diffusers 0.30.2__py3-none-any.whl → 0.31.0__py3-none-any.whl

diffusers/schedulers/scheduling_euler_discrete.py

@@ -20,11 +20,14 @@ import numpy as np
 import torch
 
 from ..configuration_utils import ConfigMixin, register_to_config
-from ..utils import BaseOutput, logging
+from ..utils import BaseOutput, is_scipy_available, logging
 from ..utils.torch_utils import randn_tensor
 from .scheduling_utils import KarrasDiffusionSchedulers, SchedulerMixin
 
 
+if is_scipy_available():
+    import scipy.stats
+
 logger = logging.get_logger(__name__)  # pylint: disable=invalid-name
 
 
@@ -158,6 +161,11 @@ class EulerDiscreteScheduler(SchedulerMixin, ConfigMixin):
         use_karras_sigmas (`bool`, *optional*, defaults to `False`):
             Whether to use Karras sigmas for step sizes in the noise schedule during the sampling process. If `True`,
             the sigmas are determined according to a sequence of noise levels {σi}.
+        use_exponential_sigmas (`bool`, *optional*, defaults to `False`):
+            Whether to use exponential sigmas for step sizes in the noise schedule during the sampling process.
+        use_beta_sigmas (`bool`, *optional*, defaults to `False`):
+            Whether to use beta sigmas for step sizes in the noise schedule during the sampling process. Refer to [Beta
+            Sampling is All You Need](https://huggingface.co/papers/2407.12173) for more information.
         timestep_spacing (`str`, defaults to `"linspace"`):
             The way the timesteps should be scaled. Refer to Table 2 of the [Common Diffusion Noise Schedules and
             Sample Steps are Flawed](https://huggingface.co/papers/2305.08891) for more information.
@@ -186,6 +194,8 @@ class EulerDiscreteScheduler(SchedulerMixin, ConfigMixin):
         prediction_type: str = "epsilon",
         interpolation_type: str = "linear",
         use_karras_sigmas: Optional[bool] = False,
+        use_exponential_sigmas: Optional[bool] = False,
+        use_beta_sigmas: Optional[bool] = False,
         sigma_min: Optional[float] = None,
         sigma_max: Optional[float] = None,
         timestep_spacing: str = "linspace",
@@ -194,6 +204,12 @@ class EulerDiscreteScheduler(SchedulerMixin, ConfigMixin):
         rescale_betas_zero_snr: bool = False,
         final_sigmas_type: str = "zero",  # can be "zero" or "sigma_min"
     ):
+        if self.config.use_beta_sigmas and not is_scipy_available():
+            raise ImportError("Make sure to install scipy if you want to use beta sigmas.")
+        if sum([self.config.use_beta_sigmas, self.config.use_exponential_sigmas, self.config.use_karras_sigmas]) > 1:
+            raise ValueError(
+                "Only one of `config.use_beta_sigmas`, `config.use_exponential_sigmas`, `config.use_karras_sigmas` can be used."
+            )
         if trained_betas is not None:
             self.betas = torch.tensor(trained_betas, dtype=torch.float32)
         elif beta_schedule == "linear":
@@ -235,6 +251,8 @@ class EulerDiscreteScheduler(SchedulerMixin, ConfigMixin):
 
         self.is_scale_input_called = False
         self.use_karras_sigmas = use_karras_sigmas
+        self.use_exponential_sigmas = use_exponential_sigmas
+        self.use_beta_sigmas = use_beta_sigmas
 
         self._step_index = None
         self._begin_index = None
@@ -332,6 +350,10 @@ class EulerDiscreteScheduler(SchedulerMixin, ConfigMixin):
             raise ValueError("Can only pass one of `num_inference_steps` or `timesteps` or `sigmas`.")
         if timesteps is not None and self.config.use_karras_sigmas:
             raise ValueError("Cannot set `timesteps` with `config.use_karras_sigmas = True`.")
+        if timesteps is not None and self.config.use_exponential_sigmas:
+            raise ValueError("Cannot set `timesteps` with `config.use_exponential_sigmas = True`.")
+        if timesteps is not None and self.config.use_beta_sigmas:
+            raise ValueError("Cannot set `timesteps` with `config.use_beta_sigmas = True`.")
         if (
             timesteps is not None
             and self.config.timestep_type == "continuous"
@@ -396,6 +418,14 @@ class EulerDiscreteScheduler(SchedulerMixin, ConfigMixin):
                 sigmas = self._convert_to_karras(in_sigmas=sigmas, num_inference_steps=self.num_inference_steps)
                 timesteps = np.array([self._sigma_to_t(sigma, log_sigmas) for sigma in sigmas])
 
+            elif self.config.use_exponential_sigmas:
+                sigmas = self._convert_to_exponential(in_sigmas=sigmas, num_inference_steps=self.num_inference_steps)
+                timesteps = np.array([self._sigma_to_t(sigma, log_sigmas) for sigma in sigmas])
+
+            elif self.config.use_beta_sigmas:
+                sigmas = self._convert_to_beta(in_sigmas=sigmas, num_inference_steps=self.num_inference_steps)
+                timesteps = np.array([self._sigma_to_t(sigma, log_sigmas) for sigma in sigmas])
+
             if self.config.final_sigmas_type == "sigma_min":
                 sigma_last = ((1 - self.alphas_cumprod[0]) / self.alphas_cumprod[0]) ** 0.5
             elif self.config.final_sigmas_type == "zero":
@@ -468,6 +498,59 @@ class EulerDiscreteScheduler(SchedulerMixin, ConfigMixin):
         sigmas = (max_inv_rho + ramp * (min_inv_rho - max_inv_rho)) ** rho
         return sigmas
 
+    # Copied from https://github.com/crowsonkb/k-diffusion/blob/686dbad0f39640ea25c8a8c6a6e56bb40eacefa2/k_diffusion/sampling.py#L26
+    def _convert_to_exponential(self, in_sigmas: torch.Tensor, num_inference_steps: int) -> torch.Tensor:
+        """Constructs an exponential noise schedule."""
+
+        # Hack to make sure that other schedulers which copy this function don't break
+        # TODO: Add this logic to the other schedulers
+        if hasattr(self.config, "sigma_min"):
+            sigma_min = self.config.sigma_min
+        else:
+            sigma_min = None
+
+        if hasattr(self.config, "sigma_max"):
+            sigma_max = self.config.sigma_max
+        else:
+            sigma_max = None
+
+        sigma_min = sigma_min if sigma_min is not None else in_sigmas[-1].item()
+        sigma_max = sigma_max if sigma_max is not None else in_sigmas[0].item()
+
+        sigmas = torch.linspace(math.log(sigma_max), math.log(sigma_min), num_inference_steps).exp()
+        return sigmas
+
+    def _convert_to_beta(
+        self, in_sigmas: torch.Tensor, num_inference_steps: int, alpha: float = 0.6, beta: float = 0.6
+    ) -> torch.Tensor:
+        """From "Beta Sampling is All You Need" [arXiv:2407.12173] (Lee et. al, 2024)"""
+
+        # Hack to make sure that other schedulers which copy this function don't break
+        # TODO: Add this logic to the other schedulers
+        if hasattr(self.config, "sigma_min"):
+            sigma_min = self.config.sigma_min
+        else:
+            sigma_min = None
+
+        if hasattr(self.config, "sigma_max"):
+            sigma_max = self.config.sigma_max
+        else:
+            sigma_max = None
+
+        sigma_min = sigma_min if sigma_min is not None else in_sigmas[-1].item()
+        sigma_max = sigma_max if sigma_max is not None else in_sigmas[0].item()
+
+        sigmas = torch.Tensor(
+            [
+                sigma_min + (ppf * (sigma_max - sigma_min))
+                for ppf in [
+                    scipy.stats.beta.ppf(timestep, alpha, beta)
+                    for timestep in 1 - np.linspace(0, 1, num_inference_steps)
+                ]
+            ]
+        )
+        return sigmas
+
     def index_for_timestep(self, timestep, schedule_timesteps=None):
         if schedule_timesteps is None:
             schedule_timesteps = self.timesteps
@@ -555,14 +638,13 @@ class EulerDiscreteScheduler(SchedulerMixin, ConfigMixin):
 
         gamma = min(s_churn / (len(self.sigmas) - 1), 2**0.5 - 1) if s_tmin <= sigma <= s_tmax else 0.0
 
-        noise = randn_tensor(
-            model_output.shape, dtype=model_output.dtype, device=model_output.device, generator=generator
-        )
-
-        eps = noise * s_noise
         sigma_hat = sigma * (gamma + 1)
 
         if gamma > 0:
+            noise = randn_tensor(
+                model_output.shape, dtype=model_output.dtype, device=model_output.device, generator=generator
+            )
+            eps = noise * s_noise
             sample = sample + eps * (sigma_hat**2 - sigma**2) ** 0.5
 
         # 1. compute predicted original sample (x_0) from sigma-scaled predicted noise
@@ -594,7 +676,10 @@ class EulerDiscreteScheduler(SchedulerMixin, ConfigMixin):
         self._step_index += 1
 
         if not return_dict:
-            return (prev_sample,)
+            return (
+                prev_sample,
+                pred_original_sample,
+            )
 
         return EulerDiscreteSchedulerOutput(prev_sample=prev_sample, pred_original_sample=pred_original_sample)
 

diffusers/schedulers/scheduling_flow_match_heun_discrete.py

@@ -266,14 +266,13 @@ class FlowMatchHeunDiscreteScheduler(SchedulerMixin, ConfigMixin):
 
         gamma = min(s_churn / (len(self.sigmas) - 1), 2**0.5 - 1) if s_tmin <= sigma <= s_tmax else 0.0
 
-        noise = randn_tensor(
-            model_output.shape, dtype=model_output.dtype, device=model_output.device, generator=generator
-        )
-
-        eps = noise * s_noise
         sigma_hat = sigma * (gamma + 1)
 
         if gamma > 0:
+            noise = randn_tensor(
+                model_output.shape, dtype=model_output.dtype, device=model_output.device, generator=generator
+            )
+            eps = noise * s_noise
             sample = sample + eps * (sigma_hat**2 - sigma**2) ** 0.5
 
         if self.state_in_first_order:

diffusers/schedulers/scheduling_heun_discrete.py

@@ -13,13 +13,38 @@
 # limitations under the License.
 
 import math
+from dataclasses import dataclass
 from typing import List, Optional, Tuple, Union
 
 import numpy as np
 import torch
 
 from ..configuration_utils import ConfigMixin, register_to_config
-from .scheduling_utils import KarrasDiffusionSchedulers, SchedulerMixin, SchedulerOutput
+from ..utils import BaseOutput, is_scipy_available
+from .scheduling_utils import KarrasDiffusionSchedulers, SchedulerMixin
+
+
+if is_scipy_available():
+    import scipy.stats
+
+
+@dataclass
+# Copied from diffusers.schedulers.scheduling_ddpm.DDPMSchedulerOutput with DDPM->HeunDiscrete
+class HeunDiscreteSchedulerOutput(BaseOutput):
+    """
+    Output class for the scheduler's `step` function output.
+
+    Args:
+        prev_sample (`torch.Tensor` of shape `(batch_size, num_channels, height, width)` for images):
+            Computed sample `(x_{t-1})` of previous timestep. `prev_sample` should be used as next model input in the
+            denoising loop.
+        pred_original_sample (`torch.Tensor` of shape `(batch_size, num_channels, height, width)` for images):
+            The predicted denoised sample `(x_{0})` based on the model output from the current timestep.
+            `pred_original_sample` can be used to preview progress or for guidance.
+    """
+
+    prev_sample: torch.Tensor
+    pred_original_sample: Optional[torch.Tensor] = None
 
 
 # Copied from diffusers.schedulers.scheduling_ddpm.betas_for_alpha_bar
@@ -97,6 +122,11 @@ class HeunDiscreteScheduler(SchedulerMixin, ConfigMixin):
         use_karras_sigmas (`bool`, *optional*, defaults to `False`):
             Whether to use Karras sigmas for step sizes in the noise schedule during the sampling process. If `True`,
             the sigmas are determined according to a sequence of noise levels {σi}.
+        use_exponential_sigmas (`bool`, *optional*, defaults to `False`):
+            Whether to use exponential sigmas for step sizes in the noise schedule during the sampling process.
+        use_beta_sigmas (`bool`, *optional*, defaults to `False`):
+            Whether to use beta sigmas for step sizes in the noise schedule during the sampling process. Refer to [Beta
+            Sampling is All You Need](https://huggingface.co/papers/2407.12173) for more information.
         timestep_spacing (`str`, defaults to `"linspace"`):
             The way the timesteps should be scaled. Refer to Table 2 of the [Common Diffusion Noise Schedules and
             Sample Steps are Flawed](https://huggingface.co/papers/2305.08891) for more information.
@@ -117,11 +147,19 @@ class HeunDiscreteScheduler(SchedulerMixin, ConfigMixin):
         trained_betas: Optional[Union[np.ndarray, List[float]]] = None,
         prediction_type: str = "epsilon",
         use_karras_sigmas: Optional[bool] = False,
+        use_exponential_sigmas: Optional[bool] = False,
+        use_beta_sigmas: Optional[bool] = False,
         clip_sample: Optional[bool] = False,
         clip_sample_range: float = 1.0,
         timestep_spacing: str = "linspace",
         steps_offset: int = 0,
     ):
+        if self.config.use_beta_sigmas and not is_scipy_available():
+            raise ImportError("Make sure to install scipy if you want to use beta sigmas.")
+        if sum([self.config.use_beta_sigmas, self.config.use_exponential_sigmas, self.config.use_karras_sigmas]) > 1:
+            raise ValueError(
+                "Only one of `config.use_beta_sigmas`, `config.use_exponential_sigmas`, `config.use_karras_sigmas` can be used."
+            )
         if trained_betas is not None:
             self.betas = torch.tensor(trained_betas, dtype=torch.float32)
         elif beta_schedule == "linear":
@@ -251,6 +289,10 @@ class HeunDiscreteScheduler(SchedulerMixin, ConfigMixin):
             raise ValueError("Can only pass one of `num_inference_steps` or `custom_timesteps`.")
         if timesteps is not None and self.config.use_karras_sigmas:
             raise ValueError("Cannot use `timesteps` with `config.use_karras_sigmas = True`")
+        if timesteps is not None and self.config.use_exponential_sigmas:
+            raise ValueError("Cannot set `timesteps` with `config.use_exponential_sigmas = True`.")
+        if timesteps is not None and self.config.use_beta_sigmas:
+            raise ValueError("Cannot set `timesteps` with `config.use_beta_sigmas = True`.")
 
         num_inference_steps = num_inference_steps or len(timesteps)
         self.num_inference_steps = num_inference_steps
@@ -286,6 +328,12 @@ class HeunDiscreteScheduler(SchedulerMixin, ConfigMixin):
         if self.config.use_karras_sigmas:
             sigmas = self._convert_to_karras(in_sigmas=sigmas, num_inference_steps=self.num_inference_steps)
             timesteps = np.array([self._sigma_to_t(sigma, log_sigmas) for sigma in sigmas])
+        elif self.config.use_exponential_sigmas:
+            sigmas = self._convert_to_exponential(in_sigmas=sigmas, num_inference_steps=self.num_inference_steps)
+            timesteps = np.array([self._sigma_to_t(sigma, log_sigmas) for sigma in sigmas])
+        elif self.config.use_beta_sigmas:
+            sigmas = self._convert_to_beta(in_sigmas=sigmas, num_inference_steps=self.num_inference_steps)
+            timesteps = np.array([self._sigma_to_t(sigma, log_sigmas) for sigma in sigmas])
 
         sigmas = np.concatenate([sigmas, [0.0]]).astype(np.float32)
         sigmas = torch.from_numpy(sigmas).to(device=device)
@@ -354,6 +402,60 @@ class HeunDiscreteScheduler(SchedulerMixin, ConfigMixin):
         sigmas = (max_inv_rho + ramp * (min_inv_rho - max_inv_rho)) ** rho
         return sigmas
 
+    # Copied from diffusers.schedulers.scheduling_euler_discrete.EulerDiscreteScheduler._convert_to_exponential
+    def _convert_to_exponential(self, in_sigmas: torch.Tensor, num_inference_steps: int) -> torch.Tensor:
+        """Constructs an exponential noise schedule."""
+
+        # Hack to make sure that other schedulers which copy this function don't break
+        # TODO: Add this logic to the other schedulers
+        if hasattr(self.config, "sigma_min"):
+            sigma_min = self.config.sigma_min
+        else:
+            sigma_min = None
+
+        if hasattr(self.config, "sigma_max"):
+            sigma_max = self.config.sigma_max
+        else:
+            sigma_max = None
+
+        sigma_min = sigma_min if sigma_min is not None else in_sigmas[-1].item()
+        sigma_max = sigma_max if sigma_max is not None else in_sigmas[0].item()
+
+        sigmas = torch.linspace(math.log(sigma_max), math.log(sigma_min), num_inference_steps).exp()
+        return sigmas
+
+    # Copied from diffusers.schedulers.scheduling_euler_discrete.EulerDiscreteScheduler._convert_to_beta
+    def _convert_to_beta(
+        self, in_sigmas: torch.Tensor, num_inference_steps: int, alpha: float = 0.6, beta: float = 0.6
+    ) -> torch.Tensor:
+        """From "Beta Sampling is All You Need" [arXiv:2407.12173] (Lee et. al, 2024)"""
+
+        # Hack to make sure that other schedulers which copy this function don't break
+        # TODO: Add this logic to the other schedulers
+        if hasattr(self.config, "sigma_min"):
+            sigma_min = self.config.sigma_min
+        else:
+            sigma_min = None
+
+        if hasattr(self.config, "sigma_max"):
+            sigma_max = self.config.sigma_max
+        else:
+            sigma_max = None
+
+        sigma_min = sigma_min if sigma_min is not None else in_sigmas[-1].item()
+        sigma_max = sigma_max if sigma_max is not None else in_sigmas[0].item()
+
+        sigmas = torch.Tensor(
+            [
+                sigma_min + (ppf * (sigma_max - sigma_min))
+                for ppf in [
+                    scipy.stats.beta.ppf(timestep, alpha, beta)
+                    for timestep in 1 - np.linspace(0, 1, num_inference_steps)
+                ]
+            ]
+        )
+        return sigmas
+
     @property
     def state_in_first_order(self):
         return self.dt is None
@@ -373,7 +475,7 @@ class HeunDiscreteScheduler(SchedulerMixin, ConfigMixin):
         timestep: Union[float, torch.Tensor],
         sample: Union[torch.Tensor, np.ndarray],
         return_dict: bool = True,
-    ) -> Union[SchedulerOutput, Tuple]:
+    ) -> Union[HeunDiscreteSchedulerOutput, Tuple]:
         """
         Predict the sample from the previous timestep by reversing the SDE. This function propagates the diffusion
         process from the learned model outputs (most often the predicted noise).
@@ -386,12 +488,13 @@ class HeunDiscreteScheduler(SchedulerMixin, ConfigMixin):
             sample (`torch.Tensor`):
                 A current instance of a sample created by the diffusion process.
             return_dict (`bool`):
-                Whether or not to return a [`~schedulers.scheduling_utils.SchedulerOutput`] or tuple.
+                Whether or not to return a [`~schedulers.scheduling_heun_discrete.HeunDiscreteSchedulerOutput`] or
+                tuple.
 
         Returns:
-            [`~schedulers.scheduling_utils.SchedulerOutput`] or `tuple`:
-                If return_dict is `True`, [`~schedulers.scheduling_utils.SchedulerOutput`] is returned, otherwise a
-                tuple is returned where the first element is the sample tensor.
+            [`~schedulers.scheduling_heun_discrete.HeunDiscreteSchedulerOutput`] or `tuple`:
+                If return_dict is `True`, [`~schedulers.scheduling_heun_discrete.HeunDiscreteSchedulerOutput`] is
+                returned, otherwise a tuple is returned where the first element is the sample tensor.
         """
         if self.step_index is None:
             self._init_step_index(timestep)
@@ -462,9 +565,12 @@ class HeunDiscreteScheduler(SchedulerMixin, ConfigMixin):
         self._step_index += 1
 
         if not return_dict:
-            return (prev_sample,)
+            return (
+                prev_sample,
+                pred_original_sample,
+            )
 
-        return SchedulerOutput(prev_sample=prev_sample)
+        return HeunDiscreteSchedulerOutput(prev_sample=prev_sample, pred_original_sample=pred_original_sample)
 
     # Copied from diffusers.schedulers.scheduling_euler_discrete.EulerDiscreteScheduler.add_noise
     def add_noise(

diffusers/schedulers/scheduling_k_dpm_2_ancestral_discrete.py

@@ -13,14 +13,39 @@
 # limitations under the License.
 
 import math
+from dataclasses import dataclass
 from typing import List, Optional, Tuple, Union
 
 import numpy as np
 import torch
 
 from ..configuration_utils import ConfigMixin, register_to_config
+from ..utils import BaseOutput, is_scipy_available
 from ..utils.torch_utils import randn_tensor
-from .scheduling_utils import KarrasDiffusionSchedulers, SchedulerMixin, SchedulerOutput
+from .scheduling_utils import KarrasDiffusionSchedulers, SchedulerMixin
+
+
+if is_scipy_available():
+    import scipy.stats
+
+
+@dataclass
+# Copied from diffusers.schedulers.scheduling_ddpm.DDPMSchedulerOutput with DDPM->KDPM2AncestralDiscrete
+class KDPM2AncestralDiscreteSchedulerOutput(BaseOutput):
+    """
+    Output class for the scheduler's `step` function output.
+
+    Args:
+        prev_sample (`torch.Tensor` of shape `(batch_size, num_channels, height, width)` for images):
+            Computed sample `(x_{t-1})` of previous timestep. `prev_sample` should be used as next model input in the
+            denoising loop.
+        pred_original_sample (`torch.Tensor` of shape `(batch_size, num_channels, height, width)` for images):
+            The predicted denoised sample `(x_{0})` based on the model output from the current timestep.
+            `pred_original_sample` can be used to preview progress or for guidance.
+    """
+
+    prev_sample: torch.Tensor
+    pred_original_sample: Optional[torch.Tensor] = None
 
 
 # Copied from diffusers.schedulers.scheduling_ddpm.betas_for_alpha_bar
@@ -91,6 +116,11 @@ class KDPM2AncestralDiscreteScheduler(SchedulerMixin, ConfigMixin):
         use_karras_sigmas (`bool`, *optional*, defaults to `False`):
             Whether to use Karras sigmas for step sizes in the noise schedule during the sampling process. If `True`,
             the sigmas are determined according to a sequence of noise levels {σi}.
+        use_exponential_sigmas (`bool`, *optional*, defaults to `False`):
+            Whether to use exponential sigmas for step sizes in the noise schedule during the sampling process.
+        use_beta_sigmas (`bool`, *optional*, defaults to `False`):
+            Whether to use beta sigmas for step sizes in the noise schedule during the sampling process. Refer to [Beta
+            Sampling is All You Need](https://huggingface.co/papers/2407.12173) for more information.
         prediction_type (`str`, defaults to `epsilon`, *optional*):
             Prediction type of the scheduler function; can be `epsilon` (predicts the noise of the diffusion process),
             `sample` (directly predicts the noisy sample`) or `v_prediction` (see section 2.4 of [Imagen
@@ -114,10 +144,18 @@ class KDPM2AncestralDiscreteScheduler(SchedulerMixin, ConfigMixin):
         beta_schedule: str = "linear",
         trained_betas: Optional[Union[np.ndarray, List[float]]] = None,
         use_karras_sigmas: Optional[bool] = False,
+        use_exponential_sigmas: Optional[bool] = False,
+        use_beta_sigmas: Optional[bool] = False,
         prediction_type: str = "epsilon",
         timestep_spacing: str = "linspace",
         steps_offset: int = 0,
     ):
+        if self.config.use_beta_sigmas and not is_scipy_available():
+            raise ImportError("Make sure to install scipy if you want to use beta sigmas.")
+        if sum([self.config.use_beta_sigmas, self.config.use_exponential_sigmas, self.config.use_karras_sigmas]) > 1:
+            raise ValueError(
+                "Only one of `config.use_beta_sigmas`, `config.use_exponential_sigmas`, `config.use_karras_sigmas` can be used."
+            )
         if trained_betas is not None:
             self.betas = torch.tensor(trained_betas, dtype=torch.float32)
         elif beta_schedule == "linear":
@@ -250,6 +288,12 @@ class KDPM2AncestralDiscreteScheduler(SchedulerMixin, ConfigMixin):
         if self.config.use_karras_sigmas:
             sigmas = self._convert_to_karras(in_sigmas=sigmas, num_inference_steps=num_inference_steps)
             timesteps = np.array([self._sigma_to_t(sigma, log_sigmas) for sigma in sigmas]).round()
+        elif self.config.use_exponential_sigmas:
+            sigmas = self._convert_to_exponential(in_sigmas=sigmas, num_inference_steps=self.num_inference_steps)
+            timesteps = np.array([self._sigma_to_t(sigma, log_sigmas) for sigma in sigmas])
+        elif self.config.use_beta_sigmas:
+            sigmas = self._convert_to_beta(in_sigmas=sigmas, num_inference_steps=self.num_inference_steps)
+            timesteps = np.array([self._sigma_to_t(sigma, log_sigmas) for sigma in sigmas])
 
         self.log_sigmas = torch.from_numpy(log_sigmas).to(device)
         sigmas = np.concatenate([sigmas, [0.0]]).astype(np.float32)
@@ -346,6 +390,60 @@ class KDPM2AncestralDiscreteScheduler(SchedulerMixin, ConfigMixin):
         sigmas = (max_inv_rho + ramp * (min_inv_rho - max_inv_rho)) ** rho
         return sigmas
 
+    # Copied from diffusers.schedulers.scheduling_euler_discrete.EulerDiscreteScheduler._convert_to_exponential
+    def _convert_to_exponential(self, in_sigmas: torch.Tensor, num_inference_steps: int) -> torch.Tensor:
+        """Constructs an exponential noise schedule."""
+
+        # Hack to make sure that other schedulers which copy this function don't break
+        # TODO: Add this logic to the other schedulers
+        if hasattr(self.config, "sigma_min"):
+            sigma_min = self.config.sigma_min
+        else:
+            sigma_min = None
+
+        if hasattr(self.config, "sigma_max"):
+            sigma_max = self.config.sigma_max
+        else:
+            sigma_max = None
+
+        sigma_min = sigma_min if sigma_min is not None else in_sigmas[-1].item()
+        sigma_max = sigma_max if sigma_max is not None else in_sigmas[0].item()
+
+        sigmas = torch.linspace(math.log(sigma_max), math.log(sigma_min), num_inference_steps).exp()
+        return sigmas
+
+    # Copied from diffusers.schedulers.scheduling_euler_discrete.EulerDiscreteScheduler._convert_to_beta
+    def _convert_to_beta(
+        self, in_sigmas: torch.Tensor, num_inference_steps: int, alpha: float = 0.6, beta: float = 0.6
+    ) -> torch.Tensor:
+        """From "Beta Sampling is All You Need" [arXiv:2407.12173] (Lee et. al, 2024)"""
+
+        # Hack to make sure that other schedulers which copy this function don't break
+        # TODO: Add this logic to the other schedulers
+        if hasattr(self.config, "sigma_min"):
+            sigma_min = self.config.sigma_min
+        else:
+            sigma_min = None
+
+        if hasattr(self.config, "sigma_max"):
+            sigma_max = self.config.sigma_max
+        else:
+            sigma_max = None
+
+        sigma_min = sigma_min if sigma_min is not None else in_sigmas[-1].item()
+        sigma_max = sigma_max if sigma_max is not None else in_sigmas[0].item()
+
+        sigmas = torch.Tensor(
+            [
+                sigma_min + (ppf * (sigma_max - sigma_min))
+                for ppf in [
+                    scipy.stats.beta.ppf(timestep, alpha, beta)
+                    for timestep in 1 - np.linspace(0, 1, num_inference_steps)
+                ]
+            ]
+        )
+        return sigmas
+
     @property
     def state_in_first_order(self):
         return self.sample is None
@@ -381,7 +479,7 @@ class KDPM2AncestralDiscreteScheduler(SchedulerMixin, ConfigMixin):
         sample: Union[torch.Tensor, np.ndarray],
         generator: Optional[torch.Generator] = None,
         return_dict: bool = True,
-    ) -> Union[SchedulerOutput, Tuple]:
+    ) -> Union[KDPM2AncestralDiscreteSchedulerOutput, Tuple]:
         """
         Predict the sample from the previous timestep by reversing the SDE. This function propagates the diffusion
         process from the learned model outputs (most often the predicted noise).
@@ -396,12 +494,14 @@ class KDPM2AncestralDiscreteScheduler(SchedulerMixin, ConfigMixin):
             generator (`torch.Generator`, *optional*):
                 A random number generator.
             return_dict (`bool`):
-                Whether or not to return a [`~schedulers.scheduling_utils.SchedulerOutput`] or tuple.
+                Whether or not to return a
+                [`~schedulers.scheduling_k_dpm_2_ancestral_discrete.KDPM2AncestralDiscreteSchedulerOutput`] or tuple.
 
         Returns:
-            [`~schedulers.scheduling_utils.SchedulerOutput`] or `tuple`:
-                If return_dict is `True`, [`~schedulers.scheduling_ddim.SchedulerOutput`] is returned, otherwise a
-                tuple is returned where the first element is the sample tensor.
+            [`~schedulers.scheduling_k_dpm_2_ancestral_discrete.KDPM2AncestralDiscreteSchedulerOutput`] or `tuple`:
+                If return_dict is `True`,
+                [`~schedulers.scheduling_k_dpm_2_ancestral_discrete.KDPM2AncestralDiscreteSchedulerOutput`] is
+                returned, otherwise a tuple is returned where the first element is the sample tensor.
         """
         if self.step_index is None:
             self._init_step_index(timestep)
@@ -424,9 +524,6 @@ class KDPM2AncestralDiscreteScheduler(SchedulerMixin, ConfigMixin):
         gamma = 0
         sigma_hat = sigma * (gamma + 1)  # Note: sigma_hat == sigma for now
 
-        device = model_output.device
-        noise = randn_tensor(model_output.shape, dtype=model_output.dtype, device=device, generator=generator)
-
         # 1. compute predicted original sample (x_0) from sigma-scaled predicted noise
         if self.config.prediction_type == "epsilon":
             sigma_input = sigma_hat if self.state_in_first_order else sigma_interpol
@@ -464,15 +561,23 @@ class KDPM2AncestralDiscreteScheduler(SchedulerMixin, ConfigMixin):
             self.sample = None
 
             prev_sample = sample + derivative * dt
+            noise = randn_tensor(
+                model_output.shape, dtype=model_output.dtype, device=model_output.device, generator=generator
+            )
             prev_sample = prev_sample + noise * sigma_up
 
         # upon completion increase step index by one
         self._step_index += 1
 
         if not return_dict:
-            return (prev_sample,)
+            return (
+                prev_sample,
+                pred_original_sample,
+            )
 
-        return SchedulerOutput(prev_sample=prev_sample)
+        return KDPM2AncestralDiscreteSchedulerOutput(
+            prev_sample=prev_sample, pred_original_sample=pred_original_sample
+        )
 
     # Copied from diffusers.schedulers.scheduling_euler_discrete.EulerDiscreteScheduler.add_noise
     def add_noise(