diffusers 0.33.1__py3-none-any.whl → 0.35.0__py3-none-any.whl

diffusers/pipelines/easyanimate/pipeline_easyanimate_control.py

@@ -177,7 +177,7 @@ def rescale_noise_cfg(noise_cfg, noise_pred_text, guidance_rescale=0.0):
     r"""
     Rescales `noise_cfg` tensor based on `guidance_rescale` to improve image quality and fix overexposure. Based on
     Section 3.4 from [Common Diffusion Noise Schedules and Sample Steps are
-    Flawed](https://arxiv.org/pdf/2305.08891.pdf).
+    Flawed](https://huggingface.co/papers/2305.08891).
 
     Args:
         noise_cfg (`torch.Tensor`):
@@ -515,7 +515,7 @@ class EasyAnimateControlPipeline(DiffusionPipeline):
     def prepare_extra_step_kwargs(self, generator, eta):
         # prepare extra kwargs for the scheduler step, since not all schedulers have the same signature
         # eta (η) is only used with the DDIMScheduler, it will be ignored for other schedulers.
-        # eta corresponds to η in DDIM paper: https://arxiv.org/abs/2010.02502
+        # eta corresponds to η in DDIM paper: https://huggingface.co/papers/2010.02502
         # and should be between [0, 1]
 
         accepts_eta = "eta" in set(inspect.signature(self.scheduler.step).parameters.keys())
@@ -653,7 +653,7 @@ class EasyAnimateControlPipeline(DiffusionPipeline):
         return self._guidance_rescale
 
     # here `guidance_scale` is defined analog to the guidance weight `w` of equation (2)
-    # of the Imagen paper: https://arxiv.org/pdf/2205.11487.pdf . `guidance_scale = 1`
+    # of the Imagen paper: https://huggingface.co/papers/2205.11487 . `guidance_scale = 1`
     # corresponds to doing no classifier free guidance.
     @property
     def do_classifier_free_guidance(self):
@@ -956,7 +956,7 @@ class EasyAnimateControlPipeline(DiffusionPipeline):
                     noise_pred = noise_pred_uncond + guidance_scale * (noise_pred_text - noise_pred_uncond)
 
                 if self.do_classifier_free_guidance and guidance_rescale > 0.0:
-                    # Based on 3.4. in https://arxiv.org/pdf/2305.08891.pdf
+                    # Based on 3.4. in https://huggingface.co/papers/2305.08891
                     noise_pred = rescale_noise_cfg(noise_pred, noise_pred_text, guidance_rescale=guidance_rescale)
 
                 # compute the previous noisy sample x_t -> x_t-1

diffusers/pipelines/easyanimate/pipeline_easyanimate_inpaint.py

@@ -199,7 +199,7 @@ def rescale_noise_cfg(noise_cfg, noise_pred_text, guidance_rescale=0.0):
     r"""
     Rescales `noise_cfg` tensor based on `guidance_rescale` to improve image quality and fix overexposure. Based on
     Section 3.4 from [Common Diffusion Noise Schedules and Sample Steps are
-    Flawed](https://arxiv.org/pdf/2305.08891.pdf).
+    Flawed](https://huggingface.co/papers/2305.08891).
 
     Args:
         noise_cfg (`torch.Tensor`):
@@ -557,7 +557,7 @@ class EasyAnimateInpaintPipeline(DiffusionPipeline):
     def prepare_extra_step_kwargs(self, generator, eta):
         # prepare extra kwargs for the scheduler step, since not all schedulers have the same signature
         # eta (η) is only used with the DDIMScheduler, it will be ignored for other schedulers.
-        # eta corresponds to η in DDIM paper: https://arxiv.org/abs/2010.02502
+        # eta corresponds to η in DDIM paper: https://huggingface.co/papers/2010.02502
         # and should be between [0, 1]
 
         accepts_eta = "eta" in set(inspect.signature(self.scheduler.step).parameters.keys())
@@ -771,7 +771,7 @@ class EasyAnimateInpaintPipeline(DiffusionPipeline):
         return self._guidance_rescale
 
     # here `guidance_scale` is defined analog to the guidance weight `w` of equation (2)
-    # of the Imagen paper: https://arxiv.org/pdf/2205.11487.pdf . `guidance_scale = 1`
+    # of the Imagen paper: https://huggingface.co/papers/2205.11487 . `guidance_scale = 1`
     # corresponds to doing no classifier free guidance.
     @property
     def do_classifier_free_guidance(self):
@@ -849,7 +849,7 @@ class EasyAnimateInpaintPipeline(DiffusionPipeline):
             num_images_per_prompt (`int`, *optional*, defaults to 1):
                 The number of images to generate per prompt.
             eta (`float`, *optional*, defaults to 0.0):
-                A parameter defined in the [DDIM](https://arxiv.org/abs/2010.02502) paper. Only applies to the
+                A parameter defined in the [DDIM](https://huggingface.co/papers/2010.02502) paper. Only applies to the
                 [`~schedulers.DDIMScheduler`] and is ignored in other schedulers. It adjusts noise level during the
                 inference process.
             generator (`torch.Generator` or `List[torch.Generator]`, *optional*):
@@ -883,7 +883,8 @@ class EasyAnimateInpaintPipeline(DiffusionPipeline):
                 inputs will be passed, facilitating enhanced logging or monitoring of the generation process.
             guidance_rescale (`float`, *optional*, defaults to 0.0):
                 Rescale parameter for adjusting noise configuration based on guidance rescale. Based on findings from
-                [Common Diffusion Noise Schedules and Sample Steps are Flawed](https://arxiv.org/pdf/2305.08891.pdf).
+                [Common Diffusion Noise Schedules and Sample Steps are
+                Flawed](https://huggingface.co/papers/2305.08891).
             strength (`float`, *optional*, defaults to 1.0):
                 Affects the overall styling or quality of the generated output. Values closer to 1 usually provide
                 direct adherence to prompts.
@@ -1180,7 +1181,7 @@ class EasyAnimateInpaintPipeline(DiffusionPipeline):
                     noise_pred = noise_pred_uncond + guidance_scale * (noise_pred_text - noise_pred_uncond)
 
                 if self.do_classifier_free_guidance and guidance_rescale > 0.0:
-                    # Based on 3.4. in https://arxiv.org/pdf/2305.08891.pdf
+                    # Based on 3.4. in https://huggingface.co/papers/2305.08891
                     noise_pred = rescale_noise_cfg(noise_pred, noise_pred_text, guidance_rescale=guidance_rescale)
 
                 # compute the previous noisy sample x_t -> x_t-1

diffusers/pipelines/flux/__init__.py

@@ -33,6 +33,8 @@ else:
     _import_structure["pipeline_flux_fill"] = ["FluxFillPipeline"]
     _import_structure["pipeline_flux_img2img"] = ["FluxImg2ImgPipeline"]
     _import_structure["pipeline_flux_inpaint"] = ["FluxInpaintPipeline"]
+    _import_structure["pipeline_flux_kontext"] = ["FluxKontextPipeline"]
+    _import_structure["pipeline_flux_kontext_inpaint"] = ["FluxKontextInpaintPipeline"]
     _import_structure["pipeline_flux_prior_redux"] = ["FluxPriorReduxPipeline"]
 if TYPE_CHECKING or DIFFUSERS_SLOW_IMPORT:
     try:
@@ -52,6 +54,8 @@ if TYPE_CHECKING or DIFFUSERS_SLOW_IMPORT:
         from .pipeline_flux_fill import FluxFillPipeline
         from .pipeline_flux_img2img import FluxImg2ImgPipeline
         from .pipeline_flux_inpaint import FluxInpaintPipeline
+        from .pipeline_flux_kontext import FluxKontextPipeline
+        from .pipeline_flux_kontext_inpaint import FluxKontextInpaintPipeline
         from .pipeline_flux_prior_redux import FluxPriorReduxPipeline
 else:
     import sys

diffusers/pipelines/flux/modeling_flux.py

@@ -1,4 +1,4 @@
-# Copyright 2024 Black Forest Labs and The HuggingFace Team. All rights reserved.
+# Copyright 2025 Black Forest Labs and 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.

diffusers/pipelines/flux/pipeline_flux.py

@@ -1,4 +1,4 @@
-# Copyright 2024 Black Forest Labs and The HuggingFace Team. All rights reserved.
+# Copyright 2025 Black Forest Labs and 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.
@@ -310,7 +310,7 @@ class FluxPipeline(
     def encode_prompt(
         self,
         prompt: Union[str, List[str]],
-        prompt_2: Union[str, List[str]],
+        prompt_2: Optional[Union[str, List[str]]] = None,
         device: Optional[torch.device] = None,
         num_images_per_prompt: int = 1,
         prompt_embeds: Optional[torch.FloatTensor] = None,
@@ -490,14 +490,6 @@ class FluxPipeline(
                 f" {negative_prompt_embeds}. Please make sure to only forward one of the two."
             )
 
-        if prompt_embeds is not None and negative_prompt_embeds is not None:
-            if prompt_embeds.shape != negative_prompt_embeds.shape:
-                raise ValueError(
-                    "`prompt_embeds` and `negative_prompt_embeds` must have the same shape when passed directly, but"
-                    f" got: `prompt_embeds` {prompt_embeds.shape} != `negative_prompt_embeds`"
-                    f" {negative_prompt_embeds.shape}."
-                )
-
         if prompt_embeds is not None and pooled_prompt_embeds is None:
             raise ValueError(
                 "If `prompt_embeds` are provided, `pooled_prompt_embeds` also have to be passed. Make sure to generate `pooled_prompt_embeds` from the same text encoder that was used to generate `prompt_embeds`."
@@ -682,7 +674,8 @@ class FluxPipeline(
                 The prompt or prompts not to guide the image generation to be sent to `tokenizer_2` and
                 `text_encoder_2`. If not defined, `negative_prompt` is used in all the text-encoders.
             true_cfg_scale (`float`, *optional*, defaults to 1.0):
-                When > 1.0 and a provided `negative_prompt`, enables true classifier-free guidance.
+                True classifier-free guidance (guidance scale) is enabled when `true_cfg_scale` > 1 and
+                `negative_prompt` is provided.
             height (`int`, *optional*, defaults to self.unet.config.sample_size * self.vae_scale_factor):
                 The height in pixels of the generated image. This is set to 1024 by default for the best results.
             width (`int`, *optional*, defaults to self.unet.config.sample_size * self.vae_scale_factor):
@@ -695,11 +688,11 @@ class FluxPipeline(
                 their `set_timesteps` method. If not defined, the default behavior when `num_inference_steps` is passed
                 will be used.
             guidance_scale (`float`, *optional*, defaults to 3.5):
-                Guidance scale as defined in [Classifier-Free Diffusion Guidance](https://arxiv.org/abs/2207.12598).
-                `guidance_scale` is defined as `w` of equation 2. of [Imagen
-                Paper](https://arxiv.org/pdf/2205.11487.pdf). Guidance scale is enabled by setting `guidance_scale >
-                1`. Higher guidance scale encourages to generate images that are closely linked to the text `prompt`,
-                usually at the expense of lower image quality.
+                Embedded guiddance scale is enabled by setting `guidance_scale` > 1. Higher `guidance_scale` encourages
+                a model to generate images more aligned with `prompt` at the expense of lower image quality.
+
+                Guidance-distilled models approximates true classifer-free guidance for `guidance_scale` > 1. Refer to
+                the [paper](https://huggingface.co/papers/2210.03142) to learn more.
             num_images_per_prompt (`int`, *optional*, defaults to 1):
                 The number of images to generate per prompt.
             generator (`torch.Generator` or `List[torch.Generator]`, *optional*):
@@ -708,7 +701,7 @@ class FluxPipeline(
             latents (`torch.FloatTensor`, *optional*):
                 Pre-generated noisy latents, sampled from a Gaussian distribution, to be used as inputs for image
                 generation. Can be used to tweak the same generation with different prompts. If not provided, a latents
-                tensor will ge generated by sampling using the supplied random `generator`.
+                tensor will be generated by sampling using the supplied random `generator`.
             prompt_embeds (`torch.FloatTensor`, *optional*):
                 Pre-generated text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt weighting. If not
                 provided, text embeddings will be generated from `prompt` input argument.
@@ -821,7 +814,7 @@ class FluxPipeline(
             (
                 negative_prompt_embeds,
                 negative_pooled_prompt_embeds,
-                _,
+                negative_text_ids,
             ) = self.encode_prompt(
                 prompt=negative_prompt,
                 prompt_2=negative_prompt_2,
@@ -848,6 +841,8 @@ class FluxPipeline(
 
         # 5. Prepare timesteps
         sigmas = np.linspace(1.0, 1 / num_inference_steps, num_inference_steps) if sigmas is None else sigmas
+        if hasattr(self.scheduler.config, "use_flow_sigmas") and self.scheduler.config.use_flow_sigmas:
+            sigmas = None
         image_seq_len = latents.shape[1]
         mu = calculate_shift(
             image_seq_len,
@@ -906,6 +901,9 @@ class FluxPipeline(
             )
 
         # 6. Denoising loop
+        # We set the index here to remove DtoH sync, helpful especially during compilation.
+        # Check out more details here: https://github.com/huggingface/diffusers/pull/11696
+        self.scheduler.set_begin_index(0)
         with self.progress_bar(total=num_inference_steps) as progress_bar:
             for i, t in enumerate(timesteps):
                 if self.interrupt:
@@ -917,32 +915,35 @@ class FluxPipeline(
                 # broadcast to batch dimension in a way that's compatible with ONNX/Core ML
                 timestep = t.expand(latents.shape[0]).to(latents.dtype)
 
-                noise_pred = self.transformer(
-                    hidden_states=latents,
-                    timestep=timestep / 1000,
-                    guidance=guidance,
-                    pooled_projections=pooled_prompt_embeds,
-                    encoder_hidden_states=prompt_embeds,
-                    txt_ids=text_ids,
-                    img_ids=latent_image_ids,
-                    joint_attention_kwargs=self.joint_attention_kwargs,
-                    return_dict=False,
-                )[0]
-
-                if do_true_cfg:
-                    if negative_image_embeds is not None:
-                        self._joint_attention_kwargs["ip_adapter_image_embeds"] = negative_image_embeds
-                    neg_noise_pred = self.transformer(
+                with self.transformer.cache_context("cond"):
+                    noise_pred = self.transformer(
                         hidden_states=latents,
                         timestep=timestep / 1000,
                         guidance=guidance,
-                        pooled_projections=negative_pooled_prompt_embeds,
-                        encoder_hidden_states=negative_prompt_embeds,
+                        pooled_projections=pooled_prompt_embeds,
+                        encoder_hidden_states=prompt_embeds,
                         txt_ids=text_ids,
                         img_ids=latent_image_ids,
                         joint_attention_kwargs=self.joint_attention_kwargs,
                         return_dict=False,
                     )[0]
+
+                if do_true_cfg:
+                    if negative_image_embeds is not None:
+                        self._joint_attention_kwargs["ip_adapter_image_embeds"] = negative_image_embeds
+
+                    with self.transformer.cache_context("uncond"):
+                        neg_noise_pred = self.transformer(
+                            hidden_states=latents,
+                            timestep=timestep / 1000,
+                            guidance=guidance,
+                            pooled_projections=negative_pooled_prompt_embeds,
+                            encoder_hidden_states=negative_prompt_embeds,
+                            txt_ids=negative_text_ids,
+                            img_ids=latent_image_ids,
+                            joint_attention_kwargs=self.joint_attention_kwargs,
+                            return_dict=False,
+                        )[0]
                     noise_pred = neg_noise_pred + true_cfg_scale * (noise_pred - neg_noise_pred)
 
                 # compute the previous noisy sample x_t -> x_t-1

diffusers/pipelines/flux/pipeline_flux_control.py

@@ -1,4 +1,4 @@
-# Copyright 2024 Black Forest Labs and The HuggingFace Team. All rights reserved.
+# Copyright 2025 Black Forest Labs and 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.
@@ -163,9 +163,9 @@ class FluxControlPipeline(
     TextualInversionLoaderMixin,
 ):
     r"""
-    The Flux pipeline for controllable text-to-image generation.
+    The Flux pipeline for controllable text-to-image generation with image conditions.
 
-    Reference: https://blackforestlabs.ai/announcing-black-forest-labs/
+    Reference: https://bfl.ai/flux-1-tools
 
     Args:
         transformer ([`FluxTransformer2DModel`]):
@@ -324,7 +324,7 @@ class FluxControlPipeline(
     def encode_prompt(
         self,
         prompt: Union[str, List[str]],
-        prompt_2: Union[str, List[str]],
+        prompt_2: Optional[Union[str, List[str]]] = None,
         device: Optional[torch.device] = None,
         num_images_per_prompt: int = 1,
         prompt_embeds: Optional[torch.FloatTensor] = None,
@@ -661,11 +661,11 @@ class FluxControlPipeline(
                 their `set_timesteps` method. If not defined, the default behavior when `num_inference_steps` is passed
                 will be used.
             guidance_scale (`float`, *optional*, defaults to 3.5):
-                Guidance scale as defined in [Classifier-Free Diffusion Guidance](https://arxiv.org/abs/2207.12598).
-                `guidance_scale` is defined as `w` of equation 2. of [Imagen
-                Paper](https://arxiv.org/pdf/2205.11487.pdf). Guidance scale is enabled by setting `guidance_scale >
-                1`. Higher guidance scale encourages to generate images that are closely linked to the text `prompt`,
-                usually at the expense of lower image quality.
+                Embedded guidance scale is enabled by setting `guidance_scale` > 1. Higher `guidance_scale` encourages
+                a model to generate images more aligned with prompt at the expense of lower image quality.
+
+                Guidance-distilled models approximates true classifier-free guidance for `guidance_scale` > 1. Refer to
+                the [paper](https://huggingface.co/papers/2210.03142) to learn more.
             num_images_per_prompt (`int`, *optional*, defaults to 1):
                 The number of images to generate per prompt.
             generator (`torch.Generator` or `List[torch.Generator]`, *optional*):

diffusers/pipelines/flux/pipeline_flux_control_img2img.py

@@ -1,4 +1,4 @@
-# Copyright 2024 Black Forest Labs and The HuggingFace Team. All rights reserved.
+# Copyright 2025 Black Forest Labs and 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.
@@ -335,7 +335,7 @@ class FluxControlImg2ImgPipeline(DiffusionPipeline, FluxLoraLoaderMixin, FromSin
     def encode_prompt(
         self,
         prompt: Union[str, List[str]],
-        prompt_2: Union[str, List[str]],
+        prompt_2: Optional[Union[str, List[str]]] = None,
         device: Optional[torch.device] = None,
         num_images_per_prompt: int = 1,
         prompt_embeds: Optional[torch.FloatTensor] = None,
@@ -699,11 +699,11 @@ class FluxControlImg2ImgPipeline(DiffusionPipeline, FluxLoraLoaderMixin, FromSin
                 their `set_timesteps` method. If not defined, the default behavior when `num_inference_steps` is passed
                 will be used.
             guidance_scale (`float`, *optional*, defaults to 7.0):
-                Guidance scale as defined in [Classifier-Free Diffusion Guidance](https://arxiv.org/abs/2207.12598).
-                `guidance_scale` is defined as `w` of equation 2. of [Imagen
-                Paper](https://arxiv.org/pdf/2205.11487.pdf). Guidance scale is enabled by setting `guidance_scale >
-                1`. Higher guidance scale encourages to generate images that are closely linked to the text `prompt`,
-                usually at the expense of lower image quality.
+                Guidance scale as defined in [Classifier-Free Diffusion
+                Guidance](https://huggingface.co/papers/2207.12598). `guidance_scale` is defined as `w` of equation 2.
+                of [Imagen Paper](https://huggingface.co/papers/2205.11487). Guidance scale is enabled by setting
+                `guidance_scale > 1`. Higher guidance scale encourages to generate images that are closely linked to
+                the text `prompt`, usually at the expense of lower image quality.
             num_images_per_prompt (`int`, *optional*, defaults to 1):
                 The number of images to generate per prompt.
             generator (`torch.Generator` or `List[torch.Generator]`, *optional*):

diffusers/pipelines/flux/pipeline_flux_control_inpaint.py

@@ -1,4 +1,4 @@
-# Copyright 2024 Black Forest Labs and The HuggingFace Team. All rights reserved.
+# Copyright 2025 Black Forest Labs and 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.
@@ -374,7 +374,7 @@ class FluxControlInpaintPipeline(
     def encode_prompt(
         self,
         prompt: Union[str, List[str]],
-        prompt_2: Union[str, List[str]],
+        prompt_2: Optional[Union[str, List[str]]] = None,
         device: Optional[torch.device] = None,
         num_images_per_prompt: int = 1,
         prompt_embeds: Optional[torch.FloatTensor] = None,
@@ -857,11 +857,11 @@ class FluxControlInpaintPipeline(
                 their `set_timesteps` method. If not defined, the default behavior when `num_inference_steps` is passed
                 will be used.
             guidance_scale (`float`, *optional*, defaults to 7.0):
-                Guidance scale as defined in [Classifier-Free Diffusion Guidance](https://arxiv.org/abs/2207.12598).
-                `guidance_scale` is defined as `w` of equation 2. of [Imagen
-                Paper](https://arxiv.org/pdf/2205.11487.pdf). Guidance scale is enabled by setting `guidance_scale >
-                1`. Higher guidance scale encourages to generate images that are closely linked to the text `prompt`,
-                usually at the expense of lower image quality.
+                Guidance scale as defined in [Classifier-Free Diffusion
+                Guidance](https://huggingface.co/papers/2207.12598). `guidance_scale` is defined as `w` of equation 2.
+                of [Imagen Paper](https://huggingface.co/papers/2205.11487). Guidance scale is enabled by setting
+                `guidance_scale > 1`. Higher guidance scale encourages to generate images that are closely linked to
+                the text `prompt`, usually at the expense of lower image quality.
             num_images_per_prompt (`int`, *optional*, defaults to 1):
                 The number of images to generate per prompt.
             generator (`torch.Generator` or `List[torch.Generator]`, *optional*):

diffusers/pipelines/flux/pipeline_flux_controlnet.py

@@ -1,4 +1,4 @@
-# Copyright 2024 Black Forest Labs, The HuggingFace Team and The InstantX Team. All rights reserved.
+# Copyright 2025 Black Forest Labs, The HuggingFace Team and The InstantX 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.
@@ -341,7 +341,7 @@ class FluxControlNetPipeline(DiffusionPipeline, FluxLoraLoaderMixin, FromSingleF
     def encode_prompt(
         self,
         prompt: Union[str, List[str]],
-        prompt_2: Union[str, List[str]],
+        prompt_2: Optional[Union[str, List[str]]] = None,
         device: Optional[torch.device] = None,
         num_images_per_prompt: int = 1,
         prompt_embeds: Optional[torch.FloatTensor] = None,
@@ -733,11 +733,11 @@ class FluxControlNetPipeline(DiffusionPipeline, FluxLoraLoaderMixin, FromSingleF
                 their `set_timesteps` method. If not defined, the default behavior when `num_inference_steps` is passed
                 will be used.
             guidance_scale (`float`, *optional*, defaults to 7.0):
-                Guidance scale as defined in [Classifier-Free Diffusion Guidance](https://arxiv.org/abs/2207.12598).
-                `guidance_scale` is defined as `w` of equation 2. of [Imagen
-                Paper](https://arxiv.org/pdf/2205.11487.pdf). Guidance scale is enabled by setting `guidance_scale >
-                1`. Higher guidance scale encourages to generate images that are closely linked to the text `prompt`,
-                usually at the expense of lower image quality.
+                Guidance scale as defined in [Classifier-Free Diffusion
+                Guidance](https://huggingface.co/papers/2207.12598). `guidance_scale` is defined as `w` of equation 2.
+                of [Imagen Paper](https://huggingface.co/papers/2205.11487). Guidance scale is enabled by setting
+                `guidance_scale > 1`. Higher guidance scale encourages to generate images that are closely linked to
+                the text `prompt`, usually at the expense of lower image quality.
             control_guidance_start (`float` or `List[float]`, *optional*, defaults to 0.0):
                 The percentage of total steps at which the ControlNet starts applying.
             control_guidance_end (`float` or `List[float]`, *optional*, defaults to 1.0):

diffusers/pipelines/flux/pipeline_flux_controlnet_image_to_image.py

@@ -335,7 +335,7 @@ class FluxControlNetImg2ImgPipeline(DiffusionPipeline, FluxLoraLoaderMixin, From
     def encode_prompt(
         self,
         prompt: Union[str, List[str]],
-        prompt_2: Union[str, List[str]],
+        prompt_2: Optional[Union[str, List[str]]] = None,
         device: Optional[torch.device] = None,
         num_images_per_prompt: int = 1,
         prompt_embeds: Optional[torch.FloatTensor] = None,
@@ -687,7 +687,8 @@ class FluxControlNetImg2ImgPipeline(DiffusionPipeline, FluxLoraLoaderMixin, From
                 their `set_timesteps` method. If not defined, the default behavior when `num_inference_steps` is passed
                 will be used.
             guidance_scale (`float`, *optional*, defaults to 7.0):
-                Guidance scale as defined in [Classifier-Free Diffusion Guidance](https://arxiv.org/abs/2207.12598).
+                Guidance scale as defined in [Classifier-Free Diffusion
+                Guidance](https://huggingface.co/papers/2207.12598).
             control_mode (`int` or `List[int]`, *optional*):
                 The mode for the ControlNet. If multiple ControlNets are used, this should be a list.
             controlnet_conditioning_scale (`float` or `List[float]`, *optional*, defaults to 1.0):
@@ -800,17 +801,20 @@ class FluxControlNetImg2ImgPipeline(DiffusionPipeline, FluxLoraLoaderMixin, From
             )
             height, width = control_image.shape[-2:]
 
-            control_image = retrieve_latents(self.vae.encode(control_image), generator=generator)
-            control_image = (control_image - self.vae.config.shift_factor) * self.vae.config.scaling_factor
+            # xlab controlnet has a input_hint_block and instantx controlnet does not
+            controlnet_blocks_repeat = False if self.controlnet.input_hint_block is None else True
+            if self.controlnet.input_hint_block is None:
+                control_image = retrieve_latents(self.vae.encode(control_image), generator=generator)
+                control_image = (control_image - self.vae.config.shift_factor) * self.vae.config.scaling_factor
 
-            height_control_image, width_control_image = control_image.shape[2:]
-            control_image = self._pack_latents(
-                control_image,
-                batch_size * num_images_per_prompt,
-                num_channels_latents,
-                height_control_image,
-                width_control_image,
-            )
+                height_control_image, width_control_image = control_image.shape[2:]
+                control_image = self._pack_latents(
+                    control_image,
+                    batch_size * num_images_per_prompt,
+                    num_channels_latents,
+                    height_control_image,
+                    width_control_image,
+                )
 
             if control_mode is not None:
                 control_mode = torch.tensor(control_mode).to(device, dtype=torch.long)
@@ -819,7 +823,9 @@ class FluxControlNetImg2ImgPipeline(DiffusionPipeline, FluxLoraLoaderMixin, From
         elif isinstance(self.controlnet, FluxMultiControlNetModel):
             control_images = []
 
-            for control_image_ in control_image:
+            # xlab controlnet has a input_hint_block and instantx controlnet does not
+            controlnet_blocks_repeat = False if self.controlnet.nets[0].input_hint_block is None else True
+            for i, control_image_ in enumerate(control_image):
                 control_image_ = self.prepare_image(
                     image=control_image_,
                     width=width,
@@ -831,17 +837,18 @@ class FluxControlNetImg2ImgPipeline(DiffusionPipeline, FluxLoraLoaderMixin, From
                 )
                 height, width = control_image_.shape[-2:]
 
-                control_image_ = retrieve_latents(self.vae.encode(control_image_), generator=generator)
-                control_image_ = (control_image_ - self.vae.config.shift_factor) * self.vae.config.scaling_factor
+                if self.controlnet.nets[0].input_hint_block is None:
+                    control_image_ = retrieve_latents(self.vae.encode(control_image_), generator=generator)
+                    control_image_ = (control_image_ - self.vae.config.shift_factor) * self.vae.config.scaling_factor
 
-                height_control_image, width_control_image = control_image_.shape[2:]
-                control_image_ = self._pack_latents(
-                    control_image_,
-                    batch_size * num_images_per_prompt,
-                    num_channels_latents,
-                    height_control_image,
-                    width_control_image,
-                )
+                    height_control_image, width_control_image = control_image_.shape[2:]
+                    control_image_ = self._pack_latents(
+                        control_image_,
+                        batch_size * num_images_per_prompt,
+                        num_channels_latents,
+                        height_control_image,
+                        width_control_image,
+                    )
 
                 control_images.append(control_image_)
 
@@ -955,6 +962,7 @@ class FluxControlNetImg2ImgPipeline(DiffusionPipeline, FluxLoraLoaderMixin, From
                     img_ids=latent_image_ids,
                     joint_attention_kwargs=self.joint_attention_kwargs,
                     return_dict=False,
+                    controlnet_blocks_repeat=controlnet_blocks_repeat,
                 )[0]
 
                 latents_dtype = latents.dtype

diffusers/pipelines/flux/pipeline_flux_controlnet_inpainting.py

@@ -346,7 +346,7 @@ class FluxControlNetInpaintPipeline(DiffusionPipeline, FluxLoraLoaderMixin, From
     def encode_prompt(
         self,
         prompt: Union[str, List[str]],
-        prompt_2: Union[str, List[str]],
+        prompt_2: Optional[Union[str, List[str]]] = None,
         device: Optional[torch.device] = None,
         num_images_per_prompt: int = 1,
         prompt_embeds: Optional[torch.FloatTensor] = None,
@@ -801,7 +801,8 @@ class FluxControlNetInpaintPipeline(DiffusionPipeline, FluxLoraLoaderMixin, From
                 their `set_timesteps` method. If not defined, the default behavior when `num_inference_steps` is passed
                 will be used.
             guidance_scale (`float`, *optional*, defaults to 7.0):
-                Guidance scale as defined in [Classifier-Free Diffusion Guidance](https://arxiv.org/abs/2207.12598).
+                Guidance scale as defined in [Classifier-Free Diffusion
+                Guidance](https://huggingface.co/papers/2207.12598).
             control_guidance_start (`float` or `List[float]`, *optional*, defaults to 0.0):
                 The percentage of total steps at which the ControlNet starts applying.
             control_guidance_end (`float` or `List[float]`, *optional*, defaults to 1.0):

diffusers/pipelines/flux/pipeline_flux_fill.py

@@ -1,4 +1,4 @@
-# Copyright 2024 Black Forest Labs and The HuggingFace Team. All rights reserved.
+# Copyright 2025 Black Forest Labs and 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.
@@ -419,7 +419,7 @@ class FluxFillPipeline(
     def encode_prompt(
         self,
         prompt: Union[str, List[str]],
-        prompt_2: Union[str, List[str]],
+        prompt_2: Optional[Union[str, List[str]]] = None,
         device: Optional[torch.device] = None,
         num_images_per_prompt: int = 1,
         prompt_embeds: Optional[torch.FloatTensor] = None,
@@ -794,11 +794,11 @@ class FluxFillPipeline(
                 their `set_timesteps` method. If not defined, the default behavior when `num_inference_steps` is passed
                 will be used.
             guidance_scale (`float`, *optional*, defaults to 30.0):
-                Guidance scale as defined in [Classifier-Free Diffusion Guidance](https://arxiv.org/abs/2207.12598).
-                `guidance_scale` is defined as `w` of equation 2. of [Imagen
-                Paper](https://arxiv.org/pdf/2205.11487.pdf). Guidance scale is enabled by setting `guidance_scale >
-                1`. Higher guidance scale encourages to generate images that are closely linked to the text `prompt`,
-                usually at the expense of lower image quality.
+                Guidance scale as defined in [Classifier-Free Diffusion
+                Guidance](https://huggingface.co/papers/2207.12598). `guidance_scale` is defined as `w` of equation 2.
+                of [Imagen Paper](https://huggingface.co/papers/2205.11487). Guidance scale is enabled by setting
+                `guidance_scale > 1`. Higher guidance scale encourages to generate images that are closely linked to
+                the text `prompt`, usually at the expense of lower image quality.
             num_images_per_prompt (`int`, *optional*, defaults to 1):
                 The number of images to generate per prompt.
             generator (`torch.Generator` or `List[torch.Generator]`, *optional*):

diffusers/pipelines/flux/pipeline_flux_img2img.py

@@ -1,4 +1,4 @@
-# Copyright 2024 Black Forest Labs and The HuggingFace Team. All rights reserved.
+# Copyright 2025 Black Forest Labs and 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.
@@ -333,7 +333,7 @@ class FluxImg2ImgPipeline(DiffusionPipeline, FluxLoraLoaderMixin, FromSingleFile
     def encode_prompt(
         self,
         prompt: Union[str, List[str]],
-        prompt_2: Union[str, List[str]],
+        prompt_2: Optional[Union[str, List[str]]] = None,
         device: Optional[torch.device] = None,
         num_images_per_prompt: int = 1,
         prompt_embeds: Optional[torch.FloatTensor] = None,
@@ -607,6 +607,39 @@ class FluxImg2ImgPipeline(DiffusionPipeline, FluxLoraLoaderMixin, FromSingleFile
 
         return latents
 
+    # Copied from diffusers.pipelines.flux.pipeline_flux.FluxPipeline.enable_vae_slicing
+    def enable_vae_slicing(self):
+        r"""
+        Enable sliced VAE decoding. When this option is enabled, the VAE will split the input tensor in slices to
+        compute decoding in several steps. This is useful to save some memory and allow larger batch sizes.
+        """
+        self.vae.enable_slicing()
+
+    # Copied from diffusers.pipelines.flux.pipeline_flux.FluxPipeline.disable_vae_slicing
+    def disable_vae_slicing(self):
+        r"""
+        Disable sliced VAE decoding. If `enable_vae_slicing` was previously enabled, this method will go back to
+        computing decoding in one step.
+        """
+        self.vae.disable_slicing()
+
+    # Copied from diffusers.pipelines.flux.pipeline_flux.FluxPipeline.enable_vae_tiling
+    def enable_vae_tiling(self):
+        r"""
+        Enable tiled VAE decoding. When this option is enabled, the VAE will split the input tensor into tiles to
+        compute decoding and encoding in several steps. This is useful for saving a large amount of memory and to allow
+        processing larger images.
+        """
+        self.vae.enable_tiling()
+
+    # Copied from diffusers.pipelines.flux.pipeline_flux.FluxPipeline.disable_vae_tiling
+    def disable_vae_tiling(self):
+        r"""
+        Disable tiled VAE decoding. If `enable_vae_tiling` was previously enabled, this method will go back to
+        computing decoding in one step.
+        """
+        self.vae.disable_tiling()
+
     def prepare_latents(
         self,
         image,
@@ -741,11 +774,11 @@ class FluxImg2ImgPipeline(DiffusionPipeline, FluxLoraLoaderMixin, FromSingleFile
                 their `set_timesteps` method. If not defined, the default behavior when `num_inference_steps` is passed
                 will be used.
             guidance_scale (`float`, *optional*, defaults to 7.0):
-                Guidance scale as defined in [Classifier-Free Diffusion Guidance](https://arxiv.org/abs/2207.12598).
-                `guidance_scale` is defined as `w` of equation 2. of [Imagen
-                Paper](https://arxiv.org/pdf/2205.11487.pdf). Guidance scale is enabled by setting `guidance_scale >
-                1`. Higher guidance scale encourages to generate images that are closely linked to the text `prompt`,
-                usually at the expense of lower image quality.
+                Guidance scale as defined in [Classifier-Free Diffusion
+                Guidance](https://huggingface.co/papers/2207.12598). `guidance_scale` is defined as `w` of equation 2.
+                of [Imagen Paper](https://huggingface.co/papers/2205.11487). Guidance scale is enabled by setting
+                `guidance_scale > 1`. Higher guidance scale encourages to generate images that are closely linked to
+                the text `prompt`, usually at the expense of lower image quality.
             num_images_per_prompt (`int`, *optional*, defaults to 1):
                 The number of images to generate per prompt.
             generator (`torch.Generator` or `List[torch.Generator]`, *optional*):