Chroma inpainting pipeline

[rodjjo]

I'd like someone to add an inpaiting pipeline for Chroma

I created a working one here: see the gist file

Working example:

"""
ChromaInpaintPipeline implements a text-guided image inpainting pipeline for the lodestones/Chroma1-HD model,
based on the ChromaPipeline from Hugging Face Diffusers:contentReference[oaicite:0]{index=0} and the Stable Diffusion inpainting approach:contentReference[oaicite:1]{index=1}.
"""
import gc
import time
import torch
from typing import List, Union, Optional

from diffusers.models import AutoencoderKL
from diffusers.schedulers import FlowMatchEulerDiscreteScheduler
from diffusers.models.transformers import ChromaTransformer2DModel
from transformers import T5EncoderModel, T5TokenizerFast
from diffusers.pipelines.pipeline_utils import DiffusionPipeline, ImagePipelineOutput
from diffusers.utils.torch_utils import randn_tensor
from PIL import Image

from transformers import (
 CLIPImageProcessor,
 CLIPVisionModelWithProjection,
)

import numpy as np
import torch.nn.functional as F
from tqdm.auto import tqdm # Add progress bar library


# 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.
# 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.

import gc
import inspect
from typing import Any, Callable, Dict, List, Optional, Union

import numpy as np
import PIL.Image
import torch
from transformers import (
 CLIPImageProcessor,
 CLIPVisionModelWithProjection,
 T5EncoderModel,
 T5TokenizerFast,
)

from diffusers.image_processor import PipelineImageInput, VaeImageProcessor
from diffusers.loaders import FluxIPAdapterMixin, FluxLoraLoaderMixin, TextualInversionLoaderMixin
from diffusers.models.autoencoders import AutoencoderKL
from diffusers.models.transformers import ChromaTransformer2DModel
from diffusers.schedulers import DDIMScheduler
from diffusers.utils import (
 USE_PEFT_BACKEND,
 is_torch_xla_available,
 logging,
 replace_example_docstring,
 scale_lora_layers,
 unscale_lora_layers,
)
from diffusers.utils.torch_utils import randn_tensor
from diffusers.pipelines.pipeline_utils import DiffusionPipeline
from diffusers.pipelines.chroma.pipeline_output import ChromaPipelineOutput

from diffusers.schedulers import DDIMScheduler
from diffusers.models.transformers import ChromaTransformer2DModel


if is_torch_xla_available():
 import torch_xla.core.xla_model as xm

 XLA_AVAILABLE = True
else:
 XLA_AVAILABLE = False


logger = logging.get_logger(__name__) # pylint: disable=invalid-name

EXAMPLE_DOC_STRING = """
 Examples:
 ```py
 >>> import torch
 >>> from diffusers import FluxInpaintPipeline
 >>> from diffusers.utils import load_image

 >>> pipe = FluxInpaintPipeline.from_pretrained("black-forest-labs/FLUX.1-schnell", torch_dtype=torch.bfloat16)
 >>> pipe.to("cuda")
 >>> prompt = "Face of a yellow cat, high resolution, sitting on a park bench"
 >>> img_url = "https://raw.githubusercontent.com/CompVis/latent-diffusion/main/data/inpainting_examples/overture-creations-5sI6fQgYIuo.png"
 >>> mask_url = "https://raw.githubusercontent.com/CompVis/latent-diffusion/main/data/inpainting_examples/overture-creations-5sI6fQgYIuo_mask.png"
 >>> source = load_image(img_url)
 >>> mask = load_image(mask_url)
 >>> image = pipe(prompt=prompt, image=source, mask_image=mask).images[0]
 >>> image.save("flux_inpainting.png")
 ```
"""


# Copied from diffusers.pipelines.flux.pipeline_flux.calculate_shift
def calculate_shift(
 image_seq_len,
 base_seq_len: int = 256,
 max_seq_len: int = 4096,
 base_shift: float = 0.5,
 max_shift: float = 1.15,
):
 m = (max_shift - base_shift) / (max_seq_len - base_seq_len)
 b = base_shift - m * base_seq_len
 mu = image_seq_len * m + b
 return mu


# Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion_img2img.retrieve_latents
def retrieve_latents(
 encoder_output: torch.Tensor, generator: Optional[torch.Generator] = None, sample_mode: str = "sample"
):
 if hasattr(encoder_output, "latent_dist") and sample_mode == "sample":
 return encoder_output.latent_dist.sample(generator)
 elif hasattr(encoder_output, "latent_dist") and sample_mode == "argmax":
 return encoder_output.latent_dist.mode()
 elif hasattr(encoder_output, "latents"):
 return encoder_output.latents
 else:
 raise AttributeError("Could not access latents of provided encoder_output")


# Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion.retrieve_timesteps
def retrieve_timesteps(
 scheduler,
 num_inference_steps: Optional[int] = None,
 device: Optional[Union[str, torch.device]] = None,
 timesteps: Optional[List[int]] = None,
 sigmas: Optional[List[float]] = None,
 **kwargs,
):
 r"""
 Calls the scheduler's `set_timesteps` method and retrieves timesteps from the scheduler after the call. Handles
 custom timesteps. Any kwargs will be supplied to `scheduler.set_timesteps`.

 Args:
 scheduler (`SchedulerMixin`):
 The scheduler to get timesteps from.
 num_inference_steps (`int`):
 The number of diffusion steps used when generating samples with a pre-trained model. If used, `timesteps`
 must be `None`.
 device (`str` or `torch.device`, *optional*):
 The device to which the timesteps should be moved to. If `None`, the timesteps are not moved.
 timesteps (`List[int]`, *optional*):
 Custom timesteps used to override the timestep spacing strategy of the scheduler. If `timesteps` is passed,
 `num_inference_steps` and `sigmas` must be `None`.
 sigmas (`List[float]`, *optional*):
 Custom sigmas used to override the timestep spacing strategy of the scheduler. If `sigmas` is passed,
 `num_inference_steps` and `timesteps` must be `None`.

 Returns:
 `Tuple[torch.Tensor, int]`: A tuple where the first element is the timestep schedule from the scheduler and the
 second element is the number of inference steps.
 """
 if timesteps is not None and sigmas is not None:
 raise ValueError("Only one of `timesteps` or `sigmas` can be passed. Please choose one to set custom values")
 if timesteps is not None:
 accepts_timesteps = "timesteps" in set(inspect.signature(scheduler.set_timesteps).parameters.keys())
 if not accepts_timesteps:
 raise ValueError(
 f"The current scheduler class {scheduler.__class__}'s `set_timesteps` does not support custom"
 f" timestep schedules. Please check whether you are using the correct scheduler."
 )
 scheduler.set_timesteps(timesteps=timesteps, device=device, **kwargs)
 timesteps = scheduler.timesteps
 num_inference_steps = len(timesteps)
 elif sigmas is not None:
 accept_sigmas = "sigmas" in set(inspect.signature(scheduler.set_timesteps).parameters.keys())
 if not accept_sigmas:
 raise ValueError(
 f"The current scheduler class {scheduler.__class__}'s `set_timesteps` does not support custom"
 f" sigmas schedules. Please check whether you are using the correct scheduler."
 )
 scheduler.set_timesteps(sigmas=sigmas, device=device, **kwargs)
 timesteps = scheduler.timesteps
 num_inference_steps = len(timesteps)
 else:
 scheduler.set_timesteps(num_inference_steps, device=device, **kwargs)
 timesteps = scheduler.timesteps
 return timesteps, num_inference_steps


class ChromaInpaintPipeline(DiffusionPipeline, FluxLoraLoaderMixin, FluxIPAdapterMixin):
 r"""
 The Flux pipeline for image inpainting.

 Reference: https://blackforestlabs.ai/announcing-black-forest-labs/

 Args:
 transformer ([`ChromaTransformer2DModel`]):
 Conditional Transformer (MMDiT) architecture to denoise the encoded image latents.
 scheduler ([`DDIMScheduler`]):
 A scheduler to be used in combination with `transformer` to denoise the encoded image latents.
 vae ([`AutoencoderKL`]):
 Variational Auto-Encoder (VAE) Model to encode and decode images to and from latent representations.
 text_encoder ([`CLIPTextModel`]):
 [CLIP](https://huggingface.co/docs/transformers/model_doc/clip#transformers.CLIPTextModel), specifically
 the [clip-vit-large-patch14](https://huggingface.co/openai/clip-vit-large-patch14) variant.
 text_encoder_2 ([`T5EncoderModel`]):
 [T5](https://huggingface.co/docs/transformers/en/model_doc/t5#transformers.T5EncoderModel), specifically
 the [google/t5-v1_1-xxl](https://huggingface.co/google/t5-v1_1-xxl) variant.
 tokenizer (`CLIPTokenizer`):
 Tokenizer of class
 [CLIPTokenizer](https://huggingface.co/docs/transformers/en/model_doc/clip#transformers.CLIPTokenizer).
 tokenizer_2 (`T5TokenizerFast`):
 Second Tokenizer of class
 [T5TokenizerFast](https://huggingface.co/docs/transformers/en/model_doc/t5#transformers.T5TokenizerFast).
 """

 model_cpu_offload_seq = "text_encoder->text_encoder_2->image_encoder->transformer->vae"
 _optional_components = ["image_encoder", "feature_extractor"]
 _callback_tensor_inputs = ["latents", "prompt_embeds"]

 def __init__(
 self,
 scheduler: FlowMatchEulerDiscreteScheduler,
 vae: AutoencoderKL,
 text_encoder: T5EncoderModel,
 tokenizer: T5TokenizerFast,
 transformer: ChromaTransformer2DModel,
 image_encoder: CLIPVisionModelWithProjection = None,
 feature_extractor: CLIPImageProcessor = None,
 ):
 super().__init__()

 self.register_modules(
 vae=vae,
 text_encoder=text_encoder,
 tokenizer=tokenizer,
 transformer=transformer,
 scheduler=scheduler,
 image_encoder=image_encoder,
 feature_extractor=feature_extractor,
 )
 self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1) if getattr(self, "vae", None) else 8
 self.latent_channels = self.vae.config.latent_channels if getattr(self, "vae", None) else 16

 # Flux latents are turned into 2x2 patches and packed. This means the latent width and height has to be divisible
 # by the patch size. So the vae scale factor is multiplied by the patch size to account for this
 self.image_processor = VaeImageProcessor(vae_scale_factor=self.vae_scale_factor * 2)
 self.default_sample_size = 128

 self.mask_processor = VaeImageProcessor(
 vae_scale_factor=self.vae_scale_factor * 2,
 vae_latent_channels=self.latent_channels,
 do_normalize=False,
 do_binarize=True,
 do_convert_grayscale=True,
 )

 def _get_t5_prompt_embeds(
 self,
 prompt: Union[str, List[str]] = None,
 num_images_per_prompt: int = 1,
 max_sequence_length: int = 512,
 device: Optional[torch.device] = None,
 dtype: Optional[torch.dtype] = None,
 ):
 device = device or self._execution_device
 dtype = dtype or self.text_encoder.dtype

 prompt = [prompt] if isinstance(prompt, str) else prompt
 batch_size = len(prompt)

 if isinstance(self, TextualInversionLoaderMixin):
 prompt = self.maybe_convert_prompt(prompt, self.tokenizer)

 text_inputs = self.tokenizer(
 prompt,
 padding="max_length",
 max_length=max_sequence_length,
 truncation=True,
 return_length=False,
 return_overflowing_tokens=False,
 return_tensors="pt",
 )
 text_input_ids = text_inputs.input_ids
 tokenizer_mask = text_inputs.attention_mask

 tokenizer_mask_device = tokenizer_mask.to(device)

 prompt_embeds = self.text_encoder(
 text_input_ids.to(device),
 output_hidden_states=False,
 attention_mask=tokenizer_mask_device,
 )[0]

 prompt_embeds = prompt_embeds.to(dtype=dtype, device=device)

 seq_lengths = tokenizer_mask_device.sum(dim=1)
 mask_indices = torch.arange(tokenizer_mask_device.size(1), device=device).unsqueeze(0).expand(batch_size, -1)
 attention_mask = (mask_indices <= seq_lengths.unsqueeze(1)).to(dtype=dtype, device=device)

 _, seq_len, _ = prompt_embeds.shape

 # duplicate text embeddings and attention mask for each generation per prompt, using mps friendly method
 prompt_embeds = prompt_embeds.repeat(1, num_images_per_prompt, 1)
 prompt_embeds = prompt_embeds.view(batch_size * num_images_per_prompt, seq_len, -1)

 attention_mask = attention_mask.repeat(1, num_images_per_prompt)
 attention_mask = attention_mask.view(batch_size * num_images_per_prompt, seq_len)

 return prompt_embeds, attention_mask

 def encode_prompt(
 self,
 prompt: Union[str, List[str]],
 negative_prompt: Union[str, List[str]] = None,
 device: Optional[torch.device] = None,
 num_images_per_prompt: int = 1,
 prompt_embeds: Optional[torch.Tensor] = None,
 negative_prompt_embeds: Optional[torch.Tensor] = None,
 prompt_attention_mask: Optional[torch.Tensor] = None,
 negative_prompt_attention_mask: Optional[torch.Tensor] = None,
 do_classifier_free_guidance: bool = True,
 max_sequence_length: int = 256,
 lora_scale: Optional[float] = None,
 ):
 r"""

 Args:
 prompt (`str` or `List[str]`, *optional*):
 prompt to be encoded
 negative_prompt (`str` or `List[str]`, *optional*):
 The prompt not to guide the image generation. If not defined, one has to pass `negative_prompt_embeds`
 instead. Ignored when not using guidance (i.e., ignored if `guidance_scale` is less than `1`).
 device: (`torch.device`):
 torch device
 num_images_per_prompt (`int`):
 number of images that should be generated per prompt
 prompt_embeds (`torch.Tensor`, *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.
 lora_scale (`float`, *optional*):
 A lora scale that will be applied to all LoRA layers of the text encoder if LoRA layers are loaded.
 """
 device = device or self._execution_device

 # set lora scale so that monkey patched LoRA
 # function of text encoder can correctly access it
 if lora_scale is not None and isinstance(self, FluxLoraLoaderMixin):
 self._lora_scale = lora_scale

 # dynamically adjust the LoRA scale
 if self.text_encoder is not None and USE_PEFT_BACKEND:
 scale_lora_layers(self.text_encoder, lora_scale)

 prompt = [prompt] if isinstance(prompt, str) else prompt

 if prompt is not None:
 batch_size = len(prompt)
 else:
 batch_size = prompt_embeds.shape[0]

 if prompt_embeds is None:
 prompt_embeds, prompt_attention_mask = self._get_t5_prompt_embeds(
 prompt=prompt,
 num_images_per_prompt=num_images_per_prompt,
 max_sequence_length=max_sequence_length,
 device=device,
 )

 dtype = self.text_encoder.dtype if self.text_encoder is not None else self.transformer.dtype
 text_ids = torch.zeros(prompt_embeds.shape[1], 3).to(device=device, dtype=dtype)
 negative_text_ids = None

 if do_classifier_free_guidance:
 if negative_prompt_embeds is None:
 negative_prompt = negative_prompt or ""
 negative_prompt = (
 batch_size * [negative_prompt] if isinstance(negative_prompt, str) else negative_prompt
 )

 if prompt is not None and type(prompt) is not type(negative_prompt):
 raise TypeError(
 f"`negative_prompt` should be the same type to `prompt`, but got {type(negative_prompt)} !="
 f" {type(prompt)}."
 )
 elif batch_size != len(negative_prompt):
 raise ValueError(
 f"`negative_prompt`: {negative_prompt} has batch size {len(negative_prompt)}, but `prompt`:"
 f" {prompt} has batch size {batch_size}. Please make sure that passed `negative_prompt` matches"
 " the batch size of `prompt`."
 )

 negative_prompt_embeds, negative_prompt_attention_mask = self._get_t5_prompt_embeds(
 prompt=negative_prompt,
 num_images_per_prompt=num_images_per_prompt,
 max_sequence_length=max_sequence_length,
 device=device,
 )

 negative_text_ids = torch.zeros(negative_prompt_embeds.shape[1], 3).to(device=device, dtype=dtype)

 if self.text_encoder is not None:
 if isinstance(self, FluxLoraLoaderMixin) and USE_PEFT_BACKEND:
 # Retrieve the original scale by scaling back the LoRA layers
 unscale_lora_layers(self.text_encoder, lora_scale)

 return (
 prompt_embeds,
 text_ids,
 prompt_attention_mask,
 negative_prompt_embeds,
 negative_text_ids,
 negative_prompt_attention_mask,
 )

 # Copied from diffusers.pipelines.flux.pipeline_flux.FluxPipeline.encode_image
 def encode_image(self, image, device, num_images_per_prompt):
 dtype = next(self.image_encoder.parameters()).dtype

 if not isinstance(image, torch.Tensor):
 image = self.feature_extractor(image, return_tensors="pt").pixel_values

 image = image.to(device=device, dtype=dtype)
 image_embeds = self.image_encoder(image).image_embeds
 image_embeds = image_embeds.repeat_interleave(num_images_per_prompt, dim=0)
 return image_embeds

 # Copied from diffusers.pipelines.flux.pipeline_flux.FluxPipeline.prepare_ip_adapter_image_embeds
 def prepare_ip_adapter_image_embeds(
 self, ip_adapter_image, ip_adapter_image_embeds, device, num_images_per_prompt
 ):
 image_embeds = []
 if ip_adapter_image_embeds is None:
 if not isinstance(ip_adapter_image, list):
 ip_adapter_image = [ip_adapter_image]

 if len(ip_adapter_image) != self.transformer.encoder_hid_proj.num_ip_adapters:
 raise ValueError(
 f"`ip_adapter_image` must have same length as the number of IP Adapters. Got {len(ip_adapter_image)} images and {self.transformer.encoder_hid_proj.num_ip_adapters} IP Adapters."
 )

 for single_ip_adapter_image in ip_adapter_image:
 single_image_embeds = self.encode_image(single_ip_adapter_image, device, 1)
 image_embeds.append(single_image_embeds[None, :])
 else:
 if not isinstance(ip_adapter_image_embeds, list):
 ip_adapter_image_embeds = [ip_adapter_image_embeds]

 if len(ip_adapter_image_embeds) != self.transformer.encoder_hid_proj.num_ip_adapters:
 raise ValueError(
 f"`ip_adapter_image_embeds` must have same length as the number of IP Adapters. Got {len(ip_adapter_image_embeds)} image embeds and {self.transformer.encoder_hid_proj.num_ip_adapters} IP Adapters."
 )

 for single_image_embeds in ip_adapter_image_embeds:
 image_embeds.append(single_image_embeds)

 ip_adapter_image_embeds = []
 for single_image_embeds in image_embeds:
 single_image_embeds = torch.cat([single_image_embeds] * num_images_per_prompt, dim=0)
 single_image_embeds = single_image_embeds.to(device=device)
 ip_adapter_image_embeds.append(single_image_embeds)

 return ip_adapter_image_embeds

 # Copied from diffusers.pipelines.stable_diffusion_3.pipeline_stable_diffusion_3_inpaint.StableDiffusion3InpaintPipeline._encode_vae_image
 def _encode_vae_image(self, image: torch.Tensor, generator: torch.Generator):
 if isinstance(generator, list):
 image_latents = [
 retrieve_latents(self.vae.encode(image[i : i + 1]), generator=generator[i])
 for i in range(image.shape[0])
 ]
 image_latents = torch.cat(image_latents, dim=0)
 else:
 image_latents = retrieve_latents(self.vae.encode(image), generator=generator)

 image_latents = (image_latents - self.vae.config.shift_factor) * self.vae.config.scaling_factor
 
 return image_latents

 # Copied from diffusers.pipelines.stable_diffusion_3.pipeline_stable_diffusion_3_img2img.StableDiffusion3Img2ImgPipeline.get_timesteps
 def get_timesteps(self, num_inference_steps, strength, device):
 # get the original timestep using init_timestep
 init_timestep = min(num_inference_steps * strength, num_inference_steps)

 t_start = int(max(num_inference_steps - init_timestep, 0))
 timesteps = self.scheduler.timesteps[t_start * self.scheduler.order :]
 if hasattr(self.scheduler, "set_begin_index"):
 self.scheduler.set_begin_index(t_start * self.scheduler.order)

 return timesteps, num_inference_steps - t_start
 
 def check_inputs(
 self,
 prompt,
 height,
 width,
 strength,
 negative_prompt=None,
 prompt_embeds=None,
 negative_prompt_embeds=None,
 prompt_attention_mask=None,
 negative_prompt_attention_mask=None,
 callback_on_step_end_tensor_inputs=None,
 max_sequence_length=None,
 image=None,
 mask_image=None,
 padding_mask_crop=None,
 ):
 if strength < 0 or strength > 1:
 raise ValueError(f"The value of strength should in [0.0, 1.0] but is {strength}")

 if height % (self.vae_scale_factor * 2) != 0 or width % (self.vae_scale_factor * 2) != 0:
 logger.warning(
 f"`height` and `width` have to be divisible by {self.vae_scale_factor * 2} but are {height} and {width}. Dimensions will be resized accordingly"
 )

 if callback_on_step_end_tensor_inputs is not None and not all(
 k in self._callback_tensor_inputs for k in callback_on_step_end_tensor_inputs
 ):
 raise ValueError(
 f"`callback_on_step_end_tensor_inputs` has to be in {self._callback_tensor_inputs}, but found {[k for k in callback_on_step_end_tensor_inputs if k not in self._callback_tensor_inputs]}"
 )

 if prompt is not None and prompt_embeds is not None:
 raise ValueError(
 f"Cannot forward both `prompt`: {prompt} and `prompt_embeds`: {prompt_embeds}. Please make sure to"
 " only forward one of the two."
 )
 elif prompt is None and prompt_embeds is None:
 raise ValueError(
 "Provide either `prompt` or `prompt_embeds`. Cannot leave both `prompt` and `prompt_embeds` undefined."
 )
 elif prompt is not None and (not isinstance(prompt, str) and not isinstance(prompt, list)):
 raise ValueError(f"`prompt` has to be of type `str` or `list` but is {type(prompt)}")

 if negative_prompt is not None and negative_prompt_embeds is not None:
 raise ValueError(
 f"Cannot forward both `negative_prompt`: {negative_prompt} and `negative_prompt_embeds`:"
 f" {negative_prompt_embeds}. Please make sure to only forward one of the two."
 )

 if prompt_embeds is not None and prompt_attention_mask is None:
 raise ValueError("Cannot provide `prompt_embeds` without also providing `prompt_attention_mask")

 if negative_prompt_embeds is not None and negative_prompt_attention_mask is None:
 raise ValueError(
 "Cannot provide `negative_prompt_embeds` without also providing `negative_prompt_attention_mask"
 )

 if max_sequence_length is not None and max_sequence_length > 512:
 raise ValueError(f"`max_sequence_length` cannot be greater than 512 but is {max_sequence_length}")
 
 if padding_mask_crop is not None:
 if not isinstance(image, PIL.Image.Image):
 raise ValueError(
 f"The image should be a PIL image when inpainting mask crop, but is of type {type(image)}."
 )
 if not isinstance(mask_image, PIL.Image.Image):
 raise ValueError(
 f"The mask image should be a PIL image when inpainting mask crop, but is of type"
 f" {type(mask_image)}."
 )


 def check_inputs(
 self,
 prompt,
 image,
 mask_image,
 strength,
 height,
 width,
 negative_prompt=None,
 prompt_embeds=None,
 negative_prompt_embeds=None,
 pooled_prompt_embeds=None,
 negative_pooled_prompt_embeds=None,
 callback_on_step_end_tensor_inputs=None,
 padding_mask_crop=None,
 max_sequence_length=None,
 prompt_attention_mask=None,
 negative_prompt_attention_mask=None,
 ):
 if strength < 0 or strength > 1:
 raise ValueError(f"The value of strength should in [0.0, 1.0] but is {strength}")

 if height % (self.vae_scale_factor * 2) != 0 or width % (self.vae_scale_factor * 2) != 0:
 logger.warning(
 f"`height` and `width` have to be divisible by {self.vae_scale_factor * 2} but are {height} and {width}. Dimensions will be resized accordingly"
 )

 if callback_on_step_end_tensor_inputs is not None and not all(
 k in self._callback_tensor_inputs for k in callback_on_step_end_tensor_inputs
 ):
 raise ValueError(
 f"`callback_on_step_end_tensor_inputs` has to be in {self._callback_tensor_inputs}, but found {[k for k in callback_on_step_end_tensor_inputs if k not in self._callback_tensor_inputs]}"
 )

 if prompt is not None and prompt_embeds is not None:
 raise ValueError(
 f"Cannot forward both `prompt`: {prompt} and `prompt_embeds`: {prompt_embeds}. Please make sure to"
 " only forward one of the two."
 )
 elif prompt is None and prompt_embeds is None:
 raise ValueError(
 "Provide either `prompt` or `prompt_embeds`. Cannot leave both `prompt` and `prompt_embeds` undefined."
 )
 elif prompt is not None and (not isinstance(prompt, str) and not isinstance(prompt, list)):
 raise ValueError(f"`prompt` has to be of type `str` or `list` but is {type(prompt)}")

 if negative_prompt is not None and negative_prompt_embeds is not None:
 raise ValueError(
 f"Cannot forward both `negative_prompt`: {negative_prompt} and `negative_prompt_embeds`:"
 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`."
 )
 if negative_prompt_embeds is not None and negative_pooled_prompt_embeds is None:
 raise ValueError(
 "If `negative_prompt_embeds` are provided, `negative_pooled_prompt_embeds` also have to be passed. Make sure to generate `negative_pooled_prompt_embeds` from the same text encoder that was used to generate `negative_prompt_embeds`."
 )
 
 if prompt_embeds is not None and prompt_attention_mask is None:
 raise ValueError("Cannot provide `prompt_embeds` without also providing `prompt_attention_mask")

 if negative_prompt_embeds is not None and negative_prompt_attention_mask is None:
 raise ValueError(
 "Cannot provide `negative_prompt_embeds` without also providing `negative_prompt_attention_mask"
 )

 if padding_mask_crop is not None:
 if not isinstance(image, PIL.Image.Image):
 raise ValueError(
 f"The image should be a PIL image when inpainting mask crop, but is of type {type(image)}."
 )
 if not isinstance(mask_image, PIL.Image.Image):
 raise ValueError(
 f"The mask image should be a PIL image when inpainting mask crop, but is of type"
 f" {type(mask_image)}."
 )
 if output_type != "pil":
 raise ValueError(f"The output type should be PIL when inpainting mask crop, but is {output_type}.")

 if max_sequence_length is not None and max_sequence_length > 512:
 raise ValueError(f"`max_sequence_length` cannot be greater than 512 but is {max_sequence_length}")

 @staticmethod
 # Copied from diffusers.pipelines.flux.pipeline_flux.FluxPipeline._prepare_latent_image_ids
 def _prepare_latent_image_ids(batch_size, height, width, device, dtype):
 latent_image_ids = torch.zeros(height, width, 3)
 latent_image_ids[..., 1] = latent_image_ids[..., 1] + torch.arange(height)[:, None]
 latent_image_ids[..., 2] = latent_image_ids[..., 2] + torch.arange(width)[None, :]

 latent_image_id_height, latent_image_id_width, latent_image_id_channels = latent_image_ids.shape

 latent_image_ids = latent_image_ids.reshape(
 latent_image_id_height * latent_image_id_width, latent_image_id_channels
 )

 return latent_image_ids.to(device=device, dtype=dtype)

 @staticmethod
 # Copied from diffusers.pipelines.flux.pipeline_flux.FluxPipeline._pack_latents
 def _pack_latents(latents, batch_size, num_channels_latents, height, width):
 latents = latents.view(batch_size, num_channels_latents, height // 2, 2, width // 2, 2)
 latents = latents.permute(0, 2, 4, 1, 3, 5)
 latents = latents.reshape(batch_size, (height // 2) * (width // 2), num_channels_latents * 4)

 return latents

 @staticmethod
 # Copied from diffusers.pipelines.flux.pipeline_flux.FluxPipeline._unpack_latents
 def _unpack_latents(latents, height, width, vae_scale_factor):
 batch_size, num_patches, channels = latents.shape

 # VAE applies 8x compression on images but we must also account for packing which requires
 # latent height and width to be divisible by 2.
 height = 2 * (int(height) // (vae_scale_factor * 2))
 width = 2 * (int(width) // (vae_scale_factor * 2))

 latents = latents.view(batch_size, height // 2, width // 2, channels // 4, 2, 2)
 latents = latents.permute(0, 3, 1, 4, 2, 5)

 latents = latents.reshape(batch_size, channels // (2 * 2), height, width)

 return latents

 def prepare_latents(
 self,
 image,
 timestep,
 batch_size,
 num_channels_latents,
 height,
 width,
 dtype,
 device,
 generator,
 latents=None,
 ):
 if isinstance(generator, list) and len(generator) != batch_size:
 raise ValueError(
 f"You have passed a list of generators of length {len(generator)}, but requested an effective batch"
 f" size of {batch_size}. Make sure the batch size matches the length of the generators."
 )

 # VAE applies 8x compression on images but we must also account for packing which requires
 # latent height and width to be divisible by 2.
 height = 2 * (int(height) // (self.vae_scale_factor * 2))
 width = 2 * (int(width) // (self.vae_scale_factor * 2))
 shape = (batch_size, num_channels_latents, height, width)
 latent_image_ids = self._prepare_latent_image_ids(batch_size, height // 2, width // 2, device, dtype)

 image = image.to(device=device, dtype=dtype)
 if image.shape[1] != self.latent_channels:
 image_latents = self._encode_vae_image(image=image, generator=generator)
 else:
 image_latents = image

 if batch_size > image_latents.shape[0] and batch_size % image_latents.shape[0] == 0:
 # expand init_latents for batch_size
 additional_image_per_prompt = batch_size // image_latents.shape[0]
 image_latents = torch.cat([image_latents] * additional_image_per_prompt, dim=0)
 elif batch_size > image_latents.shape[0] and batch_size % image_latents.shape[0] != 0:
 raise ValueError(
 f"Cannot duplicate `image` of batch size {image_latents.shape[0]} to {batch_size} text prompts."
 )
 else:
 image_latents = torch.cat([image_latents], dim=0)

 if latents is None:
 noise = randn_tensor(shape, generator=generator, device=device, dtype=dtype)
 latents = self.scheduler.scale_noise(image_latents, timestep, noise)
 else:
 noise = latents.to(device)
 latents = noise

 noise = self._pack_latents(noise, batch_size, num_channels_latents, height, width)
 image_latents = self._pack_latents(image_latents, batch_size, num_channels_latents, height, width)
 latents = self._pack_latents(latents, batch_size, num_channels_latents, height, width)
 return latents, noise, image_latents, latent_image_ids

 def prepare_mask_latents(
 self,
 mask,
 masked_image,
 batch_size,
 num_channels_latents,
 num_images_per_prompt,
 height,
 width,
 dtype,
 device,
 generator,
 ):
 # VAE applies 8x compression on images but we must also account for packing which requires
 # latent height and width to be divisible by 2.
 height = 2 * (int(height) // (self.vae_scale_factor * 2))
 width = 2 * (int(width) // (self.vae_scale_factor * 2))
 # resize the mask to latents shape as we concatenate the mask to the latents
 # we do that before converting to dtype to avoid breaking in case we're using cpu_offload
 # and half precision
 mask = torch.nn.functional.interpolate(mask, size=(height, width))
 mask = mask.to(device=device, dtype=dtype)

 batch_size = batch_size * num_images_per_prompt

 masked_image = masked_image.to(device=device, dtype=dtype)

 if masked_image.shape[1] == 16:
 masked_image_latents = masked_image
 else:
 masked_image_latents = retrieve_latents(self.vae.encode(masked_image), generator=generator)

 masked_image_latents = (
 masked_image_latents - self.vae.config.shift_factor
 ) * self.vae.config.scaling_factor

 # duplicate mask and masked_image_latents for each generation per prompt, using mps friendly method
 if mask.shape[0] < batch_size:
 if not batch_size % mask.shape[0] == 0:
 raise ValueError(
 "The passed mask and the required batch size don't match. Masks are supposed to be duplicated to"
 f" a total batch size of {batch_size}, but {mask.shape[0]} masks were passed. Make sure the number"
 " of masks that you pass is divisible by the total requested batch size."
 )
 mask = mask.repeat(batch_size // mask.shape[0], 1, 1, 1)
 if masked_image_latents.shape[0] < batch_size:
 if not batch_size % masked_image_latents.shape[0] == 0:
 raise ValueError(
 "The passed images and the required batch size don't match. Images are supposed to be duplicated"
 f" to a total batch size of {batch_size}, but {masked_image_latents.shape[0]} images were passed."
 " Make sure the number of images that you pass is divisible by the total requested batch size."
 )
 masked_image_latents = masked_image_latents.repeat(batch_size // masked_image_latents.shape[0], 1, 1, 1)

 # aligning device to prevent device errors when concating it with the latent model input
 masked_image_latents = masked_image_latents.to(device=device, dtype=dtype)
 masked_image_latents = self._pack_latents(
 masked_image_latents,
 batch_size,
 num_channels_latents,
 height,
 width,
 )
 mask = self._pack_latents(
 mask.repeat(1, num_channels_latents, 1, 1),
 batch_size,
 num_channels_latents,
 height,
 width,
 )

 return mask, masked_image_latents

 @property
 def guidance_scale(self):
 return self._guidance_scale

 @property
 def joint_attention_kwargs(self):
 return self._joint_attention_kwargs

 @property
 def num_timesteps(self):
 return self._num_timesteps

 @property
 def interrupt(self):
 return self._interrupt
 
 def _prepare_attention_mask(
 self,
 batch_size,
 sequence_length,
 dtype,
 attention_mask=None,
 ):
 if attention_mask is None:
 return attention_mask

 # Extend the prompt attention mask to account for image tokens in the final sequence
 attention_mask = torch.cat(
 [attention_mask, torch.ones(batch_size, sequence_length, device=attention_mask.device)],
 dim=1,
 )
 attention_mask = attention_mask.to(dtype)

 return attention_mask
 
 @property
 def do_classifier_free_guidance(self):
 return self._guidance_scale > 1

 @replace_example_docstring(EXAMPLE_DOC_STRING)
 @torch.no_grad()
 def __call__(
 self,
 prompt: Union[str, List[str]] = None,
 negative_prompt: Union[str, List[str]] = None,
 true_cfg_scale: float = 1.0,
 image: PipelineImageInput = None,
 mask_image: PipelineImageInput = None,
 masked_image_latents: PipelineImageInput = None,
 height: Optional[int] = None,
 width: Optional[int] = None,
 padding_mask_crop: Optional[int] = None,
 strength: float = 0.6,
 num_inference_steps: int = 28,
 sigmas: Optional[List[float]] = None,
 guidance_scale: float = 7.0,
 num_images_per_prompt: Optional[int] = 1,
 generator: Optional[Union[torch.Generator, List[torch.Generator]]] = None,
 latents: Optional[torch.FloatTensor] = None,
 prompt_embeds: Optional[torch.FloatTensor] = None,
 pooled_prompt_embeds: Optional[torch.FloatTensor] = None,
 ip_adapter_image: Optional[PipelineImageInput] = None,
 ip_adapter_image_embeds: Optional[List[torch.Tensor]] = None,
 negative_ip_adapter_image: Optional[PipelineImageInput] = None,
 negative_ip_adapter_image_embeds: Optional[List[torch.Tensor]] = None,
 negative_prompt_embeds: Optional[torch.FloatTensor] = None,
 negative_pooled_prompt_embeds: Optional[torch.FloatTensor] = None,
 output_type: Optional[str] = "pil",
 return_dict: bool = True,
 joint_attention_kwargs: Optional[Dict[str, Any]] = None,
 callback_on_step_end: Optional[Callable[[int, int, Dict], None]] = None,
 callback_on_step_end_tensor_inputs: List[str] = ["latents"],
 max_sequence_length: int = 256,
 prompt_attention_mask: Optional[torch.Tensor] = None,
 negative_prompt_attention_mask: Optional[torch.Tensor] = None,
 ):
 r"""
 Function invoked when calling the pipeline for generation.

 Args:
 prompt (`str` or `List[str]`, *optional*):
 The prompt or prompts to guide the image generation. If not defined, one has to pass `prompt_embeds`.
 instead.
 negative_prompt (`str` or `List[str]`, *optional*):
 The prompt or prompts not to guide the image generation. If not defined, one has to pass
 `negative_prompt_embeds` instead. Ignored when not using guidance (i.e., ignored if `guidance_scale` is
 not greater than `1`).
 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):
 The width in pixels of the generated image. This is set to 1024 by default for the best results.
 num_inference_steps (`int`, *optional*, defaults to 35):
 The number of denoising steps. More denoising steps usually lead to a higher quality image at the
 expense of slower inference.
 sigmas (`List[float]`, *optional*):
 Custom sigmas to use for the denoising process with schedulers which support a `sigmas` argument in
 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://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.
 strength (`float, *optional*, defaults to 0.9):
 Conceptually, indicates how much to transform the reference image. Must be between 0 and 1. image will
 be used as a starting point, adding more noise to it the larger the strength. The number of denoising
 steps depends on the amount of noise initially added. When strength is 1, added noise will be maximum
 and the denoising process will run for the full number of iterations specified in num_inference_steps.
 A value of 1, therefore, essentially ignores image.
 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*):
 One or a list of [torch generator(s)](https://pytorch.org/docs/stable/generated/torch.Generator.html)
 to make generation deterministic.
 latents (`torch.Tensor`, *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 be generated by sampling using the supplied random `generator`.
 prompt_embeds (`torch.Tensor`, *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.
 ip_adapter_image: (`PipelineImageInput`, *optional*): Optional image input to work with IP Adapters.
 ip_adapter_image_embeds (`List[torch.Tensor]`, *optional*):
 Pre-generated image embeddings for IP-Adapter. It should be a list of length same as number of
 IP-adapters. Each element should be a tensor of shape `(batch_size, num_images, emb_dim)`. If not
 provided, embeddings are computed from the `ip_adapter_image` input argument.
 negative_ip_adapter_image:
 (`PipelineImageInput`, *optional*): Optional image input to work with IP Adapters.
 negative_ip_adapter_image_embeds (`List[torch.Tensor]`, *optional*):
 Pre-generated image embeddings for IP-Adapter. It should be a list of length same as number of
 IP-adapters. Each element should be a tensor of shape `(batch_size, num_images, emb_dim)`. If not
 provided, embeddings are computed from the `ip_adapter_image` input argument.
 negative_prompt_embeds (`torch.Tensor`, *optional*):
 Pre-generated negative text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt
 weighting. If not provided, negative_prompt_embeds will be generated from `negative_prompt` input
 argument.
 prompt_attention_mask (torch.Tensor, *optional*):
 Attention mask for the prompt embeddings. Used to mask out padding tokens in the prompt sequence.
 Chroma requires a single padding token remain unmasked. Please refer to
 https://huggingface.co/lodestones/Chroma#tldr-masking-t5-padding-tokens-enhanced-fidelity-and-increased-stability-during-training
 negative_prompt_attention_mask (torch.Tensor, *optional*):
 Attention mask for the negative prompt embeddings. Used to mask out padding tokens in the negative
 prompt sequence. Chroma requires a single padding token remain unmasked. PLease refer to
 https://huggingface.co/lodestones/Chroma#tldr-masking-t5-padding-tokens-enhanced-fidelity-and-increased-stability-during-training
 output_type (`str`, *optional*, defaults to `"pil"`):
 The output format of the generate image. Choose between
 [PIL](https://pillow.readthedocs.io/en/stable/): `PIL.Image.Image` or `np.array`.
 return_dict (`bool`, *optional*, defaults to `True`):
 Whether or not to return a [`~pipelines.flux.ChromaPipelineOutput`] instead of a plain tuple.
 joint_attention_kwargs (`dict`, *optional*):
 A kwargs dictionary that if specified is passed along to the `AttentionProcessor` as defined under
 `self.processor` in
 [diffusers.models.attention_processor](https://github.com/huggingface/diffusers/blob/main/src/diffusers/models/attention_processor.py).
 callback_on_step_end (`Callable`, *optional*):
 A function that calls at the end of each denoising steps during the inference. The function is called
 with the following arguments: `callback_on_step_end(self: DiffusionPipeline, step: int, timestep: int,
 callback_kwargs: Dict)`. `callback_kwargs` will include a list of all tensors as specified by
 `callback_on_step_end_tensor_inputs`.
 callback_on_step_end_tensor_inputs (`List`, *optional*):
 The list of tensor inputs for the `callback_on_step_end` function. The tensors specified in the list
 will be passed as `callback_kwargs` argument. You will only be able to include variables listed in the
 `._callback_tensor_inputs` attribute of your pipeline class.
 max_sequence_length (`int` defaults to 512): Maximum sequence length to use with the `prompt`.

 Examples:

 Returns:
 [`~pipelines.chroma.ChromaPipelineOutput`] or `tuple`: [`~pipelines.chroma.ChromaPipelineOutput`] if
 `return_dict` is True, otherwise a `tuple`. When returning a tuple, the first element is a list with the
 generated images.
 """

 height = height or self.default_sample_size * self.vae_scale_factor
 width = width or self.default_sample_size * self.vae_scale_factor

 # 1. Check inputs. Raise error if not correct
 self.check_inputs(
 prompt=prompt,
 height=height,
 width=width,
 strength=strength,
 negative_prompt=negative_prompt,
 prompt_embeds=prompt_embeds,
 negative_prompt_embeds=negative_prompt_embeds,
 prompt_attention_mask=prompt_attention_mask,
 negative_prompt_attention_mask=negative_prompt_attention_mask,
 callback_on_step_end_tensor_inputs=callback_on_step_end_tensor_inputs,
 max_sequence_length=max_sequence_length,
 image=image,
 mask_image=mask_image,
 padding_mask_crop=padding_mask_crop,
 )

 self._guidance_scale = guidance_scale
 self._joint_attention_kwargs = joint_attention_kwargs
 self._current_timestep = None
 self._interrupt = False

 # 2. Preprocess mask and image
 if padding_mask_crop is not None:
 crops_coords = self.mask_processor.get_crop_region(mask_image, width, height, pad=padding_mask_crop)
 resize_mode = "fill"
 else:
 crops_coords = None
 resize_mode = "default"

 original_image = image
 init_image = self.image_processor.preprocess(
 image, height=height, width=width, crops_coords=crops_coords, resize_mode=resize_mode
 )
 init_image = init_image.to(dtype=torch.float32)

 # 3. Define call parameters
 if prompt is not None and isinstance(prompt, str):
 batch_size = 1
 elif prompt is not None and isinstance(prompt, list):
 batch_size = len(prompt)
 else:
 batch_size = prompt_embeds.shape[0]

 device = self._execution_device
 lora_scale = (
 self.joint_attention_kwargs.get("scale", None) if self.joint_attention_kwargs is not None else None
 )

 self.vae.to('cpu')
 gc.collect()
 torch.cuda.empty_cache()
 self.text_encoder.to(device)

 (
 prompt_embeds,
 text_ids,
 prompt_attention_mask,
 negative_prompt_embeds,
 negative_text_ids,
 negative_prompt_attention_mask,
 ) = self.encode_prompt(
 prompt=prompt,
 negative_prompt=negative_prompt,
 prompt_embeds=prompt_embeds,
 negative_prompt_embeds=negative_prompt_embeds,
 prompt_attention_mask=prompt_attention_mask,
 negative_prompt_attention_mask=negative_prompt_attention_mask,
 do_classifier_free_guidance=self.do_classifier_free_guidance,
 device=device,
 num_images_per_prompt=num_images_per_prompt,
 max_sequence_length=max_sequence_length,
 lora_scale=lora_scale,
 )

 self.text_encoder.to('cpu')
 gc.collect()
 torch.cuda.empty_cache()
 self.vae.to(device)

 # 4. Prepare timesteps
 sigmas = np.linspace(1.0, 1 / num_inference_steps, num_inference_steps) if sigmas is None else sigmas
 image_seq_len = (int(height) // self.vae_scale_factor // 2) * (int(width) // self.vae_scale_factor // 2)
 mu = calculate_shift(
 image_seq_len,
 self.scheduler.config.get("base_image_seq_len", 256),
 self.scheduler.config.get("max_image_seq_len", 4096),
 self.scheduler.config.get("base_shift", 0.5),
 self.scheduler.config.get("max_shift", 1.15),
 )
 timesteps, num_inference_steps = retrieve_timesteps(
 self.scheduler,
 num_inference_steps,
 device,
 sigmas=sigmas,
 mu=mu,
 )
 timesteps, num_inference_steps = self.get_timesteps(num_inference_steps, strength, device)

 num_warmup_steps = max(len(timesteps) - num_inference_steps * self.scheduler.order, 0)
 self._num_timesteps = len(timesteps)

 if num_inference_steps < 1:
 raise ValueError(
 f"After adjusting the num_inference_steps by strength parameter: {strength}, the number of pipeline"
 f"steps is {num_inference_steps} which is < 1 and not appropriate for this pipeline."
 )
 latent_timestep = timesteps[:1].repeat(batch_size * num_images_per_prompt)

 # 5. Prepare latent variables
 num_channels_latents = self.transformer.config.in_channels // 4
 num_channels_transformer = self.transformer.config.in_channels

 latents, noise, image_latents, latent_image_ids = self.prepare_latents(
 init_image,
 latent_timestep,
 batch_size * num_images_per_prompt,
 num_channels_latents,
 height,
 width,
 prompt_embeds.dtype,
 device,
 generator,
 latents,
 )

 mask_condition = self.mask_processor.preprocess(
 mask_image, height=height, width=width, resize_mode=resize_mode, crops_coords=crops_coords
 )

 if masked_image_latents is None:
 masked_image = init_image * (mask_condition < 0.5)
 else:
 masked_image = masked_image_latents

 mask, masked_image_latents = self.prepare_mask_latents(
 mask_condition,
 masked_image,
 batch_size,
 num_channels_latents,
 num_images_per_prompt,
 height,
 width,
 prompt_embeds.dtype,
 device,
 generator,
 )

 num_warmup_steps = max(len(timesteps) - num_inference_steps * self.scheduler.order, 0)
 self._num_timesteps = len(timesteps)

 # handle guidance
 if self.transformer.config.guidance_embeds:
 guidance = torch.full([1], guidance_scale, device=device, dtype=torch.float32)
 guidance = guidance.expand(latents.shape[0])
 else:
 guidance = None

 if (ip_adapter_image is not None or ip_adapter_image_embeds is not None) and (
 negative_ip_adapter_image is None and negative_ip_adapter_image_embeds is None
 ):
 negative_ip_adapter_image = np.zeros((width, height, 3), dtype=np.uint8)
 elif (ip_adapter_image is None and ip_adapter_image_embeds is None) and (
 negative_ip_adapter_image is not None or negative_ip_adapter_image_embeds is not None
 ):
 ip_adapter_image = np.zeros((width, height, 3), dtype=np.uint8)

 if self.joint_attention_kwargs is None:
 self._joint_attention_kwargs = {}

 image_embeds = None
 negative_image_embeds = None
 if ip_adapter_image is not None or ip_adapter_image_embeds is not None:
 image_embeds = self.prepare_ip_adapter_image_embeds(
 ip_adapter_image,
 ip_adapter_image_embeds,
 device,
 batch_size * num_images_per_prompt,
 )
 if negative_ip_adapter_image is not None or negative_ip_adapter_image_embeds is not None:
 negative_image_embeds = self.prepare_ip_adapter_image_embeds(
 negative_ip_adapter_image,
 negative_ip_adapter_image_embeds,
 device,
 batch_size * num_images_per_prompt,
 )

 self.vae.to('cpu')
 gc.collect()
 torch.cuda.empty_cache()

 attention_mask = self._prepare_attention_mask(
 batch_size=latents.shape[0],
 sequence_length=image_seq_len,
 dtype=latents.dtype,
 attention_mask=prompt_attention_mask,
 )
 negative_attention_mask = self._prepare_attention_mask(
 batch_size=latents.shape[0],
 sequence_length=image_seq_len,
 dtype=latents.dtype,
 attention_mask=negative_prompt_attention_mask,
 )

 # 6. Denoising loop
 with self.progress_bar(total=num_inference_steps) as progress_bar:
 for i, t in enumerate(timesteps):
 if self.interrupt:
 continue

 self._current_timestep = t
 # broadcast to batch dimension in a way that's compatible with ONNX/Core ML
 timestep = t.expand(latents.shape[0])

 if image_embeds is not None:
 self._joint_attention_kwargs["ip_adapter_image_embeds"] = image_embeds

 noise_pred = self.transformer(
 hidden_states=latents,
 timestep=timestep / 1000,
 encoder_hidden_states=prompt_embeds,
 txt_ids=text_ids,
 img_ids=latent_image_ids,
 attention_mask=attention_mask,
 joint_attention_kwargs=self.joint_attention_kwargs,
 return_dict=False,
 )[0]

 if self.do_classifier_free_guidance:
 if negative_image_embeds is not None:
 self._joint_attention_kwargs["ip_adapter_image_embeds"] = negative_image_embeds

 noise_pred_uncond = self.transformer(
 hidden_states=latents,
 timestep=timestep / 1000,
 encoder_hidden_states=negative_prompt_embeds,
 txt_ids=negative_text_ids,
 img_ids=latent_image_ids,
 attention_mask=negative_attention_mask,
 joint_attention_kwargs=self.joint_attention_kwargs,
 return_dict=False,
 )[0]
 noise_pred = noise_pred_uncond + guidance_scale * (noise_pred - noise_pred_uncond)

 # compute the previous noisy sample x_t -> x_t-1
 latents_dtype = latents.dtype
 latents = self.scheduler.step(noise_pred, t, latents, return_dict=False)[0]

 # for 64 channel transformer only.
 init_latents_proper = image_latents
 init_mask = mask

 if i < len(timesteps) - 1:
 noise_timestep = timesteps[i + 1]
 init_latents_proper = self.scheduler.scale_noise(
 init_latents_proper, torch.tensor([noise_timestep]), noise
 )

 latents = (1 - init_mask) * init_latents_proper + init_mask * latents

 if latents.dtype != latents_dtype:
 if torch.backends.mps.is_available():
 # some platforms (eg. apple mps) misbehave due to a pytorch bug: https://github.com/pytorch/pytorch/pull/99272
 latents = latents.to(latents_dtype)

 if callback_on_step_end is not None:
 callback_kwargs = {}
 for k in callback_on_step_end_tensor_inputs:
 callback_kwargs[k] = locals()[k]
 callback_outputs = callback_on_step_end(self, i, t, callback_kwargs)

 latents = callback_outputs.pop("latents", latents)
 prompt_embeds = callback_outputs.pop("prompt_embeds", prompt_embeds)

 # call the callback, if provided
 if i == len(timesteps) - 1 or ((i + 1) > num_warmup_steps and (i + 1) % self.scheduler.order == 0):
 progress_bar.update()

 if XLA_AVAILABLE:
 xm.mark_step()

 self._current_timestep = None

 if output_type == "latent":
 image = latents
 else:
 self.vae.to(device)
 
 latents = self._unpack_latents(latents, height, width, self.vae_scale_factor)
 latents = (latents / self.vae.config.scaling_factor) + self.vae.config.shift_factor
 image = self.vae.decode(latents, return_dict=False)[0]
 image = self.image_processor.postprocess(image, output_type=output_type)

 self.vae.to('cpu')
 gc.collect()
 torch.cuda.empty_cache()

 # Offload all models
 self.maybe_free_model_hooks()

 if not return_dict:
 return (image,)

 return ChromaPipelineOutput(images=image)

[rodjjo]

Maybe @Ednaordinary could help :)

[Ednaordinary]

I can take a look

[rodjjo]

Thanks in advance !

[DN6]

@rodjjo It seems like you have a working pipeline based on the code provided, why not open a PR for it? And could you share some results here?

[rodjjo]

@DN6 I just modified https://github.com/huggingface/diffusers/blob/main/src/diffusers/pipelines/chroma/pipeline_chroma_img2img.py incorporating the inpaiting logic from https://github.com/huggingface/diffusers/blob/main/src/diffusers/pipelines/flux/pipeline_flux_inpaint.py

Sorry for not opening a pull request, I think it would be faster for who worked on it to implement it, that's why I suggested @Ednaordinary as I saw a pull request with that name related with Chroma's pipeline.

I think the inpainting is going to work similar to flux inpaiting pipeline with those limitations we have there too.

It's good for low strenght, to ajdust generated images without changing other parts, as we know the model itself was not designed for inpaiting. It works good for small mdofications and fixes on the generated images, like flux inpaiting pipeline.

[rodjjo]

Model of the test Chroma1-Fast

The original image apple:

Orange Inpaint with strength 75%:

Orange Inpaint with strength 85%:

[NulliferBones]

Is there any way to get this working currently in comfyui?

[Ednaordinary]

Sorry for not opening a pull request, I think it would be faster for who worked on it to implement it, that's why I suggested @Ednaordinary as I saw a pull request with that name related with Chroma's pipeline.

unfortunately I'm rather busy, I just started college this semester. If you have a working version (which it seems you do), I recommend you PR as it's your work. I took a look over the difference between this and the current chroma pipeline and it looks fine to me, but I can take a closer look once a PR is open

[Ednaordinary]

Is there any way to get this working currently in comfyui?

I don't believe this is the correct place for this (do correct me if I'm wrong, DN6 or someone else)

[NulliferBones]

Is there any way to get this working currently in comfyui?

I don't believe this is the correct place for this (do correct me if I'm wrong, DN6 or someone else)

I was referring to the OP as it seems he has it working. So I also assumed he did this within comfy as well. 🤷‍♂️

[rodjjo]

I will open a PR next week.

[hameerabbasi]

@rodjjo Is it okay if I open a PR with you as the author?

[VantomPayne]

Is this still being worked on?

[rodjjo]

Is it okay if I open a PR with you as the author?

I did not have time to work on a PR, feel free to do @hameerabbasi I can be the author, I just create a testing code, I PR involves a lot more things (tests, docs, etc).