forked from huggingface/diffusers
-
Notifications
You must be signed in to change notification settings - Fork 0
Expand file tree
/
Copy pathhub_utils.py
More file actions
452 lines (405 loc) 路 18.8 KB
/
Copy pathhub_utils.py
File metadata and controls
452 lines (405 loc) 路 18.8 KB
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
# coding=utf-8
# Copyright 2023 The HuggingFace Inc. team.
#
# 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
import sys
import tempfile
import traceback
import warnings
from pathlib import Path
from typing import Dict, Optional, Union
from uuid import uuid4
from huggingface_hub import (
ModelCard,
ModelCardData,
create_repo,
get_full_repo_name,
hf_hub_download,
upload_folder,
)
from huggingface_hub.constants import HF_HUB_CACHE, HF_HUB_DISABLE_TELEMETRY, HF_HUB_OFFLINE
from huggingface_hub.file_download import REGEX_COMMIT_HASH
from huggingface_hub.utils import (
EntryNotFoundError,
RepositoryNotFoundError,
RevisionNotFoundError,
is_jinja_available,
validate_hf_hub_args,
)
from packaging import version
from requests import HTTPError
from .. import __version__
from .constants import (
DEPRECATED_REVISION_ARGS,
HUGGINGFACE_CO_RESOLVE_ENDPOINT,
SAFETENSORS_WEIGHTS_NAME,
WEIGHTS_NAME,
)
from .import_utils import (
ENV_VARS_TRUE_VALUES,
_flax_version,
_jax_version,
_onnxruntime_version,
_torch_version,
is_flax_available,
is_onnx_available,
is_torch_available,
)
from .logging import get_logger
logger = get_logger(__name__)
MODEL_CARD_TEMPLATE_PATH = Path(__file__).parent / "model_card_template.md"
SESSION_ID = uuid4().hex
def http_user_agent(user_agent: Union[Dict, str, None] = None) -> str:
"""
Formats a user-agent string with basic info about a request.
"""
ua = f"diffusers/{__version__}; python/{sys.version.split()[0]}; session_id/{SESSION_ID}"
if HF_HUB_DISABLE_TELEMETRY or HF_HUB_OFFLINE:
return ua + "; telemetry/off"
if is_torch_available():
ua += f"; torch/{_torch_version}"
if is_flax_available():
ua += f"; jax/{_jax_version}"
ua += f"; flax/{_flax_version}"
if is_onnx_available():
ua += f"; onnxruntime/{_onnxruntime_version}"
# CI will set this value to True
if os.environ.get("DIFFUSERS_IS_CI", "").upper() in ENV_VARS_TRUE_VALUES:
ua += "; is_ci/true"
if isinstance(user_agent, dict):
ua += "; " + "; ".join(f"{k}/{v}" for k, v in user_agent.items())
elif isinstance(user_agent, str):
ua += "; " + user_agent
return ua
def create_model_card(args, model_name):
if not is_jinja_available():
raise ValueError(
"Modelcard rendering is based on Jinja templates."
" Please make sure to have `jinja` installed before using `create_model_card`."
" To install it, please run `pip install Jinja2`."
)
if hasattr(args, "local_rank") and args.local_rank not in [-1, 0]:
return
hub_token = args.hub_token if hasattr(args, "hub_token") else None
repo_name = get_full_repo_name(model_name, token=hub_token)
model_card = ModelCard.from_template(
card_data=ModelCardData( # Card metadata object that will be converted to YAML block
language="en",
license="apache-2.0",
library_name="diffusers",
tags=[],
datasets=args.dataset_name,
metrics=[],
),
template_path=MODEL_CARD_TEMPLATE_PATH,
model_name=model_name,
repo_name=repo_name,
dataset_name=args.dataset_name if hasattr(args, "dataset_name") else None,
learning_rate=args.learning_rate,
train_batch_size=args.train_batch_size,
eval_batch_size=args.eval_batch_size,
gradient_accumulation_steps=(
args.gradient_accumulation_steps if hasattr(args, "gradient_accumulation_steps") else None
),
adam_beta1=args.adam_beta1 if hasattr(args, "adam_beta1") else None,
adam_beta2=args.adam_beta2 if hasattr(args, "adam_beta2") else None,
adam_weight_decay=args.adam_weight_decay if hasattr(args, "adam_weight_decay") else None,
adam_epsilon=args.adam_epsilon if hasattr(args, "adam_epsilon") else None,
lr_scheduler=args.lr_scheduler if hasattr(args, "lr_scheduler") else None,
lr_warmup_steps=args.lr_warmup_steps if hasattr(args, "lr_warmup_steps") else None,
ema_inv_gamma=args.ema_inv_gamma if hasattr(args, "ema_inv_gamma") else None,
ema_power=args.ema_power if hasattr(args, "ema_power") else None,
ema_max_decay=args.ema_max_decay if hasattr(args, "ema_max_decay") else None,
mixed_precision=args.mixed_precision,
)
card_path = os.path.join(args.output_dir, "README.md")
model_card.save(card_path)
def extract_commit_hash(resolved_file: Optional[str], commit_hash: Optional[str] = None):
"""
Extracts the commit hash from a resolved filename toward a cache file.
"""
if resolved_file is None or commit_hash is not None:
return commit_hash
resolved_file = str(Path(resolved_file).as_posix())
search = re.search(r"snapshots/([^/]+)/", resolved_file)
if search is None:
return None
commit_hash = search.groups()[0]
return commit_hash if REGEX_COMMIT_HASH.match(commit_hash) else None
# Old default cache path, potentially to be migrated.
# This logic was more or less taken from `transformers`, with the following differences:
# - Diffusers doesn't use custom environment variables to specify the cache path.
# - There is no need to migrate the cache format, just move the files to the new location.
hf_cache_home = os.path.expanduser(
os.getenv("HF_HOME", os.path.join(os.getenv("XDG_CACHE_HOME", "~/.cache"), "huggingface"))
)
old_diffusers_cache = os.path.join(hf_cache_home, "diffusers")
def move_cache(old_cache_dir: Optional[str] = None, new_cache_dir: Optional[str] = None) -> None:
if new_cache_dir is None:
new_cache_dir = HF_HUB_CACHE
if old_cache_dir is None:
old_cache_dir = old_diffusers_cache
old_cache_dir = Path(old_cache_dir).expanduser()
new_cache_dir = Path(new_cache_dir).expanduser()
for old_blob_path in old_cache_dir.glob("**/blobs/*"):
if old_blob_path.is_file() and not old_blob_path.is_symlink():
new_blob_path = new_cache_dir / old_blob_path.relative_to(old_cache_dir)
new_blob_path.parent.mkdir(parents=True, exist_ok=True)
os.replace(old_blob_path, new_blob_path)
try:
os.symlink(new_blob_path, old_blob_path)
except OSError:
logger.warning(
"Could not create symlink between old cache and new cache. If you use an older version of diffusers again, files will be re-downloaded."
)
# At this point, old_cache_dir contains symlinks to the new cache (it can still be used).
cache_version_file = os.path.join(HF_HUB_CACHE, "version_diffusers_cache.txt")
if not os.path.isfile(cache_version_file):
cache_version = 0
else:
with open(cache_version_file) as f:
try:
cache_version = int(f.read())
except ValueError:
cache_version = 0
if cache_version < 1:
old_cache_is_not_empty = os.path.isdir(old_diffusers_cache) and len(os.listdir(old_diffusers_cache)) > 0
if old_cache_is_not_empty:
logger.warning(
"The cache for model files in Diffusers v0.14.0 has moved to a new location. Moving your "
"existing cached models. This is a one-time operation, you can interrupt it or run it "
"later by calling `diffusers.utils.hub_utils.move_cache()`."
)
try:
move_cache()
except Exception as e:
trace = "\n".join(traceback.format_tb(e.__traceback__))
logger.error(
f"There was a problem when trying to move your cache:\n\n{trace}\n{e.__class__.__name__}: {e}\n\nPlease "
"file an issue at https://github.com/huggingface/diffusers/issues/new/choose, copy paste this whole "
"message and we will do our best to help."
)
if cache_version < 1:
try:
os.makedirs(HF_HUB_CACHE, exist_ok=True)
with open(cache_version_file, "w") as f:
f.write("1")
except Exception:
logger.warning(
f"There was a problem when trying to write in your cache folder ({HF_HUB_CACHE}). Please, ensure "
"the directory exists and can be written to."
)
def _add_variant(weights_name: str, variant: Optional[str] = None) -> str:
if variant is not None:
splits = weights_name.split(".")
splits = splits[:-1] + [variant] + splits[-1:]
weights_name = ".".join(splits)
return weights_name
@validate_hf_hub_args
def _get_model_file(
pretrained_model_name_or_path: Union[str, Path],
*,
weights_name: str,
subfolder: Optional[str] = None,
cache_dir: Optional[str] = None,
force_download: bool = False,
proxies: Optional[Dict] = None,
resume_download: bool = False,
local_files_only: bool = False,
token: Optional[str] = None,
user_agent: Optional[Union[Dict, str]] = None,
revision: Optional[str] = None,
commit_hash: Optional[str] = None,
):
pretrained_model_name_or_path = str(pretrained_model_name_or_path)
if os.path.isfile(pretrained_model_name_or_path):
return pretrained_model_name_or_path
elif os.path.isdir(pretrained_model_name_or_path):
if os.path.isfile(os.path.join(pretrained_model_name_or_path, weights_name)):
# Load from a PyTorch checkpoint
model_file = os.path.join(pretrained_model_name_or_path, weights_name)
return model_file
elif subfolder is not None and os.path.isfile(
os.path.join(pretrained_model_name_or_path, subfolder, weights_name)
):
model_file = os.path.join(pretrained_model_name_or_path, subfolder, weights_name)
return model_file
else:
raise EnvironmentError(
f"Error no file named {weights_name} found in directory {pretrained_model_name_or_path}."
)
else:
# 1. First check if deprecated way of loading from branches is used
if (
revision in DEPRECATED_REVISION_ARGS
and (weights_name == WEIGHTS_NAME or weights_name == SAFETENSORS_WEIGHTS_NAME)
and version.parse(version.parse(__version__).base_version) >= version.parse("0.22.0")
):
try:
model_file = hf_hub_download(
pretrained_model_name_or_path,
filename=_add_variant(weights_name, revision),
cache_dir=cache_dir,
force_download=force_download,
proxies=proxies,
resume_download=resume_download,
local_files_only=local_files_only,
token=token,
user_agent=user_agent,
subfolder=subfolder,
revision=revision or commit_hash,
)
warnings.warn(
f"Loading the variant {revision} from {pretrained_model_name_or_path} via `revision='{revision}'` is deprecated. Loading instead from `revision='main'` with `variant={revision}`. Loading model variants via `revision='{revision}'` will be removed in diffusers v1. Please use `variant='{revision}'` instead.",
FutureWarning,
)
return model_file
except: # noqa: E722
warnings.warn(
f"You are loading the variant {revision} from {pretrained_model_name_or_path} via `revision='{revision}'`. This behavior is deprecated and will be removed in diffusers v1. One should use `variant='{revision}'` instead. However, it appears that {pretrained_model_name_or_path} currently does not have a {_add_variant(weights_name, revision)} file in the 'main' branch of {pretrained_model_name_or_path}. \n The Diffusers team and community would be very grateful if you could open an issue: https://github.com/huggingface/diffusers/issues/new with the title '{pretrained_model_name_or_path} is missing {_add_variant(weights_name, revision)}' so that the correct variant file can be added.",
FutureWarning,
)
try:
# 2. Load model file as usual
model_file = hf_hub_download(
pretrained_model_name_or_path,
filename=weights_name,
cache_dir=cache_dir,
force_download=force_download,
proxies=proxies,
resume_download=resume_download,
local_files_only=local_files_only,
token=token,
user_agent=user_agent,
subfolder=subfolder,
revision=revision or commit_hash,
)
return model_file
except RepositoryNotFoundError:
raise EnvironmentError(
f"{pretrained_model_name_or_path} is not a local folder and is not a valid model identifier "
"listed on 'https://huggingface.co/models'\nIf this is a private repository, make sure to pass a "
"token having permission to this repo with `token` or log in with `huggingface-cli "
"login`."
)
except RevisionNotFoundError:
raise EnvironmentError(
f"{revision} is not a valid git identifier (branch name, tag name or commit id) that exists for "
"this model name. Check the model page at "
f"'https://huggingface.co/{pretrained_model_name_or_path}' for available revisions."
)
except EntryNotFoundError:
raise EnvironmentError(
f"{pretrained_model_name_or_path} does not appear to have a file named {weights_name}."
)
except HTTPError as err:
raise EnvironmentError(
f"There was a specific connection error when trying to load {pretrained_model_name_or_path}:\n{err}"
)
except ValueError:
raise EnvironmentError(
f"We couldn't connect to '{HUGGINGFACE_CO_RESOLVE_ENDPOINT}' to load this model, couldn't find it"
f" in the cached files and it looks like {pretrained_model_name_or_path} is not the path to a"
f" directory containing a file named {weights_name} or"
" \nCheckout your internet connection or see how to run the library in"
" offline mode at 'https://huggingface.co/docs/diffusers/installation#offline-mode'."
)
except EnvironmentError:
raise EnvironmentError(
f"Can't load the model for '{pretrained_model_name_or_path}'. If you were trying to load it from "
"'https://huggingface.co/models', make sure you don't have a local directory with the same name. "
f"Otherwise, make sure '{pretrained_model_name_or_path}' is the correct path to a directory "
f"containing a file named {weights_name}"
)
class PushToHubMixin:
"""
A Mixin to push a model, scheduler, or pipeline to the Hugging Face Hub.
"""
def _upload_folder(
self,
working_dir: Union[str, os.PathLike],
repo_id: str,
token: Optional[str] = None,
commit_message: Optional[str] = None,
create_pr: bool = False,
):
"""
Uploads all files in `working_dir` to `repo_id`.
"""
if commit_message is None:
if "Model" in self.__class__.__name__:
commit_message = "Upload model"
elif "Scheduler" in self.__class__.__name__:
commit_message = "Upload scheduler"
else:
commit_message = f"Upload {self.__class__.__name__}"
logger.info(f"Uploading the files of {working_dir} to {repo_id}.")
return upload_folder(
repo_id=repo_id, folder_path=working_dir, token=token, commit_message=commit_message, create_pr=create_pr
)
def push_to_hub(
self,
repo_id: str,
commit_message: Optional[str] = None,
private: Optional[bool] = None,
token: Optional[str] = None,
create_pr: bool = False,
safe_serialization: bool = True,
variant: Optional[str] = None,
) -> str:
"""
Upload model, scheduler, or pipeline files to the 馃 Hugging Face Hub.
Parameters:
repo_id (`str`):
The name of the repository you want to push your model, scheduler, or pipeline files to. It should
contain your organization name when pushing to an organization. `repo_id` can also be a path to a local
directory.
commit_message (`str`, *optional*):
Message to commit while pushing. Default to `"Upload {object}"`.
private (`bool`, *optional*):
Whether or not the repository created should be private.
token (`str`, *optional*):
The token to use as HTTP bearer authorization for remote files. The token generated when running
`huggingface-cli login` (stored in `~/.huggingface`).
create_pr (`bool`, *optional*, defaults to `False`):
Whether or not to create a PR with the uploaded files or directly commit.
safe_serialization (`bool`, *optional*, defaults to `True`):
Whether or not to convert the model weights to the `safetensors` format.
variant (`str`, *optional*):
If specified, weights are saved in the format `pytorch_model.<variant>.bin`.
Examples:
```python
from diffusers import UNet2DConditionModel
unet = UNet2DConditionModel.from_pretrained("stabilityai/stable-diffusion-2", subfolder="unet")
# Push the `unet` to your namespace with the name "my-finetuned-unet".
unet.push_to_hub("my-finetuned-unet")
# Push the `unet` to an organization with the name "my-finetuned-unet".
unet.push_to_hub("your-org/my-finetuned-unet")
```
"""
repo_id = create_repo(repo_id, private=private, token=token, exist_ok=True).repo_id
# Save all files.
save_kwargs = {"safe_serialization": safe_serialization}
if "Scheduler" not in self.__class__.__name__:
save_kwargs.update({"variant": variant})
with tempfile.TemporaryDirectory() as tmpdir:
self.save_pretrained(tmpdir, **save_kwargs)
return self._upload_folder(
tmpdir,
repo_id,
token=token,
commit_message=commit_message,
create_pr=create_pr,
)