kanshan/diffusers
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Ruff: apply same rules as in transformers (#2827)

Browse Source 
* Apply same ruff settings as in transformers

See https://github.com/huggingface/transformers/blob/main/pyproject.toml
Co-authored-by: Aaron Gokaslan <aaronGokaslan@gmail.com>

* Apply new style rules

* Style

Co-authored-by: Aaron Gokaslan <aaronGokaslan@gmail.com>

* style

* remove list, ruff wouldn't auto fix.

---------

Co-authored-by: Aaron Gokaslan <aaronGokaslan@gmail.com>
This commit is contained in:
Pedro Cuenca 2023-03-27 16:18:57 +02:00 committed by GitHub
parent abb22b4eeb
commit 1d7b4b60b7
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
45 changed files with 209 additions and 213 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

20
examples/community/checkpoint_merger.py
View File

@ -199,24 +199,20 @@ class CheckpointMergerPipeline(DiffusionPipeline):
if not attr.startswith("_"):
checkpoint_path_1 = os.path.join(cached_folders[1], attr)
if os.path.exists(checkpoint_path_1):
files = list(
(
*glob.glob(os.path.join(checkpoint_path_1, "*.safetensors")),
*glob.glob(os.path.join(checkpoint_path_1, "*.bin")),
)
)
files = [
*glob.glob(os.path.join(checkpoint_path_1, "*.safetensors")),
*glob.glob(os.path.join(checkpoint_path_1, "*.bin")),
]
checkpoint_path_1 = files[0] if len(files) > 0 else None
if len(cached_folders) < 3:
checkpoint_path_2 = None
else:
checkpoint_path_2 = os.path.join(cached_folders[2], attr)
if os.path.exists(checkpoint_path_2):
files = list(
(
*glob.glob(os.path.join(checkpoint_path_2, "*.safetensors")),
*glob.glob(os.path.join(checkpoint_path_2, "*.bin")),
)
)
files = [
*glob.glob(os.path.join(checkpoint_path_2, "*.safetensors")),
*glob.glob(os.path.join(checkpoint_path_2, "*.bin")),
]
checkpoint_path_2 = files[0] if len(files) > 0 else None
# For an attr if both checkpoint_path_1 and 2 are None, ignore.
# If atleast one is present, deal with it according to interp method, of course only if the state_dict keys match.

2
examples/community/imagic_stable_diffusion.py
View File

@ -48,7 +48,7 @@ logger = logging.get_logger(__name__) # pylint: disable=invalid-name
def preprocess(image):
w, h = image.size
w, h = map(lambda x: x - x % 32, (w, h)) # resize to integer multiple of 32
w, h = (x - x % 32 for x in (w, h)) # resize to integer multiple of 32
image = image.resize((w, h), resample=PIL_INTERPOLATION["lanczos"])
image = np.array(image).astype(np.float32) / 255.0
image = image[None].transpose(0, 3, 1, 2)

4
examples/community/lpw_stable_diffusion.py
View File

@ -376,7 +376,7 @@ def get_weighted_text_embeddings(
def preprocess_image(image):
w, h = image.size
w, h = map(lambda x: x - x % 32, (w, h)) # resize to integer multiple of 32
w, h = (x - x % 32 for x in (w, h)) # resize to integer multiple of 32
image = image.resize((w, h), resample=PIL_INTERPOLATION["lanczos"])
image = np.array(image).astype(np.float32) / 255.0
image = image[None].transpose(0, 3, 1, 2)
@ -387,7 +387,7 @@ def preprocess_image(image):
def preprocess_mask(mask, scale_factor=8):
mask = mask.convert("L")
w, h = mask.size
w, h = map(lambda x: x - x % 32, (w, h)) # resize to integer multiple of 32
w, h = (x - x % 32 for x in (w, h)) # resize to integer multiple of 32
mask = mask.resize((w // scale_factor, h // scale_factor), resample=PIL_INTERPOLATION["nearest"])
mask = np.array(mask).astype(np.float32) / 255.0
mask = np.tile(mask, (4, 1, 1))

4
examples/community/lpw_stable_diffusion_onnx.py
View File

@ -403,7 +403,7 @@ def get_weighted_text_embeddings(
def preprocess_image(image):
w, h = image.size
w, h = map(lambda x: x - x % 32, (w, h)) # resize to integer multiple of 32
w, h = (x - x % 32 for x in (w, h)) # resize to integer multiple of 32
image = image.resize((w, h), resample=PIL_INTERPOLATION["lanczos"])
image = np.array(image).astype(np.float32) / 255.0
image = image[None].transpose(0, 3, 1, 2)
@ -413,7 +413,7 @@ def preprocess_image(image):
def preprocess_mask(mask, scale_factor=8):
mask = mask.convert("L")
w, h = mask.size
w, h = map(lambda x: x - x % 32, (w, h)) # resize to integer multiple of 32
w, h = (x - x % 32 for x in (w, h)) # resize to integer multiple of 32
mask = mask.resize((w // scale_factor, h // scale_factor), resample=PIL_INTERPOLATION["nearest"])
mask = np.array(mask).astype(np.float32) / 255.0
mask = np.tile(mask, (4, 1, 1))

2
examples/community/stable_unclip.py
View File

@ -46,7 +46,7 @@ class StableUnCLIPPipeline(DiffusionPipeline):
):
super().__init__()
decoder_pipe_kwargs = dict(image_encoder=None) if decoder_pipe_kwargs is None else decoder_pipe_kwargs
decoder_pipe_kwargs = {"image_encoder": None} if decoder_pipe_kwargs is None else decoder_pipe_kwargs
decoder_pipe_kwargs["torch_dtype"] = decoder_pipe_kwargs.get("torch_dtype", None) or prior.dtype

2
examples/instruct_pix2pix/train_instruct_pix2pix.py
View File

@ -673,7 +673,7 @@ def main():
examples["edited_pixel_values"] = edited_images
# Preprocess the captions.
captions = [caption for caption in examples[edit_prompt_column]]
captions = list(examples[edit_prompt_column])
examples["input_ids"] = tokenize_captions(captions)
return examples

22
examples/rl/run_diffuser_locomotion.py
View File

@ -4,17 +4,17 @@ import tqdm
from diffusers.experimental import ValueGuidedRLPipeline
config = dict(
n_samples=64,
horizon=32,
num_inference_steps=20,
n_guide_steps=2, # can set to 0 for faster sampling, does not use value network
scale_grad_by_std=True,
scale=0.1,
eta=0.0,
t_grad_cutoff=2,
device="cpu",
)
config = {
"n_samples": 64,
"horizon": 32,
"num_inference_steps": 20,
"n_guide_steps": 2, # can set to 0 for faster sampling, does not use value network
"scale_grad_by_std": True,
"scale": 0.1,
"eta": 0.0,
"t_grad_cutoff": 2,
"device": "cpu",
}
if __name__ == "__main__":

4
pyproject.toml
View File

@ -4,8 +4,8 @@ target-version = ['py37']
[tool.ruff]
# Never enforce `E501` (line length violations).
ignore = ["E501", "E741", "W605"]
select = ["E", "F", "I", "W"]
ignore = ["C901", "E501", "E741", "W605"]
select = ["C", "E", "F", "I", "W"]
line-length = 119
# Ignore import violations in all `__init__.py` files.

2
scripts/convert_ddpm_original_checkpoint_to_diffusers.py
View File

@ -404,7 +404,7 @@ if __name__ == "__main__":
config = json.loads(f.read())
# unet case
key_prefix_set = set(key.split(".")[0] for key in checkpoint.keys())
key_prefix_set = {key.split(".")[0] for key in checkpoint.keys()}
if "encoder" in key_prefix_set and "decoder" in key_prefix_set:
converted_checkpoint = convert_vq_autoenc_checkpoint(checkpoint, config)
else:

80
scripts/convert_models_diffuser_to_diffusers.py
View File

@ -24,29 +24,29 @@ def unet(hor):
up_block_types = ("UpResnetBlock1D", "UpResnetBlock1D", "UpResnetBlock1D")
model = torch.load(f"/Users/bglickenhaus/Documents/diffuser/temporal_unet-hopper-mediumv2-hor{hor}.torch")
state_dict = model.state_dict()
config = dict(
down_block_types=down_block_types,
block_out_channels=block_out_channels,
up_block_types=up_block_types,
layers_per_block=1,
use_timestep_embedding=True,
out_block_type="OutConv1DBlock",
norm_num_groups=8,
downsample_each_block=False,
in_channels=14,
out_channels=14,
extra_in_channels=0,
time_embedding_type="positional",
flip_sin_to_cos=False,
freq_shift=1,
sample_size=65536,
mid_block_type="MidResTemporalBlock1D",
act_fn="mish",
)
config = {
"down_block_types": down_block_types,
"block_out_channels": block_out_channels,
"up_block_types": up_block_types,
"layers_per_block": 1,
"use_timestep_embedding": True,
"out_block_type": "OutConv1DBlock",
"norm_num_groups": 8,
"downsample_each_block": False,
"in_channels": 14,
"out_channels": 14,
"extra_in_channels": 0,
"time_embedding_type": "positional",
"flip_sin_to_cos": False,
"freq_shift": 1,
"sample_size": 65536,
"mid_block_type": "MidResTemporalBlock1D",
"act_fn": "mish",
}
hf_value_function = UNet1DModel(**config)
print(f"length of state dict: {len(state_dict.keys())}")
print(f"length of value function dict: {len(hf_value_function.state_dict().keys())}")
mapping = dict((k, hfk) for k, hfk in zip(model.state_dict().keys(), hf_value_function.state_dict().keys()))
mapping = dict(zip(model.state_dict().keys(), hf_value_function.state_dict().keys()))
for k, v in mapping.items():
state_dict[v] = state_dict.pop(k)
hf_value_function.load_state_dict(state_dict)
@ -57,25 +57,25 @@ def unet(hor):
def value_function():
config = dict(
in_channels=14,
down_block_types=("DownResnetBlock1D", "DownResnetBlock1D", "DownResnetBlock1D", "DownResnetBlock1D"),
up_block_types=(),
out_block_type="ValueFunction",
mid_block_type="ValueFunctionMidBlock1D",
block_out_channels=(32, 64, 128, 256),
layers_per_block=1,
downsample_each_block=True,
sample_size=65536,
out_channels=14,
extra_in_channels=0,
time_embedding_type="positional",
use_timestep_embedding=True,
flip_sin_to_cos=False,
freq_shift=1,
norm_num_groups=8,
act_fn="mish",
)
config = {
"in_channels": 14,
"down_block_types": ("DownResnetBlock1D", "DownResnetBlock1D", "DownResnetBlock1D", "DownResnetBlock1D"),
"up_block_types": (),
"out_block_type": "ValueFunction",
"mid_block_type": "ValueFunctionMidBlock1D",
"block_out_channels": (32, 64, 128, 256),
"layers_per_block": 1,
"downsample_each_block": True,
"sample_size": 65536,
"out_channels": 14,
"extra_in_channels": 0,
"time_embedding_type": "positional",
"use_timestep_embedding": True,
"flip_sin_to_cos": False,
"freq_shift": 1,
"norm_num_groups": 8,
"act_fn": "mish",
}
model = torch.load("/Users/bglickenhaus/Documents/diffuser/value_function-hopper-mediumv2-hor32.torch")
state_dict = model
@ -83,7 +83,7 @@ def value_function():
print(f"length of state dict: {len(state_dict.keys())}")
print(f"length of value function dict: {len(hf_value_function.state_dict().keys())}")
mapping = dict((k, hfk) for k, hfk in zip(state_dict.keys(), hf_value_function.state_dict().keys()))
mapping = dict(zip(state_dict.keys(), hf_value_function.state_dict().keys()))
for k, v in mapping.items():
state_dict[v] = state_dict.pop(k)

68
scripts/convert_original_audioldm_to_diffusers.py
View File

@ -246,19 +246,19 @@ def create_unet_diffusers_config(original_config, image_size: int):
)
class_embeddings_concat = unet_params.extra_film_use_concat if "extra_film_use_concat" in unet_params else None
config = dict(
sample_size=image_size // vae_scale_factor,
in_channels=unet_params.in_channels,
out_channels=unet_params.out_channels,
down_block_types=tuple(down_block_types),
up_block_types=tuple(up_block_types),
block_out_channels=tuple(block_out_channels),
layers_per_block=unet_params.num_res_blocks,
cross_attention_dim=cross_attention_dim,
class_embed_type=class_embed_type,
projection_class_embeddings_input_dim=projection_class_embeddings_input_dim,
class_embeddings_concat=class_embeddings_concat,
)
config = {
"sample_size": image_size // vae_scale_factor,
"in_channels": unet_params.in_channels,
"out_channels": unet_params.out_channels,
"down_block_types": tuple(down_block_types),
"up_block_types": tuple(up_block_types),
"block_out_channels": tuple(block_out_channels),
"layers_per_block": unet_params.num_res_blocks,
"cross_attention_dim": cross_attention_dim,
"class_embed_type": class_embed_type,
"projection_class_embeddings_input_dim": projection_class_embeddings_input_dim,
"class_embeddings_concat": class_embeddings_concat,
}
return config
@ -278,17 +278,17 @@ def create_vae_diffusers_config(original_config, checkpoint, image_size: int):
scaling_factor = checkpoint["scale_factor"] if "scale_by_std" in original_config.model.params else 0.18215
config = dict(
sample_size=image_size,
in_channels=vae_params.in_channels,
out_channels=vae_params.out_ch,
down_block_types=tuple(down_block_types),
up_block_types=tuple(up_block_types),
block_out_channels=tuple(block_out_channels),
latent_channels=vae_params.z_channels,
layers_per_block=vae_params.num_res_blocks,
scaling_factor=float(scaling_factor),
)
config = {
"sample_size": image_size,
"in_channels": vae_params.in_channels,
"out_channels": vae_params.out_ch,
"down_block_types": tuple(down_block_types),
"up_block_types": tuple(up_block_types),
"block_out_channels": tuple(block_out_channels),
"latent_channels": vae_params.z_channels,
"layers_per_block": vae_params.num_res_blocks,
"scaling_factor": float(scaling_factor),
}
return config
@ -670,18 +670,18 @@ def create_transformers_vocoder_config(original_config):
"""
vocoder_params = original_config.model.params.vocoder_config.params
config = dict(
model_in_dim=vocoder_params.num_mels,
sampling_rate=vocoder_params.sampling_rate,
upsample_initial_channel=vocoder_params.upsample_initial_channel,
upsample_rates=list(vocoder_params.upsample_rates),
upsample_kernel_sizes=list(vocoder_params.upsample_kernel_sizes),
resblock_kernel_sizes=list(vocoder_params.resblock_kernel_sizes),
resblock_dilation_sizes=[
config = {
"model_in_dim": vocoder_params.num_mels,
"sampling_rate": vocoder_params.sampling_rate,
"upsample_initial_channel": vocoder_params.upsample_initial_channel,
"upsample_rates": list(vocoder_params.upsample_rates),
"upsample_kernel_sizes": list(vocoder_params.upsample_kernel_sizes),
"resblock_kernel_sizes": list(vocoder_params.resblock_kernel_sizes),
"resblock_dilation_sizes": [
list(resblock_dilation) for resblock_dilation in vocoder_params.resblock_dilation_sizes
],
normalize_before=False,
)
"normalize_before": False,
}
return config

64
scripts/convert_versatile_diffusion_to_diffusers.py
View File

@ -280,17 +280,17 @@ def create_image_unet_diffusers_config(unet_params):
if not all(n == unet_params.num_noattn_blocks[0] for n in unet_params.num_noattn_blocks):
raise ValueError("Not all num_res_blocks are equal, which is not supported in this script.")
config = dict(
sample_size=None,
in_channels=unet_params.input_channels,
out_channels=unet_params.output_channels,
down_block_types=tuple(down_block_types),
up_block_types=tuple(up_block_types),
block_out_channels=tuple(block_out_channels),
layers_per_block=unet_params.num_noattn_blocks[0],
cross_attention_dim=unet_params.context_dim,
attention_head_dim=unet_params.num_heads,
)
config = {
"sample_size": None,
"in_channels": unet_params.input_channels,
"out_channels": unet_params.output_channels,
"down_block_types": tuple(down_block_types),
"up_block_types": tuple(up_block_types),
"block_out_channels": tuple(block_out_channels),
"layers_per_block": unet_params.num_noattn_blocks[0],
"cross_attention_dim": unet_params.context_dim,
"attention_head_dim": unet_params.num_heads,
}
return config
@ -319,17 +319,17 @@ def create_text_unet_diffusers_config(unet_params):
if not all(n == unet_params.num_noattn_blocks[0] for n in unet_params.num_noattn_blocks):
raise ValueError("Not all num_res_blocks are equal, which is not supported in this script.")
config = dict(
sample_size=None,
in_channels=(unet_params.input_channels, 1, 1),
out_channels=(unet_params.output_channels, 1, 1),
down_block_types=tuple(down_block_types),
up_block_types=tuple(up_block_types),
block_out_channels=tuple(block_out_channels),
layers_per_block=unet_params.num_noattn_blocks[0],
cross_attention_dim=unet_params.context_dim,
attention_head_dim=unet_params.num_heads,
)
config = {
"sample_size": None,
"in_channels": (unet_params.input_channels, 1, 1),
"out_channels": (unet_params.output_channels, 1, 1),
"down_block_types": tuple(down_block_types),
"up_block_types": tuple(up_block_types),
"block_out_channels": tuple(block_out_channels),
"layers_per_block": unet_params.num_noattn_blocks[0],
"cross_attention_dim": unet_params.context_dim,
"attention_head_dim": unet_params.num_heads,
}
return config
@ -343,16 +343,16 @@ def create_vae_diffusers_config(vae_params):
down_block_types = ["DownEncoderBlock2D"] * len(block_out_channels)
up_block_types = ["UpDecoderBlock2D"] * len(block_out_channels)
config = dict(
sample_size=vae_params.resolution,
in_channels=vae_params.in_channels,
out_channels=vae_params.out_ch,
down_block_types=tuple(down_block_types),
up_block_types=tuple(up_block_types),
block_out_channels=tuple(block_out_channels),
latent_channels=vae_params.z_channels,
layers_per_block=vae_params.num_res_blocks,
)
config = {
"sample_size": vae_params.resolution,
"in_channels": vae_params.in_channels,
"out_channels": vae_params.out_ch,
"down_block_types": tuple(down_block_types),
"up_block_types": tuple(up_block_types),
"block_out_channels": tuple(block_out_channels),
"latent_channels": vae_params.z_channels,
"layers_per_block": vae_params.num_res_blocks,
}
return config

4
src/diffusers/configuration_utils.py
View File

@ -420,7 +420,7 @@ class ConfigMixin:
@classmethod
def extract_init_dict(cls, config_dict, **kwargs):
# 0. Copy origin config dict
original_dict = {k: v for k, v in config_dict.items()}
original_dict = dict(config_dict.items())
# 1. Retrieve expected config attributes from __init__ signature
expected_keys = cls._get_init_keys(cls)
@ -610,7 +610,7 @@ def flax_register_to_config(cls):
)
# Ignore private kwargs in the init. Retrieve all passed attributes
init_kwargs = {k: v for k, v in kwargs.items()}
init_kwargs = dict(kwargs.items())
# Retrieve default values
fields = dataclasses.fields(self)

4
src/diffusers/experimental/rl/value_guided_sampling.py
View File

@ -52,13 +52,13 @@ class ValueGuidedRLPipeline(DiffusionPipeline):
self.scheduler = scheduler
self.env = env
self.data = env.get_dataset()
self.means = dict()
self.means = {}
for key in self.data.keys():
try:
self.means[key] = self.data[key].mean()
except: # noqa: E722
pass
self.stds = dict()
self.stds = {}
for key in self.data.keys():
try:
self.stds[key] = self.data[key].std()

2
src/diffusers/image_processor.py
View File

@ -99,7 +99,7 @@ class VaeImageProcessor(ConfigMixin):
Resize a PIL image. Both height and width will be downscaled to the next integer multiple of `vae_scale_factor`
"""
w, h = images.size
w, h = map(lambda x: x - x % self.vae_scale_factor, (w, h)) # resize to integer multiple of vae_scale_factor
w, h = (x - x % self.vae_scale_factor for x in (w, h)) # resize to integer multiple of vae_scale_factor
images = images.resize((w, h), resample=PIL_INTERPOLATION[self.resample])
return images

2
src/diffusers/loaders.py
View File

@ -37,7 +37,7 @@ class AttnProcsLayers(torch.nn.Module):
def __init__(self, state_dict: Dict[str, torch.Tensor]):
super().__init__()
self.layers = torch.nn.ModuleList(state_dict.values())
self.mapping = {k: v for k, v in enumerate(state_dict.keys())}
self.mapping = dict(enumerate(state_dict.keys()))
self.rev_mapping = {v: k for k, v in enumerate(state_dict.keys())}
# we add a hook to state_dict() and load_state_dict() so that the

2
src/diffusers/models/modeling_utils.py
View File

@ -647,7 +647,7 @@ class ModelMixin(torch.nn.Module):
):
# Retrieve missing & unexpected_keys
model_state_dict = model.state_dict()
loaded_keys = [k for k in state_dict.keys()]
loaded_keys = list(state_dict.keys())
expected_keys = list(model_state_dict.keys())

2
src/diffusers/pipelines/alt_diffusion/pipeline_alt_diffusion_img2img.py
View File

@ -74,7 +74,7 @@ def preprocess(image):
if isinstance(image[0], PIL.Image.Image):
w, h = image[0].size
w, h = map(lambda x: x - x % 8, (w, h)) # resize to integer multiple of 8
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)

6
src/diffusers/pipelines/audio_diffusion/pipeline_audio_diffusion.py
View File

@ -201,12 +201,12 @@ class AudioDiffusionPipeline(DiffusionPipeline):
images = images.cpu().permute(0, 2, 3, 1).numpy()
images = (images * 255).round().astype("uint8")
images = list(
map(lambda _: Image.fromarray(_[:, :, 0]), images)
(Image.fromarray(_[:, :, 0]) for _ in images)
if images.shape[3] == 1
else map(lambda _: Image.fromarray(_, mode="RGB").convert("L"), images)
else (Image.fromarray(_, mode="RGB").convert("L") for _ in images)
)
audios = list(map(lambda _: self.mel.image_to_audio(_), images))
audios = [self.mel.image_to_audio(_) for _ in images]
if not return_dict:
return images, (self.mel.get_sample_rate(), audios)

2
src/diffusers/pipelines/latent_diffusion/pipeline_latent_diffusion_superresolution.py
View File

@ -21,7 +21,7 @@ from ..pipeline_utils import DiffusionPipeline, ImagePipelineOutput
def preprocess(image):
w, h = image.size
w, h = map(lambda x: x - x % 32, (w, h)) # resize to integer multiple of 32
w, h = (x - x % 32 for x in (w, h)) # resize to integer multiple of 32
image = image.resize((w, h), resample=PIL_INTERPOLATION["lanczos"])
image = np.array(image).astype(np.float32) / 255.0
image = image[None].transpose(0, 3, 1, 2)

2
src/diffusers/pipelines/pipeline_flax_utils.py
View File

@ -491,7 +491,7 @@ class FlaxDiffusionPipeline(ConfigMixin):
parameters = inspect.signature(obj.__init__).parameters
required_parameters = {k: v for k, v in parameters.items() if v.default == inspect._empty}
optional_parameters = set({k for k, v in parameters.items() if v.default != inspect._empty})
expected_modules = set(required_parameters.keys()) - set(["self"])
expected_modules = set(required_parameters.keys()) - {"self"}
return expected_modules, optional_parameters
@property

20
src/diffusers/pipelines/pipeline_utils.py
View File

@ -204,11 +204,11 @@ def variant_compatible_siblings(filenames, variant=None) -> Union[List[os.PathLi
non_variant_file_regex = re.compile(f"{'|'.join(weight_names)}")
if variant is not None:
variant_filenames = set(f for f in filenames if variant_file_regex.match(f.split("/")[-1]) is not None)
variant_filenames = {f for f in filenames if variant_file_regex.match(f.split("/")[-1]) is not None}
else:
variant_filenames = set()
non_variant_filenames = set(f for f in filenames if non_variant_file_regex.match(f.split("/")[-1]) is not None)
non_variant_filenames = {f for f in filenames if non_variant_file_regex.match(f.split("/")[-1]) is not None}
usable_filenames = set(variant_filenames)
for f in non_variant_filenames:
@ -225,7 +225,7 @@ def warn_deprecated_model_variant(pretrained_model_name_or_path, use_auth_token,
use_auth_token=use_auth_token,
revision=None,
)
filenames = set(sibling.rfilename for sibling in info.siblings)
filenames = {sibling.rfilename for sibling in info.siblings}
comp_model_filenames, _ = variant_compatible_siblings(filenames, variant=revision)
comp_model_filenames = [".".join(f.split(".")[:1] + f.split(".")[2:]) for f in comp_model_filenames]
@ -1115,7 +1115,7 @@ class DiffusionPipeline(ConfigMixin):
# retrieve all folder_names that contain relevant files
folder_names = [k for k, v in config_dict.items() if isinstance(v, list)]
filenames = set(sibling.rfilename for sibling in info.siblings)
filenames = {sibling.rfilename for sibling in info.siblings}
model_filenames, variant_filenames = variant_compatible_siblings(filenames, variant=variant)
# if the whole pipeline is cached we don't have to ping the Hub
@ -1126,7 +1126,7 @@ class DiffusionPipeline(ConfigMixin):
pretrained_model_name, use_auth_token, variant, revision, model_filenames
)
model_folder_names = set([os.path.split(f)[0] for f in model_filenames])
model_folder_names = {os.path.split(f)[0] for f in model_filenames}
# all filenames compatible with variant will be added
allow_patterns = list(model_filenames)
@ -1157,8 +1157,8 @@ class DiffusionPipeline(ConfigMixin):
elif use_safetensors and is_safetensors_compatible(model_filenames, variant=variant):
ignore_patterns = ["*.bin", "*.msgpack"]
safetensors_variant_filenames = set([f for f in variant_filenames if f.endswith(".safetensors")])
safetensors_model_filenames = set([f for f in model_filenames if f.endswith(".safetensors")])
safetensors_variant_filenames = {f for f in variant_filenames if f.endswith(".safetensors")}
safetensors_model_filenames = {f for f in model_filenames if f.endswith(".safetensors")}
if (
len(safetensors_variant_filenames) > 0
and safetensors_model_filenames != safetensors_variant_filenames
@ -1169,8 +1169,8 @@ class DiffusionPipeline(ConfigMixin):
else:
ignore_patterns = ["*.safetensors", "*.msgpack"]
bin_variant_filenames = set([f for f in variant_filenames if f.endswith(".bin")])
bin_model_filenames = set([f for f in model_filenames if f.endswith(".bin")])
bin_variant_filenames = {f for f in variant_filenames if f.endswith(".bin")}
bin_model_filenames = {f for f in model_filenames if f.endswith(".bin")}
if len(bin_variant_filenames) > 0 and bin_model_filenames != bin_variant_filenames:
logger.warn(
f"\nA mixture of {variant} and non-{variant} filenames will be loaded.\nLoaded {variant} filenames:\n[{', '.join(bin_variant_filenames)}]\nLoaded non-{variant} filenames:\n[{', '.join(bin_model_filenames - bin_variant_filenames)}\nIf this behavior is not expected, please check your folder structure."
@ -1215,7 +1215,7 @@ class DiffusionPipeline(ConfigMixin):
parameters = inspect.signature(obj.__init__).parameters
required_parameters = {k: v for k, v in parameters.items() if v.default == inspect._empty}
optional_parameters = set({k for k, v in parameters.items() if v.default != inspect._empty})
expected_modules = set(required_parameters.keys()) - set(["self"])
expected_modules = set(required_parameters.keys()) - {"self"}
return expected_modules, optional_parameters
@property

4
src/diffusers/pipelines/repaint/pipeline_repaint.py
View File

@ -37,7 +37,7 @@ def _preprocess_image(image: Union[List, PIL.Image.Image, torch.Tensor]):
if isinstance(image[0], PIL.Image.Image):
w, h = image[0].size
w, h = map(lambda x: x - x % 8, (w, h)) # resize to integer multiple of 8
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)
@ -58,7 +58,7 @@ def _preprocess_mask(mask: Union[List, PIL.Image.Image, torch.Tensor]):
if isinstance(mask[0], PIL.Image.Image):
w, h = mask[0].size
w, h = map(lambda x: x - x % 32, (w, h)) # resize to integer multiple of 32
w, h = (x - x % 32 for x in (w, h)) # resize to integer multiple of 32
mask = [np.array(m.convert("L").resize((w, h), resample=PIL_INTERPOLATION["nearest"]))[None, :] for m in mask]
mask = np.concatenate(mask, axis=0)
mask = mask.astype(np.float32) / 255.0

2
src/diffusers/pipelines/spectrogram_diffusion/midi_utils.py
View File

@ -166,7 +166,7 @@ class Codec:
self._shift_range = EventRange(type="shift", min_value=0, max_value=max_shift_steps)
self._event_ranges = [self._shift_range] + event_ranges
# Ensure all event types have unique names.
assert len(self._event_ranges) == len(set([er.type for er in self._event_ranges]))
assert len(self._event_ranges) == len({er.type for er in self._event_ranges})
@property
def num_classes(self) -> int:

44
src/diffusers/pipelines/stable_diffusion/convert_from_ckpt.py
View File

@ -274,18 +274,18 @@ def create_unet_diffusers_config(original_config, image_size: int, controlnet=Fa
else:
raise NotImplementedError(f"Unknown conditional unet num_classes config: {unet_params.num_classes}")
config = dict(
sample_size=image_size // vae_scale_factor,
in_channels=unet_params.in_channels,
down_block_types=tuple(down_block_types),
block_out_channels=tuple(block_out_channels),
layers_per_block=unet_params.num_res_blocks,
cross_attention_dim=unet_params.context_dim,
attention_head_dim=head_dim,
use_linear_projection=use_linear_projection,
class_embed_type=class_embed_type,
projection_class_embeddings_input_dim=projection_class_embeddings_input_dim,
)
config = {
"sample_size": image_size // vae_scale_factor,
"in_channels": unet_params.in_channels,
"down_block_types": tuple(down_block_types),
"block_out_channels": tuple(block_out_channels),
"layers_per_block": unet_params.num_res_blocks,
"cross_attention_dim": unet_params.context_dim,
"attention_head_dim": head_dim,
"use_linear_projection": use_linear_projection,
"class_embed_type": class_embed_type,
"projection_class_embeddings_input_dim": projection_class_embeddings_input_dim,
}
if not controlnet:
config["out_channels"] = unet_params.out_channels
@ -305,16 +305,16 @@ def create_vae_diffusers_config(original_config, image_size: int):
down_block_types = ["DownEncoderBlock2D"] * len(block_out_channels)
up_block_types = ["UpDecoderBlock2D"] * len(block_out_channels)
config = dict(
sample_size=image_size,
in_channels=vae_params.in_channels,
out_channels=vae_params.out_ch,
down_block_types=tuple(down_block_types),
up_block_types=tuple(up_block_types),
block_out_channels=tuple(block_out_channels),
latent_channels=vae_params.z_channels,
layers_per_block=vae_params.num_res_blocks,
)
config = {
"sample_size": image_size,
"in_channels": vae_params.in_channels,
"out_channels": vae_params.out_ch,
"down_block_types": tuple(down_block_types),
"up_block_types": tuple(up_block_types),
"block_out_channels": tuple(block_out_channels),
"latent_channels": vae_params.z_channels,
"layers_per_block": vae_params.num_res_blocks,
}
return config

2
src/diffusers/pipelines/stable_diffusion/pipeline_cycle_diffusion.py
View File

@ -44,7 +44,7 @@ def preprocess(image):
if isinstance(image[0], PIL.Image.Image):
w, h = image[0].size
w, h = map(lambda x: x - x % 8, (w, h)) # resize to integer multiple of 8
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)

2
src/diffusers/pipelines/stable_diffusion/pipeline_flax_stable_diffusion_controlnet.py
View File

@ -530,7 +530,7 @@ def unshard(x: jnp.ndarray):
def preprocess(image, dtype):
image = image.convert("RGB")
w, h = image.size
w, h = map(lambda x: x - x % 64, (w, h)) # resize to integer multiple of 64
w, h = (x - x % 64 for x in (w, h)) # resize to integer multiple of 64
image = image.resize((w, h), resample=PIL_INTERPOLATION["lanczos"])
image = jnp.array(image).astype(dtype) / 255.0
image = image[None].transpose(0, 3, 1, 2)

2
src/diffusers/pipelines/stable_diffusion/pipeline_flax_stable_diffusion_img2img.py
View File

@ -520,7 +520,7 @@ def unshard(x: jnp.ndarray):
def preprocess(image, dtype):
w, h = image.size
w, h = map(lambda x: x - x % 32, (w, h)) # resize to integer multiple of 32
w, h = (x - x % 32 for x in (w, h)) # resize to integer multiple of 32
image = image.resize((w, h), resample=PIL_INTERPOLATION["lanczos"])
image = jnp.array(image).astype(dtype) / 255.0
image = image[None].transpose(0, 3, 1, 2)

4
src/diffusers/pipelines/stable_diffusion/pipeline_flax_stable_diffusion_inpaint.py
View File

@ -563,7 +563,7 @@ def unshard(x: jnp.ndarray):
def preprocess_image(image, dtype):
w, h = image.size
w, h = map(lambda x: x - x % 32, (w, h)) # resize to integer multiple of 32
w, h = (x - x % 32 for x in (w, h)) # resize to integer multiple of 32
image = image.resize((w, h), resample=PIL_INTERPOLATION["lanczos"])
image = jnp.array(image).astype(dtype) / 255.0
image = image[None].transpose(0, 3, 1, 2)
@ -572,7 +572,7 @@ def preprocess_image(image, dtype):
def preprocess_mask(mask, dtype):
w, h = mask.size
w, h = map(lambda x: x - x % 32, (w, h)) # resize to integer multiple of 32
w, h = (x - x % 32 for x in (w, h)) # resize to integer multiple of 32
mask = mask.resize((w, h))
mask = jnp.array(mask.convert("L")).astype(dtype) / 255.0
mask = jnp.expand_dims(mask, axis=(0, 1))

2
src/diffusers/pipelines/stable_diffusion/pipeline_onnx_stable_diffusion_img2img.py
View File

@ -40,7 +40,7 @@ def preprocess(image):
if isinstance(image[0], PIL.Image.Image):
w, h = image[0].size
w, h = map(lambda x: x - x % 64, (w, h)) # resize to integer multiple of 64
w, h = (x - x % 64 for x in (w, h)) # resize to integer multiple of 64
image = [np.array(i.resize((w, h), resample=PIL_INTERPOLATION["lanczos"]))[None, :] for i in image]
image = np.concatenate(image, axis=0)

4
src/diffusers/pipelines/stable_diffusion/pipeline_onnx_stable_diffusion_inpaint_legacy.py
View File

@ -19,7 +19,7 @@ logger = logging.get_logger(__name__) # pylint: disable=invalid-name
def preprocess(image):
w, h = image.size
w, h = map(lambda x: x - x % 32, (w, h)) # resize to integer multiple of 32
w, h = (x - x % 32 for x in (w, h)) # resize to integer multiple of 32
image = image.resize((w, h), resample=PIL.Image.LANCZOS)
image = np.array(image).astype(np.float32) / 255.0
image = image[None].transpose(0, 3, 1, 2)
@ -29,7 +29,7 @@ def preprocess(image):
def preprocess_mask(mask, scale_factor=8):
mask = mask.convert("L")
w, h = mask.size
w, h = map(lambda x: x - x % 32, (w, h)) # resize to integer multiple of 32
w, h = (x - x % 32 for x in (w, h)) # resize to integer multiple of 32
mask = mask.resize((w // scale_factor, h // scale_factor), resample=PIL.Image.NEAREST)
mask = np.array(mask).astype(np.float32) / 255.0
mask = np.tile(mask, (4, 1, 1))

2
src/diffusers/pipelines/stable_diffusion/pipeline_onnx_stable_diffusion_upscale.py
View File

@ -31,7 +31,7 @@ def preprocess(image):
if isinstance(image[0], PIL.Image.Image):
w, h = image[0].size
w, h = map(lambda x: x - x % 64, (w, h)) # resize to integer multiple of 32
w, h = (x - x % 64 for x in (w, h)) # resize to integer multiple of 32
image = [np.array(i.resize((w, h)))[None, :] for i in image]
image = np.concatenate(image, axis=0)

4
src/diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_depth2img.py
View File

@ -41,7 +41,7 @@ def preprocess(image):
if isinstance(image[0], PIL.Image.Image):
w, h = image[0].size
w, h = map(lambda x: x - x % 8, (w, h)) # resize to integer multiple of 8
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)
@ -442,7 +442,7 @@ class StableDiffusionDepth2ImgPipeline(DiffusionPipeline):
if isinstance(image, PIL.Image.Image):
image = [image]
else:
image = [img for img in image]
image = list(image)
if isinstance(image[0], PIL.Image.Image):
width, height = image[0].size

2
src/diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_img2img.py
View File

@ -78,7 +78,7 @@ def preprocess(image):
if isinstance(image[0], PIL.Image.Image):
w, h = image[0].size
w, h = map(lambda x: x - x % 8, (w, h)) # resize to integer multiple of 8
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)

6
src/diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_inpaint_legacy.py
View File

@ -42,7 +42,7 @@ logger = logging.get_logger(__name__)
def preprocess_image(image):
w, h = image.size
w, h = map(lambda x: x - x % 8, (w, h)) # resize to integer multiple of 8
w, h = (x - x % 8 for x in (w, h)) # resize to integer multiple of 8
image = image.resize((w, h), resample=PIL_INTERPOLATION["lanczos"])
image = np.array(image).astype(np.float32) / 255.0
image = image[None].transpose(0, 3, 1, 2)
@ -54,7 +54,7 @@ def preprocess_mask(mask, scale_factor=8):
if not isinstance(mask, torch.FloatTensor):
mask = mask.convert("L")
w, h = mask.size
w, h = map(lambda x: x - x % 8, (w, h)) # resize to integer multiple of 8
w, h = (x - x % 8 for x in (w, h)) # resize to integer multiple of 8
mask = mask.resize((w // scale_factor, h // scale_factor), resample=PIL_INTERPOLATION["nearest"])
mask = np.array(mask).astype(np.float32) / 255.0
mask = np.tile(mask, (4, 1, 1))
@ -76,7 +76,7 @@ def preprocess_mask(mask, scale_factor=8):
# (potentially) reduce mask channel dimension from 3 to 1 for broadcasting to latent shape
mask = mask.mean(dim=1, keepdim=True)
h, w = mask.shape[-2:]
h, w = map(lambda x: x - x % 8, (h, w)) # resize to integer multiple of 8
h, w = (x - x % 8 for x in (h, w)) # resize to integer multiple of 8
mask = torch.nn.functional.interpolate(mask, (h // scale_factor, w // scale_factor))
return mask

2
src/diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_instruct_pix2pix.py
View File

@ -47,7 +47,7 @@ def preprocess(image):
if isinstance(image[0], PIL.Image.Image):
w, h = image[0].size
w, h = map(lambda x: x - x % 8, (w, h)) # resize to integer multiple of 8
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)

2
src/diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_latent_upscale.py
View File

@ -38,7 +38,7 @@ def preprocess(image):
if isinstance(image[0], PIL.Image.Image):
w, h = image[0].size
w, h = map(lambda x: x - x % 64, (w, h)) # resize to integer multiple of 64
w, h = (x - x % 64 for x in (w, h)) # resize to integer multiple of 64
image = [np.array(i.resize((w, h)))[None, :] for i in image]
image = np.concatenate(image, axis=0)

2
src/diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_pix2pix_zero.py
View File

@ -180,7 +180,7 @@ def preprocess(image):
if isinstance(image[0], PIL.Image.Image):
w, h = image[0].size
w, h = map(lambda x: x - x % 8, (w, h)) # resize to integer multiple of 8
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)

2
src/diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_upscale.py
View File

@ -37,7 +37,7 @@ def preprocess(image):
if isinstance(image[0], PIL.Image.Image):
w, h = image[0].size
w, h = map(lambda x: x - x % 64, (w, h)) # resize to integer multiple of 64
w, h = (x - x % 64 for x in (w, h)) # resize to integer multiple of 64
image = [np.array(i.resize((w, h)))[None, :] for i in image]
image = np.concatenate(image, axis=0)

2
src/diffusers/pipelines/versatile_diffusion/pipeline_versatile_diffusion_image_variation.py
View File

@ -134,7 +134,7 @@ class VersatileDiffusionImageVariationPipeline(DiffusionPipeline):
return embeds
if isinstance(prompt, torch.Tensor) and len(prompt.shape) == 4:
prompt = [p for p in prompt]
prompt = list(prompt)
batch_size = len(prompt) if isinstance(prompt, list) else 1

2
src/diffusers/utils/outputs.py
View File

@ -84,7 +84,7 @@ class BaseOutput(OrderedDict):
def __getitem__(self, k):
if isinstance(k, str):
inner_dict = {k: v for (k, v) in self.items()}
inner_dict = dict(self.items())
return inner_dict[k]
else:
return self.to_tuple()[k]

2
utils/check_doc_toc.py
View File

@ -43,7 +43,7 @@ def clean_doc_toc(doc_list):
new_doc = []
for duplicate_key in duplicates:
titles = list(set(doc["title"] for doc in doc_list if doc["local"] == duplicate_key))
titles = list({doc["title"] for doc in doc_list if doc["local"] == duplicate_key})
if len(titles) > 1:
raise ValueError(
f"{duplicate_key} is present several times in the documentation table of content at "

4
utils/check_repo.py
View File

@ -219,7 +219,7 @@ def check_model_list():
# Get the models from the directory structure of `src/transformers/models/`
models = [model for model in dir(diffusers.models) if not model.startswith("__")]
missing_models = sorted(list(set(_models).difference(models)))
missing_models = sorted(set(_models).difference(models))
if missing_models:
raise Exception(
f"The following models should be included in {models_dir}/__init__.py: {','.join(missing_models)}."
@ -429,7 +429,7 @@ def get_all_auto_configured_models():
for attr_name in dir(diffusers.models.auto.modeling_flax_auto):
if attr_name.startswith("FLAX_MODEL_") and attr_name.endswith("MAPPING_NAMES"):
result = result | set(get_values(getattr(diffusers.models.auto.modeling_flax_auto, attr_name)))
return [cls for cls in result]
return list(result)
def ignore_unautoclassed(model_name):

2
utils/overwrite_expected_slice.py
View File

@ -67,7 +67,7 @@ def overwrite_file(file, class_name, test_name, correct_line, done_test):
def main(correct, fail=None):
if fail is not None:
with open(fail, "r") as f:
test_failures = set([l.strip() for l in f.readlines()])
test_failures = {l.strip() for l in f.readlines()}
else:
test_failures = None

2
utils/stale.py
View File

@ -38,7 +38,7 @@ def main():
open_issues = repo.get_issues(state="open")
for issue in open_issues:
comments = sorted([comment for comment in issue.get_comments()], key=lambda i: i.created_at, reverse=True)
comments = sorted(issue.get_comments(), key=lambda i: i.created_at, reverse=True)
last_comment = comments[0] if len(comments) > 0 else None
if (
last_comment is not None