apply @deltheil suggestions

src/refiners/foundationals/clip/text_encoder.py

@@ -154,7 +154,7 @@ class CLIPTextEncoderL(CLIPTextEncoder):
 
     Note:
         We replace the GeLU activation function with an approximate GeLU to comply with the original CLIP implementation
-        of OpenAI (https://github.com/openai/CLIP/blob/main/clip/model.py#L166)
+        of OpenAI (https://github.com/openai/CLIP/blob/a1d0717/clip/model.py#L166)
 
     See [[arXiv:2103.00020] Learning Transferable Visual Models From Natural Language Supervision](https://arxiv.org/abs/2103.00020)
     for more details.

src/refiners/foundationals/latent_diffusion/auto_encoder.py

@@ -238,8 +238,6 @@ def decode(self, x: Tensor) -> Tensor:
         x = decoder(x / self.encoder_scale)
         return x
 
-    # backward-compatibility alias
-    # TODO: deprecate this method
     def image_to_latents(self, image: Image.Image) -> Tensor:
         return self.images_to_latents([image])
 
@@ -261,7 +259,6 @@ def images_to_latents(self, images: list[Image.Image]) -> Tensor:
     def decode_latents(self, x: Tensor) -> Image.Image:
         return self.latents_to_image(x)
 
-    # TODO: deprecated this method ?
     def latents_to_image(self, x: Tensor) -> Image.Image:
         if x.shape[0] != 1:
             raise ValueError(f"Expected batch size of 1, got {x.shape[0]}")

src/refiners/foundationals/latent_diffusion/solvers/solver.py

@@ -121,6 +121,10 @@ def add_noise(
     def remove_noise(self, x: Tensor, noise: Tensor, step: int) -> Tensor:
         """Remove noise from the input tensor using the current step of the diffusion process.
 
+        Note:
+            See [[arXiv:2006.11239] Denoising Diffusion Probabilistic Models, Equation 15](https://arxiv.org/abs/2006.11239)
+            and [[arXiv:2210.00939] Improving Sample Quality of Diffusion Models Using Self-Attention Guidance](https://arxiv.org/abs/2210.00939).
+
         Args:
             x: The input tensor to remove noise from.
             noise: The noise tensor to remove from the input tensor.
@@ -132,9 +136,6 @@ def remove_noise(self, x: Tensor, noise: Tensor, step: int) -> Tensor:
         timestep = self.timesteps[step]
         cumulative_scale_factors = self.cumulative_scale_factors[timestep]
         noise_stds = self.noise_std[timestep]
-        # See equation (15) from https://arxiv.org/pdf/2006.11239.pdf.
-        # Useful to preview progress or for guidance
-        # See also https://arxiv.org/pdf/2210.00939.pdf (self-attention guidance)
         denoised_x = (x - noise_stds * noise) / cumulative_scale_factors
         return denoised_x
 
@@ -196,6 +197,9 @@ def scale_model_input(self, x: Tensor, step: int) -> Tensor:
             This method should only be overridden by solvers that
             need to scale the input according to the current timestep.
 
+            By default, this method does not scale the input.
+            (scale=1)
+
         Args:
             x: The input tensor to scale.
             step: The current step of the diffusion process.

src/refiners/foundationals/latent_diffusion/stable_diffusion_1/model.py

@@ -96,6 +96,9 @@ def set_unet_context(self, *, timestep: Tensor, clip_text_embedding: Tensor, **_
     def set_self_attention_guidance(self, enable: bool, scale: float = 1.0) -> None:
         """Set whether to enable self-attention guidance.
 
+        See [[arXiv:2210.00939] Improving Sample Quality of Diffusion Models Using Self-Attention Guidance](https://arxiv.org/abs/2210.00939)
+        for more details.
+
         Args:
             enable: Whether to enable self-attention guidance.
             scale: The scale to use.
@@ -114,7 +117,7 @@ def has_self_attention_guidance(self) -> bool:
         return self._find_sag_adapter() is not None
 
     def _find_sag_adapter(self) -> SD1SAGAdapter | None:
-        """Finds the self-attention guidance adapter."""
+        """Finds the self-attention guidance adapter, if any."""
         for p in self.unet.get_parents():
             if isinstance(p, SD1SAGAdapter):
                 return p

src/refiners/foundationals/latent_diffusion/stable_diffusion_xl/model.py

@@ -140,6 +140,9 @@ def forward(
     def set_self_attention_guidance(self, enable: bool, scale: float = 1.0) -> None:
         """Sets the self-attention guidance.
 
+        See [[arXiv:2210.00939] Improving Sample Quality of Diffusion Models Using Self-Attention Guidance](https://arxiv.org/abs/2210.00939)
+        for more details.
+
         Args:
             enable: Whether to enable self-attention guidance or not.
             scale: The scale to use.
@@ -158,7 +161,7 @@ def has_self_attention_guidance(self) -> bool:
         return self._find_sag_adapter() is not None
 
     def _find_sag_adapter(self) -> SDXLSAGAdapter | None:
-        """Finds the self-attention guidance adapter."""
+        """Finds the self-attention guidance adapter, if any."""
         for p in self.unet.get_parents():
             if isinstance(p, SDXLSAGAdapter):
                 return p

src/refiners/foundationals/segment_anything/model.py

@@ -154,10 +154,10 @@ def image_size(self) -> int:
         return w
 
     def compute_target_size(self, size: tuple[int, int]) -> tuple[int, int]:
-        """Compute the target size for a given size.
+        """Compute the target size as expected by the image encoder.
 
         Args:
-            size: The size of the image.
+            size: The size of the input image.
 
         Returns:
             The target height.
@@ -171,7 +171,7 @@ def compute_target_size(self, size: tuple[int, int]) -> tuple[int, int]:
         return (newh, neww)
 
     def preprocess_image(self, image: Image.Image, target_size: tuple[int, int]) -> Tensor:
-        """Preprocess an image.
+        """Preprocess an image without distorting its aspect ratio.
 
         Args:
             image: The image to preprocess.