more merge fix

src/diffusers/guiders/classifier_free_guidance_plus_plus.py

@@ -1,115 +0,0 @@
-# Copyright 2024 The HuggingFace 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.
-# 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.
-
-from typing import Dict, Optional, Union, Tuple, List
-
-import torch
-
-from .guider_utils import BaseGuidance, rescale_noise_cfg, _default_prepare_inputs
-
-
-class CFGPlusPlusGuidance(BaseGuidance):
-    """
-    CFG++: https://huggingface.co/papers/2406.08070
-    
-    Args:
-        guidance_scale (`float`, defaults to `0.7`):
-            The scale parameter for classifier-free guidance. Higher values result in stronger conditioning on the text
-            prompt, while lower values allow for more freedom in generation. Higher values may lead to saturation and
-            deterioration of image quality.
-        guidance_rescale (`float`, defaults to `0.0`):
-            The rescale factor applied to the noise predictions. This is used to improve image quality and fix
-            overexposure. Based on Section 3.4 from [Common Diffusion Noise Schedules and Sample Steps are
-            Flawed](https://huggingface.co/papers/2305.08891).
-        use_original_formulation (`bool`, defaults to `False`):
-            Whether to use the original formulation of classifier-free guidance as proposed in the paper. By default,
-            we use the diffusers-native implementation that has been in the codebase for a long time. See
-            [~guiders.classifier_free_guidance.ClassifierFreeGuidance] for more details.
-        start (`float`, defaults to `0.0`):
-            The fraction of the total number of denoising steps after which guidance starts.
-        stop (`float`, defaults to `1.0`):
-            The fraction of the total number of denoising steps after which guidance stops.
-    """
-
-    _input_predictions = ["pred_cond", "pred_uncond"]
-
-    def __init__(
-        self,
-        guidance_scale: float = 0.7,
-        guidance_rescale: float = 0.0,
-        use_original_formulation: bool = False,
-        start: float = 0.0,
-        stop: float = 1.0,
-    ):
-        super().__init__(start, stop)
-
-        self.guidance_scale = guidance_scale
-        self.guidance_rescale = guidance_rescale
-        self.use_original_formulation = use_original_formulation
-
-    def prepare_inputs(self, denoiser: torch.nn.Module, *args: Union[Tuple[torch.Tensor], List[torch.Tensor]]) -> Tuple[List[torch.Tensor], ...]:
-        return _default_prepare_inputs(denoiser, self.num_conditions, *args)
-
-    def prepare_outputs(self, denoiser: torch.nn.Module, pred: torch.Tensor) -> None:
-        self._num_outputs_prepared += 1
-        if self._num_outputs_prepared > self.num_conditions:
-            raise ValueError(f"Expected {self.num_conditions} outputs, but prepare_outputs called more times.")
-        key = self._input_predictions[self._num_outputs_prepared - 1]
-        self._preds[key] = pred
-
-    def forward(self, pred_cond: torch.Tensor, pred_uncond: Optional[torch.Tensor] = None) -> torch.Tensor:
-        pred = None
-
-        if not self._is_cfgpp_enabled():
-            pred = pred_cond
-        else:
-            shift = pred_cond - pred_uncond
-            pred = pred_cond if self.use_original_formulation else pred_uncond
-            pred = pred + self.guidance_scale * shift
-
-        if self.guidance_rescale > 0.0:
-            pred = rescale_noise_cfg(pred, pred_cond, self.guidance_rescale)
-
-        return pred
-
-    @property
-    def is_conditional(self) -> bool:
-        return self._num_outputs_prepared == 0
-
-    @property
-    def num_conditions(self) -> int:
-        num_conditions = 1
-        if self._is_cfgpp_enabled():
-            num_conditions += 1
-        return num_conditions
-
-    @property
-    def outputs(self) -> Dict[str, torch.Tensor]:
-        scheduler_step_kwargs = {}
-        if self._is_cfgpp_enabled():
-            scheduler_step_kwargs["_use_cfgpp"] = True
-            scheduler_step_kwargs["_model_output_uncond"] = self._preds.get("pred_uncond")
-        return self._preds, scheduler_step_kwargs
-
-    def _is_cfgpp_enabled(self) -> bool:
-        if not self._enabled:
-            return False
-        
-        is_within_range = True
-        if self._num_inference_steps is not None:
-            skip_start_step = int(self._start * self._num_inference_steps)
-            skip_stop_step = int(self._stop * self._num_inference_steps)
-            is_within_range = skip_start_step <= self._step < skip_stop_step
-        
-        return is_within_range