length {len(generator)}, but requested an effective batch"
f" size of {batch_size}. Make sure the batch size matches the length of the generators."
)
if latents is None:
latents = randn_tensor(shape, generator=generator, device=device, dtype=dtype)
else:
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
def prepare_image_latents(self, image, batch_size, dtype, device, generator=None):
if not isinstance(image, (torch.Tensor, PIL.Image.Image, list)):
raise ValueError(
f"`image` has to be of type `torch.Tensor`, `PIL.Image.Image` or list but is {type(image)}"
)
image = image.to(device=device, dtype=dtype)
if image.shape[1] == 4:
latents = image
else:
if isinstance(generator, list) and len(generator) != batch_size:
raise ValueError(
f"You have passed a list of generators of length {len(generator)}, but requested an effective batch"
f" size of {batch_size}. Make sure the batch size matches the length of the generators."
)
if isinstance(generator, list):
latents = [
self.vae.encode(image[i : i + 1]).latent_dist.sample(generator[i]) for i in range(batch_size)
]
latents = torch.cat(latents, dim=0)
else:
latents = self.vae.encode(image).latent_dist.sample(generator)
latents = self.vae.config.scaling_factor * latents
if batch_size != latents.shape[0]:
if batch_size % latents.shape[0] == 0:
# expand image_latents for batch_size
deprecation_message = (
f"You have passed {batch_size} text prompts (`prompt`), but only {latents.shape[0]} initial"
" images (`image`). Initial images are now duplicating to match the number of text prompts. Note"
" that this behavior is deprecated and will be removed in a version 1.0.0. Please make sure to update"
" your script to pass as many initial images as text prompts to suppress this warning."
)
deprecate("len(prompt) != len(image)", "1.0.0", deprecation_message, standard_warn=False)
additional_latents_per_image = batch_size // latents.shape[0]
latents = torch.cat([latents] * additional_latents_per_image, dim=0)
else:
raise ValueError(
f"Cannot duplicate `image` of batch size {latents.shape[0]} to {batch_size} text prompts."
)
else:
latents = torch.cat([latents], dim=0)
return latents
def get_epsilon(self, model_output: torch.Tensor, sample: torch.Tensor, timestep: int):
pred_type = self.inverse_scheduler.config.prediction_type
alpha_prod_t = self.inverse_scheduler.alphas_cumprod[timestep]
beta_prod_t = 1 - alpha_prod_t
if pred_type == "epsilon":
return model_output
elif pred_type == "sample":
return (sample - alpha_prod_t ** (0.5) * model_output) / beta_prod_t ** (0.5)
elif pred_type == "v_prediction":
return (alpha_prod_t**0.5) * model_output + (beta_prod_t**0.5) * sample
else:
raise ValueError(
f"prediction_type given as {pred_type} must be one of `epsilon`, `sample`, or `v_prediction`"
)
@torch.no_grad()
@replace_example_docstring(EXAMPLE_DOC_STRING)
def generate_mask(
self,
image: Union[torch.FloatTensor, PIL.Image.Image] = None,
target_prompt: Optional[Union[str, List[str]]] = None,
target_negative_prompt: Optional[Union[str, List[str]]] = None,
target_prompt_embeds: Optional[torch.FloatTensor] = None,
target_negative_prompt_embeds: Optional[torch.FloatTensor] = None,
source_prompt: Optional[Union[str, List[str]]] = None,
source_negative_prompt: Optional[Union[str, List[str]]] = None,
source_prompt_embeds: Optional[torch.FloatTensor] = None,
source_negative_prompt_embeds: Optional[torch.FloatTensor] = None,
num_maps_per_mask: Optional[int] = 10,
mask_encode_strength: Optional[float] = 0.5,
mask_thresholding_ratio: Optional[float] = 3.0,
num_inference_steps: int = 50,
guidance_scale: float = 7.5,
generator: Optional[Union[torch.Generator, List[torch.Generator]]] = None,
output_type: Optional[str] = "np",
cross_attention_kwargs: Optional[Dict[str, Any]] = None,
):
r"""
Generate a latent mask given a mask prompt, a target prompt, and an image.
Args:
image (`PIL.Image.Image`):
`Image` or tensor representing an image batch to be used for computing the mask.
target_prompt (`str` or `List[str]`, *optional*):
The prompt or prompts to guide semantic mask generation. If not defined, you need to pass
`prompt_embeds`.
target_negative_prompt (`str` or `List[str]`, *optional*):
The prompt or prompts to guide what to not include in image generation. If not defined, you need to
pass `negative_prompt_embeds` instead. Ignored when not using guidance (`guidance_scale < 1`).
target_prompt_embeds (`torch.FloatTensor`, *optional*):
Pre-generated text embeddings. Can be used to easily tweak text inputs (prompt weighting). If not
provided, text embeddings are generated from the `prompt` input argument.
target_negative_prompt_embeds (`torch.FloatTensor`, *optional*):
Pre-generated negative text embeddings. Can be used to easily tweak text inputs (prompt weighting). If
not provided, `negative_prompt_embeds` are generated from the `negative_prompt` input argument.
source_prompt (`str` or `List[str]`, *optional*):
The prompt or prompts to guide semantic mask generation using DiffEdit. If not defined, you need to
pass `source_prompt_embeds` or `source_image` instead.
source_negative_prompt (`str` or `List[str]`, *optional*):
The prompt or prompts to guide semantic mask generation away from using DiffEdit. If not defined, you
need to pass `source_negative_prompt_embeds` or `source_image` instead.
source_prompt_embeds (`torch.FloatTensor`, *optional*):
Pre-generated text embeddings to guide the semantic mask generation. Can be used to easily tweak text
inputs (prompt weighting). If not provided, text embeddings are generated from `source_prompt` input
argument.
source_negative_prompt_embeds (`torch.FloatTensor`, *optional*):
Pre-generated text embeddings to negatively guide the semantic mask generation. Can be used to easily
tweak text inputs (prompt weighting). If not provided, text embeddings are generated from
`source_negative_prompt` input argument.
num_maps_per_mask (`int`, *optional*, defaults to 10):
The number of noise maps sampled to generate the semantic mask using DiffEdit.
mask_encode_strength (`float`, *optional*, defaults to 0.5):
The strength of the noise maps sampled to generate the semantic mask using DiffEdit. Must be between 0
and 1.
mask_thresholding_ratio (`float`, *optional*, defaults to 3.0):
The maximum multiple of the mean absolute difference used to clamp the semantic guidance map before
mask binarization.
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):
A higher guidance scale value encourages the model to generate images closely linked to the text
`prompt` at the expense of lower image quality. Guidance scale is enabled when `guidance_scale > 1`.
generator (`torch.Generator` or `List[torch.Generator]`, *optional*):
A [`torch.Generator`](https://pytorch.org/docs/stable/generated/torch.Generator.html) to make
generation deterministic.
output_type (`str`, *optional*, defaults to `"pil"`):
The output format of the generated image. Choose between `PIL.Image` or `np.array`.
cross_attention_kwargs (`dict`, *optional*):
A kwargs dictionary that if specified is passed along to the
[`~models.attention_processor.AttnProcessor`] as defined in
[`self.processor`](https://github.com/huggingface/diffusers/blob/main/src/diffusers/models/attention_processor.py).
Examples:
Returns:
`List[PIL.Image.Image]` or `np.array`:
When returning a `List[PIL.Image.Image]`, the list consists of a batch of single-channel binary images
with dimensions `(height // self.vae_scale_factor, width // self.vae_scale_factor)`. If it's
`np.array`, the shape is `(batch_size, height // self.vae_scale_factor, width //
self.vae_scale_factor)`.
"""
# 1. Check inputs (Provide dummy argument for callback_steps)
self.check_inputs(
target_prompt,
mask_encode_strength,
1,
target_negative_prompt,
target_prompt_embeds,
target_negative_prompt_embeds,
)
self.check_source_inputs(
source_prompt,
source_negative_prompt,
source_prompt_embeds,
source_negative_prompt_embeds,
)
if (num_maps_per_mask is None) or (
num_maps_per_mask is not None and (not isinstance(num_maps_per_mask, int) or num_maps_per_mask <= 0)
):
raise ValueError(
f"`num_maps_per_mask` has to be a positive integer but is {num_maps_per_mask} of type"
f" {type(num_maps_per_mask)}."
)
if mask_thresholding_ratio is None or mask_thresholding_ratio <= 0:
raise ValueError(
f"`mask_thresholding_ratio` has to be positive but is {mask_thresholding_ratio} of type"
f" {type(mask_thresholding_ratio)}."
)
# 2. Define call parameters
if target_prompt is not None and isinstance(target_prompt, str):
batch_size = 1
elif target_prompt is not None and isinstance(target_prompt, list):
batch_size = len(target_prompt)
else:
batch_size = target_prompt_embeds.shape[0]
if cross_attention_kwargs is None:
cross_attention_kwargs = {}
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 prompts
(cross_attention_kwargs.get("scale", None) if cross_attention_kwargs is not None else None)
target_negative_prompt_embeds, target_prompt_embeds = self.encode_prompt(
target_prompt,
device,
num_maps_per_mask,
do_classifier_free_guidance,
target_negative_prompt,
prompt_embeds=target_prompt_embeds,
negative_prompt_embeds=target_negative_prompt_embeds,
)
# For classifier free guidance, we need to do two forward passes.
# Here we concatenate the unconditional and text embeddings into a single batch
# to avoid doing two forward passes
if do_classifier_free_guidance:
target_prompt_embeds = torch.cat([target_negative_prompt_embeds, target_prompt_embeds])
source_negative_prompt_embeds, source_prompt_embeds = self.encode_prompt(
source_prompt,
device,
num_maps_per_mask,
do_classifier_free_guidance,
source_negative_prompt,
prompt_embeds=source_prompt_embeds,
negative_prompt_embeds=source_negative_prompt_embeds,
)
if do_classifier_free_guidance:
source_prompt_embeds = torch.cat([source_negative_prompt_embeds, source_prompt_embeds])
# 4. Preprocess image
image = self.image_processor.preprocess(image).repeat_interleave(num_maps_per_mask, dim=0)
# 5. Set timesteps
self.scheduler.set_timesteps(num_inference_steps, device=device)
timesteps, _ = self.get_timesteps(num_inference_steps, mask_encode_strength, device)
encode_timestep = timesteps[0]
# 6. Prepare image latents and add noise with specified strength
image_latents = self.prepare_image_latents(
image, batch_size * num_maps_per_mask, self.vae.dtype, device, generator
)
noise = randn_tensor(image_latents.shape, generator=generator, device=device, dtype=self.vae.dtype)
image_latents = self.scheduler.add_noise(image_latents, noise, encode_timestep)
latent_model_input = torch.cat([image_latents] * (4 if do_classifier_free_guidance else 2))
latent_model_input = self.scheduler.scale_model_input(latent_model_input, encode_timestep)
# 7. Predict the noise residual
prompt_embeds = torch.cat([source_prompt_embeds, target_prompt_embeds])
noise_pred = self.unet(
latent_model_input,
encode_timestep,
encoder_hidden_states=prompt_embeds,
cross_attention_kwargs=cross_attention_kwargs,
).sample
if do_classifier_free_guidance:
noise_pred_neg_src, noise_pred_source, noise_pred_uncond, noise_pred_target = noise_pred.chunk(4)
noise_pred_source = noise_pred_neg_src + guidance_scale * (noise_pred_source - noise_pred_neg_src)
noise_pred_target = noise_pred_uncond + guidance_scale * (noise_pred_target - noise_pred_uncond)
else:
noise_pred_source, noise_pred_target = noise_pred.chunk(2)
# 8. Compute the mask from the absolute difference of predicted noise residuals
# TODO: Consider smoothing mask guidance map
mask_guidance_map = (
torch.abs(noise_pred_target - noise_pred_source)
.reshape(batch_size, num_maps_per_mask, *noise_pred_target.shape[-3:])
.mean([1, 2])
)
clamp_magnitude = mask_guidance_map.mean() * mask_thresholding_ratio
semantic_mask_image = mask_guidance_map.clamp(0, clamp_magnitude) / clamp_magnitude
semantic_mask_image = torch.where(semantic_mask_image <= 0.5, 0, 1)
mask_image = semantic_mask_image.cpu().numpy()
# 9. Convert to Numpy array or PIL.
if output_type == "pil":
mask_image = self.image_processor.numpy_to_pil(mask_image)
# Offload all models
self.maybe_free_model_hooks()
return mask_image
@torch.no_grad()
@replace_example_docstring(EXAMPLE_INVERT_DOC_STRING)
def invert(
self,
prompt: Optional[Union[str, List[str]]] = None,
image: Union[torch.FloatTensor, PIL.Image.Image] = None,
num_inference_steps: int = 50,
inpaint_strength: float = 0.8,
guidance_scale: float = 7.5,
negative_prompt: Optional[Union[str, List[str]]] = None,
generator: Optional[Union[torch.Generator, List[torch.Generator]]] = None,
prompt_embeds: Optional[torch.FloatTensor] = None,
negative_prompt_embeds: Optional[torch.FloatTensor] = None,
decode_latents: bool = False,
output_type: Optional[str] = "pil",
return_dict: bool = True,
callback: Optional[Callable[[int, int, torch.FloatTensor], None]] = None,
callback_steps: Optional[int] = 1,
cross_attention_kwargs: Optional[Dict[str, Any]] = None,
lambda_auto_corr: float = 20.0,
lambda_kl: float = 20.0,
num_reg_steps: int = 0,
num_auto_corr_rolls: int = 5,
):
r"""
Generate inverted latents given a prompt and image.
Args:
prompt (`str` or `List[str]`, *optional*):
The prompt or prompts to guide image generation. If not defined, you need to pass `prompt_embeds`.
image (`PIL.Image.Image`):
`Image` or tensor representing an image batch to produce the inverted latents guided by `prompt`.
inpaint_strength (`float`, *optional*, defaults to 0.8):
Indicates extent of the noising process to run latent inversion. Must be between 0 and 1. When
`inpaint_strength` is 1, the inversion process is run for the full number of iterations specified in
`num_inference_steps`. `image` is used as a reference for the inversion process, and adding more noise
increases `inpaint_strength`. If `inpaint_strength` is 0, no inpainting occurs.
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):
A higher guidance scale value encourages the model to generate images closely linked to the text
`prompt` at the expense of lower image quality. Guidance scale is enabled when `guidance_scale > 1`.
negative_prompt (`str` or `List[str]`, *optional*):
The prompt or prompts to guide what to not include in image generation. If not defined, you need to
pass `negative_prompt_embeds` instead. Ignored when not using guidance (`guidance_scale < 1`).
generator (`torch.Generator`, *optional*):
A [`torch.Generator`](https://pytorch.org/docs/stable/generated/torch.Generator.html) to make
generation deterministic.
prompt_embeds (`torch.FloatTensor`, *optional*):
Pre-generated text embeddings. Can be used to easily tweak text inputs (prompt weighting). If not
provided, text embeddings are generated from the `prompt` input argument.
negative_prompt_embeds (`torch.FloatTensor`, *optional*):
Pre-generated negative text embeddings. Can be used to easily tweak text inputs (prompt weighting). If
not provided, `negative_prompt_embeds` are generated from the `negative_prompt` input argument.
decode_latents (`bool`, *optional*, defaults to `False`):
Whether or not to decode the inverted latents into a generated image. Setting this argument to `True`
decodes all inverted latents for each timestep into a list of generated images.
output_type (`str`, *optional*, defaults to `"pil"`):
The output format of the generated image. Choose between `PIL.Image` or `np.array`.
return_dict (`bool`, *optional*, defaults to `True`):
Whether or not to return a [`~pipelines.stable_diffusion.DiffEditInversionPipelineOutput`] instead of a
plain tuple.
callback (`Callable`, *optional*):
A function that calls every `callback_steps` steps during inference. The function is called with the
following arguments: `callback(step: int, timestep: int, latents: torch.FloatTensor)`.
callback_steps (`int`, *optional*, defaults to 1):
The frequency at which the `callback` function is called. If not specified, the callback is called at
every step.
cross_attention_kwargs (`dict`, *optional*):
A kwargs dictionary that if specified is passed along to the
[`~models.attention_processor.AttnProcessor`] as defined in
[`self.processor`](https://github.com/huggingface/diffusers/blob/main/src/diffusers/models/attention_processor.py).
lambda_auto_corr (`float`, *optional*, defaults to 20.0):
Lambda parameter to control auto correction.
lambda_kl (`float`, *optional*, defaults to 20.0):
Lambda parameter to control Kullback-Leibler divergence output.
num_reg_steps (`int`, *optional*, defaults to 0):
Number of regularization loss steps.
num_auto_corr_rolls (`int`, *optional*, defaults to 5):
Number of auto correction roll steps.
Examples:
Returns:
[`~pipelines.stable_diffusion.pipeline_stable_diffusion_diffedit.DiffEditInversionPipelineOutput`] or
`tuple`:
If `return_dict` is `True`,
[`~pipelines.stable_diffusion.pipeline_stable_diffusion_diffedit.DiffEditInversionPipelineOutput`] is
returned, otherwise a `tuple` is returned where the first element is the inverted latents tensors
ordered by increasing noise, and the second is the corresponding decoded images if `decode_latents` is
`True`, otherwise `None`.
"""
# 1. Check inputs
self.check_inputs(
prompt,
inpaint_strength,
callback_steps,
negative_prompt,
prompt_embeds,
negative_prompt_embeds,
)
if image is None:
raise ValueError("`image` input cannot be undefined.")
# 2. Define call parameters
if prompt is not None and isinstance(prompt, str):
batch_size = 1
elif prompt is not None and isinstance(prompt, list):
batch_size = len(prompt)
else:
batch_size = prompt_embeds.shape[0]
if cross_attention_kwargs is None:
cross_attention_kwargs = {}
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. Preprocess image
image = self.image_processor.preprocess(image)
# 4. Prepare latent variables
num_images_per_prompt = 1
latents = self.prepare_image_latents(
image, batch_size * num_images_per_prompt, self.vae.dtype, device, generator
)
# 5. Encode input prompt
prompt_embeds, negative_prompt_embeds = self.encode_prompt(
prompt,
device,
num_images_per_prompt,
do_classifier_free_guidance,
negative_prompt,
prompt_embeds=prompt_embeds,
negative_prompt_embeds=negative_prompt_embeds,
)
# For classifier free guidance, we need to do two forward passes.
# Here we concatenate the unconditional and text embeddings into a single batch
# to avoid doing two forward passes
if do_classifier_free_guidance:
prompt_embeds = torch.cat([negative_prompt_embeds, prompt_embeds])
# 6. Prepare timesteps
self.inverse_scheduler.set_timesteps(num_inference_steps, device=device)
timesteps, num_inference_steps = self.get_inverse_timesteps(num_inference_steps, inpaint_strength, device)
# 7. Noising loop where we obtain the intermediate noised latent image for each timestep.
num_warmup_steps = len(timesteps) - num_inference_steps * self.inverse_scheduler.order
inverted_latents = []
with self.progress_bar(total=num_inference_steps) as progress_bar:
for i, t in enumerate(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.inverse_scheduler.scale_model_input(latent_model_input, t)
# predict the noise residual
noise_pred = self.unet(
latent_model_input,
t,
encoder_hidden_states=prompt_embeds,
cross_attention_kwargs=cross_attention_kwargs,
).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)
# regularization of the noise prediction (not in original code or paper but borrowed from Pix2PixZero)
if num_reg_steps > 0:
with torch.enable_grad():
for _ in range(num_reg_steps):
if lambda_auto_corr > 0:
for _ in range(num_auto_corr_rolls):
var = torch.autograd.Variable(noise_pred.detach().clone(), requires_grad=True)
# Derive epsilon from model output before regularizing to IID standard normal
var_epsilon = self.get_epsilon(var, latent_model_input.detach(), t)
l_ac = auto_corr_loss(var_epsilon, generator=generator)
l_ac.backward()
grad = var.grad.detach() / num_auto_corr_rolls
noise_pred = noise_pred - lambda_auto_corr * grad
if lambda_kl > 0:
var = torch.autograd.Variable(noise_pred.detach().clone(), requires_grad=True)
# Derive epsilon from model output before regularizing to IID standard normal
var_epsilon = self.get_epsilon(var, latent_model_input.detach(), t)
l_kld = kl_divergence(var_epsilon)
l_kld.backward()
grad = var.grad.detach()
noise_pred = noise_pred - lambda_kl * grad
noise_pred = noise_pred.detach()
# compute the previous noisy sample x_t -> x_t-1
latents = self.inverse_scheduler.step(noise_pred, t, latents).prev_sample
inverted_latents.append(latents.detach().clone())
# call the callback, if provided
if i == len(timesteps) - 1 or (
(i + 1) > num_warmup_steps and (i + 1) % self.inverse_scheduler.order == 0
):
progress_bar.update()
if callback is not None and i % callback_steps == 0:
step_idx = i // getattr(self.scheduler, "order", 1)
callback(step_idx, t, latents)
assert len(inverted_latents) == len(timesteps)
latents = torch.stack(list(reversed(inverted_latents)), 1)
# 8. Post-processing
image = None
if decode_latents:
image = self.decode_latents(latents.flatten(0, 1))
# 9. Convert to PIL.
if decode_latents and output_type == "pil":
image = self.image_processor.numpy_to_pil(image)
# Offload all models
self.maybe_free_model_hooks()
if not return_dict:
return (latents, image)
return DiffEditInversionPipelineOutput(latents=latents, images=image)
@torch.no_grad()
@replace_example_docstring(EXAMPLE_DOC_STRING)
def __call__(
self,
prompt: Optional[Union[str, List[str]]] = None,
mask_image: Union[torch.FloatTensor, PIL.Image.Image] = None,
image_latents: Union[torch.FloatTensor, PIL.Image.Image] = None,
inpaint_strength: Optional[float] = 0.8,
num_inference_steps: int = 50,
guidance_scale: float = 7.5,
negative_prompt: Optional[Union[str, List[str]]] = None,
num_images_per_prompt: Optional[int] = 1,
eta: float = 0.0,
generator: Optional[Union[torch.Generator, List[torch.Generator]]] = None,
latents: Optional[torch.FloatTensor] = None,
prompt_embeds: Optional[torch.FloatTensor] = None,
negative_prompt_embeds: Optional[torch.FloatTensor] = None,
output_type: Optional[str] = "pil",
return_dict: bool = True,
callback: Optional[Callable[[int, int, torch.FloatTensor], None]] = None,
callback_steps: int = 1,
cross_attention_kwargs: Optional[Dict[str, Any]] = None,
clip_skip: int = None,
):
r"""
The call function to the pipeline for generation.
Args:
prompt (`str` or `List[str]`, *optional*):
The prompt or prompts to guide image generation. If not defined, you need to pass `prompt_embeds`.
mask_image (`PIL.Image.Image`):
`Image` or tensor representing an image batch to mask the generated image. White pixels in the mask are
repainted, while black pixels are preserved. If `mask_image` is a PIL image, it is 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, 1, H, W)`.
image_latents (`PIL.Image.Image` or `torch.FloatTensor`):
Partially noised image latents from the inversion process to be used as inputs for image generation.
inpaint_strength (`float`, *optional*, defaults to 0.8):
Indicates extent to inpaint the masked area. Must be between 0 and 1. When `inpaint_strength` is 1, the
denoising process is run on the masked area for the full number of iterations specified in
`num_inference_steps`. `image_latents` is used as a reference for the masked area, and adding more
noise to a region increases `inpaint_strength`. If `inpaint_strength` is 0, no inpainting occurs.
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):
A higher guidance scale value encourages the model to generate images closely linked to the text
`prompt` at the expense of lower image quality. Guidance scale is enabled when `guidance_scale > 1`.
negative_prompt (`str` or `List[str]`, *optional*):
The prompt or prompts to guide what to not include in image generation. If not defined, you need to
pass `negative_prompt_embeds` instead. Ignored when not using guidance (`guidance_scale < 1`).
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) from the [DDIM](https://arxiv.org/abs/2010.02502) paper. Only applies
to the [`~schedulers.DDIMScheduler`], and is ignored in other schedulers.
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 is generated by sampling using the supplied random `generator`.
prompt_embeds (`torch.FloatTensor`, *optional*):
Pre-generated text embeddings. Can be used to easily tweak text inputs (prompt weighting). If not
provided, text embeddings are generated from the `prompt` input argument.
negative_prompt_embeds (`torch.FloatTensor`, *optional*):
Pre-generated negative text embeddings. Can be used to easily tweak text inputs (prompt weighting). If
not provided, `negative_prompt_embeds` are generated from the `negative_prompt` input argument.
output_type (`str`, *optional*, defaults to `"pil"`):
The output format of the generated image. Choose between `PIL.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 calls every `callback_steps` steps during inference. The function is called with the
following arguments: `callback(step: int, timestep: int, latents: torch.FloatTensor)`.
callback_steps (`int`, *optional*, defaults to 1):
The frequency at which the `callback` function is called. If not specified, the callback is called at
every step.
cross_attention_kwargs (`dict`, *optional*):
A kwargs dictionary that if specified is passed along to the [`AttentionProcessor`] as defined in
[`self.processor`](https://github.com/huggingface/diffusers/blob/main/src/diffusers/models/attention_processor.py).
clip_skip (`int`, *optional*):
Number of layers to be skipped from CLIP while computing the prompt embeddings. A value of 1 means that
the output of the pre-final layer will be used for computing the prompt embeddings.
Examples:
Returns:
[`~pipelines.stable_diffusion.StableDiffusionPipelineOutput`] or `tuple`:
If `return_dict` is `True`, [`~pipelines.stable_diffusion.StableDiffusionPipelineOutput`] is returned,
otherwise a `tuple` is returned where the first element is a list with the generated images and the
second element is a list of `bool`s indicating whether the corresponding generated image contains
"not-safe-for-work" (nsfw) content.
"""
# 1. Check inputs
self.check_inputs(
prompt,
inpaint_strength,
callback_steps,
negative_prompt,
prompt_embeds,
negative_prompt_embeds,
)
if mask_image is None:
raise ValueError(
"`mask_image` input cannot be undefined. Use `generate_mask()` to compute `mask_image` from text prompts."
)
if image_latents is None:
raise ValueError(
"`image_latents` input cannot be undefined. Use `invert()` to compute `image_latents` from input images."
)
# 2. Define call parameters
if prompt is not None and isinstance(prompt, str):
batch_size = 1
elif prompt is not None and isinstance(prompt, list):
batch_size = len(prompt)
else:
batch_size = prompt_embeds.shape[0]
if cross_attention_kwargs is None:
cross_attention_kwargs = {}
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_encoder_lora_scale = (
cross_attention_kwargs.get("scale", None) if cross_attention_kwargs is not None else None
)
prompt_embeds, negative_prompt_embeds = self.encode_prompt(
prompt,
device,
num_images_per_prompt,
do_classifier_free_guidance,
negative_prompt,
prompt_embeds=prompt_embeds,
negative_prompt_embeds=negative_prompt_embeds,
lora_scale=text_encoder_lora_scale,
clip_skip=clip_skip,
)
# For classifier free guidance, we need to do two forward passes.
# Here we concatenate the unconditional and text embeddings into a single batch
# to avoid doing two forward passes
if do_classifier_free_guidance:
prompt_embeds = torch.cat([negative_prompt_embeds, prompt_embeds])
# 4. Preprocess mask
mask_image = preprocess_mask(mask_image, batch_size)
latent_height, latent_width = mask_image.shape[-2:]
mask_image = torch.cat([mask_image] * num_images_per_prompt)
mask_image = mask_image.to(device=device, dtype=prompt_embeds.dtype)
# 5. Set timesteps
self.scheduler.set_timesteps(num_inference_steps, device=device)
timesteps, num_inference_steps = self.get_timesteps(num_inference_steps, inpaint_strength, device)
# 6. Preprocess image latents
if isinstance(image_latents, list) and any(isinstance(l, torch.Tensor) and l.ndim == 5 for l in image_latents):
image_latents = torch.cat(image_latents).detach()
elif isinstance(image_latents, torch.Tensor) and image_latents.ndim == 5:
image_latents = image_latents.detach()
else:
image_latents = self.image_processor.preprocess(image_latents).detach()
latent_shape = (self.vae.config.latent_channels, latent_height, latent_width)
if image_latents.shape[-3:] != latent_shape:
raise ValueError(
f"Each latent image in `image_latents` must have shape {latent_shape}, "
f"but has shape {image_latents.shape[-3:]}"
)
if image_latents.ndim == 4:
image_latents = image_latents.reshape(batch_size, len(timesteps), *latent_shape)
if image_latents.shape[:2] != (batch_size, len(timesteps)):
raise ValueError(
f"`image_latents` must have batch size {batch_size} with latent images from {len(timesteps)}"
f" timesteps, but has batch size {image_latents.shape[0]} with latent images from"
f" {image_latents.shape[1]} timesteps."
)
image_latents = image_latents.transpose(0, 1).repeat_interleave(num_images_per_prompt, dim=1)
image_latents = image_latents.to(device=device, dtype=prompt_embeds.dtype)
# 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
latents = image_latents[0].clone()
num_warmup_steps = len(timesteps) - num_inference_steps * self.scheduler.order
with self.progress_bar(total=num_inference_steps) as progress_bar:
for i, t in enumerate(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=prompt_embeds,
cross_attention_kwargs=cross_attention_kwargs,
).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
# mask with inverted latents from appropriate timestep - use original image latent for last step
latents = latents * mask_image + image_latents[i] * (1 - mask_image)
# call the callback, if provided
if i == len(timesteps) - 1 or ((i + 1) > num_warmup_steps and (i + 1) % self.scheduler.order == 0):
progress_bar.update()
if callback is not None and i % callback_steps == 0:
step_idx = i // getattr(self.scheduler, "order", 1)
callback(step_idx, t, latents)
if not output_type == "latent":
image = self.vae.decode(latents / self.vae.config.scaling_factor, return_dict=False)[0]
image, has_nsfw_concept = self.run_safety_checker(image, device, prompt_embeds.dtype)
else:
image = latents
has_nsfw_concept = None
if has_nsfw_concept is None:
do_denormalize = [True] * image.shape[0]
else:
do_denormalize = [not has_nsfw for has_nsfw in has_nsfw_concept]
image = self.image_processor.postprocess(image, output_type=output_type, do_denormalize=do_denormalize)
# Offload all models
self.maybe_free_model_hooks()
if not return_dict:
return (image, has_nsfw_concept)
return StableDiffusionPipelineOutput(images=image, nsfw_content_detected=has_nsfw_concept)
|