File size: 74,664 Bytes
5fe7310 |
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 1534 1535 1536 1537 1538 1539 1540 |
# coding=utf-8
# Copyright 2024 suzukimain
#
# 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 os
import re
from collections import OrderedDict
from dataclasses import asdict, dataclass
from typing import Union
import requests
from huggingface_hub import hf_api, hf_hub_download
from huggingface_hub.file_download import http_get
from huggingface_hub.utils import validate_hf_hub_args
from diffusers.loaders.single_file_utils import (
VALID_URL_PREFIXES,
_extract_repo_id_and_weights_name,
infer_diffusers_model_type,
load_single_file_checkpoint,
)
from diffusers.pipelines.auto_pipeline import (
AutoPipelineForImage2Image,
AutoPipelineForInpainting,
AutoPipelineForText2Image,
)
from diffusers.pipelines.controlnet import (
StableDiffusionControlNetImg2ImgPipeline,
StableDiffusionControlNetInpaintPipeline,
StableDiffusionControlNetPipeline,
)
from diffusers.pipelines.pipeline_utils import DiffusionPipeline
from diffusers.pipelines.stable_diffusion import (
StableDiffusionImg2ImgPipeline,
StableDiffusionInpaintPipeline,
StableDiffusionPipeline,
)
from diffusers.pipelines.stable_diffusion_xl import (
StableDiffusionXLImg2ImgPipeline,
StableDiffusionXLInpaintPipeline,
StableDiffusionXLPipeline,
)
from diffusers.utils import logging
logger = logging.get_logger(__name__)
SINGLE_FILE_CHECKPOINT_TEXT2IMAGE_PIPELINE_MAPPING = OrderedDict(
[
("xl_base", StableDiffusionXLPipeline),
("xl_refiner", StableDiffusionXLPipeline),
("xl_inpaint", None),
("playground-v2-5", StableDiffusionXLPipeline),
("upscale", None),
("inpainting", None),
("inpainting_v2", None),
("controlnet", StableDiffusionControlNetPipeline),
("v2", StableDiffusionPipeline),
("v1", StableDiffusionPipeline),
]
)
SINGLE_FILE_CHECKPOINT_IMAGE2IMAGE_PIPELINE_MAPPING = OrderedDict(
[
("xl_base", StableDiffusionXLImg2ImgPipeline),
("xl_refiner", StableDiffusionXLImg2ImgPipeline),
("xl_inpaint", None),
("playground-v2-5", StableDiffusionXLImg2ImgPipeline),
("upscale", None),
("inpainting", None),
("inpainting_v2", None),
("controlnet", StableDiffusionControlNetImg2ImgPipeline),
("v2", StableDiffusionImg2ImgPipeline),
("v1", StableDiffusionImg2ImgPipeline),
]
)
SINGLE_FILE_CHECKPOINT_INPAINT_PIPELINE_MAPPING = OrderedDict(
[
("xl_base", None),
("xl_refiner", None),
("xl_inpaint", StableDiffusionXLInpaintPipeline),
("playground-v2-5", None),
("upscale", None),
("inpainting", StableDiffusionInpaintPipeline),
("inpainting_v2", StableDiffusionInpaintPipeline),
("controlnet", StableDiffusionControlNetInpaintPipeline),
("v2", None),
("v1", None),
]
)
CONFIG_FILE_LIST = [
"pytorch_model.bin",
"pytorch_model.fp16.bin",
"diffusion_pytorch_model.bin",
"diffusion_pytorch_model.fp16.bin",
"diffusion_pytorch_model.safetensors",
"diffusion_pytorch_model.fp16.safetensors",
"diffusion_pytorch_model.ckpt",
"diffusion_pytorch_model.fp16.ckpt",
"diffusion_pytorch_model.non_ema.bin",
"diffusion_pytorch_model.non_ema.safetensors",
]
DIFFUSERS_CONFIG_DIR = ["safety_checker", "unet", "vae", "text_encoder", "text_encoder_2"]
INPAINT_PIPELINE_KEYS = [
"xl_inpaint",
"inpainting",
"inpainting_v2",
]
EXTENSION = [".safetensors", ".ckpt", ".bin"]
CACHE_HOME = os.path.expanduser("~/.cache")
@dataclass
class RepoStatus:
r"""
Data class for storing repository status information.
Attributes:
repo_id (`str`):
The name of the repository.
repo_hash (`str`):
The hash of the repository.
version (`str`):
The version ID of the repository.
"""
repo_id: str = ""
repo_hash: str = ""
version: str = ""
@dataclass
class ModelStatus:
r"""
Data class for storing model status information.
Attributes:
search_word (`str`):
The search word used to find the model.
download_url (`str`):
The URL to download the model.
file_name (`str`):
The name of the model file.
local (`bool`):
Whether the model exists locally
"""
search_word: str = ""
download_url: str = ""
file_name: str = ""
local: bool = False
@dataclass
class SearchResult:
r"""
Data class for storing model data.
Attributes:
model_path (`str`):
The path to the model.
loading_method (`str`):
The type of loading method used for the model ( None or 'from_single_file' or 'from_pretrained')
checkpoint_format (`str`):
The format of the model checkpoint (`single_file` or `diffusers`).
repo_status (`RepoStatus`):
The status of the repository.
model_status (`ModelStatus`):
The status of the model.
"""
model_path: str = ""
loading_method: Union[str, None] = None
checkpoint_format: Union[str, None] = None
repo_status: RepoStatus = RepoStatus()
model_status: ModelStatus = ModelStatus()
@validate_hf_hub_args
def load_pipeline_from_single_file(pretrained_model_or_path, pipeline_mapping, **kwargs):
r"""
Instantiate a [`DiffusionPipeline`] from pretrained pipeline weights saved in the `.ckpt` or `.safetensors`
format. The pipeline is set in evaluation mode (`model.eval()`) by default.
Parameters:
pretrained_model_or_path (`str` or `os.PathLike`, *optional*):
Can be either:
- A link to the `.ckpt` file (for example
`"https://huggingface.co/<repo_id>/blob/main/<path_to_file>.ckpt"`) on the Hub.
- A path to a *file* containing all pipeline weights.
pipeline_mapping (`dict`):
A mapping of model types to their corresponding pipeline classes. This is used to determine
which pipeline class to instantiate based on the model type inferred from the checkpoint.
torch_dtype (`str` or `torch.dtype`, *optional*):
Override the default `torch.dtype` and load the model with another dtype.
force_download (`bool`, *optional*, defaults to `False`):
Whether or not to force the (re-)download of the model weights and configuration files, overriding the
cached versions if they exist.
cache_dir (`Union[str, os.PathLike]`, *optional*):
Path to a directory where a downloaded pretrained model configuration is cached if the standard cache
is not used.
proxies (`Dict[str, str]`, *optional*):
A dictionary of proxy servers to use by protocol or endpoint, for example, `{'http': 'foo.bar:3128',
'http://hostname': 'foo.bar:4012'}`. The proxies are used on each request.
local_files_only (`bool`, *optional*, defaults to `False`):
Whether to only load local model weights and configuration files or not. If set to `True`, the model
won't be downloaded from the Hub.
token (`str` or *bool*, *optional*):
The token to use as HTTP bearer authorization for remote files. If `True`, the token generated from
`diffusers-cli login` (stored in `~/.huggingface`) is used.
revision (`str`, *optional*, defaults to `"main"`):
The specific model version to use. It can be a branch name, a tag name, a commit id, or any identifier
allowed by Git.
original_config_file (`str`, *optional*):
The path to the original config file that was used to train the model. If not provided, the config file
will be inferred from the checkpoint file.
config (`str`, *optional*):
Can be either:
- A string, the *repo id* (for example `CompVis/ldm-text2im-large-256`) of a pretrained pipeline
hosted on the Hub.
- A path to a *directory* (for example `./my_pipeline_directory/`) containing the pipeline
component configs in Diffusers format.
checkpoint (`dict`, *optional*):
The loaded state dictionary of the model.
kwargs (remaining dictionary of keyword arguments, *optional*):
Can be used to overwrite load and saveable variables (the pipeline components of the specific pipeline
class). The overwritten components are passed directly to the pipelines `__init__` method. See example
below for more information.
"""
# Load the checkpoint from the provided link or path
checkpoint = load_single_file_checkpoint(pretrained_model_or_path)
# Infer the model type from the loaded checkpoint
model_type = infer_diffusers_model_type(checkpoint)
# Get the corresponding pipeline class from the pipeline mapping
pipeline_class = pipeline_mapping[model_type]
# For tasks not supported by this pipeline
if pipeline_class is None:
raise ValueError(
f"{model_type} is not supported in this pipeline."
"For `Text2Image`, please use `AutoPipelineForText2Image.from_pretrained`, "
"for `Image2Image` , please use `AutoPipelineForImage2Image.from_pretrained`, "
"and `inpaint` is only supported in `AutoPipelineForInpainting.from_pretrained`"
)
else:
# Instantiate and return the pipeline with the loaded checkpoint and any additional kwargs
return pipeline_class.from_single_file(pretrained_model_or_path, **kwargs)
def get_keyword_types(keyword):
r"""
Determine the type and loading method for a given keyword.
Parameters:
keyword (`str`):
The input keyword to classify.
Returns:
`dict`: A dictionary containing the model format, loading method,
and various types and extra types flags.
"""
# Initialize the status dictionary with default values
status = {
"checkpoint_format": None,
"loading_method": None,
"type": {
"other": False,
"hf_url": False,
"hf_repo": False,
"civitai_url": False,
"local": False,
},
"extra_type": {
"url": False,
"missing_model_index": None,
},
}
# Check if the keyword is an HTTP or HTTPS URL
status["extra_type"]["url"] = bool(re.search(r"^(https?)://", keyword))
# Check if the keyword is a file
if os.path.isfile(keyword):
status["type"]["local"] = True
status["checkpoint_format"] = "single_file"
status["loading_method"] = "from_single_file"
# Check if the keyword is a directory
elif os.path.isdir(keyword):
status["type"]["local"] = True
status["checkpoint_format"] = "diffusers"
status["loading_method"] = "from_pretrained"
if not os.path.exists(os.path.join(keyword, "model_index.json")):
status["extra_type"]["missing_model_index"] = True
# Check if the keyword is a Civitai URL
elif keyword.startswith("https://civitai.com/"):
status["type"]["civitai_url"] = True
status["checkpoint_format"] = "single_file"
status["loading_method"] = None
# Check if the keyword starts with any valid URL prefixes
elif any(keyword.startswith(prefix) for prefix in VALID_URL_PREFIXES):
repo_id, weights_name = _extract_repo_id_and_weights_name(keyword)
if weights_name:
status["type"]["hf_url"] = True
status["checkpoint_format"] = "single_file"
status["loading_method"] = "from_single_file"
else:
status["type"]["hf_repo"] = True
status["checkpoint_format"] = "diffusers"
status["loading_method"] = "from_pretrained"
# Check if the keyword matches a Hugging Face repository format
elif re.match(r"^[^/]+/[^/]+$", keyword):
status["type"]["hf_repo"] = True
status["checkpoint_format"] = "diffusers"
status["loading_method"] = "from_pretrained"
# If none of the above apply
else:
status["type"]["other"] = True
status["checkpoint_format"] = None
status["loading_method"] = None
return status
def file_downloader(
url,
save_path,
**kwargs,
) -> None:
"""
Downloads a file from a given URL and saves it to the specified path.
parameters:
url (`str`):
The URL of the file to download.
save_path (`str`):
The local path where the file will be saved.
resume (`bool`, *optional*, defaults to `False`):
Whether to resume an incomplete download.
headers (`dict`, *optional*, defaults to `None`):
Dictionary of HTTP Headers to send with the request.
proxies (`dict`, *optional*, defaults to `None`):
Dictionary mapping protocol to the URL of the proxy passed to `requests.request`.
force_download (`bool`, *optional*, defaults to `False`):
Whether to force the download even if the file already exists.
displayed_filename (`str`, *optional*):
The filename of the file that is being downloaded. Value is used only to display a nice progress bar. If
not set, the filename is guessed from the URL or the `Content-Disposition` header.
returns:
None
"""
# Get optional parameters from kwargs, with their default values
resume = kwargs.pop("resume", False)
headers = kwargs.pop("headers", None)
proxies = kwargs.pop("proxies", None)
force_download = kwargs.pop("force_download", False)
displayed_filename = kwargs.pop("displayed_filename", None)
# Default mode for file writing and initial file size
mode = "wb"
file_size = 0
# Create directory
os.makedirs(os.path.dirname(save_path), exist_ok=True)
# Check if the file already exists at the save path
if os.path.exists(save_path):
if not force_download:
# If the file exists and force_download is False, skip the download
logger.warning(f"File already exists: {save_path}, skipping download.")
return None
elif resume:
# If resuming, set mode to append binary and get current file size
mode = "ab"
file_size = os.path.getsize(save_path)
# Open the file in the appropriate mode (write or append)
with open(save_path, mode) as model_file:
# Call the http_get function to perform the file download
return http_get(
url=url,
temp_file=model_file,
resume_size=file_size,
displayed_filename=displayed_filename,
headers=headers,
proxies=proxies,
**kwargs,
)
def search_huggingface(search_word: str, **kwargs) -> Union[str, SearchResult, None]:
r"""
Downloads a model from Hugging Face.
Parameters:
search_word (`str`):
The search query string.
revision (`str`, *optional*):
The specific version of the model to download.
checkpoint_format (`str`, *optional*, defaults to `"single_file"`):
The format of the model checkpoint.
download (`bool`, *optional*, defaults to `False`):
Whether to download the model.
force_download (`bool`, *optional*, defaults to `False`):
Whether to force the download if the model already exists.
include_params (`bool`, *optional*, defaults to `False`):
Whether to include parameters in the returned data.
pipeline_tag (`str`, *optional*):
Tag to filter models by pipeline.
token (`str`, *optional*):
API token for Hugging Face authentication.
gated (`bool`, *optional*, defaults to `False` ):
A boolean to filter models on the Hub that are gated or not.
skip_error (`bool`, *optional*, defaults to `False`):
Whether to skip errors and return None.
Returns:
`Union[str, SearchResult, None]`: The model path or SearchResult or None.
"""
# Extract additional parameters from kwargs
revision = kwargs.pop("revision", None)
checkpoint_format = kwargs.pop("checkpoint_format", "single_file")
download = kwargs.pop("download", False)
force_download = kwargs.pop("force_download", False)
include_params = kwargs.pop("include_params", False)
pipeline_tag = kwargs.pop("pipeline_tag", None)
token = kwargs.pop("token", None)
gated = kwargs.pop("gated", False)
skip_error = kwargs.pop("skip_error", False)
# Get the type and loading method for the keyword
search_word_status = get_keyword_types(search_word)
if search_word_status["type"]["hf_repo"]:
if download:
model_path = DiffusionPipeline.download(
search_word,
revision=revision,
token=token,
force_download=force_download,
**kwargs,
)
else:
model_path = search_word
elif search_word_status["type"]["hf_url"]:
repo_id, weights_name = _extract_repo_id_and_weights_name(search_word)
if download:
model_path = hf_hub_download(
repo_id=repo_id,
filename=weights_name,
force_download=force_download,
token=token,
)
else:
model_path = search_word
elif search_word_status["type"]["local"]:
model_path = search_word
elif search_word_status["type"]["civitai_url"]:
if skip_error:
return None
else:
raise ValueError("The URL for Civitai is invalid with `for_hf`. Please use `for_civitai` instead.")
else:
# Get model data from HF API
hf_models = hf_api.list_models(
search=search_word,
direction=-1,
limit=100,
fetch_config=True,
pipeline_tag=pipeline_tag,
full=True,
gated=gated,
token=token,
)
model_dicts = [asdict(value) for value in list(hf_models)]
file_list = []
hf_repo_info = {}
hf_security_info = {}
model_path = ""
repo_id, file_name = "", ""
diffusers_model_exists = False
# Loop through models to find a suitable candidate
for repo_info in model_dicts:
repo_id = repo_info["id"]
file_list = []
hf_repo_info = hf_api.model_info(repo_id=repo_id, securityStatus=True)
# Lists files with security issues.
hf_security_info = hf_repo_info.security_repo_status
exclusion = [issue["path"] for issue in hf_security_info["filesWithIssues"]]
# Checks for multi-folder diffusers model or valid files (models with security issues are excluded).
if hf_security_info["scansDone"]:
for info in repo_info["siblings"]:
file_path = info["rfilename"]
if "model_index.json" == file_path and checkpoint_format in ["diffusers", "all"]:
diffusers_model_exists = True
break
elif (
any(file_path.endswith(ext) for ext in EXTENSION)
and not any(config in file_path for config in CONFIG_FILE_LIST)
and not any(exc in file_path for exc in exclusion)
and os.path.basename(os.path.dirname(file_path)) not in DIFFUSERS_CONFIG_DIR
):
file_list.append(file_path)
# Exit from the loop if a multi-folder diffusers model or valid file is found
if diffusers_model_exists or file_list:
break
else:
# Handle case where no models match the criteria
if skip_error:
return None
else:
raise ValueError("No models matching your criteria were found on huggingface.")
if diffusers_model_exists:
if download:
model_path = DiffusionPipeline.download(
repo_id,
token=token,
**kwargs,
)
else:
model_path = repo_id
elif file_list:
# Sort and find the safest model
file_name = next(
(model for model in sorted(file_list, reverse=True) if re.search(r"(?i)[-_](safe|sfw)", model)),
file_list[0],
)
if download:
model_path = hf_hub_download(
repo_id=repo_id,
filename=file_name,
revision=revision,
token=token,
force_download=force_download,
)
if file_name:
download_url = f"https://huggingface.co/{repo_id}/blob/main/{file_name}"
else:
download_url = f"https://huggingface.co/{repo_id}"
output_info = get_keyword_types(model_path)
if include_params:
return SearchResult(
model_path=model_path or download_url,
loading_method=output_info["loading_method"],
checkpoint_format=output_info["checkpoint_format"],
repo_status=RepoStatus(repo_id=repo_id, repo_hash=hf_repo_info.sha, version=revision),
model_status=ModelStatus(
search_word=search_word,
download_url=download_url,
file_name=file_name,
local=download,
),
)
else:
return model_path
def search_civitai(search_word: str, **kwargs) -> Union[str, SearchResult, None]:
r"""
Downloads a model from Civitai.
Parameters:
search_word (`str`):
The search query string.
model_type (`str`, *optional*, defaults to `Checkpoint`):
The type of model to search for.
base_model (`str`, *optional*):
The base model to filter by.
download (`bool`, *optional*, defaults to `False`):
Whether to download the model.
force_download (`bool`, *optional*, defaults to `False`):
Whether to force the download if the model already exists.
token (`str`, *optional*):
API token for Civitai authentication.
include_params (`bool`, *optional*, defaults to `False`):
Whether to include parameters in the returned data.
cache_dir (`str`, `Path`, *optional*):
Path to the folder where cached files are stored.
resume (`bool`, *optional*, defaults to `False`):
Whether to resume an incomplete download.
skip_error (`bool`, *optional*, defaults to `False`):
Whether to skip errors and return None.
Returns:
`Union[str, SearchResult, None]`: The model path or ` SearchResult` or None.
"""
# Extract additional parameters from kwargs
model_type = kwargs.pop("model_type", "Checkpoint")
download = kwargs.pop("download", False)
base_model = kwargs.pop("base_model", None)
force_download = kwargs.pop("force_download", False)
token = kwargs.pop("token", None)
include_params = kwargs.pop("include_params", False)
resume = kwargs.pop("resume", False)
cache_dir = kwargs.pop("cache_dir", None)
skip_error = kwargs.pop("skip_error", False)
# Initialize additional variables with default values
model_path = ""
repo_name = ""
repo_id = ""
version_id = ""
models_list = []
selected_repo = {}
selected_model = {}
selected_version = {}
civitai_cache_dir = cache_dir or os.path.join(CACHE_HOME, "Civitai")
# Set up parameters and headers for the CivitAI API request
params = {
"query": search_word,
"types": model_type,
"sort": "Most Downloaded",
"limit": 20,
}
if base_model is not None:
params["baseModel"] = base_model
headers = {}
if token:
headers["Authorization"] = f"Bearer {token}"
try:
# Make the request to the CivitAI API
response = requests.get("https://civitai.com/api/v1/models", params=params, headers=headers)
response.raise_for_status()
except requests.exceptions.HTTPError as err:
raise requests.HTTPError(f"Could not get elements from the URL: {err}")
else:
try:
data = response.json()
except AttributeError:
if skip_error:
return None
else:
raise ValueError("Invalid JSON response")
# Sort repositories by download count in descending order
sorted_repos = sorted(data["items"], key=lambda x: x["stats"]["downloadCount"], reverse=True)
for selected_repo in sorted_repos:
repo_name = selected_repo["name"]
repo_id = selected_repo["id"]
# Sort versions within the selected repo by download count
sorted_versions = sorted(
selected_repo["modelVersions"], key=lambda x: x["stats"]["downloadCount"], reverse=True
)
for selected_version in sorted_versions:
version_id = selected_version["id"]
models_list = []
for model_data in selected_version["files"]:
# Check if the file passes security scans and has a valid extension
file_name = model_data["name"]
if (
model_data["pickleScanResult"] == "Success"
and model_data["virusScanResult"] == "Success"
and any(file_name.endswith(ext) for ext in EXTENSION)
and os.path.basename(os.path.dirname(file_name)) not in DIFFUSERS_CONFIG_DIR
):
file_status = {
"filename": file_name,
"download_url": model_data["downloadUrl"],
}
models_list.append(file_status)
if models_list:
# Sort the models list by filename and find the safest model
sorted_models = sorted(models_list, key=lambda x: x["filename"], reverse=True)
selected_model = next(
(
model_data
for model_data in sorted_models
if bool(re.search(r"(?i)[-_](safe|sfw)", model_data["filename"]))
),
sorted_models[0],
)
break
else:
continue
break
# Exception handling when search candidates are not found
if not selected_model:
if skip_error:
return None
else:
raise ValueError("No model found. Please try changing the word you are searching for.")
# Define model file status
file_name = selected_model["filename"]
download_url = selected_model["download_url"]
# Handle file download and setting model information
if download:
# The path where the model is to be saved.
model_path = os.path.join(str(civitai_cache_dir), str(repo_id), str(version_id), str(file_name))
# Download Model File
file_downloader(
url=download_url,
save_path=model_path,
resume=resume,
force_download=force_download,
displayed_filename=file_name,
headers=headers,
**kwargs,
)
else:
model_path = download_url
output_info = get_keyword_types(model_path)
if not include_params:
return model_path
else:
return SearchResult(
model_path=model_path,
loading_method=output_info["loading_method"],
checkpoint_format=output_info["checkpoint_format"],
repo_status=RepoStatus(repo_id=repo_name, repo_hash=repo_id, version=version_id),
model_status=ModelStatus(
search_word=search_word,
download_url=download_url,
file_name=file_name,
local=output_info["type"]["local"],
),
)
class EasyPipelineForText2Image(AutoPipelineForText2Image):
r"""
[`AutoPipelineForText2Image`] is a generic pipeline class that instantiates a text-to-image pipeline class. The
specific underlying pipeline class is automatically selected from either the
[`~AutoPipelineForText2Image.from_pretrained`] or [`~AutoPipelineForText2Image.from_pipe`] methods.
This class cannot be instantiated using `__init__()` (throws an error).
Class attributes:
- **config_name** (`str`) -- The configuration filename that stores the class and module names of all the
diffusion pipeline's components.
"""
config_name = "model_index.json"
def __init__(self, *args, **kwargs):
# EnvironmentError is returned
super().__init__()
@classmethod
@validate_hf_hub_args
def from_huggingface(cls, pretrained_model_link_or_path, **kwargs):
r"""
Parameters:
pretrained_model_or_path (`str` or `os.PathLike`, *optional*):
Can be either:
- A keyword to search for Hugging Face (for example `Stable Diffusion`)
- Link to `.ckpt` or `.safetensors` file (for example
`"https://huggingface.co/<repo_id>/blob/main/<path_to_file>.safetensors"`) on the Hub.
- A string, the *repo id* (for example `CompVis/ldm-text2im-large-256`) of a pretrained pipeline
hosted on the Hub.
- A path to a *directory* (for example `./my_pipeline_directory/`) containing pipeline weights
saved using
[`~DiffusionPipeline.save_pretrained`].
checkpoint_format (`str`, *optional*, defaults to `"single_file"`):
The format of the model checkpoint.
pipeline_tag (`str`, *optional*):
Tag to filter models by pipeline.
torch_dtype (`str` or `torch.dtype`, *optional*):
Override the default `torch.dtype` and load the model with another dtype. If "auto" is passed, the
dtype is automatically derived from the model's weights.
force_download (`bool`, *optional*, defaults to `False`):
Whether or not to force the (re-)download of the model weights and configuration files, overriding the
cached versions if they exist.
cache_dir (`Union[str, os.PathLike]`, *optional*):
Path to a directory where a downloaded pretrained model configuration is cached if the standard cache
is not used.
proxies (`Dict[str, str]`, *optional*):
A dictionary of proxy servers to use by protocol or endpoint, for example, `{'http': 'foo.bar:3128',
'http://hostname': 'foo.bar:4012'}`. The proxies are used on each request.
output_loading_info(`bool`, *optional*, defaults to `False`):
Whether or not to also return a dictionary containing missing keys, unexpected keys and error messages.
local_files_only (`bool`, *optional*, defaults to `False`):
Whether to only load local model weights and configuration files or not. If set to `True`, the model
won't be downloaded from the Hub.
token (`str` or *bool*, *optional*):
The token to use as HTTP bearer authorization for remote files. If `True`, the token generated from
`diffusers-cli login` (stored in `~/.huggingface`) is used.
revision (`str`, *optional*, defaults to `"main"`):
The specific model version to use. It can be a branch name, a tag name, a commit id, or any identifier
allowed by Git.
custom_revision (`str`, *optional*, defaults to `"main"`):
The specific model version to use. It can be a branch name, a tag name, or a commit id similar to
`revision` when loading a custom pipeline from the Hub. It can be a 🤗 Diffusers version when loading a
custom pipeline from GitHub, otherwise it defaults to `"main"` when loading from the Hub.
mirror (`str`, *optional*):
Mirror source to resolve accessibility issues if you’re downloading a model in China. We do not
guarantee the timeliness or safety of the source, and you should refer to the mirror site for more
information.
device_map (`str` or `Dict[str, Union[int, str, torch.device]]`, *optional*):
A map that specifies where each submodule should go. It doesn’t need to be defined for each
parameter/buffer name; once a given module name is inside, every submodule of it will be sent to the
same device.
Set `device_map="auto"` to have 🤗 Accelerate automatically compute the most optimized `device_map`. For
more information about each option see [designing a device
map](https://hf.co/docs/accelerate/main/en/usage_guides/big_modeling#designing-a-device-map).
max_memory (`Dict`, *optional*):
A dictionary device identifier for the maximum memory. Will default to the maximum memory available for
each GPU and the available CPU RAM if unset.
offload_folder (`str` or `os.PathLike`, *optional*):
The path to offload weights if device_map contains the value `"disk"`.
offload_state_dict (`bool`, *optional*):
If `True`, temporarily offloads the CPU state dict to the hard drive to avoid running out of CPU RAM if
the weight of the CPU state dict + the biggest shard of the checkpoint does not fit. Defaults to `True`
when there is some disk offload.
low_cpu_mem_usage (`bool`, *optional*, defaults to `True` if torch version >= 1.9.0 else `False`):
Speed up model loading only loading the pretrained weights and not initializing the weights. This also
tries to not use more than 1x model size in CPU memory (including peak memory) while loading the model.
Only supported for PyTorch >= 1.9.0. If you are using an older version of PyTorch, setting this
argument to `True` will raise an error.
use_safetensors (`bool`, *optional*, defaults to `None`):
If set to `None`, the safetensors weights are downloaded if they're available **and** if the
safetensors library is installed. If set to `True`, the model is forcibly loaded from safetensors
weights. If set to `False`, safetensors weights are not loaded.
gated (`bool`, *optional*, defaults to `False` ):
A boolean to filter models on the Hub that are gated or not.
kwargs (remaining dictionary of keyword arguments, *optional*):
Can be used to overwrite load and saveable variables (the pipeline components of the specific pipeline
class). The overwritten components are passed directly to the pipelines `__init__` method. See example
below for more information.
variant (`str`, *optional*):
Load weights from a specified variant filename such as `"fp16"` or `"ema"`. This is ignored when
loading `from_flax`.
<Tip>
To use private or [gated](https://huggingface.co/docs/hub/models-gated#gated-models) models, log-in with
`huggingface-cli login`.
</Tip>
Examples:
```py
>>> from diffusers import AutoPipelineForText2Image
>>> pipeline = AutoPipelineForText2Image.from_huggingface("stable-diffusion-v1-5")
>>> image = pipeline(prompt).images[0]
```
"""
# Update kwargs to ensure the model is downloaded and parameters are included
_status = {
"download": True,
"include_params": True,
"skip_error": False,
"pipeline_tag": "text-to-image",
}
kwargs.update(_status)
# Search for the model on Hugging Face and get the model status
hf_model_status = search_huggingface(pretrained_model_link_or_path, **kwargs)
logger.warning(f"checkpoint_path: {hf_model_status.model_status.download_url}")
checkpoint_path = hf_model_status.model_path
# Check the format of the model checkpoint
if hf_model_status.checkpoint_format == "single_file":
# Load the pipeline from a single file checkpoint
return load_pipeline_from_single_file(
pretrained_model_or_path=checkpoint_path,
pipeline_mapping=SINGLE_FILE_CHECKPOINT_TEXT2IMAGE_PIPELINE_MAPPING,
**kwargs,
)
else:
return cls.from_pretrained(checkpoint_path, **kwargs)
@classmethod
def from_civitai(cls, pretrained_model_link_or_path, **kwargs):
r"""
Parameters:
pretrained_model_or_path (`str` or `os.PathLike`, *optional*):
Can be either:
- A keyword to search for Hugging Face (for example `Stable Diffusion`)
- Link to `.ckpt` or `.safetensors` file (for example
`"https://huggingface.co/<repo_id>/blob/main/<path_to_file>.safetensors"`) on the Hub.
- A string, the *repo id* (for example `CompVis/ldm-text2im-large-256`) of a pretrained pipeline
hosted on the Hub.
- A path to a *directory* (for example `./my_pipeline_directory/`) containing pipeline weights
saved using
[`~DiffusionPipeline.save_pretrained`].
model_type (`str`, *optional*, defaults to `Checkpoint`):
The type of model to search for. (for example `Checkpoint`, `TextualInversion`, `LORA`, `Controlnet`)
base_model (`str`, *optional*):
The base model to filter by.
cache_dir (`str`, `Path`, *optional*):
Path to the folder where cached files are stored.
resume (`bool`, *optional*, defaults to `False`):
Whether to resume an incomplete download.
torch_dtype (`str` or `torch.dtype`, *optional*):
Override the default `torch.dtype` and load the model with another dtype. If "auto" is passed, the
dtype is automatically derived from the model's weights.
force_download (`bool`, *optional*, defaults to `False`):
Whether or not to force the (re-)download of the model weights and configuration files, overriding the
cached versions if they exist.
output_loading_info(`bool`, *optional*, defaults to `False`):
Whether or not to also return a dictionary containing missing keys, unexpected keys and error messages.
local_files_only (`bool`, *optional*, defaults to `False`):
Whether to only load local model weights and configuration files or not. If set to `True`, the model
won't be downloaded from the Hub.
token (`str`, *optional*):
The token to use as HTTP bearer authorization for remote files.
device_map (`str` or `Dict[str, Union[int, str, torch.device]]`, *optional*):
A map that specifies where each submodule should go. It doesn’t need to be defined for each
parameter/buffer name; once a given module name is inside, every submodule of it will be sent to the
same device.
Set `device_map="auto"` to have 🤗 Accelerate automatically compute the most optimized `device_map`. For
more information about each option see [designing a device
map](https://hf.co/docs/accelerate/main/en/usage_guides/big_modeling#designing-a-device-map).
max_memory (`Dict`, *optional*):
A dictionary device identifier for the maximum memory. Will default to the maximum memory available for
each GPU and the available CPU RAM if unset.
offload_folder (`str` or `os.PathLike`, *optional*):
The path to offload weights if device_map contains the value `"disk"`.
offload_state_dict (`bool`, *optional*):
If `True`, temporarily offloads the CPU state dict to the hard drive to avoid running out of CPU RAM if
the weight of the CPU state dict + the biggest shard of the checkpoint does not fit. Defaults to `True`
when there is some disk offload.
low_cpu_mem_usage (`bool`, *optional*, defaults to `True` if torch version >= 1.9.0 else `False`):
Speed up model loading only loading the pretrained weights and not initializing the weights. This also
tries to not use more than 1x model size in CPU memory (including peak memory) while loading the model.
Only supported for PyTorch >= 1.9.0. If you are using an older version of PyTorch, setting this
argument to `True` will raise an error.
use_safetensors (`bool`, *optional*, defaults to `None`):
If set to `None`, the safetensors weights are downloaded if they're available **and** if the
safetensors library is installed. If set to `True`, the model is forcibly loaded from safetensors
weights. If set to `False`, safetensors weights are not loaded.
kwargs (remaining dictionary of keyword arguments, *optional*):
Can be used to overwrite load and saveable variables (the pipeline components of the specific pipeline
class). The overwritten components are passed directly to the pipelines `__init__` method. See example
below for more information.
<Tip>
To use private or [gated](https://huggingface.co/docs/hub/models-gated#gated-models) models, log-in with
`huggingface-cli login`.
</Tip>
Examples:
```py
>>> from diffusers import AutoPipelineForText2Image
>>> pipeline = AutoPipelineForText2Image.from_huggingface("stable-diffusion-v1-5")
>>> image = pipeline(prompt).images[0]
```
"""
# Update kwargs to ensure the model is downloaded and parameters are included
_status = {
"download": True,
"include_params": True,
"skip_error": False,
"model_type": "Checkpoint",
}
kwargs.update(_status)
# Search for the model on Civitai and get the model status
model_status = search_civitai(pretrained_model_link_or_path, **kwargs)
logger.warning(f"checkpoint_path: {model_status.model_status.download_url}")
checkpoint_path = model_status.model_path
# Load the pipeline from a single file checkpoint
return load_pipeline_from_single_file(
pretrained_model_or_path=checkpoint_path,
pipeline_mapping=SINGLE_FILE_CHECKPOINT_TEXT2IMAGE_PIPELINE_MAPPING,
**kwargs,
)
class EasyPipelineForImage2Image(AutoPipelineForImage2Image):
r"""
[`AutoPipelineForImage2Image`] is a generic pipeline class that instantiates an image-to-image pipeline class. The
specific underlying pipeline class is automatically selected from either the
[`~AutoPipelineForImage2Image.from_pretrained`] or [`~AutoPipelineForImage2Image.from_pipe`] methods.
This class cannot be instantiated using `__init__()` (throws an error).
Class attributes:
- **config_name** (`str`) -- The configuration filename that stores the class and module names of all the
diffusion pipeline's components.
"""
config_name = "model_index.json"
def __init__(self, *args, **kwargs):
# EnvironmentError is returned
super().__init__()
@classmethod
@validate_hf_hub_args
def from_huggingface(cls, pretrained_model_link_or_path, **kwargs):
r"""
Parameters:
pretrained_model_or_path (`str` or `os.PathLike`, *optional*):
Can be either:
- A keyword to search for Hugging Face (for example `Stable Diffusion`)
- Link to `.ckpt` or `.safetensors` file (for example
`"https://huggingface.co/<repo_id>/blob/main/<path_to_file>.safetensors"`) on the Hub.
- A string, the *repo id* (for example `CompVis/ldm-text2im-large-256`) of a pretrained pipeline
hosted on the Hub.
- A path to a *directory* (for example `./my_pipeline_directory/`) containing pipeline weights
saved using
[`~DiffusionPipeline.save_pretrained`].
checkpoint_format (`str`, *optional*, defaults to `"single_file"`):
The format of the model checkpoint.
pipeline_tag (`str`, *optional*):
Tag to filter models by pipeline.
torch_dtype (`str` or `torch.dtype`, *optional*):
Override the default `torch.dtype` and load the model with another dtype. If "auto" is passed, the
dtype is automatically derived from the model's weights.
force_download (`bool`, *optional*, defaults to `False`):
Whether or not to force the (re-)download of the model weights and configuration files, overriding the
cached versions if they exist.
cache_dir (`Union[str, os.PathLike]`, *optional*):
Path to a directory where a downloaded pretrained model configuration is cached if the standard cache
is not used.
proxies (`Dict[str, str]`, *optional*):
A dictionary of proxy servers to use by protocol or endpoint, for example, `{'http': 'foo.bar:3128',
'http://hostname': 'foo.bar:4012'}`. The proxies are used on each request.
output_loading_info(`bool`, *optional*, defaults to `False`):
Whether or not to also return a dictionary containing missing keys, unexpected keys and error messages.
local_files_only (`bool`, *optional*, defaults to `False`):
Whether to only load local model weights and configuration files or not. If set to `True`, the model
won't be downloaded from the Hub.
token (`str` or *bool*, *optional*):
The token to use as HTTP bearer authorization for remote files. If `True`, the token generated from
`diffusers-cli login` (stored in `~/.huggingface`) is used.
revision (`str`, *optional*, defaults to `"main"`):
The specific model version to use. It can be a branch name, a tag name, a commit id, or any identifier
allowed by Git.
custom_revision (`str`, *optional*, defaults to `"main"`):
The specific model version to use. It can be a branch name, a tag name, or a commit id similar to
`revision` when loading a custom pipeline from the Hub. It can be a 🤗 Diffusers version when loading a
custom pipeline from GitHub, otherwise it defaults to `"main"` when loading from the Hub.
mirror (`str`, *optional*):
Mirror source to resolve accessibility issues if you’re downloading a model in China. We do not
guarantee the timeliness or safety of the source, and you should refer to the mirror site for more
information.
device_map (`str` or `Dict[str, Union[int, str, torch.device]]`, *optional*):
A map that specifies where each submodule should go. It doesn’t need to be defined for each
parameter/buffer name; once a given module name is inside, every submodule of it will be sent to the
same device.
Set `device_map="auto"` to have 🤗 Accelerate automatically compute the most optimized `device_map`. For
more information about each option see [designing a device
map](https://hf.co/docs/accelerate/main/en/usage_guides/big_modeling#designing-a-device-map).
max_memory (`Dict`, *optional*):
A dictionary device identifier for the maximum memory. Will default to the maximum memory available for
each GPU and the available CPU RAM if unset.
offload_folder (`str` or `os.PathLike`, *optional*):
The path to offload weights if device_map contains the value `"disk"`.
offload_state_dict (`bool`, *optional*):
If `True`, temporarily offloads the CPU state dict to the hard drive to avoid running out of CPU RAM if
the weight of the CPU state dict + the biggest shard of the checkpoint does not fit. Defaults to `True`
when there is some disk offload.
low_cpu_mem_usage (`bool`, *optional*, defaults to `True` if torch version >= 1.9.0 else `False`):
Speed up model loading only loading the pretrained weights and not initializing the weights. This also
tries to not use more than 1x model size in CPU memory (including peak memory) while loading the model.
Only supported for PyTorch >= 1.9.0. If you are using an older version of PyTorch, setting this
argument to `True` will raise an error.
use_safetensors (`bool`, *optional*, defaults to `None`):
If set to `None`, the safetensors weights are downloaded if they're available **and** if the
safetensors library is installed. If set to `True`, the model is forcibly loaded from safetensors
weights. If set to `False`, safetensors weights are not loaded.
gated (`bool`, *optional*, defaults to `False` ):
A boolean to filter models on the Hub that are gated or not.
kwargs (remaining dictionary of keyword arguments, *optional*):
Can be used to overwrite load and saveable variables (the pipeline components of the specific pipeline
class). The overwritten components are passed directly to the pipelines `__init__` method. See example
below for more information.
variant (`str`, *optional*):
Load weights from a specified variant filename such as `"fp16"` or `"ema"`. This is ignored when
loading `from_flax`.
<Tip>
To use private or [gated](https://huggingface.co/docs/hub/models-gated#gated-models) models, log-in with
`huggingface-cli login`.
</Tip>
Examples:
```py
>>> from diffusers import AutoPipelineForText2Image
>>> pipeline = AutoPipelineForText2Image.from_huggingface("stable-diffusion-v1-5")
>>> image = pipeline(prompt).images[0]
```
"""
# Update kwargs to ensure the model is downloaded and parameters are included
_parmas = {
"download": True,
"include_params": True,
"skip_error": False,
"pipeline_tag": "image-to-image",
}
kwargs.update(_parmas)
# Search for the model on Hugging Face and get the model status
model_status = search_huggingface(pretrained_model_link_or_path, **kwargs)
logger.warning(f"checkpoint_path: {model_status.model_status.download_url}")
checkpoint_path = model_status.model_path
# Check the format of the model checkpoint
if model_status.checkpoint_format == "single_file":
# Load the pipeline from a single file checkpoint
return load_pipeline_from_single_file(
pretrained_model_or_path=checkpoint_path,
pipeline_mapping=SINGLE_FILE_CHECKPOINT_IMAGE2IMAGE_PIPELINE_MAPPING,
**kwargs,
)
else:
return cls.from_pretrained(checkpoint_path, **kwargs)
@classmethod
def from_civitai(cls, pretrained_model_link_or_path, **kwargs):
r"""
Parameters:
pretrained_model_or_path (`str` or `os.PathLike`, *optional*):
Can be either:
- A keyword to search for Hugging Face (for example `Stable Diffusion`)
- Link to `.ckpt` or `.safetensors` file (for example
`"https://huggingface.co/<repo_id>/blob/main/<path_to_file>.safetensors"`) on the Hub.
- A string, the *repo id* (for example `CompVis/ldm-text2im-large-256`) of a pretrained pipeline
hosted on the Hub.
- A path to a *directory* (for example `./my_pipeline_directory/`) containing pipeline weights
saved using
[`~DiffusionPipeline.save_pretrained`].
model_type (`str`, *optional*, defaults to `Checkpoint`):
The type of model to search for. (for example `Checkpoint`, `TextualInversion`, `LORA`, `Controlnet`)
base_model (`str`, *optional*):
The base model to filter by.
cache_dir (`str`, `Path`, *optional*):
Path to the folder where cached files are stored.
resume (`bool`, *optional*, defaults to `False`):
Whether to resume an incomplete download.
torch_dtype (`str` or `torch.dtype`, *optional*):
Override the default `torch.dtype` and load the model with another dtype. If "auto" is passed, the
dtype is automatically derived from the model's weights.
force_download (`bool`, *optional*, defaults to `False`):
Whether or not to force the (re-)download of the model weights and configuration files, overriding the
cached versions if they exist.
output_loading_info(`bool`, *optional*, defaults to `False`):
Whether or not to also return a dictionary containing missing keys, unexpected keys and error messages.
local_files_only (`bool`, *optional*, defaults to `False`):
Whether to only load local model weights and configuration files or not. If set to `True`, the model
won't be downloaded from the Hub.
token (`str`, *optional*):
The token to use as HTTP bearer authorization for remote files.
device_map (`str` or `Dict[str, Union[int, str, torch.device]]`, *optional*):
A map that specifies where each submodule should go. It doesn’t need to be defined for each
parameter/buffer name; once a given module name is inside, every submodule of it will be sent to the
same device.
Set `device_map="auto"` to have 🤗 Accelerate automatically compute the most optimized `device_map`. For
more information about each option see [designing a device
map](https://hf.co/docs/accelerate/main/en/usage_guides/big_modeling#designing-a-device-map).
max_memory (`Dict`, *optional*):
A dictionary device identifier for the maximum memory. Will default to the maximum memory available for
each GPU and the available CPU RAM if unset.
offload_folder (`str` or `os.PathLike`, *optional*):
The path to offload weights if device_map contains the value `"disk"`.
offload_state_dict (`bool`, *optional*):
If `True`, temporarily offloads the CPU state dict to the hard drive to avoid running out of CPU RAM if
the weight of the CPU state dict + the biggest shard of the checkpoint does not fit. Defaults to `True`
when there is some disk offload.
low_cpu_mem_usage (`bool`, *optional*, defaults to `True` if torch version >= 1.9.0 else `False`):
Speed up model loading only loading the pretrained weights and not initializing the weights. This also
tries to not use more than 1x model size in CPU memory (including peak memory) while loading the model.
Only supported for PyTorch >= 1.9.0. If you are using an older version of PyTorch, setting this
argument to `True` will raise an error.
use_safetensors (`bool`, *optional*, defaults to `None`):
If set to `None`, the safetensors weights are downloaded if they're available **and** if the
safetensors library is installed. If set to `True`, the model is forcibly loaded from safetensors
weights. If set to `False`, safetensors weights are not loaded.
kwargs (remaining dictionary of keyword arguments, *optional*):
Can be used to overwrite load and saveable variables (the pipeline components of the specific pipeline
class). The overwritten components are passed directly to the pipelines `__init__` method. See example
below for more information.
<Tip>
To use private or [gated](https://huggingface.co/docs/hub/models-gated#gated-models) models, log-in with
`huggingface-cli login`.
</Tip>
Examples:
```py
>>> from diffusers import AutoPipelineForText2Image
>>> pipeline = AutoPipelineForText2Image.from_huggingface("stable-diffusion-v1-5")
>>> image = pipeline(prompt).images[0]
```
"""
# Update kwargs to ensure the model is downloaded and parameters are included
_status = {
"download": True,
"include_params": True,
"skip_error": False,
"model_type": "Checkpoint",
}
kwargs.update(_status)
# Search for the model on Civitai and get the model status
model_status = search_civitai(pretrained_model_link_or_path, **kwargs)
logger.warning(f"checkpoint_path: {model_status.model_status.download_url}")
checkpoint_path = model_status.model_path
# Load the pipeline from a single file checkpoint
return load_pipeline_from_single_file(
pretrained_model_or_path=checkpoint_path,
pipeline_mapping=SINGLE_FILE_CHECKPOINT_IMAGE2IMAGE_PIPELINE_MAPPING,
**kwargs,
)
class EasyPipelineForInpainting(AutoPipelineForInpainting):
r"""
[`AutoPipelineForInpainting`] is a generic pipeline class that instantiates an inpainting pipeline class. The
specific underlying pipeline class is automatically selected from either the
[`~AutoPipelineForInpainting.from_pretrained`] or [`~AutoPipelineForInpainting.from_pipe`] methods.
This class cannot be instantiated using `__init__()` (throws an error).
Class attributes:
- **config_name** (`str`) -- The configuration filename that stores the class and module names of all the
diffusion pipeline's components.
"""
config_name = "model_index.json"
def __init__(self, *args, **kwargs):
# EnvironmentError is returned
super().__init__()
@classmethod
@validate_hf_hub_args
def from_huggingface(cls, pretrained_model_link_or_path, **kwargs):
r"""
Parameters:
pretrained_model_or_path (`str` or `os.PathLike`, *optional*):
Can be either:
- A keyword to search for Hugging Face (for example `Stable Diffusion`)
- Link to `.ckpt` or `.safetensors` file (for example
`"https://huggingface.co/<repo_id>/blob/main/<path_to_file>.safetensors"`) on the Hub.
- A string, the *repo id* (for example `CompVis/ldm-text2im-large-256`) of a pretrained pipeline
hosted on the Hub.
- A path to a *directory* (for example `./my_pipeline_directory/`) containing pipeline weights
saved using
[`~DiffusionPipeline.save_pretrained`].
checkpoint_format (`str`, *optional*, defaults to `"single_file"`):
The format of the model checkpoint.
pipeline_tag (`str`, *optional*):
Tag to filter models by pipeline.
torch_dtype (`str` or `torch.dtype`, *optional*):
Override the default `torch.dtype` and load the model with another dtype. If "auto" is passed, the
dtype is automatically derived from the model's weights.
force_download (`bool`, *optional*, defaults to `False`):
Whether or not to force the (re-)download of the model weights and configuration files, overriding the
cached versions if they exist.
cache_dir (`Union[str, os.PathLike]`, *optional*):
Path to a directory where a downloaded pretrained model configuration is cached if the standard cache
is not used.
proxies (`Dict[str, str]`, *optional*):
A dictionary of proxy servers to use by protocol or endpoint, for example, `{'http': 'foo.bar:3128',
'http://hostname': 'foo.bar:4012'}`. The proxies are used on each request.
output_loading_info(`bool`, *optional*, defaults to `False`):
Whether or not to also return a dictionary containing missing keys, unexpected keys and error messages.
local_files_only (`bool`, *optional*, defaults to `False`):
Whether to only load local model weights and configuration files or not. If set to `True`, the model
won't be downloaded from the Hub.
token (`str` or *bool*, *optional*):
The token to use as HTTP bearer authorization for remote files. If `True`, the token generated from
`diffusers-cli login` (stored in `~/.huggingface`) is used.
revision (`str`, *optional*, defaults to `"main"`):
The specific model version to use. It can be a branch name, a tag name, a commit id, or any identifier
allowed by Git.
custom_revision (`str`, *optional*, defaults to `"main"`):
The specific model version to use. It can be a branch name, a tag name, or a commit id similar to
`revision` when loading a custom pipeline from the Hub. It can be a 🤗 Diffusers version when loading a
custom pipeline from GitHub, otherwise it defaults to `"main"` when loading from the Hub.
mirror (`str`, *optional*):
Mirror source to resolve accessibility issues if you’re downloading a model in China. We do not
guarantee the timeliness or safety of the source, and you should refer to the mirror site for more
information.
device_map (`str` or `Dict[str, Union[int, str, torch.device]]`, *optional*):
A map that specifies where each submodule should go. It doesn’t need to be defined for each
parameter/buffer name; once a given module name is inside, every submodule of it will be sent to the
same device.
Set `device_map="auto"` to have 🤗 Accelerate automatically compute the most optimized `device_map`. For
more information about each option see [designing a device
map](https://hf.co/docs/accelerate/main/en/usage_guides/big_modeling#designing-a-device-map).
max_memory (`Dict`, *optional*):
A dictionary device identifier for the maximum memory. Will default to the maximum memory available for
each GPU and the available CPU RAM if unset.
offload_folder (`str` or `os.PathLike`, *optional*):
The path to offload weights if device_map contains the value `"disk"`.
offload_state_dict (`bool`, *optional*):
If `True`, temporarily offloads the CPU state dict to the hard drive to avoid running out of CPU RAM if
the weight of the CPU state dict + the biggest shard of the checkpoint does not fit. Defaults to `True`
when there is some disk offload.
low_cpu_mem_usage (`bool`, *optional*, defaults to `True` if torch version >= 1.9.0 else `False`):
Speed up model loading only loading the pretrained weights and not initializing the weights. This also
tries to not use more than 1x model size in CPU memory (including peak memory) while loading the model.
Only supported for PyTorch >= 1.9.0. If you are using an older version of PyTorch, setting this
argument to `True` will raise an error.
use_safetensors (`bool`, *optional*, defaults to `None`):
If set to `None`, the safetensors weights are downloaded if they're available **and** if the
safetensors library is installed. If set to `True`, the model is forcibly loaded from safetensors
weights. If set to `False`, safetensors weights are not loaded.
gated (`bool`, *optional*, defaults to `False` ):
A boolean to filter models on the Hub that are gated or not.
kwargs (remaining dictionary of keyword arguments, *optional*):
Can be used to overwrite load and saveable variables (the pipeline components of the specific pipeline
class). The overwritten components are passed directly to the pipelines `__init__` method. See example
below for more information.
variant (`str`, *optional*):
Load weights from a specified variant filename such as `"fp16"` or `"ema"`. This is ignored when
loading `from_flax`.
<Tip>
To use private or [gated](https://huggingface.co/docs/hub/models-gated#gated-models) models, log-in with
`huggingface-cli login`.
</Tip>
Examples:
```py
>>> from diffusers import AutoPipelineForText2Image
>>> pipeline = AutoPipelineForText2Image.from_huggingface("stable-diffusion-v1-5")
>>> image = pipeline(prompt).images[0]
```
"""
# Update kwargs to ensure the model is downloaded and parameters are included
_status = {
"download": True,
"include_params": True,
"skip_error": False,
"pipeline_tag": "image-to-image",
}
kwargs.update(_status)
# Search for the model on Hugging Face and get the model status
model_status = search_huggingface(pretrained_model_link_or_path, **kwargs)
logger.warning(f"checkpoint_path: {model_status.model_status.download_url}")
checkpoint_path = model_status.model_path
# Check the format of the model checkpoint
if model_status.checkpoint_format == "single_file":
# Load the pipeline from a single file checkpoint
return load_pipeline_from_single_file(
pretrained_model_or_path=checkpoint_path,
pipeline_mapping=SINGLE_FILE_CHECKPOINT_INPAINT_PIPELINE_MAPPING,
**kwargs,
)
else:
return cls.from_pretrained(checkpoint_path, **kwargs)
@classmethod
def from_civitai(cls, pretrained_model_link_or_path, **kwargs):
r"""
Parameters:
pretrained_model_or_path (`str` or `os.PathLike`, *optional*):
Can be either:
- A keyword to search for Hugging Face (for example `Stable Diffusion`)
- Link to `.ckpt` or `.safetensors` file (for example
`"https://huggingface.co/<repo_id>/blob/main/<path_to_file>.safetensors"`) on the Hub.
- A string, the *repo id* (for example `CompVis/ldm-text2im-large-256`) of a pretrained pipeline
hosted on the Hub.
- A path to a *directory* (for example `./my_pipeline_directory/`) containing pipeline weights
saved using
[`~DiffusionPipeline.save_pretrained`].
model_type (`str`, *optional*, defaults to `Checkpoint`):
The type of model to search for. (for example `Checkpoint`, `TextualInversion`, `LORA`, `Controlnet`)
base_model (`str`, *optional*):
The base model to filter by.
cache_dir (`str`, `Path`, *optional*):
Path to the folder where cached files are stored.
resume (`bool`, *optional*, defaults to `False`):
Whether to resume an incomplete download.
torch_dtype (`str` or `torch.dtype`, *optional*):
Override the default `torch.dtype` and load the model with another dtype. If "auto" is passed, the
dtype is automatically derived from the model's weights.
force_download (`bool`, *optional*, defaults to `False`):
Whether or not to force the (re-)download of the model weights and configuration files, overriding the
cached versions if they exist.
output_loading_info(`bool`, *optional*, defaults to `False`):
Whether or not to also return a dictionary containing missing keys, unexpected keys and error messages.
local_files_only (`bool`, *optional*, defaults to `False`):
Whether to only load local model weights and configuration files or not. If set to `True`, the model
won't be downloaded from the Hub.
token (`str`, *optional*):
The token to use as HTTP bearer authorization for remote files.
device_map (`str` or `Dict[str, Union[int, str, torch.device]]`, *optional*):
A map that specifies where each submodule should go. It doesn’t need to be defined for each
parameter/buffer name; once a given module name is inside, every submodule of it will be sent to the
same device.
Set `device_map="auto"` to have 🤗 Accelerate automatically compute the most optimized `device_map`. For
more information about each option see [designing a device
map](https://hf.co/docs/accelerate/main/en/usage_guides/big_modeling#designing-a-device-map).
max_memory (`Dict`, *optional*):
A dictionary device identifier for the maximum memory. Will default to the maximum memory available for
each GPU and the available CPU RAM if unset.
offload_folder (`str` or `os.PathLike`, *optional*):
The path to offload weights if device_map contains the value `"disk"`.
offload_state_dict (`bool`, *optional*):
If `True`, temporarily offloads the CPU state dict to the hard drive to avoid running out of CPU RAM if
the weight of the CPU state dict + the biggest shard of the checkpoint does not fit. Defaults to `True`
when there is some disk offload.
low_cpu_mem_usage (`bool`, *optional*, defaults to `True` if torch version >= 1.9.0 else `False`):
Speed up model loading only loading the pretrained weights and not initializing the weights. This also
tries to not use more than 1x model size in CPU memory (including peak memory) while loading the model.
Only supported for PyTorch >= 1.9.0. If you are using an older version of PyTorch, setting this
argument to `True` will raise an error.
use_safetensors (`bool`, *optional*, defaults to `None`):
If set to `None`, the safetensors weights are downloaded if they're available **and** if the
safetensors library is installed. If set to `True`, the model is forcibly loaded from safetensors
weights. If set to `False`, safetensors weights are not loaded.
kwargs (remaining dictionary of keyword arguments, *optional*):
Can be used to overwrite load and saveable variables (the pipeline components of the specific pipeline
class). The overwritten components are passed directly to the pipelines `__init__` method. See example
below for more information.
<Tip>
To use private or [gated](https://huggingface.co/docs/hub/models-gated#gated-models) models, log-in with
`huggingface-cli login`.
</Tip>
Examples:
```py
>>> from diffusers import AutoPipelineForText2Image
>>> pipeline = AutoPipelineForText2Image.from_huggingface("stable-diffusion-v1-5")
>>> image = pipeline(prompt).images[0]
```
"""
# Update kwargs to ensure the model is downloaded and parameters are included
_status = {
"download": True,
"include_params": True,
"skip_error": False,
"model_type": "Checkpoint",
}
kwargs.update(_status)
# Search for the model on Civitai and get the model status
model_status = search_civitai(pretrained_model_link_or_path, **kwargs)
logger.warning(f"checkpoint_path: {model_status.model_status.download_url}")
checkpoint_path = model_status.model_path
# Load the pipeline from a single file checkpoint
return load_pipeline_from_single_file(
pretrained_model_or_path=checkpoint_path,
pipeline_mapping=SINGLE_FILE_CHECKPOINT_INPAINT_PIPELINE_MAPPING,
**kwargs,
)
|