diffusers 0.32.1__py3-none-any.whl → 0.33.0__py3-none-any.whl

diffusers/training_utils.py

@@ -241,14 +241,20 @@ def _set_state_dict_into_text_encoder(
     """
 
     text_encoder_state_dict = {
-        f'{k.replace(prefix, "")}': v for k, v in lora_state_dict.items() if k.startswith(prefix)
+        f"{k.replace(prefix, '')}": v for k, v in lora_state_dict.items() if k.startswith(prefix)
     }
     text_encoder_state_dict = convert_state_dict_to_peft(convert_state_dict_to_diffusers(text_encoder_state_dict))
     set_peft_model_state_dict(text_encoder, text_encoder_state_dict, adapter_name="default")
 
 
 def compute_density_for_timestep_sampling(
-    weighting_scheme: str, batch_size: int, logit_mean: float = None, logit_std: float = None, mode_scale: float = None
+    weighting_scheme: str,
+    batch_size: int,
+    logit_mean: float = None,
+    logit_std: float = None,
+    mode_scale: float = None,
+    device: Union[torch.device, str] = "cpu",
+    generator: Optional[torch.Generator] = None,
 ):
     """
     Compute the density for sampling the timesteps when doing SD3 training.
@@ -258,14 +264,13 @@ def compute_density_for_timestep_sampling(
     SD3 paper reference: https://arxiv.org/abs/2403.03206v1.
     """
     if weighting_scheme == "logit_normal":
-        # See 3.1 in the SD3 paper ($rf/lognorm(0.00,1.00)$).
-        u = torch.normal(mean=logit_mean, std=logit_std, size=(batch_size,), device="cpu")
+        u = torch.normal(mean=logit_mean, std=logit_std, size=(batch_size,), device=device, generator=generator)
         u = torch.nn.functional.sigmoid(u)
     elif weighting_scheme == "mode":
-        u = torch.rand(size=(batch_size,), device="cpu")
+        u = torch.rand(size=(batch_size,), device=device, generator=generator)
         u = 1 - u - mode_scale * (torch.cos(math.pi * u / 2) ** 2 - 1 + u)
     else:
-        u = torch.rand(size=(batch_size,), device="cpu")
+        u = torch.rand(size=(batch_size,), device=device, generator=generator)
     return u
 
 
@@ -299,6 +304,8 @@ def free_memory():
         torch.mps.empty_cache()
     elif is_torch_npu_available():
         torch_npu.npu.empty_cache()
+    elif hasattr(torch, "xpu") and torch.xpu.is_available():
+        torch.xpu.empty_cache()
 
 
 # Adapted from torch-ema https://github.com/fadel/pytorch_ema/blob/master/torch_ema/ema.py#L14
@@ -576,7 +583,7 @@ class EMAModel:
         """
 
         if self.temp_stored_params is None:
-            raise RuntimeError("This ExponentialMovingAverage has no `store()`ed weights " "to `restore()`")
+            raise RuntimeError("This ExponentialMovingAverage has no `store()`ed weights to `restore()`")
         if self.foreach:
             torch._foreach_copy_(
                 [param.data for param in parameters], [c_param.data for c_param in self.temp_stored_params]

diffusers/utils/__init__.py

@@ -1,4 +1,4 @@
-# Copyright 2024 The HuggingFace Inc. team. All rights reserved.
+# Copyright 2025 The HuggingFace Inc. team. All rights reserved.
 #
 # Licensed under the Apache License, Version 2.0 (the "License");
 # you may not use this file except in compliance with the License.
@@ -70,6 +70,7 @@ from .import_utils import (
     is_gguf_available,
     is_gguf_version,
     is_google_colab,
+    is_hf_hub_version,
     is_inflect_available,
     is_invisible_watermark_available,
     is_k_diffusion_available,
@@ -78,6 +79,9 @@ from .import_utils import (
     is_matplotlib_available,
     is_note_seq_available,
     is_onnx_available,
+    is_opencv_available,
+    is_optimum_quanto_available,
+    is_optimum_quanto_version,
     is_peft_available,
     is_peft_version,
     is_safetensors_available,
@@ -91,6 +95,7 @@ from .import_utils import (
     is_torch_xla_available,
     is_torch_xla_version,
     is_torchao_available,
+    is_torchao_version,
     is_torchsde_available,
     is_torchvision_available,
     is_transformers_available,
@@ -100,7 +105,7 @@ from .import_utils import (
     is_xformers_available,
     requires_backends,
 )
-from .loading_utils import get_module_from_name, load_image, load_video
+from .loading_utils import get_module_from_name, get_submodule_by_name, load_image, load_video
 from .logging import get_logger
 from .outputs import BaseOutput
 from .peft_utils import (
@@ -115,13 +120,16 @@ from .peft_utils import (
     unscale_lora_layers,
 )
 from .pil_utils import PIL_INTERPOLATION, make_image_grid, numpy_to_pil, pt_to_pil
+from .remote_utils import remote_decode
 from .state_dict_utils import (
     convert_all_state_dict_to_peft,
     convert_state_dict_to_diffusers,
     convert_state_dict_to_kohya,
     convert_state_dict_to_peft,
     convert_unet_state_dict_to_peft,
+    state_dict_all_zero,
 )
+from .typing_utils import _get_detailed_type, _is_valid_type
 
 
 logger = get_logger(__name__)

diffusers/utils/constants.py

@@ -1,4 +1,4 @@
-# Copyright 2024 The HuggingFace Inc. team. All rights reserved.
+# Copyright 2025 The HuggingFace Inc. team. All rights reserved.
 #
 # Licensed under the Apache License, Version 2.0 (the "License");
 # you may not use this file except in compliance with the License.
@@ -40,6 +40,7 @@ HUGGINGFACE_CO_RESOLVE_ENDPOINT = os.environ.get("HF_ENDPOINT", "https://hugging
 DIFFUSERS_DYNAMIC_MODULE_NAME = "diffusers_modules"
 HF_MODULES_CACHE = os.getenv("HF_MODULES_CACHE", os.path.join(HF_HOME, "modules"))
 DEPRECATED_REVISION_ARGS = ["fp16", "non-ema"]
+DIFFUSERS_REQUEST_TIMEOUT = 60
 
 # Below should be `True` if the current version of `peft` and `transformers` are compatible with
 # PEFT backend. Will automatically fall back to PEFT backend if the correct versions of the libraries are
@@ -56,3 +57,14 @@ USE_PEFT_BACKEND = _required_peft_version and _required_transformers_version
 
 if USE_PEFT_BACKEND and _CHECK_PEFT:
     dep_version_check("peft")
+
+
+DECODE_ENDPOINT_SD_V1 = "https://q1bj3bpq6kzilnsu.us-east-1.aws.endpoints.huggingface.cloud/"
+DECODE_ENDPOINT_SD_XL = "https://x2dmsqunjd6k9prw.us-east-1.aws.endpoints.huggingface.cloud/"
+DECODE_ENDPOINT_FLUX = "https://whhx50ex1aryqvw6.us-east-1.aws.endpoints.huggingface.cloud/"
+DECODE_ENDPOINT_HUNYUAN_VIDEO = "https://o7ywnmrahorts457.us-east-1.aws.endpoints.huggingface.cloud/"
+
+
+ENCODE_ENDPOINT_SD_V1 = "https://qc6479g0aac6qwy9.us-east-1.aws.endpoints.huggingface.cloud/"
+ENCODE_ENDPOINT_SD_XL = "https://xjqqhmyn62rog84g.us-east-1.aws.endpoints.huggingface.cloud/"
+ENCODE_ENDPOINT_FLUX = "https://ptccx55jz97f9zgo.us-east-1.aws.endpoints.huggingface.cloud/"

diffusers/utils/deprecation_utils.py

@@ -40,7 +40,7 @@ def deprecate(*args, take_from: Optional[Union[Dict, Any]] = None, standard_warn
         line_number = call_frame.lineno
         function = call_frame.function
         key, value = next(iter(deprecated_kwargs.items()))
-        raise TypeError(f"{function} in {filename} line {line_number-1} got an unexpected keyword argument `{key}`")
+        raise TypeError(f"{function} in {filename} line {line_number - 1} got an unexpected keyword argument `{key}`")
 
     if len(values) == 0:
         return

diffusers/utils/dummy_bitsandbytes_objects.py

@@ -0,0 +1,17 @@
+# This file is autogenerated by the command `make fix-copies`, do not edit.
+from ..utils import DummyObject, requires_backends
+
+
+class BitsAndBytesConfig(metaclass=DummyObject):
+    _backends = ["bitsandbytes"]
+
+    def __init__(self, *args, **kwargs):
+        requires_backends(self, ["bitsandbytes"])
+
+    @classmethod
+    def from_config(cls, *args, **kwargs):
+        requires_backends(cls, ["bitsandbytes"])
+
+    @classmethod
+    def from_pretrained(cls, *args, **kwargs):
+        requires_backends(cls, ["bitsandbytes"])

diffusers/utils/dummy_gguf_objects.py

@@ -0,0 +1,17 @@
+# This file is autogenerated by the command `make fix-copies`, do not edit.
+from ..utils import DummyObject, requires_backends
+
+
+class GGUFQuantizationConfig(metaclass=DummyObject):
+    _backends = ["gguf"]
+
+    def __init__(self, *args, **kwargs):
+        requires_backends(self, ["gguf"])
+
+    @classmethod
+    def from_config(cls, *args, **kwargs):
+        requires_backends(cls, ["gguf"])
+
+    @classmethod
+    def from_pretrained(cls, *args, **kwargs):
+        requires_backends(cls, ["gguf"])

diffusers/utils/dummy_optimum_quanto_objects.py

@@ -0,0 +1,17 @@
+# This file is autogenerated by the command `make fix-copies`, do not edit.
+from ..utils import DummyObject, requires_backends
+
+
+class QuantoConfig(metaclass=DummyObject):
+    _backends = ["optimum_quanto"]
+
+    def __init__(self, *args, **kwargs):
+        requires_backends(self, ["optimum_quanto"])
+
+    @classmethod
+    def from_config(cls, *args, **kwargs):
+        requires_backends(cls, ["optimum_quanto"])
+
+    @classmethod
+    def from_pretrained(cls, *args, **kwargs):
+        requires_backends(cls, ["optimum_quanto"])

diffusers/utils/dummy_pt_objects.py

@@ -2,6 +2,59 @@
 from ..utils import DummyObject, requires_backends
 
 
+class FasterCacheConfig(metaclass=DummyObject):
+    _backends = ["torch"]
+
+    def __init__(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
+    @classmethod
+    def from_config(cls, *args, **kwargs):
+        requires_backends(cls, ["torch"])
+
+    @classmethod
+    def from_pretrained(cls, *args, **kwargs):
+        requires_backends(cls, ["torch"])
+
+
+class HookRegistry(metaclass=DummyObject):
+    _backends = ["torch"]
+
+    def __init__(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
+    @classmethod
+    def from_config(cls, *args, **kwargs):
+        requires_backends(cls, ["torch"])
+
+    @classmethod
+    def from_pretrained(cls, *args, **kwargs):
+        requires_backends(cls, ["torch"])
+
+
+class PyramidAttentionBroadcastConfig(metaclass=DummyObject):
+    _backends = ["torch"]
+
+    def __init__(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
+    @classmethod
+    def from_config(cls, *args, **kwargs):
+        requires_backends(cls, ["torch"])
+
+    @classmethod
+    def from_pretrained(cls, *args, **kwargs):
+        requires_backends(cls, ["torch"])
+
+
+def apply_faster_cache(*args, **kwargs):
+    requires_backends(apply_faster_cache, ["torch"])
+
+
+def apply_pyramid_attention_broadcast(*args, **kwargs):
+    requires_backends(apply_pyramid_attention_broadcast, ["torch"])
+
+
 class AllegroTransformer3DModel(metaclass=DummyObject):
     _backends = ["torch"]
 
@@ -137,6 +190,21 @@ class AutoencoderKLLTXVideo(metaclass=DummyObject):
         requires_backends(cls, ["torch"])
 
 
+class AutoencoderKLMagvit(metaclass=DummyObject):
+    _backends = ["torch"]
+
+    def __init__(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
+    @classmethod
+    def from_config(cls, *args, **kwargs):
+        requires_backends(cls, ["torch"])
+
+    @classmethod
+    def from_pretrained(cls, *args, **kwargs):
+        requires_backends(cls, ["torch"])
+
+
 class AutoencoderKLMochi(metaclass=DummyObject):
     _backends = ["torch"]
 
@@ -167,6 +235,21 @@ class AutoencoderKLTemporalDecoder(metaclass=DummyObject):
         requires_backends(cls, ["torch"])
 
 
+class AutoencoderKLWan(metaclass=DummyObject):
+    _backends = ["torch"]
+
+    def __init__(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
+    @classmethod
+    def from_config(cls, *args, **kwargs):
+        requires_backends(cls, ["torch"])
+
+    @classmethod
+    def from_pretrained(cls, *args, **kwargs):
+        requires_backends(cls, ["torch"])
+
+
 class AutoencoderOobleck(metaclass=DummyObject):
     _backends = ["torch"]
 
@@ -197,6 +280,36 @@ class AutoencoderTiny(metaclass=DummyObject):
         requires_backends(cls, ["torch"])
 
 
+class AutoModel(metaclass=DummyObject):
+    _backends = ["torch"]
+
+    def __init__(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
+    @classmethod
+    def from_config(cls, *args, **kwargs):
+        requires_backends(cls, ["torch"])
+
+    @classmethod
+    def from_pretrained(cls, *args, **kwargs):
+        requires_backends(cls, ["torch"])
+
+
+class CacheMixin(metaclass=DummyObject):
+    _backends = ["torch"]
+
+    def __init__(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
+    @classmethod
+    def from_config(cls, *args, **kwargs):
+        requires_backends(cls, ["torch"])
+
+    @classmethod
+    def from_pretrained(cls, *args, **kwargs):
+        requires_backends(cls, ["torch"])
+
+
 class CogVideoXTransformer3DModel(metaclass=DummyObject):
     _backends = ["torch"]
 
@@ -227,6 +340,36 @@ class CogView3PlusTransformer2DModel(metaclass=DummyObject):
         requires_backends(cls, ["torch"])
 
 
+class CogView4Transformer2DModel(metaclass=DummyObject):
+    _backends = ["torch"]
+
+    def __init__(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
+    @classmethod
+    def from_config(cls, *args, **kwargs):
+        requires_backends(cls, ["torch"])
+
+    @classmethod
+    def from_pretrained(cls, *args, **kwargs):
+        requires_backends(cls, ["torch"])
+
+
+class ConsisIDTransformer3DModel(metaclass=DummyObject):
+    _backends = ["torch"]
+
+    def __init__(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
+    @classmethod
+    def from_config(cls, *args, **kwargs):
+        requires_backends(cls, ["torch"])
+
+    @classmethod
+    def from_pretrained(cls, *args, **kwargs):
+        requires_backends(cls, ["torch"])
+
+
 class ConsistencyDecoderVAE(metaclass=DummyObject):
     _backends = ["torch"]
 
@@ -302,6 +445,21 @@ class DiTTransformer2DModel(metaclass=DummyObject):
         requires_backends(cls, ["torch"])
 
 
+class EasyAnimateTransformer3DModel(metaclass=DummyObject):
+    _backends = ["torch"]
+
+    def __init__(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
+    @classmethod
+    def from_config(cls, *args, **kwargs):
+        requires_backends(cls, ["torch"])
+
+    @classmethod
+    def from_pretrained(cls, *args, **kwargs):
+        requires_backends(cls, ["torch"])
+
+
 class FluxControlNetModel(metaclass=DummyObject):
     _backends = ["torch"]
 
@@ -467,6 +625,21 @@ class LTXVideoTransformer3DModel(metaclass=DummyObject):
         requires_backends(cls, ["torch"])
 
 
+class Lumina2Transformer2DModel(metaclass=DummyObject):
+    _backends = ["torch"]
+
+    def __init__(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
+    @classmethod
+    def from_config(cls, *args, **kwargs):
+        requires_backends(cls, ["torch"])
+
+    @classmethod
+    def from_pretrained(cls, *args, **kwargs):
+        requires_backends(cls, ["torch"])
+
+
 class LuminaNextDiT2DModel(metaclass=DummyObject):
     _backends = ["torch"]
 
@@ -557,6 +730,21 @@ class MultiControlNetModel(metaclass=DummyObject):
         requires_backends(cls, ["torch"])
 
 
+class OmniGenTransformer2DModel(metaclass=DummyObject):
+    _backends = ["torch"]
+
+    def __init__(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
+    @classmethod
+    def from_config(cls, *args, **kwargs):
+        requires_backends(cls, ["torch"])
+
+    @classmethod
+    def from_pretrained(cls, *args, **kwargs):
+        requires_backends(cls, ["torch"])
+
+
 class PixArtTransformer2DModel(metaclass=DummyObject):
     _backends = ["torch"]
 
@@ -722,6 +910,21 @@ class Transformer2DModel(metaclass=DummyObject):
         requires_backends(cls, ["torch"])
 
 
+class TransformerTemporalModel(metaclass=DummyObject):
+    _backends = ["torch"]
+
+    def __init__(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
+    @classmethod
+    def from_config(cls, *args, **kwargs):
+        requires_backends(cls, ["torch"])
+
+    @classmethod
+    def from_pretrained(cls, *args, **kwargs):
+        requires_backends(cls, ["torch"])
+
+
 class UNet1DModel(metaclass=DummyObject):
     _backends = ["torch"]
 
@@ -857,6 +1060,21 @@ class VQModel(metaclass=DummyObject):
         requires_backends(cls, ["torch"])
 
 
+class WanTransformer3DModel(metaclass=DummyObject):
+    _backends = ["torch"]
+
+    def __init__(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
+    @classmethod
+    def from_config(cls, *args, **kwargs):
+        requires_backends(cls, ["torch"])
+
+    @classmethod
+    def from_pretrained(cls, *args, **kwargs):
+        requires_backends(cls, ["torch"])
+
+
 def get_constant_schedule(*args, **kwargs):
     requires_backends(get_constant_schedule, ["torch"])
 
@@ -1665,6 +1883,21 @@ class SchedulerMixin(metaclass=DummyObject):
         requires_backends(cls, ["torch"])
 
 
+class SCMScheduler(metaclass=DummyObject):
+    _backends = ["torch"]
+
+    def __init__(self, *args, **kwargs):
+        requires_backends(self, ["torch"])
+
+    @classmethod
+    def from_config(cls, *args, **kwargs):
+        requires_backends(cls, ["torch"])
+
+    @classmethod
+    def from_pretrained(cls, *args, **kwargs):
+        requires_backends(cls, ["torch"])
+
+
 class ScoreSdeVeScheduler(metaclass=DummyObject):
     _backends = ["torch"]
 

diffusers/utils/dummy_torch_and_transformers_and_opencv_objects.py

@@ -0,0 +1,17 @@
+# This file is autogenerated by the command `make fix-copies`, do not edit.
+from ..utils import DummyObject, requires_backends
+
+
+class ConsisIDPipeline(metaclass=DummyObject):
+    _backends = ["torch", "transformers", "opencv"]
+
+    def __init__(self, *args, **kwargs):
+        requires_backends(self, ["torch", "transformers", "opencv"])
+
+    @classmethod
+    def from_config(cls, *args, **kwargs):
+        requires_backends(cls, ["torch", "transformers", "opencv"])
+
+    @classmethod
+    def from_pretrained(cls, *args, **kwargs):
+        requires_backends(cls, ["torch", "transformers", "opencv"])