Stable unCLIP

Stable unCLIP checkpoints are finetuned from Stable Diffusion 2.1 checkpoints to condition on CLIP image embeddings. Stable unCLIP still conditions on text embeddings. Given the two separate conditionings, stable unCLIP can be used for text guided image variation. When combined with an unCLIP prior, it can also be used for full text to image generation.

The abstract from the paper is:

Contrastive models like CLIP have been shown to learn robust representations of images that capture both semantics and style. To leverage these representations for image generation, we propose a two-stage model: a prior that generates a CLIP image embedding given a text caption, and a decoder that generates an image conditioned on the image embedding. We show that explicitly generating image representations improves image diversity with minimal loss in photorealism and caption similarity. Our decoders conditioned on image representations can also produce variations of an image that preserve both its semantics and style, while varying the non-essential details absent from the image representation. Moreover, the joint embedding space of CLIP enables language-guided image manipulations in a zero-shot fashion. We use diffusion models for the decoder and experiment with both autoregressive and diffusion models for the prior, finding that the latter are computationally more efficient and produce higher-quality samples.

Tips

Stable unCLIP takes noise_level as input during inference which determines how much noise is added to the image embeddings. A higher noise_level increases variation in the final un-noised images. By default, we do not add any additional noise to the image embeddings (noise_level = 0).

Text-to-Image Generation

Stable unCLIP can be leveraged for text-to-image generation by pipelining it with the prior model of KakaoBrain's open source DALL-E 2 replication Karlo:

import mindspore as ms
from mindone.diffusers import UnCLIPScheduler, DDPMScheduler, StableUnCLIPPipeline
from mindone.diffusers.models import PriorTransformer
from mindone.transformers import CLIPTextModelWithProjection
from transformers import CLIPTokenizer

prior_model_id = "kakaobrain/karlo-v1-alpha"
data_type = ms.float16
prior = PriorTransformer.from_pretrained(prior_model_id, subfolder="prior", mindspore_dtype=data_type)

prior_text_model_id = "openai/clip-vit-large-patch14"
prior_tokenizer = CLIPTokenizer.from_pretrained(prior_text_model_id)
prior_text_model = CLIPTextModelWithProjection.from_pretrained(prior_text_model_id, mindspore_dtype=data_type)
prior_scheduler = UnCLIPScheduler.from_pretrained(prior_model_id, subfolder="prior_scheduler")
prior_scheduler = DDPMScheduler.from_config(prior_scheduler.config)

stable_unclip_model_id = "stabilityai/stable-diffusion-2-1-unclip-small"

pipe = StableUnCLIPPipeline.from_pretrained(
    stable_unclip_model_id,
    mindspore_dtype=data_type,
    prior_tokenizer=prior_tokenizer,
    prior_text_encoder=prior_text_model,
    prior=prior,
    prior_scheduler=prior_scheduler,
)

wave_prompt = "dramatic wave, the Oceans roar, Strong wave spiral across the oceans as the waves unfurl into roaring crests; perfect wave form; perfect wave shape; dramatic wave shape; wave shape unbelievable; wave; wave shape spectacular"

image = pipe(prompt=wave_prompt)[0][0]
image

Warning

For text-to-image we use stabilityai/stable-diffusion-2-1-unclip-small as it was trained on CLIP ViT-L/14 embedding, the same as the Karlo model prior. stabilityai/stable-diffusion-2-1-unclip was trained on OpenCLIP ViT-H, so we don't recommend its use.

Text guided Image-to-Image Variation

from mindone.diffusers import StableUnCLIPImg2ImgPipeline
from mindone.diffusers.utils import load_image
import mindspore as ms

pipe = StableUnCLIPImg2ImgPipeline.from_pretrained(
    "stabilityai/stable-diffusion-2-1-unclip", mindspore_dtype=ms.float16, variant="fp16"
)

url = "https://huggingface.co/datasets/hf-internal-testing/diffusers-images/resolve/main/stable_unclip/tarsila_do_amaral.png"
init_image = load_image(url)

images = pipe(init_image)[0]
images[0].save("variation_image.png")

Optionally, you can also pass a prompt to pipe such as:

prompt = "A fantasy landscape, trending on artstation"

image = pipe(init_image, prompt=prompt)[0][0]
image

Tip

Make sure to check out the Schedulers guide to learn how to explore the tradeoff between scheduler speed and quality.

mindone.diffusers.StableUnCLIPPipeline

Bases: DiffusionPipeline, StableDiffusionMixin, TextualInversionLoaderMixin, StableDiffusionLoraLoaderMixin

Pipeline for text-to-image generation using stable unCLIP.

This model inherits from [DiffusionPipeline]. Check the superclass documentation for the generic methods implemented for all pipelines (downloading, saving, running on a particular device, etc.).

The pipeline also inherits the following loading methods

• [~loaders.TextualInversionLoaderMixin.load_textual_inversion] for loading textual inversion embeddings
• [~loaders.StableDiffusionLoraLoaderMixin.load_lora_weights] for loading LoRA weights
• [~loaders.StableDiffusionLoraLoaderMixin.save_lora_weights] for saving LoRA weights

PARAMETER	DESCRIPTION
prior_tokenizer	A [CLIPTokenizer]. TYPE: [`CLIPTokenizer`]
prior_text_encoder	Frozen [CLIPTextModelWithProjection] text-encoder. TYPE: [`CLIPTextModelWithProjection`]
prior	The canonical unCLIP prior to approximate the image embedding from the text embedding. TYPE: [`PriorTransformer`]
prior_scheduler	Scheduler used in the prior denoising process. TYPE: [`KarrasDiffusionSchedulers`]
image_normalizer	Used to normalize the predicted image embeddings before the noise is applied and un-normalize the image embeddings after the noise has been applied. TYPE: [`StableUnCLIPImageNormalizer`]
image_noising_scheduler	Noise schedule for adding noise to the predicted image embeddings. The amount of noise to add is determined by the noise_level. TYPE: [`KarrasDiffusionSchedulers`]
tokenizer	A [CLIPTokenizer]. TYPE: [`CLIPTokenizer`]
text_encoder	Frozen [CLIPTextModel] text-encoder. TYPE: [`CLIPTextModel`]
unet	A [UNet2DConditionModel] to denoise the encoded image latents. TYPE: [`UNet2DConditionModel`]
scheduler	A scheduler to be used in combination with unet to denoise the encoded image latents. TYPE: [`KarrasDiffusionSchedulers`]
vae	Variational Auto-Encoder (VAE) Model to encode and decode images to and from latent representations. TYPE: [`AutoencoderKL`]

Source code in mindone/diffusers/pipelines/stable_diffusion/pipeline_stable_unclip.py

class StableUnCLIPPipeline(
    DiffusionPipeline, StableDiffusionMixin, TextualInversionLoaderMixin, StableDiffusionLoraLoaderMixin
):
    """
    Pipeline for text-to-image generation using stable unCLIP.

    This model inherits from [`DiffusionPipeline`]. Check the superclass documentation for the generic methods
    implemented for all pipelines (downloading, saving, running on a particular device, etc.).

    The pipeline also inherits the following loading methods:
        - [`~loaders.TextualInversionLoaderMixin.load_textual_inversion`] for loading textual inversion embeddings
        - [`~loaders.StableDiffusionLoraLoaderMixin.load_lora_weights`] for loading LoRA weights
        - [`~loaders.StableDiffusionLoraLoaderMixin.save_lora_weights`] for saving LoRA weights

    Args:
        prior_tokenizer ([`CLIPTokenizer`]):
            A [`CLIPTokenizer`].
        prior_text_encoder ([`CLIPTextModelWithProjection`]):
            Frozen [`CLIPTextModelWithProjection`] text-encoder.
        prior ([`PriorTransformer`]):
            The canonical unCLIP prior to approximate the image embedding from the text embedding.
        prior_scheduler ([`KarrasDiffusionSchedulers`]):
            Scheduler used in the prior denoising process.
        image_normalizer ([`StableUnCLIPImageNormalizer`]):
            Used to normalize the predicted image embeddings before the noise is applied and un-normalize the image
            embeddings after the noise has been applied.
        image_noising_scheduler ([`KarrasDiffusionSchedulers`]):
            Noise schedule for adding noise to the predicted image embeddings. The amount of noise to add is determined
            by the `noise_level`.
        tokenizer ([`CLIPTokenizer`]):
            A [`CLIPTokenizer`].
        text_encoder ([`CLIPTextModel`]):
            Frozen [`CLIPTextModel`] text-encoder.
        unet ([`UNet2DConditionModel`]):
            A [`UNet2DConditionModel`] to denoise the encoded image latents.
        scheduler ([`KarrasDiffusionSchedulers`]):
            A scheduler to be used in combination with `unet` to denoise the encoded image latents.
        vae ([`AutoencoderKL`]):
            Variational Auto-Encoder (VAE) Model to encode and decode images to and from latent representations.
    """

    _exclude_from_cpu_offload = ["prior", "image_normalizer"]
    model_cpu_offload_seq = "text_encoder->prior_text_encoder->unet->vae"

    # prior components
    prior_tokenizer: CLIPTokenizer
    prior_text_encoder: CLIPTextModelWithProjection
    prior: PriorTransformer
    prior_scheduler: KarrasDiffusionSchedulers

    # image noising components
    image_normalizer: StableUnCLIPImageNormalizer
    image_noising_scheduler: KarrasDiffusionSchedulers

    # regular denoising components
    tokenizer: CLIPTokenizer
    text_encoder: CLIPTextModel
    unet: UNet2DConditionModel
    scheduler: KarrasDiffusionSchedulers

    vae: AutoencoderKL

    def __init__(
        self,
        # prior components
        prior_tokenizer: CLIPTokenizer,
        prior_text_encoder: CLIPTextModelWithProjection,
        prior: PriorTransformer,
        prior_scheduler: KarrasDiffusionSchedulers,
        # image noising components
        image_normalizer: StableUnCLIPImageNormalizer,
        image_noising_scheduler: KarrasDiffusionSchedulers,
        # regular denoising components
        tokenizer: CLIPTokenizer,
        text_encoder: CLIPTextModelWithProjection,
        unet: UNet2DConditionModel,
        scheduler: KarrasDiffusionSchedulers,
        # vae
        vae: AutoencoderKL,
    ):
        super().__init__()

        self.register_modules(
            prior_tokenizer=prior_tokenizer,
            prior_text_encoder=prior_text_encoder,
            prior=prior,
            prior_scheduler=prior_scheduler,
            image_normalizer=image_normalizer,
            image_noising_scheduler=image_noising_scheduler,
            tokenizer=tokenizer,
            text_encoder=text_encoder,
            unet=unet,
            scheduler=scheduler,
            vae=vae,
        )

        self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1)
        self.image_processor = VaeImageProcessor(vae_scale_factor=self.vae_scale_factor)

    # Copied from diffusers.pipelines.unclip.pipeline_unclip.UnCLIPPipeline._encode_prompt with _encode_prompt->_encode_prior_prompt, tokenizer->prior_tokenizer, text_encoder->prior_text_encoder  # noqa: E501
    def _encode_prior_prompt(
        self,
        prompt,
        num_images_per_prompt,
        do_classifier_free_guidance,
        text_model_output: Optional[Union[CLIPTextModelOutput, Tuple]] = None,
        text_attention_mask: Optional[ms.Tensor] = None,
    ):
        if text_model_output is None:
            batch_size = len(prompt) if isinstance(prompt, list) else 1
            # get prompt text embeddings
            text_inputs = self.prior_tokenizer(
                prompt,
                padding="max_length",
                max_length=self.prior_tokenizer.model_max_length,
                truncation=True,
                return_tensors="np",
            )
            text_input_ids = text_inputs.input_ids
            text_mask = ms.tensor(text_inputs.attention_mask)

            untruncated_ids = self.prior_tokenizer(prompt, padding="longest", return_tensors="np").input_ids

            if untruncated_ids.shape[-1] >= text_input_ids.shape[-1] and not np.array_equal(
                text_input_ids, untruncated_ids
            ):
                removed_text = self.prior_tokenizer.batch_decode(
                    untruncated_ids[:, self.prior_tokenizer.model_max_length - 1 : -1]
                )
                logger.warning(
                    "The following part of your input was truncated because CLIP can only handle sequences up to"
                    f" {self.prior_tokenizer.model_max_length} tokens: {removed_text}"
                )
                text_input_ids = text_input_ids[:, : self.prior_tokenizer.model_max_length]

            prior_text_encoder_output = self.prior_text_encoder(ms.tensor(text_input_ids))

            prompt_embeds = prior_text_encoder_output[0]
            text_enc_hid_states = prior_text_encoder_output[1]

        else:
            batch_size = text_model_output[0].shape[0]
            prompt_embeds, text_enc_hid_states = text_model_output[0], text_model_output[1]
            text_mask = text_attention_mask

        prompt_embeds = prompt_embeds.repeat_interleave(num_images_per_prompt, dim=0)
        text_enc_hid_states = text_enc_hid_states.repeat_interleave(num_images_per_prompt, dim=0)
        text_mask = text_mask.repeat_interleave(num_images_per_prompt, dim=0)

        if do_classifier_free_guidance:
            uncond_tokens = [""] * batch_size

            uncond_input = self.prior_tokenizer(
                uncond_tokens,
                padding="max_length",
                max_length=self.prior_tokenizer.model_max_length,
                truncation=True,
                return_tensors="np",
            )
            uncond_text_mask = ms.tensor(uncond_input.attention_mask)
            negative_prompt_embeds_prior_text_encoder_output = self.prior_text_encoder(
                ms.tensor(uncond_input.input_ids)
            )

            negative_prompt_embeds = negative_prompt_embeds_prior_text_encoder_output[0]
            uncond_text_enc_hid_states = negative_prompt_embeds_prior_text_encoder_output[1]

            # duplicate unconditional embeddings for each generation per prompt, using mps friendly method

            seq_len = negative_prompt_embeds.shape[1]
            negative_prompt_embeds = negative_prompt_embeds.tile((1, num_images_per_prompt))
            negative_prompt_embeds = negative_prompt_embeds.view(batch_size * num_images_per_prompt, seq_len)

            seq_len = uncond_text_enc_hid_states.shape[1]
            uncond_text_enc_hid_states = uncond_text_enc_hid_states.tile((1, num_images_per_prompt, 1))
            uncond_text_enc_hid_states = uncond_text_enc_hid_states.view(
                batch_size * num_images_per_prompt, seq_len, -1
            )
            uncond_text_mask = uncond_text_mask.repeat_interleave(num_images_per_prompt, dim=0)

            # done duplicates

            # For classifier free guidance, we need to do two forward passes.
            # Here we concatenate the unconditional and text embeddings into a single batch
            # to avoid doing two forward passes
            prompt_embeds = ops.cat([negative_prompt_embeds, prompt_embeds])
            text_enc_hid_states = ops.cat([uncond_text_enc_hid_states, text_enc_hid_states])

            text_mask = ops.cat([uncond_text_mask, text_mask])

        return prompt_embeds, text_enc_hid_states, text_mask

    # Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion.StableDiffusionPipeline._encode_prompt
    def _encode_prompt(
        self,
        prompt,
        device,
        num_images_per_prompt,
        do_classifier_free_guidance,
        negative_prompt=None,
        prompt_embeds: Optional[ms.Tensor] = None,
        negative_prompt_embeds: Optional[ms.Tensor] = None,
        lora_scale: Optional[float] = None,
        **kwargs,
    ):
        deprecation_message = "`_encode_prompt()` is deprecated and it will be removed in a future version. Use `encode_prompt()` instead. Also, be aware that the output format changed from a concatenated tensor to a tuple."  # noqa: E501
        deprecate("_encode_prompt()", "1.0.0", deprecation_message, standard_warn=False)

        prompt_embeds_tuple = self.encode_prompt(
            prompt=prompt,
            device=device,
            num_images_per_prompt=num_images_per_prompt,
            do_classifier_free_guidance=do_classifier_free_guidance,
            negative_prompt=negative_prompt,
            prompt_embeds=prompt_embeds,
            negative_prompt_embeds=negative_prompt_embeds,
            lora_scale=lora_scale,
            **kwargs,
        )

        # concatenate for backwards comp
        prompt_embeds = ops.cat([prompt_embeds_tuple[1], prompt_embeds_tuple[0]])

        return prompt_embeds

    # Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion.StableDiffusionPipeline.encode_prompt
    def encode_prompt(
        self,
        prompt,
        num_images_per_prompt,
        do_classifier_free_guidance,
        negative_prompt=None,
        prompt_embeds: Optional[ms.Tensor] = None,
        negative_prompt_embeds: Optional[ms.Tensor] = None,
        lora_scale: Optional[float] = None,
        clip_skip: Optional[int] = None,
    ):
        r"""
        Encodes the prompt into text encoder hidden states.

        Args:
            prompt (`str` or `List[str]`, *optional*):
                prompt to be encoded
            num_images_per_prompt (`int`):
                number of images that should be generated per prompt
            do_classifier_free_guidance (`bool`):
                whether to use classifier free guidance or not
            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
                less than `1`).
            prompt_embeds (`ms.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.
            negative_prompt_embeds (`ms.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.
            lora_scale (`float`, *optional*):
                A LoRA scale that will be applied to all LoRA layers of the text encoder if LoRA layers are loaded.
            clip_skip (`int`, *optional*):
                Number of layers to be skipped from CLIP while computing the prompt embeddings. A value of 1 means that
                the output of the pre-final layer will be used for computing the prompt embeddings.
        """
        # 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, StableDiffusionLoraLoaderMixin):
            self._lora_scale = lora_scale

            # dynamically adjust the LoRA scale
            scale_lora_layers(self.text_encoder, lora_scale)

        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]

        if prompt_embeds is None:
            # textual inversion: process multi-vector tokens if necessary
            if isinstance(self, TextualInversionLoaderMixin):
                prompt = self.maybe_convert_prompt(prompt, self.tokenizer)

            text_inputs = self.tokenizer(
                prompt,
                padding="max_length",
                max_length=self.tokenizer.model_max_length,
                truncation=True,
                return_tensors="np",
            )
            text_input_ids = text_inputs.input_ids
            untruncated_ids = self.tokenizer(prompt, padding="longest", return_tensors="np").input_ids

            if untruncated_ids.shape[-1] >= text_input_ids.shape[-1] and not np.array_equal(
                text_input_ids, untruncated_ids
            ):
                removed_text = self.tokenizer.batch_decode(untruncated_ids[:, self.tokenizer.model_max_length - 1 : -1])
                logger.warning(
                    "The following part of your input was truncated because CLIP can only handle sequences up to"
                    f" {self.tokenizer.model_max_length} tokens: {removed_text}"
                )

            if hasattr(self.text_encoder.config, "use_attention_mask") and self.text_encoder.config.use_attention_mask:
                attention_mask = ms.Tensor(text_inputs.attention_mask)
            else:
                attention_mask = None

            if clip_skip is None:
                prompt_embeds = self.text_encoder(ms.Tensor(text_input_ids), attention_mask=attention_mask)
                prompt_embeds = prompt_embeds[0]
            else:
                prompt_embeds = self.text_encoder(
                    ms.Tensor(text_input_ids), attention_mask=attention_mask, output_hidden_states=True
                )
                # Access the `hidden_states` first, that contains a tuple of
                # all the hidden states from the encoder layers. Then index into
                # the tuple to access the hidden states from the desired layer.
                prompt_embeds = prompt_embeds[-1][-(clip_skip + 1)]
                # We also need to apply the final LayerNorm here to not mess with the
                # representations. The `last_hidden_states` that we typically use for
                # obtaining the final prompt representations passes through the LayerNorm
                # layer.
                prompt_embeds = self.text_encoder.text_model.final_layer_norm(prompt_embeds)

        if self.text_encoder is not None:
            prompt_embeds_dtype = self.text_encoder.dtype
        elif self.unet is not None:
            prompt_embeds_dtype = self.unet.dtype
        else:
            prompt_embeds_dtype = prompt_embeds.dtype

        prompt_embeds = prompt_embeds.to(dtype=prompt_embeds_dtype)

        bs_embed, seq_len, _ = prompt_embeds.shape
        # duplicate text embeddings for each generation per prompt, using mps friendly method
        prompt_embeds = prompt_embeds.tile((1, num_images_per_prompt, 1))
        prompt_embeds = prompt_embeds.view(bs_embed * num_images_per_prompt, seq_len, -1)

        # get unconditional embeddings for classifier free guidance
        if do_classifier_free_guidance and negative_prompt_embeds is None:
            uncond_tokens: List[str]
            if negative_prompt is None:
                uncond_tokens = [""] * batch_size
            elif 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 isinstance(negative_prompt, str):
                uncond_tokens = [negative_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`."
                )
            else:
                uncond_tokens = negative_prompt

            # textual inversion: process multi-vector tokens if necessary
            if isinstance(self, TextualInversionLoaderMixin):
                uncond_tokens = self.maybe_convert_prompt(uncond_tokens, self.tokenizer)

            max_length = prompt_embeds.shape[1]
            uncond_input = self.tokenizer(
                uncond_tokens,
                padding="max_length",
                max_length=max_length,
                truncation=True,
                return_tensors="np",
            )

            if hasattr(self.text_encoder.config, "use_attention_mask") and self.text_encoder.config.use_attention_mask:
                attention_mask = ms.Tensor(uncond_input.attention_mask)
            else:
                attention_mask = None

            negative_prompt_embeds = self.text_encoder(
                ms.Tensor(uncond_input.input_ids),
                attention_mask=attention_mask,
            )
            negative_prompt_embeds = negative_prompt_embeds[0]

        if do_classifier_free_guidance:
            # duplicate unconditional embeddings for each generation per prompt, using mps friendly method
            seq_len = negative_prompt_embeds.shape[1]

            negative_prompt_embeds = negative_prompt_embeds.to(dtype=prompt_embeds_dtype)

            negative_prompt_embeds = negative_prompt_embeds.tile((1, num_images_per_prompt, 1))
            negative_prompt_embeds = negative_prompt_embeds.view(batch_size * num_images_per_prompt, seq_len, -1)

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

        return prompt_embeds, negative_prompt_embeds

    # Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion.StableDiffusionPipeline.decode_latents
    def decode_latents(self, latents):
        deprecation_message = "The decode_latents method is deprecated and will be removed in 1.0.0. Please use VaeImageProcessor.postprocess(...) instead"
        deprecate("decode_latents", "1.0.0", deprecation_message, standard_warn=False)

        latents = 1 / self.vae.config.scaling_factor * latents
        image = self.vae.decode(latents, return_dict=False)[0]
        image = (image / 2 + 0.5).clamp(0, 1)
        # we always cast to float32 as this does not cause significant overhead and is compatible with bfloat16
        image = image.permute(0, 2, 3, 1).float().numpy()
        return image

    # Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion.StableDiffusionPipeline.prepare_extra_step_kwargs with prepare_extra_step_kwargs->prepare_prior_extra_step_kwargs, scheduler->prior_scheduler  # noqa: E501
    def prepare_prior_extra_step_kwargs(self, generator, eta):
        # prepare extra kwargs for the prior_scheduler step, since not all prior_schedulers have the same signature
        # eta (η) is only used with the DDIMScheduler, it will be ignored for other prior_schedulers.
        # eta corresponds to η in DDIM paper: https://arxiv.org/abs/2010.02502
        # and should be between [0, 1]

        accepts_eta = "eta" in set(inspect.signature(self.prior_scheduler.step).parameters.keys())
        extra_step_kwargs = {}
        if accepts_eta:
            extra_step_kwargs["eta"] = eta

        # check if the prior_scheduler accepts generator
        accepts_generator = "generator" in set(inspect.signature(self.prior_scheduler.step).parameters.keys())
        if accepts_generator:
            extra_step_kwargs["generator"] = generator
        return extra_step_kwargs

    # Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion.StableDiffusionPipeline.prepare_extra_step_kwargs
    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
        # and should be between [0, 1]

        accepts_eta = "eta" in set(inspect.signature(self.scheduler.step).parameters.keys())
        extra_step_kwargs = {}
        if accepts_eta:
            extra_step_kwargs["eta"] = eta

        # check if the scheduler accepts generator
        accepts_generator = "generator" in set(inspect.signature(self.scheduler.step).parameters.keys())
        if accepts_generator:
            extra_step_kwargs["generator"] = generator
        return extra_step_kwargs

    def check_inputs(
        self,
        prompt,
        height,
        width,
        callback_steps,
        noise_level,
        negative_prompt=None,
        prompt_embeds=None,
        negative_prompt_embeds=None,
    ):
        if height % 8 != 0 or width % 8 != 0:
            raise ValueError(f"`height` and `width` have to be divisible by 8 but are {height} and {width}.")

        if (callback_steps is None) or (
            callback_steps is not None and (not isinstance(callback_steps, int) or callback_steps <= 0)
        ):
            raise ValueError(
                f"`callback_steps` has to be a positive integer but is {callback_steps} of type"
                f" {type(callback_steps)}."
            )

        if prompt is not None and prompt_embeds is not None:
            raise ValueError(
                "Provide either `prompt` or `prompt_embeds`. Please make sure to define only one of the two."
            )

        if prompt is None and prompt_embeds is None:
            raise ValueError(
                "Provide either `prompt` or `prompt_embeds`. Cannot leave both `prompt` and `prompt_embeds` undefined."
            )

        if 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(
                "Provide either `negative_prompt` or `negative_prompt_embeds`. Cannot leave both `negative_prompt` and `negative_prompt_embeds` undefined."
            )

        if prompt is not None and negative_prompt is not None:
            if 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)}."
                )

        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 noise_level < 0 or noise_level >= self.image_noising_scheduler.config.num_train_timesteps:
            raise ValueError(
                f"`noise_level` must be between 0 and {self.image_noising_scheduler.config.num_train_timesteps - 1}, inclusive."
            )

    # Copied from diffusers.pipelines.unclip.pipeline_unclip.UnCLIPPipeline.prepare_latents
    def prepare_latents(self, shape, dtype, generator, latents, scheduler):
        if latents is None:
            latents = randn_tensor(shape, generator=generator, dtype=dtype)
        else:
            if latents.shape != shape:
                raise ValueError(f"Unexpected latents shape, got {latents.shape}, expected {shape}")

        latents = (latents * scheduler.init_noise_sigma).to(dtype)
        return latents

    def noise_image_embeddings(
        self,
        image_embeds: ms.Tensor,
        noise_level: int,
        noise: Optional[ms.Tensor] = None,
        generator: Optional[np.random.Generator] = None,
    ):
        """
        Add noise to the image embeddings. The amount of noise is controlled by a `noise_level` input. A higher
        `noise_level` increases the variance in the final un-noised images.

        The noise is applied in two ways:
        1. A noise schedule is applied directly to the embeddings.
        2. A vector of sinusoidal time embeddings are appended to the output.

        In both cases, the amount of noise is controlled by the same `noise_level`.

        The embeddings are normalized before the noise is applied and un-normalized after the noise is applied.
        """
        if noise is None:
            noise = randn_tensor(image_embeds.shape, generator=generator, dtype=image_embeds.dtype)

        noise_level = ms.tensor([noise_level] * image_embeds.shape[0])

        image_embeds = self.image_normalizer.scale(image_embeds)

        image_embeds = self.image_noising_scheduler.add_noise(image_embeds, timesteps=noise_level, noise=noise)

        image_embeds = self.image_normalizer.unscale(image_embeds)

        noise_level = get_timestep_embedding(
            timesteps=noise_level, embedding_dim=image_embeds.shape[-1], flip_sin_to_cos=True, downscale_freq_shift=0
        )

        # `get_timestep_embeddings` does not contain any weights and will always return f32 tensors,
        # but we might actually be running in fp16. so we need to cast here.
        # there might be better ways to encapsulate this.
        noise_level = noise_level.to(image_embeds.dtype)

        image_embeds = ops.cat((image_embeds, noise_level), 1)

        return image_embeds

    def __call__(
        self,
        # regular denoising process args
        prompt: Optional[Union[str, List[str]]] = None,
        height: Optional[int] = None,
        width: Optional[int] = None,
        num_inference_steps: int = 20,
        guidance_scale: float = 10.0,
        negative_prompt: Optional[Union[str, List[str]]] = None,
        num_images_per_prompt: Optional[int] = 1,
        eta: float = 0.0,
        generator: Optional[np.random.Generator] = None,
        latents: Optional[ms.Tensor] = None,
        prompt_embeds: Optional[ms.Tensor] = None,
        negative_prompt_embeds: Optional[ms.Tensor] = None,
        output_type: Optional[str] = "pil",
        return_dict: bool = False,
        callback: Optional[Callable[[int, int, ms.Tensor], None]] = None,
        callback_steps: int = 1,
        cross_attention_kwargs: Optional[Dict[str, Any]] = None,
        noise_level: int = 0,
        # prior args
        prior_num_inference_steps: int = 25,
        prior_guidance_scale: float = 4.0,
        prior_latents: Optional[ms.Tensor] = None,
        clip_skip: Optional[int] = None,
    ):
        """
        The call function to the pipeline for generation.

        Args:
            prompt (`str` or `List[str]`, *optional*):
                The prompt or prompts to guide image generation. If not defined, you need to pass `prompt_embeds`.
            height (`int`, *optional*, defaults to `self.unet.config.sample_size * self.vae_scale_factor`):
                The height in pixels of the generated image.
            width (`int`, *optional*, defaults to `self.unet.config.sample_size * self.vae_scale_factor`):
                The width in pixels of the generated image.
            num_inference_steps (`int`, *optional*, defaults to 20):
                The number of denoising steps. More denoising steps usually lead to a higher quality image at the
                expense of slower inference.
            guidance_scale (`float`, *optional*, defaults to 10.0):
                A higher guidance scale value encourages the model to generate images closely linked to the text
                `prompt` at the expense of lower image quality. Guidance scale is enabled when `guidance_scale > 1`.
            negative_prompt (`str` or `List[str]`, *optional*):
                The prompt or prompts to guide what to not include in image generation. If not defined, you need to
                pass `negative_prompt_embeds` instead. Ignored when not using guidance (`guidance_scale < 1`).
            num_images_per_prompt (`int`, *optional*, defaults to 1):
                The number of images to generate per prompt.
            eta (`float`, *optional*, defaults to 0.0):
                Corresponds to parameter eta (η) from the [DDIM](https://arxiv.org/abs/2010.02502) paper. Only applies
                to the [`~schedulers.DDIMScheduler`], and is ignored in other schedulers.
            generator (`np.random.Generator` or `List[np.random.Generator]`, *optional*):
                A [`np.random.Generator`](https://numpy.org/doc/stable/reference/random/generator.html) to make
                generation deterministic.
            latents (`ms.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 is generated by sampling using the supplied random `generator`.
            prompt_embeds (`ms.Tensor`, *optional*):
                Pre-generated text embeddings. Can be used to easily tweak text inputs (prompt weighting). If not
                provided, text embeddings are generated from the `prompt` input argument.
            negative_prompt_embeds (`ms.Tensor`, *optional*):
                Pre-generated negative text embeddings. Can be used to easily tweak text inputs (prompt weighting). If
                not provided, `negative_prompt_embeds` are generated from the `negative_prompt` input argument.
            output_type (`str`, *optional*, defaults to `"pil"`):
                The output format of the generated image. Choose between `PIL.Image` or `np.array`.
            return_dict (`bool`, *optional*, defaults to `False`):
                Whether or not to return a [`~pipelines.ImagePipelineOutput`] instead of a plain tuple.
            callback (`Callable`, *optional*):
                A function that calls every `callback_steps` steps during inference. The function is called with the
                following arguments: `callback(step: int, timestep: int, latents: ms.Tensor)`.
            callback_steps (`int`, *optional*, defaults to 1):
                The frequency at which the `callback` function is called. If not specified, the callback is called at
                every step.
            cross_attention_kwargs (`dict`, *optional*):
                A kwargs dictionary that if specified is passed along to the [`AttentionProcessor`] as defined in
                [`self.processor`](https://github.com/huggingface/diffusers/blob/main/src/diffusers/models/attention_processor.py).
            noise_level (`int`, *optional*, defaults to `0`):
                The amount of noise to add to the image embeddings. A higher `noise_level` increases the variance in
                the final un-noised images. See [`StableUnCLIPPipeline.noise_image_embeddings`] for more details.
            prior_num_inference_steps (`int`, *optional*, defaults to 25):
                The number of denoising steps in the prior denoising process. More denoising steps usually lead to a
                higher quality image at the expense of slower inference.
            prior_guidance_scale (`float`, *optional*, defaults to 4.0):
                A higher guidance scale value encourages the model to generate images closely linked to the text
                `prompt` at the expense of lower image quality. Guidance scale is enabled when `guidance_scale > 1`.
            prior_latents (`ms.Tensor`, *optional*):
                Pre-generated noisy latents sampled from a Gaussian distribution, to be used as inputs for image
                embedding generation in the prior denoising process. Can be used to tweak the same generation with
                different prompts. If not provided, a latents tensor is generated by sampling using the supplied random
                `generator`.
            clip_skip (`int`, *optional*):
                Number of layers to be skipped from CLIP while computing the prompt embeddings. A value of 1 means that
                the output of the pre-final layer will be used for computing the prompt embeddings.
        Examples:

        Returns:
            [`~pipelines.ImagePipelineOutput`] or `tuple`:
                [`~ pipeline_utils.ImagePipelineOutput`] if `return_dict` is True, otherwise a `tuple`. When returning
                a tuple, the first element is a list with the generated images.
        """
        # 0. Default height and width to unet
        height = height or self.unet.config.sample_size * self.vae_scale_factor
        width = width or self.unet.config.sample_size * self.vae_scale_factor

        # 1. Check inputs. Raise error if not correct
        self.check_inputs(
            prompt=prompt,
            height=height,
            width=width,
            callback_steps=callback_steps,
            noise_level=noise_level,
            negative_prompt=negative_prompt,
            prompt_embeds=prompt_embeds,
            negative_prompt_embeds=negative_prompt_embeds,
        )

        # 2. 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]

        batch_size = batch_size * num_images_per_prompt

        # 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`
        # corresponds to doing no classifier free guidance.
        prior_do_classifier_free_guidance = prior_guidance_scale > 1.0

        # 3. Encode input prompt
        prior_prompt_embeds, prior_text_encoder_hidden_states, prior_text_mask = self._encode_prior_prompt(
            prompt=prompt,
            num_images_per_prompt=num_images_per_prompt,
            do_classifier_free_guidance=prior_do_classifier_free_guidance,
        )

        # 4. Prepare prior timesteps
        self.prior_scheduler.set_timesteps(prior_num_inference_steps)
        prior_timesteps_tensor = self.prior_scheduler.timesteps

        # 5. Prepare prior latent variables
        embedding_dim = self.prior.config.embedding_dim
        prior_latents = self.prepare_latents(
            (batch_size, embedding_dim),
            prior_prompt_embeds.dtype,
            generator,
            prior_latents,
            self.prior_scheduler,
        )

        # 6. Prepare extra step kwargs. TODO: Logic should ideally just be moved out of the pipeline
        prior_extra_step_kwargs = self.prepare_prior_extra_step_kwargs(generator, eta)

        # 7. Prior denoising loop
        for i, t in enumerate(self.progress_bar(prior_timesteps_tensor)):
            # expand the latents if we are doing classifier free guidance
            latent_model_input = ops.cat([prior_latents] * 2) if prior_do_classifier_free_guidance else prior_latents
            latent_model_input = self.prior_scheduler.scale_model_input(latent_model_input, t)

            predicted_image_embedding = self.prior(
                latent_model_input,
                timestep=t,
                proj_embedding=prior_prompt_embeds,
                encoder_hidden_states=prior_text_encoder_hidden_states,
                attention_mask=prior_text_mask,
            )[0]

            if prior_do_classifier_free_guidance:
                predicted_image_embedding_uncond, predicted_image_embedding_text = predicted_image_embedding.chunk(2)
                predicted_image_embedding = predicted_image_embedding_uncond + prior_guidance_scale * (
                    predicted_image_embedding_text - predicted_image_embedding_uncond
                )

            prior_latents = self.prior_scheduler.step(
                predicted_image_embedding,
                timestep=t,
                sample=prior_latents,
                **prior_extra_step_kwargs,
                return_dict=False,
            )[0]

            if callback is not None and i % callback_steps == 0:
                callback(i, t, prior_latents)

        prior_latents = self.prior.post_process_latents(prior_latents)

        image_embeds = prior_latents

        # done prior

        # 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`
        # corresponds to doing no classifier free guidance.
        do_classifier_free_guidance = guidance_scale > 1.0

        # 8. Encode input prompt
        text_encoder_lora_scale = (
            cross_attention_kwargs.get("scale", None) if cross_attention_kwargs is not None else None
        )
        prompt_embeds, negative_prompt_embeds = self.encode_prompt(
            prompt=prompt,
            num_images_per_prompt=num_images_per_prompt,
            do_classifier_free_guidance=do_classifier_free_guidance,
            negative_prompt=negative_prompt,
            prompt_embeds=prompt_embeds,
            negative_prompt_embeds=negative_prompt_embeds,
            lora_scale=text_encoder_lora_scale,
            clip_skip=clip_skip,
        )
        # For classifier free guidance, we need to do two forward passes.
        # Here we concatenate the unconditional and text embeddings into a single batch
        # to avoid doing two forward passes
        if do_classifier_free_guidance:
            prompt_embeds = ops.cat([negative_prompt_embeds, prompt_embeds])

        # 9. Prepare image embeddings
        image_embeds = self.noise_image_embeddings(
            image_embeds=image_embeds,
            noise_level=noise_level,
            generator=generator,
        )

        if do_classifier_free_guidance:
            negative_prompt_embeds = ops.zeros_like(image_embeds)

            # For classifier free guidance, we need to do two forward passes.
            # Here we concatenate the unconditional and text embeddings into a single batch
            # to avoid doing two forward passes
            image_embeds = ops.cat([negative_prompt_embeds, image_embeds])

        # 10. Prepare timesteps
        self.scheduler.set_timesteps(num_inference_steps)
        timesteps = self.scheduler.timesteps

        # 11. Prepare latent variables
        num_channels_latents = self.unet.config.in_channels
        shape = (
            batch_size,
            num_channels_latents,
            int(height) // self.vae_scale_factor,
            int(width) // self.vae_scale_factor,
        )
        latents = self.prepare_latents(
            shape=shape,
            dtype=prompt_embeds.dtype,
            generator=generator,
            latents=latents,
            scheduler=self.scheduler,
        )

        # 12. Prepare extra step kwargs. TODO: Logic should ideally just be moved out of the pipeline
        extra_step_kwargs = self.prepare_extra_step_kwargs(generator, eta)

        # 13. Denoising loop
        for i, t in enumerate(self.progress_bar(timesteps)):
            latent_model_input = ops.cat([latents] * 2) if do_classifier_free_guidance else latents
            latent_model_input = self.scheduler.scale_model_input(latent_model_input, t)

            # predict the noise residual
            noise_pred = self.unet(
                latent_model_input,
                t,
                encoder_hidden_states=prompt_embeds,
                class_labels=image_embeds,
                cross_attention_kwargs=cross_attention_kwargs,
                return_dict=False,
            )[0]

            # perform guidance
            if do_classifier_free_guidance:
                noise_pred_uncond, noise_pred_text = noise_pred.chunk(2)
                noise_pred = noise_pred_uncond + guidance_scale * (noise_pred_text - noise_pred_uncond)

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

            if callback is not None and i % callback_steps == 0:
                step_idx = i // getattr(self.scheduler, "order", 1)
                callback(step_idx, t, latents)

        if not output_type == "latent":
            image = self.vae.decode(latents / self.vae.config.scaling_factor, return_dict=False)[0]
        else:
            image = latents

        image = self.image_processor.postprocess(image, output_type=output_type)

        if not return_dict:
            return (image,)

        return ImagePipelineOutput(images=image)

mindone.diffusers.StableUnCLIPPipeline.__call__(prompt=None, height=None, width=None, num_inference_steps=20, guidance_scale=10.0, negative_prompt=None, num_images_per_prompt=1, eta=0.0, generator=None, latents=None, prompt_embeds=None, negative_prompt_embeds=None, output_type='pil', return_dict=False, callback=None, callback_steps=1, cross_attention_kwargs=None, noise_level=0, prior_num_inference_steps=25, prior_guidance_scale=4.0, prior_latents=None, clip_skip=None)

The call function to the pipeline for generation.

PARAMETER	DESCRIPTION
prompt	The prompt or prompts to guide image generation. If not defined, you need to pass prompt_embeds. TYPE: `str` or `List[str]`, *optional* DEFAULT: None
height	The height in pixels of the generated image. TYPE: `int`, *optional*, defaults to `self.unet.config.sample_size * self.vae_scale_factor` DEFAULT: None
width	The width in pixels of the generated image. TYPE: `int`, *optional*, defaults to `self.unet.config.sample_size * self.vae_scale_factor` DEFAULT: None
num_inference_steps	The number of denoising steps. More denoising steps usually lead to a higher quality image at the expense of slower inference. TYPE: `int`, *optional*, defaults to 20 DEFAULT: 20
guidance_scale	A higher guidance scale value encourages the model to generate images closely linked to the text prompt at the expense of lower image quality. Guidance scale is enabled when guidance_scale > 1. TYPE: `float`, *optional*, defaults to 10.0 DEFAULT: 10.0
negative_prompt	The prompt or prompts to guide what to not include in image generation. If not defined, you need to pass negative_prompt_embeds instead. Ignored when not using guidance (guidance_scale < 1). TYPE: `str` or `List[str]`, *optional* DEFAULT: None
num_images_per_prompt	The number of images to generate per prompt. TYPE: `int`, *optional*, defaults to 1 DEFAULT: 1
eta	Corresponds to parameter eta (η) from the DDIM paper. Only applies to the [~schedulers.DDIMScheduler], and is ignored in other schedulers. TYPE: `float`, *optional*, defaults to 0.0 DEFAULT: 0.0
generator	A np.random.Generator to make generation deterministic. TYPE: `np.random.Generator` or `List[np.random.Generator]`, *optional* DEFAULT: None
latents	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 is generated by sampling using the supplied random generator. TYPE: `ms.Tensor`, *optional* DEFAULT: None
prompt_embeds	Pre-generated text embeddings. Can be used to easily tweak text inputs (prompt weighting). If not provided, text embeddings are generated from the prompt input argument. TYPE: `ms.Tensor`, *optional* DEFAULT: None
negative_prompt_embeds	Pre-generated negative text embeddings. Can be used to easily tweak text inputs (prompt weighting). If not provided, negative_prompt_embeds are generated from the negative_prompt input argument. TYPE: `ms.Tensor`, *optional* DEFAULT: None
output_type	The output format of the generated image. Choose between PIL.Image or np.array. TYPE: `str`, *optional*, defaults to `"pil"` DEFAULT: 'pil'
return_dict	Whether or not to return a [~pipelines.ImagePipelineOutput] instead of a plain tuple. TYPE: `bool`, *optional*, defaults to `False` DEFAULT: False
callback	A function that calls every callback_steps steps during inference. The function is called with the following arguments: callback(step: int, timestep: int, latents: ms.Tensor). TYPE: `Callable`, *optional* DEFAULT: None
callback_steps	The frequency at which the callback function is called. If not specified, the callback is called at every step. TYPE: `int`, *optional*, defaults to 1 DEFAULT: 1
cross_attention_kwargs	A kwargs dictionary that if specified is passed along to the [AttentionProcessor] as defined in self.processor. TYPE: `dict`, *optional* DEFAULT: None
noise_level	The amount of noise to add to the image embeddings. A higher noise_level increases the variance in the final un-noised images. See [StableUnCLIPPipeline.noise_image_embeddings] for more details. TYPE: `int`, *optional*, defaults to `0` DEFAULT: 0
prior_num_inference_steps	The number of denoising steps in the prior denoising process. More denoising steps usually lead to a higher quality image at the expense of slower inference. TYPE: `int`, *optional*, defaults to 25 DEFAULT: 25
prior_guidance_scale	A higher guidance scale value encourages the model to generate images closely linked to the text prompt at the expense of lower image quality. Guidance scale is enabled when guidance_scale > 1. TYPE: `float`, *optional*, defaults to 4.0 DEFAULT: 4.0
prior_latents	Pre-generated noisy latents sampled from a Gaussian distribution, to be used as inputs for image embedding generation in the prior denoising process. Can be used to tweak the same generation with different prompts. If not provided, a latents tensor is generated by sampling using the supplied random generator. TYPE: `ms.Tensor`, *optional* DEFAULT: None
clip_skip	Number of layers to be skipped from CLIP while computing the prompt embeddings. A value of 1 means that the output of the pre-final layer will be used for computing the prompt embeddings. TYPE: `int`, *optional* DEFAULT: None

RETURNS	DESCRIPTION
	[~pipelines.ImagePipelineOutput] or tuple: [~ pipeline_utils.ImagePipelineOutput] if return_dict is True, otherwise a tuple. When returning a tuple, the first element is a list with the generated images.

Source code in mindone/diffusers/pipelines/stable_diffusion/pipeline_stable_unclip.py

def __call__(
    self,
    # regular denoising process args
    prompt: Optional[Union[str, List[str]]] = None,
    height: Optional[int] = None,
    width: Optional[int] = None,
    num_inference_steps: int = 20,
    guidance_scale: float = 10.0,
    negative_prompt: Optional[Union[str, List[str]]] = None,
    num_images_per_prompt: Optional[int] = 1,
    eta: float = 0.0,
    generator: Optional[np.random.Generator] = None,
    latents: Optional[ms.Tensor] = None,
    prompt_embeds: Optional[ms.Tensor] = None,
    negative_prompt_embeds: Optional[ms.Tensor] = None,
    output_type: Optional[str] = "pil",
    return_dict: bool = False,
    callback: Optional[Callable[[int, int, ms.Tensor], None]] = None,
    callback_steps: int = 1,
    cross_attention_kwargs: Optional[Dict[str, Any]] = None,
    noise_level: int = 0,
    # prior args
    prior_num_inference_steps: int = 25,
    prior_guidance_scale: float = 4.0,
    prior_latents: Optional[ms.Tensor] = None,
    clip_skip: Optional[int] = None,
):
    """
    The call function to the pipeline for generation.

    Args:
        prompt (`str` or `List[str]`, *optional*):
            The prompt or prompts to guide image generation. If not defined, you need to pass `prompt_embeds`.
        height (`int`, *optional*, defaults to `self.unet.config.sample_size * self.vae_scale_factor`):
            The height in pixels of the generated image.
        width (`int`, *optional*, defaults to `self.unet.config.sample_size * self.vae_scale_factor`):
            The width in pixels of the generated image.
        num_inference_steps (`int`, *optional*, defaults to 20):
            The number of denoising steps. More denoising steps usually lead to a higher quality image at the
            expense of slower inference.
        guidance_scale (`float`, *optional*, defaults to 10.0):
            A higher guidance scale value encourages the model to generate images closely linked to the text
            `prompt` at the expense of lower image quality. Guidance scale is enabled when `guidance_scale > 1`.
        negative_prompt (`str` or `List[str]`, *optional*):
            The prompt or prompts to guide what to not include in image generation. If not defined, you need to
            pass `negative_prompt_embeds` instead. Ignored when not using guidance (`guidance_scale < 1`).
        num_images_per_prompt (`int`, *optional*, defaults to 1):
            The number of images to generate per prompt.
        eta (`float`, *optional*, defaults to 0.0):
            Corresponds to parameter eta (η) from the [DDIM](https://arxiv.org/abs/2010.02502) paper. Only applies
            to the [`~schedulers.DDIMScheduler`], and is ignored in other schedulers.
        generator (`np.random.Generator` or `List[np.random.Generator]`, *optional*):
            A [`np.random.Generator`](https://numpy.org/doc/stable/reference/random/generator.html) to make
            generation deterministic.
        latents (`ms.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 is generated by sampling using the supplied random `generator`.
        prompt_embeds (`ms.Tensor`, *optional*):
            Pre-generated text embeddings. Can be used to easily tweak text inputs (prompt weighting). If not
            provided, text embeddings are generated from the `prompt` input argument.
        negative_prompt_embeds (`ms.Tensor`, *optional*):
            Pre-generated negative text embeddings. Can be used to easily tweak text inputs (prompt weighting). If
            not provided, `negative_prompt_embeds` are generated from the `negative_prompt` input argument.
        output_type (`str`, *optional*, defaults to `"pil"`):
            The output format of the generated image. Choose between `PIL.Image` or `np.array`.
        return_dict (`bool`, *optional*, defaults to `False`):
            Whether or not to return a [`~pipelines.ImagePipelineOutput`] instead of a plain tuple.
        callback (`Callable`, *optional*):
            A function that calls every `callback_steps` steps during inference. The function is called with the
            following arguments: `callback(step: int, timestep: int, latents: ms.Tensor)`.
        callback_steps (`int`, *optional*, defaults to 1):
            The frequency at which the `callback` function is called. If not specified, the callback is called at
            every step.
        cross_attention_kwargs (`dict`, *optional*):
            A kwargs dictionary that if specified is passed along to the [`AttentionProcessor`] as defined in
            [`self.processor`](https://github.com/huggingface/diffusers/blob/main/src/diffusers/models/attention_processor.py).
        noise_level (`int`, *optional*, defaults to `0`):
            The amount of noise to add to the image embeddings. A higher `noise_level` increases the variance in
            the final un-noised images. See [`StableUnCLIPPipeline.noise_image_embeddings`] for more details.
        prior_num_inference_steps (`int`, *optional*, defaults to 25):
            The number of denoising steps in the prior denoising process. More denoising steps usually lead to a
            higher quality image at the expense of slower inference.
        prior_guidance_scale (`float`, *optional*, defaults to 4.0):
            A higher guidance scale value encourages the model to generate images closely linked to the text
            `prompt` at the expense of lower image quality. Guidance scale is enabled when `guidance_scale > 1`.
        prior_latents (`ms.Tensor`, *optional*):
            Pre-generated noisy latents sampled from a Gaussian distribution, to be used as inputs for image
            embedding generation in the prior denoising process. Can be used to tweak the same generation with
            different prompts. If not provided, a latents tensor is generated by sampling using the supplied random
            `generator`.
        clip_skip (`int`, *optional*):
            Number of layers to be skipped from CLIP while computing the prompt embeddings. A value of 1 means that
            the output of the pre-final layer will be used for computing the prompt embeddings.
    Examples:

    Returns:
        [`~pipelines.ImagePipelineOutput`] or `tuple`:
            [`~ pipeline_utils.ImagePipelineOutput`] if `return_dict` is True, otherwise a `tuple`. When returning
            a tuple, the first element is a list with the generated images.
    """
    # 0. Default height and width to unet
    height = height or self.unet.config.sample_size * self.vae_scale_factor
    width = width or self.unet.config.sample_size * self.vae_scale_factor

    # 1. Check inputs. Raise error if not correct
    self.check_inputs(
        prompt=prompt,
        height=height,
        width=width,
        callback_steps=callback_steps,
        noise_level=noise_level,
        negative_prompt=negative_prompt,
        prompt_embeds=prompt_embeds,
        negative_prompt_embeds=negative_prompt_embeds,
    )

    # 2. 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]

    batch_size = batch_size * num_images_per_prompt

    # 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`
    # corresponds to doing no classifier free guidance.
    prior_do_classifier_free_guidance = prior_guidance_scale > 1.0

    # 3. Encode input prompt
    prior_prompt_embeds, prior_text_encoder_hidden_states, prior_text_mask = self._encode_prior_prompt(
        prompt=prompt,
        num_images_per_prompt=num_images_per_prompt,
        do_classifier_free_guidance=prior_do_classifier_free_guidance,
    )

    # 4. Prepare prior timesteps
    self.prior_scheduler.set_timesteps(prior_num_inference_steps)
    prior_timesteps_tensor = self.prior_scheduler.timesteps

    # 5. Prepare prior latent variables
    embedding_dim = self.prior.config.embedding_dim
    prior_latents = self.prepare_latents(
        (batch_size, embedding_dim),
        prior_prompt_embeds.dtype,
        generator,
        prior_latents,
        self.prior_scheduler,
    )

    # 6. Prepare extra step kwargs. TODO: Logic should ideally just be moved out of the pipeline
    prior_extra_step_kwargs = self.prepare_prior_extra_step_kwargs(generator, eta)

    # 7. Prior denoising loop
    for i, t in enumerate(self.progress_bar(prior_timesteps_tensor)):
        # expand the latents if we are doing classifier free guidance
        latent_model_input = ops.cat([prior_latents] * 2) if prior_do_classifier_free_guidance else prior_latents
        latent_model_input = self.prior_scheduler.scale_model_input(latent_model_input, t)

        predicted_image_embedding = self.prior(
            latent_model_input,
            timestep=t,
            proj_embedding=prior_prompt_embeds,
            encoder_hidden_states=prior_text_encoder_hidden_states,
            attention_mask=prior_text_mask,
        )[0]

        if prior_do_classifier_free_guidance:
            predicted_image_embedding_uncond, predicted_image_embedding_text = predicted_image_embedding.chunk(2)
            predicted_image_embedding = predicted_image_embedding_uncond + prior_guidance_scale * (
                predicted_image_embedding_text - predicted_image_embedding_uncond
            )

        prior_latents = self.prior_scheduler.step(
            predicted_image_embedding,
            timestep=t,
            sample=prior_latents,
            **prior_extra_step_kwargs,
            return_dict=False,
        )[0]

        if callback is not None and i % callback_steps == 0:
            callback(i, t, prior_latents)

    prior_latents = self.prior.post_process_latents(prior_latents)

    image_embeds = prior_latents

    # done prior

    # 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`
    # corresponds to doing no classifier free guidance.
    do_classifier_free_guidance = guidance_scale > 1.0

    # 8. Encode input prompt
    text_encoder_lora_scale = (
        cross_attention_kwargs.get("scale", None) if cross_attention_kwargs is not None else None
    )
    prompt_embeds, negative_prompt_embeds = self.encode_prompt(
        prompt=prompt,
        num_images_per_prompt=num_images_per_prompt,
        do_classifier_free_guidance=do_classifier_free_guidance,
        negative_prompt=negative_prompt,
        prompt_embeds=prompt_embeds,
        negative_prompt_embeds=negative_prompt_embeds,
        lora_scale=text_encoder_lora_scale,
        clip_skip=clip_skip,
    )
    # For classifier free guidance, we need to do two forward passes.
    # Here we concatenate the unconditional and text embeddings into a single batch
    # to avoid doing two forward passes
    if do_classifier_free_guidance:
        prompt_embeds = ops.cat([negative_prompt_embeds, prompt_embeds])

    # 9. Prepare image embeddings
    image_embeds = self.noise_image_embeddings(
        image_embeds=image_embeds,
        noise_level=noise_level,
        generator=generator,
    )

    if do_classifier_free_guidance:
        negative_prompt_embeds = ops.zeros_like(image_embeds)

        # For classifier free guidance, we need to do two forward passes.
        # Here we concatenate the unconditional and text embeddings into a single batch
        # to avoid doing two forward passes
        image_embeds = ops.cat([negative_prompt_embeds, image_embeds])

    # 10. Prepare timesteps
    self.scheduler.set_timesteps(num_inference_steps)
    timesteps = self.scheduler.timesteps

    # 11. Prepare latent variables
    num_channels_latents = self.unet.config.in_channels
    shape = (
        batch_size,
        num_channels_latents,
        int(height) // self.vae_scale_factor,
        int(width) // self.vae_scale_factor,
    )
    latents = self.prepare_latents(
        shape=shape,
        dtype=prompt_embeds.dtype,
        generator=generator,
        latents=latents,
        scheduler=self.scheduler,
    )

    # 12. Prepare extra step kwargs. TODO: Logic should ideally just be moved out of the pipeline
    extra_step_kwargs = self.prepare_extra_step_kwargs(generator, eta)

    # 13. Denoising loop
    for i, t in enumerate(self.progress_bar(timesteps)):
        latent_model_input = ops.cat([latents] * 2) if do_classifier_free_guidance else latents
        latent_model_input = self.scheduler.scale_model_input(latent_model_input, t)

        # predict the noise residual
        noise_pred = self.unet(
            latent_model_input,
            t,
            encoder_hidden_states=prompt_embeds,
            class_labels=image_embeds,
            cross_attention_kwargs=cross_attention_kwargs,
            return_dict=False,
        )[0]

        # perform guidance
        if do_classifier_free_guidance:
            noise_pred_uncond, noise_pred_text = noise_pred.chunk(2)
            noise_pred = noise_pred_uncond + guidance_scale * (noise_pred_text - noise_pred_uncond)

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

        if callback is not None and i % callback_steps == 0:
            step_idx = i // getattr(self.scheduler, "order", 1)
            callback(step_idx, t, latents)

    if not output_type == "latent":
        image = self.vae.decode(latents / self.vae.config.scaling_factor, return_dict=False)[0]
    else:
        image = latents

    image = self.image_processor.postprocess(image, output_type=output_type)

    if not return_dict:
        return (image,)

    return ImagePipelineOutput(images=image)

mindone.diffusers.StableUnCLIPPipeline.encode_prompt(prompt, num_images_per_prompt, do_classifier_free_guidance, negative_prompt=None, prompt_embeds=None, negative_prompt_embeds=None, lora_scale=None, clip_skip=None)

Encodes the prompt into text encoder hidden states.

PARAMETER	DESCRIPTION
prompt	prompt to be encoded TYPE: `str` or `List[str]`, *optional*
num_images_per_prompt	number of images that should be generated per prompt TYPE: `int`
do_classifier_free_guidance	whether to use classifier free guidance or not TYPE: `bool`
negative_prompt	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 less than 1). TYPE: `str` or `List[str]`, *optional* DEFAULT: None
prompt_embeds	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. TYPE: `ms.Tensor`, *optional* DEFAULT: None
negative_prompt_embeds	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. TYPE: `ms.Tensor`, *optional* DEFAULT: None
lora_scale	A LoRA scale that will be applied to all LoRA layers of the text encoder if LoRA layers are loaded. TYPE: `float`, *optional* DEFAULT: None
clip_skip	Number of layers to be skipped from CLIP while computing the prompt embeddings. A value of 1 means that the output of the pre-final layer will be used for computing the prompt embeddings. TYPE: `int`, *optional* DEFAULT: None

Source code in mindone/diffusers/pipelines/stable_diffusion/pipeline_stable_unclip.py

def encode_prompt(
    self,
    prompt,
    num_images_per_prompt,
    do_classifier_free_guidance,
    negative_prompt=None,
    prompt_embeds: Optional[ms.Tensor] = None,
    negative_prompt_embeds: Optional[ms.Tensor] = None,
    lora_scale: Optional[float] = None,
    clip_skip: Optional[int] = None,
):
    r"""
    Encodes the prompt into text encoder hidden states.

    Args:
        prompt (`str` or `List[str]`, *optional*):
            prompt to be encoded
        num_images_per_prompt (`int`):
            number of images that should be generated per prompt
        do_classifier_free_guidance (`bool`):
            whether to use classifier free guidance or not
        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
            less than `1`).
        prompt_embeds (`ms.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.
        negative_prompt_embeds (`ms.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.
        lora_scale (`float`, *optional*):
            A LoRA scale that will be applied to all LoRA layers of the text encoder if LoRA layers are loaded.
        clip_skip (`int`, *optional*):
            Number of layers to be skipped from CLIP while computing the prompt embeddings. A value of 1 means that
            the output of the pre-final layer will be used for computing the prompt embeddings.
    """
    # 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, StableDiffusionLoraLoaderMixin):
        self._lora_scale = lora_scale

        # dynamically adjust the LoRA scale
        scale_lora_layers(self.text_encoder, lora_scale)

    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]

    if prompt_embeds is None:
        # textual inversion: process multi-vector tokens if necessary
        if isinstance(self, TextualInversionLoaderMixin):
            prompt = self.maybe_convert_prompt(prompt, self.tokenizer)

        text_inputs = self.tokenizer(
            prompt,
            padding="max_length",
            max_length=self.tokenizer.model_max_length,
            truncation=True,
            return_tensors="np",
        )
        text_input_ids = text_inputs.input_ids
        untruncated_ids = self.tokenizer(prompt, padding="longest", return_tensors="np").input_ids

        if untruncated_ids.shape[-1] >= text_input_ids.shape[-1] and not np.array_equal(
            text_input_ids, untruncated_ids
        ):
            removed_text = self.tokenizer.batch_decode(untruncated_ids[:, self.tokenizer.model_max_length - 1 : -1])
            logger.warning(
                "The following part of your input was truncated because CLIP can only handle sequences up to"
                f" {self.tokenizer.model_max_length} tokens: {removed_text}"
            )

        if hasattr(self.text_encoder.config, "use_attention_mask") and self.text_encoder.config.use_attention_mask:
            attention_mask = ms.Tensor(text_inputs.attention_mask)
        else:
            attention_mask = None

        if clip_skip is None:
            prompt_embeds = self.text_encoder(ms.Tensor(text_input_ids), attention_mask=attention_mask)
            prompt_embeds = prompt_embeds[0]
        else:
            prompt_embeds = self.text_encoder(
                ms.Tensor(text_input_ids), attention_mask=attention_mask, output_hidden_states=True
            )
            # Access the `hidden_states` first, that contains a tuple of
            # all the hidden states from the encoder layers. Then index into
            # the tuple to access the hidden states from the desired layer.
            prompt_embeds = prompt_embeds[-1][-(clip_skip + 1)]
            # We also need to apply the final LayerNorm here to not mess with the
            # representations. The `last_hidden_states` that we typically use for
            # obtaining the final prompt representations passes through the LayerNorm
            # layer.
            prompt_embeds = self.text_encoder.text_model.final_layer_norm(prompt_embeds)

    if self.text_encoder is not None:
        prompt_embeds_dtype = self.text_encoder.dtype
    elif self.unet is not None:
        prompt_embeds_dtype = self.unet.dtype
    else:
        prompt_embeds_dtype = prompt_embeds.dtype

    prompt_embeds = prompt_embeds.to(dtype=prompt_embeds_dtype)

    bs_embed, seq_len, _ = prompt_embeds.shape
    # duplicate text embeddings for each generation per prompt, using mps friendly method
    prompt_embeds = prompt_embeds.tile((1, num_images_per_prompt, 1))
    prompt_embeds = prompt_embeds.view(bs_embed * num_images_per_prompt, seq_len, -1)

    # get unconditional embeddings for classifier free guidance
    if do_classifier_free_guidance and negative_prompt_embeds is None:
        uncond_tokens: List[str]
        if negative_prompt is None:
            uncond_tokens = [""] * batch_size
        elif 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 isinstance(negative_prompt, str):
            uncond_tokens = [negative_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`."
            )
        else:
            uncond_tokens = negative_prompt

        # textual inversion: process multi-vector tokens if necessary
        if isinstance(self, TextualInversionLoaderMixin):
            uncond_tokens = self.maybe_convert_prompt(uncond_tokens, self.tokenizer)

        max_length = prompt_embeds.shape[1]
        uncond_input = self.tokenizer(
            uncond_tokens,
            padding="max_length",
            max_length=max_length,
            truncation=True,
            return_tensors="np",
        )

        if hasattr(self.text_encoder.config, "use_attention_mask") and self.text_encoder.config.use_attention_mask:
            attention_mask = ms.Tensor(uncond_input.attention_mask)
        else:
            attention_mask = None

        negative_prompt_embeds = self.text_encoder(
            ms.Tensor(uncond_input.input_ids),
            attention_mask=attention_mask,
        )
        negative_prompt_embeds = negative_prompt_embeds[0]

    if do_classifier_free_guidance:
        # duplicate unconditional embeddings for each generation per prompt, using mps friendly method
        seq_len = negative_prompt_embeds.shape[1]

        negative_prompt_embeds = negative_prompt_embeds.to(dtype=prompt_embeds_dtype)

        negative_prompt_embeds = negative_prompt_embeds.tile((1, num_images_per_prompt, 1))
        negative_prompt_embeds = negative_prompt_embeds.view(batch_size * num_images_per_prompt, seq_len, -1)

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

    return prompt_embeds, negative_prompt_embeds

mindone.diffusers.StableUnCLIPPipeline.noise_image_embeddings(image_embeds, noise_level, noise=None, generator=None)

Add noise to the image embeddings. The amount of noise is controlled by a noise_level input. A higher noise_level increases the variance in the final un-noised images.

The noise is applied in two ways: 1. A noise schedule is applied directly to the embeddings. 2. A vector of sinusoidal time embeddings are appended to the output.

In both cases, the amount of noise is controlled by the same noise_level.

The embeddings are normalized before the noise is applied and un-normalized after the noise is applied.

Source code in mindone/diffusers/pipelines/stable_diffusion/pipeline_stable_unclip.py

def noise_image_embeddings(
    self,
    image_embeds: ms.Tensor,
    noise_level: int,
    noise: Optional[ms.Tensor] = None,
    generator: Optional[np.random.Generator] = None,
):
    """
    Add noise to the image embeddings. The amount of noise is controlled by a `noise_level` input. A higher
    `noise_level` increases the variance in the final un-noised images.

    The noise is applied in two ways:
    1. A noise schedule is applied directly to the embeddings.
    2. A vector of sinusoidal time embeddings are appended to the output.

    In both cases, the amount of noise is controlled by the same `noise_level`.

    The embeddings are normalized before the noise is applied and un-normalized after the noise is applied.
    """
    if noise is None:
        noise = randn_tensor(image_embeds.shape, generator=generator, dtype=image_embeds.dtype)

    noise_level = ms.tensor([noise_level] * image_embeds.shape[0])

    image_embeds = self.image_normalizer.scale(image_embeds)

    image_embeds = self.image_noising_scheduler.add_noise(image_embeds, timesteps=noise_level, noise=noise)

    image_embeds = self.image_normalizer.unscale(image_embeds)

    noise_level = get_timestep_embedding(
        timesteps=noise_level, embedding_dim=image_embeds.shape[-1], flip_sin_to_cos=True, downscale_freq_shift=0
    )

    # `get_timestep_embeddings` does not contain any weights and will always return f32 tensors,
    # but we might actually be running in fp16. so we need to cast here.
    # there might be better ways to encapsulate this.
    noise_level = noise_level.to(image_embeds.dtype)

    image_embeds = ops.cat((image_embeds, noise_level), 1)

    return image_embeds

mindone.diffusers.StableUnCLIPImg2ImgPipeline

Bases: DiffusionPipeline, StableDiffusionMixin, TextualInversionLoaderMixin, StableDiffusionLoraLoaderMixin

Pipeline for text-guided image-to-image generation using stable unCLIP.

This model inherits from [DiffusionPipeline]. Check the superclass documentation for the generic methods implemented for all pipelines (downloading, saving, running on a particular device, etc.).

The pipeline also inherits the following loading methods

• [~loaders.TextualInversionLoaderMixin.load_textual_inversion] for loading textual inversion embeddings
• [~loaders.StableDiffusionLoraLoaderMixin.load_lora_weights] for loading LoRA weights
• [~loaders.StableDiffusionLoraLoaderMixin.save_lora_weights] for saving LoRA weights

PARAMETER	DESCRIPTION
feature_extractor	Feature extractor for image pre-processing before being encoded. TYPE: [`CLIPImageProcessor`]
image_encoder	CLIP vision model for encoding images. TYPE: [`CLIPVisionModelWithProjection`]
image_normalizer	Used to normalize the predicted image embeddings before the noise is applied and un-normalize the image embeddings after the noise has been applied. TYPE: [`StableUnCLIPImageNormalizer`]
image_noising_scheduler	Noise schedule for adding noise to the predicted image embeddings. The amount of noise to add is determined by the noise_level. TYPE: [`KarrasDiffusionSchedulers`]
tokenizer	A [~transformers.CLIPTokenizer)]. TYPE: `~transformers.CLIPTokenizer`
text_encoder	Frozen [~transformers.CLIPTextModel] text-encoder. TYPE: [`~transformers.CLIPTextModel`]
unet	A [UNet2DConditionModel] to denoise the encoded image latents. TYPE: [`UNet2DConditionModel`]
scheduler	A scheduler to be used in combination with unet to denoise the encoded image latents. TYPE: [`KarrasDiffusionSchedulers`]
vae	Variational Auto-Encoder (VAE) Model to encode and decode images to and from latent representations. TYPE: [`AutoencoderKL`]

Source code in mindone/diffusers/pipelines/stable_diffusion/pipeline_stable_unclip_img2img.py

class StableUnCLIPImg2ImgPipeline(
    DiffusionPipeline, StableDiffusionMixin, TextualInversionLoaderMixin, StableDiffusionLoraLoaderMixin
):
    """
    Pipeline for text-guided image-to-image generation using stable unCLIP.

    This model inherits from [`DiffusionPipeline`]. Check the superclass documentation for the generic methods
    implemented for all pipelines (downloading, saving, running on a particular device, etc.).

    The pipeline also inherits the following loading methods:
        - [`~loaders.TextualInversionLoaderMixin.load_textual_inversion`] for loading textual inversion embeddings
        - [`~loaders.StableDiffusionLoraLoaderMixin.load_lora_weights`] for loading LoRA weights
        - [`~loaders.StableDiffusionLoraLoaderMixin.save_lora_weights`] for saving LoRA weights

    Args:
        feature_extractor ([`CLIPImageProcessor`]):
            Feature extractor for image pre-processing before being encoded.
        image_encoder ([`CLIPVisionModelWithProjection`]):
            CLIP vision model for encoding images.
        image_normalizer ([`StableUnCLIPImageNormalizer`]):
            Used to normalize the predicted image embeddings before the noise is applied and un-normalize the image
            embeddings after the noise has been applied.
        image_noising_scheduler ([`KarrasDiffusionSchedulers`]):
            Noise schedule for adding noise to the predicted image embeddings. The amount of noise to add is determined
            by the `noise_level`.
        tokenizer (`~transformers.CLIPTokenizer`):
            A [`~transformers.CLIPTokenizer`)].
        text_encoder ([`~transformers.CLIPTextModel`]):
            Frozen [`~transformers.CLIPTextModel`] text-encoder.
        unet ([`UNet2DConditionModel`]):
            A [`UNet2DConditionModel`] to denoise the encoded image latents.
        scheduler ([`KarrasDiffusionSchedulers`]):
            A scheduler to be used in combination with `unet` to denoise the encoded image latents.
        vae ([`AutoencoderKL`]):
            Variational Auto-Encoder (VAE) Model to encode and decode images to and from latent representations.
    """

    model_cpu_offload_seq = "text_encoder->image_encoder->unet->vae"
    _exclude_from_cpu_offload = ["image_normalizer"]

    # image encoding components
    feature_extractor: CLIPImageProcessor
    image_encoder: CLIPVisionModelWithProjection

    # image noising components
    image_normalizer: StableUnCLIPImageNormalizer
    image_noising_scheduler: KarrasDiffusionSchedulers

    # regular denoising components
    tokenizer: CLIPTokenizer
    text_encoder: CLIPTextModel
    unet: UNet2DConditionModel
    scheduler: KarrasDiffusionSchedulers

    vae: AutoencoderKL

    def __init__(
        self,
        # image encoding components
        feature_extractor: CLIPImageProcessor,
        image_encoder: CLIPVisionModelWithProjection,
        # image noising components
        image_normalizer: StableUnCLIPImageNormalizer,
        image_noising_scheduler: KarrasDiffusionSchedulers,
        # regular denoising components
        tokenizer: CLIPTokenizer,
        text_encoder: CLIPTextModel,
        unet: UNet2DConditionModel,
        scheduler: KarrasDiffusionSchedulers,
        # vae
        vae: AutoencoderKL,
    ):
        super().__init__()

        self.register_modules(
            feature_extractor=feature_extractor,
            image_encoder=image_encoder,
            image_normalizer=image_normalizer,
            image_noising_scheduler=image_noising_scheduler,
            tokenizer=tokenizer,
            text_encoder=text_encoder,
            unet=unet,
            scheduler=scheduler,
            vae=vae,
        )

        self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1)
        self.image_processor = VaeImageProcessor(vae_scale_factor=self.vae_scale_factor)

    # Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion.StableDiffusionPipeline._encode_prompt
    def _encode_prompt(
        self,
        prompt,
        num_images_per_prompt,
        do_classifier_free_guidance,
        negative_prompt=None,
        prompt_embeds: Optional[ms.Tensor] = None,
        negative_prompt_embeds: Optional[ms.Tensor] = None,
        lora_scale: Optional[float] = None,
        **kwargs,
    ):
        deprecation_message = "`_encode_prompt()` is deprecated and it will be removed in a future version. Use `encode_prompt()` instead. Also, be aware that the output format changed from a concatenated tensor to a tuple."  # noqa: E501
        deprecate("_encode_prompt()", "1.0.0", deprecation_message, standard_warn=False)

        prompt_embeds_tuple = self.encode_prompt(
            prompt=prompt,
            num_images_per_prompt=num_images_per_prompt,
            do_classifier_free_guidance=do_classifier_free_guidance,
            negative_prompt=negative_prompt,
            prompt_embeds=prompt_embeds,
            negative_prompt_embeds=negative_prompt_embeds,
            lora_scale=lora_scale,
            **kwargs,
        )

        # concatenate for backwards comp
        prompt_embeds = ops.cat([prompt_embeds_tuple[1], prompt_embeds_tuple[0]])

        return prompt_embeds

    def _encode_image(
        self,
        image,
        batch_size,
        num_images_per_prompt,
        do_classifier_free_guidance,
        noise_level,
        generator,
        image_embeds,
    ):
        dtype = next(self.image_encoder.get_parameters()).dtype

        if isinstance(image, PIL.Image.Image):
            # the image embedding should repeated so it matches the total batch size of the prompt
            repeat_by = batch_size
        else:
            # assume the image input is already properly batched and just needs to be repeated so
            # it matches the num_images_per_prompt.
            #
            # NOTE(will) this is probably missing a few number of side cases. I.e. batched/non-batched
            # `image_embeds`. If those happen to be common use cases, let's think harder about
            # what the expected dimensions of inputs should be and how we handle the encoding.
            repeat_by = num_images_per_prompt

        if image_embeds is None:
            if not isinstance(image, ms.Tensor):
                image = ms.tensor(self.feature_extractor(images=image, return_tensors="np").pixel_values)

            image = image.to(dtype=dtype)
            image_embeds = self.image_encoder(image)[0]

        image_embeds = self.noise_image_embeddings(
            image_embeds=image_embeds,
            noise_level=noise_level,
            generator=generator,
        )

        # duplicate image embeddings for each generation per prompt, using mps friendly method
        image_embeds = image_embeds.unsqueeze(1)
        bs_embed, seq_len, _ = image_embeds.shape
        image_embeds = image_embeds.tile((1, repeat_by, 1))
        image_embeds = image_embeds.view(bs_embed * repeat_by, seq_len, -1)
        image_embeds = image_embeds.squeeze(1)

        if do_classifier_free_guidance:
            negative_prompt_embeds = ops.zeros_like(image_embeds)

            # For classifier free guidance, we need to do two forward passes.
            # Here we concatenate the unconditional and text embeddings into a single batch
            # to avoid doing two forward passes
            image_embeds = ops.cat([negative_prompt_embeds, image_embeds])

        return image_embeds

    # Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion.StableDiffusionPipeline.encode_prompt
    def encode_prompt(
        self,
        prompt,
        num_images_per_prompt,
        do_classifier_free_guidance,
        negative_prompt=None,
        prompt_embeds: Optional[ms.Tensor] = None,
        negative_prompt_embeds: Optional[ms.Tensor] = None,
        lora_scale: Optional[float] = None,
        clip_skip: Optional[int] = None,
    ):
        r"""
        Encodes the prompt into text encoder hidden states.

        Args:
            prompt (`str` or `List[str]`, *optional*):
                prompt to be encoded
            num_images_per_prompt (`int`):
                number of images that should be generated per prompt
            do_classifier_free_guidance (`bool`):
                whether to use classifier free guidance or not
            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
                less than `1`).
            prompt_embeds (`ms.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.
            negative_prompt_embeds (`ms.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.
            lora_scale (`float`, *optional*):
                A LoRA scale that will be applied to all LoRA layers of the text encoder if LoRA layers are loaded.
            clip_skip (`int`, *optional*):
                Number of layers to be skipped from CLIP while computing the prompt embeddings. A value of 1 means that
                the output of the pre-final layer will be used for computing the prompt embeddings.
        """
        # 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, StableDiffusionLoraLoaderMixin):
            self._lora_scale = lora_scale

            # dynamically adjust the LoRA scale
            scale_lora_layers(self.text_encoder, lora_scale)

        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]

        if prompt_embeds is None:
            # textual inversion: process multi-vector tokens if necessary
            if isinstance(self, TextualInversionLoaderMixin):
                prompt = self.maybe_convert_prompt(prompt, self.tokenizer)

            text_inputs = self.tokenizer(
                prompt,
                padding="max_length",
                max_length=self.tokenizer.model_max_length,
                truncation=True,
                return_tensors="np",
            )
            text_input_ids = text_inputs.input_ids
            untruncated_ids = self.tokenizer(prompt, padding="longest", return_tensors="np").input_ids

            if untruncated_ids.shape[-1] >= text_input_ids.shape[-1] and not np.array_equal(
                text_input_ids, untruncated_ids
            ):
                removed_text = self.tokenizer.batch_decode(untruncated_ids[:, self.tokenizer.model_max_length - 1 : -1])
                logger.warning(
                    "The following part of your input was truncated because CLIP can only handle sequences up to"
                    f" {self.tokenizer.model_max_length} tokens: {removed_text}"
                )

            if hasattr(self.text_encoder.config, "use_attention_mask") and self.text_encoder.config.use_attention_mask:
                attention_mask = ms.Tensor(text_inputs.attention_mask)
            else:
                attention_mask = None

            if clip_skip is None:
                prompt_embeds = self.text_encoder(ms.Tensor(text_input_ids), attention_mask=attention_mask)
                prompt_embeds = prompt_embeds[0]
            else:
                prompt_embeds = self.text_encoder(
                    ms.Tensor(text_input_ids), attention_mask=attention_mask, output_hidden_states=True
                )
                # Access the `hidden_states` first, that contains a tuple of
                # all the hidden states from the encoder layers. Then index into
                # the tuple to access the hidden states from the desired layer.
                prompt_embeds = prompt_embeds[-1][-(clip_skip + 1)]
                # We also need to apply the final LayerNorm here to not mess with the
                # representations. The `last_hidden_states` that we typically use for
                # obtaining the final prompt representations passes through the LayerNorm
                # layer.
                prompt_embeds = self.text_encoder.text_model.final_layer_norm(prompt_embeds)

        if self.text_encoder is not None:
            prompt_embeds_dtype = self.text_encoder.dtype
        elif self.unet is not None:
            prompt_embeds_dtype = self.unet.dtype
        else:
            prompt_embeds_dtype = prompt_embeds.dtype

        prompt_embeds = prompt_embeds.to(dtype=prompt_embeds_dtype)

        bs_embed, seq_len, _ = prompt_embeds.shape
        # duplicate text embeddings for each generation per prompt, using mps friendly method
        prompt_embeds = prompt_embeds.tile((1, num_images_per_prompt, 1))
        prompt_embeds = prompt_embeds.view(bs_embed * num_images_per_prompt, seq_len, -1)

        # get unconditional embeddings for classifier free guidance
        if do_classifier_free_guidance and negative_prompt_embeds is None:
            uncond_tokens: List[str]
            if negative_prompt is None:
                uncond_tokens = [""] * batch_size
            elif 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 isinstance(negative_prompt, str):
                uncond_tokens = [negative_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`."
                )
            else:
                uncond_tokens = negative_prompt

            # textual inversion: process multi-vector tokens if necessary
            if isinstance(self, TextualInversionLoaderMixin):
                uncond_tokens = self.maybe_convert_prompt(uncond_tokens, self.tokenizer)

            max_length = prompt_embeds.shape[1]
            uncond_input = self.tokenizer(
                uncond_tokens,
                padding="max_length",
                max_length=max_length,
                truncation=True,
                return_tensors="np",
            )

            if hasattr(self.text_encoder.config, "use_attention_mask") and self.text_encoder.config.use_attention_mask:
                attention_mask = ms.Tensor(uncond_input.attention_mask)
            else:
                attention_mask = None

            negative_prompt_embeds = self.text_encoder(
                ms.Tensor(uncond_input.input_ids),
                attention_mask=attention_mask,
            )
            negative_prompt_embeds = negative_prompt_embeds[0]

        if do_classifier_free_guidance:
            # duplicate unconditional embeddings for each generation per prompt, using mps friendly method
            seq_len = negative_prompt_embeds.shape[1]

            negative_prompt_embeds = negative_prompt_embeds.to(dtype=prompt_embeds_dtype)

            negative_prompt_embeds = negative_prompt_embeds.tile((1, num_images_per_prompt, 1))
            negative_prompt_embeds = negative_prompt_embeds.view(batch_size * num_images_per_prompt, seq_len, -1)

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

        return prompt_embeds, negative_prompt_embeds

    # Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion.StableDiffusionPipeline.decode_latents
    def decode_latents(self, latents):
        deprecation_message = "The decode_latents method is deprecated and will be removed in 1.0.0. Please use VaeImageProcessor.postprocess(...) instead"
        deprecate("decode_latents", "1.0.0", deprecation_message, standard_warn=False)

        latents = 1 / self.vae.config.scaling_factor * latents
        image = self.vae.decode(latents, return_dict=False)[0]
        image = (image / 2 + 0.5).clamp(0, 1)
        # we always cast to float32 as this does not cause significant overhead and is compatible with bfloat16
        image = image.permute(0, 2, 3, 1).float().numpy()
        return image

    # Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion.StableDiffusionPipeline.prepare_extra_step_kwargs
    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
        # and should be between [0, 1]

        accepts_eta = "eta" in set(inspect.signature(self.scheduler.step).parameters.keys())
        extra_step_kwargs = {}
        if accepts_eta:
            extra_step_kwargs["eta"] = eta

        # check if the scheduler accepts generator
        accepts_generator = "generator" in set(inspect.signature(self.scheduler.step).parameters.keys())
        if accepts_generator:
            extra_step_kwargs["generator"] = generator
        return extra_step_kwargs

    def check_inputs(
        self,
        prompt,
        image,
        height,
        width,
        callback_steps,
        noise_level,
        negative_prompt=None,
        prompt_embeds=None,
        negative_prompt_embeds=None,
        image_embeds=None,
    ):
        if height % 8 != 0 or width % 8 != 0:
            raise ValueError(f"`height` and `width` have to be divisible by 8 but are {height} and {width}.")

        if (callback_steps is None) or (
            callback_steps is not None and (not isinstance(callback_steps, int) or callback_steps <= 0)
        ):
            raise ValueError(
                f"`callback_steps` has to be a positive integer but is {callback_steps} of type"
                f" {type(callback_steps)}."
            )

        if prompt is not None and prompt_embeds is not None:
            raise ValueError(
                "Provide either `prompt` or `prompt_embeds`. Please make sure to define only one of the two."
            )

        if prompt is None and prompt_embeds is None:
            raise ValueError(
                "Provide either `prompt` or `prompt_embeds`. Cannot leave both `prompt` and `prompt_embeds` undefined."
            )

        if 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(
                "Provide either `negative_prompt` or `negative_prompt_embeds`. Cannot leave both `negative_prompt` and `negative_prompt_embeds` undefined."
            )

        if prompt is not None and negative_prompt is not None:
            if 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)}."
                )

        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 noise_level < 0 or noise_level >= self.image_noising_scheduler.config.num_train_timesteps:
            raise ValueError(
                f"`noise_level` must be between 0 and {self.image_noising_scheduler.config.num_train_timesteps - 1}, inclusive."
            )

        if image is not None and image_embeds is not None:
            raise ValueError(
                "Provide either `image` or `image_embeds`. Please make sure to define only one of the two."
            )

        if image is None and image_embeds is None:
            raise ValueError(
                "Provide either `image` or `image_embeds`. Cannot leave both `image` and `image_embeds` undefined."
            )

        if image is not None:
            if (
                not isinstance(image, ms.Tensor)
                and not isinstance(image, PIL.Image.Image)
                and not isinstance(image, list)
            ):
                raise ValueError(
                    "`image` has to be of type `ms.Tensor` or `PIL.Image.Image` or `List[PIL.Image.Image]` but is"
                    f" {type(image)}"
                )

    # Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion.StableDiffusionPipeline.prepare_latents
    def prepare_latents(self, batch_size, num_channels_latents, height, width, dtype, generator, latents=None):
        shape = (
            batch_size,
            num_channels_latents,
            int(height) // self.vae_scale_factor,
            int(width) // self.vae_scale_factor,
        )
        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."
            )

        if latents is None:
            latents = randn_tensor(shape, generator=generator, dtype=dtype)
        else:
            latents = latents.to(dtype)

        # scale the initial noise by the standard deviation required by the scheduler
        latents = latents * self.scheduler.init_noise_sigma
        # wtf? The above line changes the dtype of latents from fp16 to fp32, so we need a casting.
        latents = latents.to(dtype=dtype)
        return latents

    # Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_unclip.StableUnCLIPPipeline.noise_image_embeddings
    def noise_image_embeddings(
        self,
        image_embeds: ms.Tensor,
        noise_level: int,
        noise: Optional[ms.Tensor] = None,
        generator: Optional[np.random.Generator] = None,
    ):
        """
        Add noise to the image embeddings. The amount of noise is controlled by a `noise_level` input. A higher
        `noise_level` increases the variance in the final un-noised images.

        The noise is applied in two ways:
        1. A noise schedule is applied directly to the embeddings.
        2. A vector of sinusoidal time embeddings are appended to the output.

        In both cases, the amount of noise is controlled by the same `noise_level`.

        The embeddings are normalized before the noise is applied and un-normalized after the noise is applied.
        """
        if noise is None:
            noise = randn_tensor(image_embeds.shape, generator=generator, dtype=image_embeds.dtype)

        # noise_level = ms.tensor([noise_level] * image_embeds.shape[0])  # shape = [1, 1], we need shape = [1]
        noise_level = ms.tensor(noise_level.tile((image_embeds.shape[0],)))

        image_embeds = self.image_normalizer.scale(image_embeds)

        image_embeds = self.image_noising_scheduler.add_noise(image_embeds, timesteps=noise_level, noise=noise)

        image_embeds = self.image_normalizer.unscale(image_embeds)

        noise_level = get_timestep_embedding(
            timesteps=noise_level, embedding_dim=image_embeds.shape[-1], flip_sin_to_cos=True, downscale_freq_shift=0
        )

        # `get_timestep_embeddings` does not contain any weights and will always return f32 tensors,
        # but we might actually be running in fp16. so we need to cast here.
        # there might be better ways to encapsulate this.
        noise_level = noise_level.to(image_embeds.dtype)

        image_embeds = ops.cat((image_embeds, noise_level), 1)

        return image_embeds

    def __call__(
        self,
        image: Union[ms.Tensor, PIL.Image.Image] = None,
        prompt: Union[str, List[str]] = None,
        height: Optional[int] = None,
        width: Optional[int] = None,
        num_inference_steps: int = 20,
        guidance_scale: float = 10,
        negative_prompt: Optional[Union[str, List[str]]] = None,
        num_images_per_prompt: Optional[int] = 1,
        eta: float = 0.0,
        generator: Optional[np.random.Generator] = None,
        latents: Optional[ms.Tensor] = None,
        prompt_embeds: Optional[ms.Tensor] = None,
        negative_prompt_embeds: Optional[ms.Tensor] = None,
        output_type: Optional[str] = "pil",
        return_dict: bool = False,
        callback: Optional[Callable[[int, int, ms.Tensor], None]] = None,
        callback_steps: int = 1,
        cross_attention_kwargs: Optional[Dict[str, Any]] = None,
        noise_level: int = 0,
        image_embeds: Optional[ms.Tensor] = None,
        clip_skip: Optional[int] = None,
    ):
        r"""
        The call function to the pipeline for generation.

        Args:
            prompt (`str` or `List[str]`, *optional*):
                The prompt or prompts to guide the image generation. If not defined, either `prompt_embeds` will be
                used or prompt is initialized to `""`.
            image (`ms.Tensor` or `PIL.Image.Image`):
                `Image` or tensor representing an image batch. The image is encoded to its CLIP embedding which the
                `unet` is conditioned on. The image is _not_ encoded by the `vae` and then used as the latents in the
                denoising process like it is in the standard Stable Diffusion text-guided image variation process.
            height (`int`, *optional*, defaults to `self.unet.config.sample_size * self.vae_scale_factor`):
                The height in pixels of the generated image.
            width (`int`, *optional*, defaults to `self.unet.config.sample_size * self.vae_scale_factor`):
                The width in pixels of the generated image.
            num_inference_steps (`int`, *optional*, defaults to 20):
                The number of denoising steps. More denoising steps usually lead to a higher quality image at the
                expense of slower inference.
            guidance_scale (`float`, *optional*, defaults to 10.0):
                A higher guidance scale value encourages the model to generate images closely linked to the text
                `prompt` at the expense of lower image quality. Guidance scale is enabled when `guidance_scale > 1`.
            negative_prompt (`str` or `List[str]`, *optional*):
                The prompt or prompts to guide what to not include in image generation. If not defined, you need to
                pass `negative_prompt_embeds` instead. Ignored when not using guidance (`guidance_scale < 1`).
            num_images_per_prompt (`int`, *optional*, defaults to 1):
                The number of images to generate per prompt.
            eta (`float`, *optional*, defaults to 0.0):
                Corresponds to parameter eta (η) from the [DDIM](https://arxiv.org/abs/2010.02502) paper. Only applies
                to the [`~schedulers.DDIMScheduler`], and is ignored in other schedulers.
            generator (`np.random.Generator` or `List[np.random.Generator]`, *optional*):
                A [`np.random.Generator`](https://numpy.org/doc/stable/reference/random/generator.html) to make
                generation deterministic.
            latents (`ms.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 is generated by sampling using the supplied random `generator`.
            prompt_embeds (`ms.Tensor`, *optional*):
                Pre-generated text embeddings. Can be used to easily tweak text inputs (prompt weighting). If not
                provided, text embeddings are generated from the `prompt` input argument.
            negative_prompt_embeds (`ms.Tensor`, *optional*):
                Pre-generated negative text embeddings. Can be used to easily tweak text inputs (prompt weighting). If
                not provided, `negative_prompt_embeds` are generated from the `negative_prompt` input argument.
            output_type (`str`, *optional*, defaults to `"pil"`):
                The output format of the generated image. Choose between `PIL.Image` or `np.array`.
            return_dict (`bool`, *optional*, defaults to `False`):
                Whether or not to return a [`~pipelines.ImagePipelineOutput`] instead of a plain tuple.
            callback (`Callable`, *optional*):
                A function that calls every `callback_steps` steps during inference. The function is called with the
                following arguments: `callback(step: int, timestep: int, latents: ms.Tensor)`.
            callback_steps (`int`, *optional*, defaults to 1):
                The frequency at which the `callback` function is called. If not specified, the callback is called at
                every step.
            cross_attention_kwargs (`dict`, *optional*):
                A kwargs dictionary that if specified is passed along to the [`AttentionProcessor`] as defined in
                [`self.processor`](https://github.com/huggingface/diffusers/blob/main/src/diffusers/models/attention_processor.py).
            noise_level (`int`, *optional*, defaults to `0`):
                The amount of noise to add to the image embeddings. A higher `noise_level` increases the variance in
                the final un-noised images. See [`StableUnCLIPPipeline.noise_image_embeddings`] for more details.
            image_embeds (`ms.Tensor`, *optional*):
                Pre-generated CLIP embeddings to condition the `unet` on. These latents are not used in the denoising
                process. If you want to provide pre-generated latents, pass them to `__call__` as `latents`.
            clip_skip (`int`, *optional*):
                Number of layers to be skipped from CLIP while computing the prompt embeddings. A value of 1 means that
                the output of the pre-final layer will be used for computing the prompt embeddings.

        Examples:

        Returns:
            [`~pipelines.ImagePipelineOutput`] or `tuple`:
                [`~ pipeline_utils.ImagePipelineOutput`] if `return_dict` is True, otherwise a `tuple`. When returning
                a tuple, the first element is a list with the generated images.
        """
        # 0. Default height and width to unet
        height = height or self.unet.config.sample_size * self.vae_scale_factor
        width = width or self.unet.config.sample_size * self.vae_scale_factor

        if prompt is None and prompt_embeds is None:
            prompt = len(image) * [""] if isinstance(image, list) else ""

        # 1. Check inputs. Raise error if not correct
        self.check_inputs(
            prompt=prompt,
            image=image,
            height=height,
            width=width,
            callback_steps=callback_steps,
            noise_level=noise_level,
            negative_prompt=negative_prompt,
            prompt_embeds=prompt_embeds,
            negative_prompt_embeds=negative_prompt_embeds,
            image_embeds=image_embeds,
        )

        # 2. 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]

        batch_size = batch_size * num_images_per_prompt

        # 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`
        # corresponds to doing no classifier free guidance.
        do_classifier_free_guidance = guidance_scale > 1.0

        # 3. Encode input prompt
        text_encoder_lora_scale = (
            cross_attention_kwargs.get("scale", None) if cross_attention_kwargs is not None else None
        )
        prompt_embeds, negative_prompt_embeds = self.encode_prompt(
            prompt=prompt,
            num_images_per_prompt=num_images_per_prompt,
            do_classifier_free_guidance=do_classifier_free_guidance,
            negative_prompt=negative_prompt,
            prompt_embeds=prompt_embeds,
            negative_prompt_embeds=negative_prompt_embeds,
            lora_scale=text_encoder_lora_scale,
        )
        # For classifier free guidance, we need to do two forward passes.
        # Here we concatenate the unconditional and text embeddings into a single batch
        # to avoid doing two forward passes
        if do_classifier_free_guidance:
            prompt_embeds = ops.cat([negative_prompt_embeds, prompt_embeds])

        # 4. Encoder input image
        noise_level = ms.tensor([noise_level])
        image_embeds = self._encode_image(
            image=image,
            batch_size=batch_size,
            num_images_per_prompt=num_images_per_prompt,
            do_classifier_free_guidance=do_classifier_free_guidance,
            noise_level=noise_level,
            generator=generator,
            image_embeds=image_embeds,
        )

        # 5. Prepare timesteps
        self.scheduler.set_timesteps(num_inference_steps)
        timesteps = self.scheduler.timesteps

        # 6. Prepare latent variables
        num_channels_latents = self.unet.config.in_channels
        if latents is None:
            latents = self.prepare_latents(
                batch_size=batch_size,
                num_channels_latents=num_channels_latents,
                height=height,
                width=width,
                dtype=prompt_embeds.dtype,
                generator=generator,
                latents=latents,
            )

        # 7. Prepare extra step kwargs. TODO: Logic should ideally just be moved out of the pipeline
        extra_step_kwargs = self.prepare_extra_step_kwargs(generator, eta)

        # 8. Denoising loop
        for i, t in enumerate(self.progress_bar(timesteps)):
            latent_model_input = ops.cat([latents] * 2) if do_classifier_free_guidance else latents
            latent_model_input = self.scheduler.scale_model_input(latent_model_input, t)

            # predict the noise residual
            noise_pred = self.unet(
                latent_model_input,
                t,
                encoder_hidden_states=prompt_embeds,
                class_labels=image_embeds,
                cross_attention_kwargs=cross_attention_kwargs,
                return_dict=False,
            )[0]

            # perform guidance
            if do_classifier_free_guidance:
                noise_pred_uncond, noise_pred_text = noise_pred.chunk(2)
                noise_pred = noise_pred_uncond + guidance_scale * (noise_pred_text - noise_pred_uncond)

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

            if callback is not None and i % callback_steps == 0:
                step_idx = i // getattr(self.scheduler, "order", 1)
                callback(step_idx, t, latents)

        # 9. Post-processing
        if not output_type == "latent":
            image = self.vae.decode(latents / self.vae.config.scaling_factor, return_dict=False)[0]
        else:
            image = latents

        image = self.image_processor.postprocess(image, output_type=output_type)

        if not return_dict:
            return (image,)

        return ImagePipelineOutput(images=image)

mindone.diffusers.StableUnCLIPImg2ImgPipeline.__call__(image=None, prompt=None, height=None, width=None, num_inference_steps=20, guidance_scale=10, negative_prompt=None, num_images_per_prompt=1, eta=0.0, generator=None, latents=None, prompt_embeds=None, negative_prompt_embeds=None, output_type='pil', return_dict=False, callback=None, callback_steps=1, cross_attention_kwargs=None, noise_level=0, image_embeds=None, clip_skip=None)

The call function to the pipeline for generation.

PARAMETER	DESCRIPTION
prompt	The prompt or prompts to guide the image generation. If not defined, either prompt_embeds will be used or prompt is initialized to "". TYPE: `str` or `List[str]`, *optional* DEFAULT: None
image	Image or tensor representing an image batch. The image is encoded to its CLIP embedding which the unet is conditioned on. The image is not encoded by the vae and then used as the latents in the denoising process like it is in the standard Stable Diffusion text-guided image variation process. TYPE: `ms.Tensor` or `PIL.Image.Image` DEFAULT: None
height	The height in pixels of the generated image. TYPE: `int`, *optional*, defaults to `self.unet.config.sample_size * self.vae_scale_factor` DEFAULT: None
width	The width in pixels of the generated image. TYPE: `int`, *optional*, defaults to `self.unet.config.sample_size * self.vae_scale_factor` DEFAULT: None
num_inference_steps	The number of denoising steps. More denoising steps usually lead to a higher quality image at the expense of slower inference. TYPE: `int`, *optional*, defaults to 20 DEFAULT: 20
guidance_scale	A higher guidance scale value encourages the model to generate images closely linked to the text prompt at the expense of lower image quality. Guidance scale is enabled when guidance_scale > 1. TYPE: `float`, *optional*, defaults to 10.0 DEFAULT: 10
negative_prompt	The prompt or prompts to guide what to not include in image generation. If not defined, you need to pass negative_prompt_embeds instead. Ignored when not using guidance (guidance_scale < 1). TYPE: `str` or `List[str]`, *optional* DEFAULT: None
num_images_per_prompt	The number of images to generate per prompt. TYPE: `int`, *optional*, defaults to 1 DEFAULT: 1
eta	Corresponds to parameter eta (η) from the DDIM paper. Only applies to the [~schedulers.DDIMScheduler], and is ignored in other schedulers. TYPE: `float`, *optional*, defaults to 0.0 DEFAULT: 0.0
generator	A np.random.Generator to make generation deterministic. TYPE: `np.random.Generator` or `List[np.random.Generator]`, *optional* DEFAULT: None
latents	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 is generated by sampling using the supplied random generator. TYPE: `ms.Tensor`, *optional* DEFAULT: None
prompt_embeds	Pre-generated text embeddings. Can be used to easily tweak text inputs (prompt weighting). If not provided, text embeddings are generated from the prompt input argument. TYPE: `ms.Tensor`, *optional* DEFAULT: None
negative_prompt_embeds	Pre-generated negative text embeddings. Can be used to easily tweak text inputs (prompt weighting). If not provided, negative_prompt_embeds are generated from the negative_prompt input argument. TYPE: `ms.Tensor`, *optional* DEFAULT: None
output_type	The output format of the generated image. Choose between PIL.Image or np.array. TYPE: `str`, *optional*, defaults to `"pil"` DEFAULT: 'pil'
return_dict	Whether or not to return a [~pipelines.ImagePipelineOutput] instead of a plain tuple. TYPE: `bool`, *optional*, defaults to `False` DEFAULT: False
callback	A function that calls every callback_steps steps during inference. The function is called with the following arguments: callback(step: int, timestep: int, latents: ms.Tensor). TYPE: `Callable`, *optional* DEFAULT: None
callback_steps	The frequency at which the callback function is called. If not specified, the callback is called at every step. TYPE: `int`, *optional*, defaults to 1 DEFAULT: 1
cross_attention_kwargs	A kwargs dictionary that if specified is passed along to the [AttentionProcessor] as defined in self.processor. TYPE: `dict`, *optional* DEFAULT: None
noise_level	The amount of noise to add to the image embeddings. A higher noise_level increases the variance in the final un-noised images. See [StableUnCLIPPipeline.noise_image_embeddings] for more details. TYPE: `int`, *optional*, defaults to `0` DEFAULT: 0
image_embeds	Pre-generated CLIP embeddings to condition the unet on. These latents are not used in the denoising process. If you want to provide pre-generated latents, pass them to __call__ as latents. TYPE: `ms.Tensor`, *optional* DEFAULT: None
clip_skip	Number of layers to be skipped from CLIP while computing the prompt embeddings. A value of 1 means that the output of the pre-final layer will be used for computing the prompt embeddings. TYPE: `int`, *optional* DEFAULT: None

RETURNS	DESCRIPTION
	[~pipelines.ImagePipelineOutput] or tuple: [~ pipeline_utils.ImagePipelineOutput] if return_dict is True, otherwise a tuple. When returning a tuple, the first element is a list with the generated images.

Source code in mindone/diffusers/pipelines/stable_diffusion/pipeline_stable_unclip_img2img.py

def __call__(
    self,
    image: Union[ms.Tensor, PIL.Image.Image] = None,
    prompt: Union[str, List[str]] = None,
    height: Optional[int] = None,
    width: Optional[int] = None,
    num_inference_steps: int = 20,
    guidance_scale: float = 10,
    negative_prompt: Optional[Union[str, List[str]]] = None,
    num_images_per_prompt: Optional[int] = 1,
    eta: float = 0.0,
    generator: Optional[np.random.Generator] = None,
    latents: Optional[ms.Tensor] = None,
    prompt_embeds: Optional[ms.Tensor] = None,
    negative_prompt_embeds: Optional[ms.Tensor] = None,
    output_type: Optional[str] = "pil",
    return_dict: bool = False,
    callback: Optional[Callable[[int, int, ms.Tensor], None]] = None,
    callback_steps: int = 1,
    cross_attention_kwargs: Optional[Dict[str, Any]] = None,
    noise_level: int = 0,
    image_embeds: Optional[ms.Tensor] = None,
    clip_skip: Optional[int] = None,
):
    r"""
    The call function to the pipeline for generation.

    Args:
        prompt (`str` or `List[str]`, *optional*):
            The prompt or prompts to guide the image generation. If not defined, either `prompt_embeds` will be
            used or prompt is initialized to `""`.
        image (`ms.Tensor` or `PIL.Image.Image`):
            `Image` or tensor representing an image batch. The image is encoded to its CLIP embedding which the
            `unet` is conditioned on. The image is _not_ encoded by the `vae` and then used as the latents in the
            denoising process like it is in the standard Stable Diffusion text-guided image variation process.
        height (`int`, *optional*, defaults to `self.unet.config.sample_size * self.vae_scale_factor`):
            The height in pixels of the generated image.
        width (`int`, *optional*, defaults to `self.unet.config.sample_size * self.vae_scale_factor`):
            The width in pixels of the generated image.
        num_inference_steps (`int`, *optional*, defaults to 20):
            The number of denoising steps. More denoising steps usually lead to a higher quality image at the
            expense of slower inference.
        guidance_scale (`float`, *optional*, defaults to 10.0):
            A higher guidance scale value encourages the model to generate images closely linked to the text
            `prompt` at the expense of lower image quality. Guidance scale is enabled when `guidance_scale > 1`.
        negative_prompt (`str` or `List[str]`, *optional*):
            The prompt or prompts to guide what to not include in image generation. If not defined, you need to
            pass `negative_prompt_embeds` instead. Ignored when not using guidance (`guidance_scale < 1`).
        num_images_per_prompt (`int`, *optional*, defaults to 1):
            The number of images to generate per prompt.
        eta (`float`, *optional*, defaults to 0.0):
            Corresponds to parameter eta (η) from the [DDIM](https://arxiv.org/abs/2010.02502) paper. Only applies
            to the [`~schedulers.DDIMScheduler`], and is ignored in other schedulers.
        generator (`np.random.Generator` or `List[np.random.Generator]`, *optional*):
            A [`np.random.Generator`](https://numpy.org/doc/stable/reference/random/generator.html) to make
            generation deterministic.
        latents (`ms.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 is generated by sampling using the supplied random `generator`.
        prompt_embeds (`ms.Tensor`, *optional*):
            Pre-generated text embeddings. Can be used to easily tweak text inputs (prompt weighting). If not
            provided, text embeddings are generated from the `prompt` input argument.
        negative_prompt_embeds (`ms.Tensor`, *optional*):
            Pre-generated negative text embeddings. Can be used to easily tweak text inputs (prompt weighting). If
            not provided, `negative_prompt_embeds` are generated from the `negative_prompt` input argument.
        output_type (`str`, *optional*, defaults to `"pil"`):
            The output format of the generated image. Choose between `PIL.Image` or `np.array`.
        return_dict (`bool`, *optional*, defaults to `False`):
            Whether or not to return a [`~pipelines.ImagePipelineOutput`] instead of a plain tuple.
        callback (`Callable`, *optional*):
            A function that calls every `callback_steps` steps during inference. The function is called with the
            following arguments: `callback(step: int, timestep: int, latents: ms.Tensor)`.
        callback_steps (`int`, *optional*, defaults to 1):
            The frequency at which the `callback` function is called. If not specified, the callback is called at
            every step.
        cross_attention_kwargs (`dict`, *optional*):
            A kwargs dictionary that if specified is passed along to the [`AttentionProcessor`] as defined in
            [`self.processor`](https://github.com/huggingface/diffusers/blob/main/src/diffusers/models/attention_processor.py).
        noise_level (`int`, *optional*, defaults to `0`):
            The amount of noise to add to the image embeddings. A higher `noise_level` increases the variance in
            the final un-noised images. See [`StableUnCLIPPipeline.noise_image_embeddings`] for more details.
        image_embeds (`ms.Tensor`, *optional*):
            Pre-generated CLIP embeddings to condition the `unet` on. These latents are not used in the denoising
            process. If you want to provide pre-generated latents, pass them to `__call__` as `latents`.
        clip_skip (`int`, *optional*):
            Number of layers to be skipped from CLIP while computing the prompt embeddings. A value of 1 means that
            the output of the pre-final layer will be used for computing the prompt embeddings.

    Examples:

    Returns:
        [`~pipelines.ImagePipelineOutput`] or `tuple`:
            [`~ pipeline_utils.ImagePipelineOutput`] if `return_dict` is True, otherwise a `tuple`. When returning
            a tuple, the first element is a list with the generated images.
    """
    # 0. Default height and width to unet
    height = height or self.unet.config.sample_size * self.vae_scale_factor
    width = width or self.unet.config.sample_size * self.vae_scale_factor

    if prompt is None and prompt_embeds is None:
        prompt = len(image) * [""] if isinstance(image, list) else ""

    # 1. Check inputs. Raise error if not correct
    self.check_inputs(
        prompt=prompt,
        image=image,
        height=height,
        width=width,
        callback_steps=callback_steps,
        noise_level=noise_level,
        negative_prompt=negative_prompt,
        prompt_embeds=prompt_embeds,
        negative_prompt_embeds=negative_prompt_embeds,
        image_embeds=image_embeds,
    )

    # 2. 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]

    batch_size = batch_size * num_images_per_prompt

    # 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`
    # corresponds to doing no classifier free guidance.
    do_classifier_free_guidance = guidance_scale > 1.0

    # 3. Encode input prompt
    text_encoder_lora_scale = (
        cross_attention_kwargs.get("scale", None) if cross_attention_kwargs is not None else None
    )
    prompt_embeds, negative_prompt_embeds = self.encode_prompt(
        prompt=prompt,
        num_images_per_prompt=num_images_per_prompt,
        do_classifier_free_guidance=do_classifier_free_guidance,
        negative_prompt=negative_prompt,
        prompt_embeds=prompt_embeds,
        negative_prompt_embeds=negative_prompt_embeds,
        lora_scale=text_encoder_lora_scale,
    )
    # For classifier free guidance, we need to do two forward passes.
    # Here we concatenate the unconditional and text embeddings into a single batch
    # to avoid doing two forward passes
    if do_classifier_free_guidance:
        prompt_embeds = ops.cat([negative_prompt_embeds, prompt_embeds])

    # 4. Encoder input image
    noise_level = ms.tensor([noise_level])
    image_embeds = self._encode_image(
        image=image,
        batch_size=batch_size,
        num_images_per_prompt=num_images_per_prompt,
        do_classifier_free_guidance=do_classifier_free_guidance,
        noise_level=noise_level,
        generator=generator,
        image_embeds=image_embeds,
    )

    # 5. Prepare timesteps
    self.scheduler.set_timesteps(num_inference_steps)
    timesteps = self.scheduler.timesteps

    # 6. Prepare latent variables
    num_channels_latents = self.unet.config.in_channels
    if latents is None:
        latents = self.prepare_latents(
            batch_size=batch_size,
            num_channels_latents=num_channels_latents,
            height=height,
            width=width,
            dtype=prompt_embeds.dtype,
            generator=generator,
            latents=latents,
        )

    # 7. Prepare extra step kwargs. TODO: Logic should ideally just be moved out of the pipeline
    extra_step_kwargs = self.prepare_extra_step_kwargs(generator, eta)

    # 8. Denoising loop
    for i, t in enumerate(self.progress_bar(timesteps)):
        latent_model_input = ops.cat([latents] * 2) if do_classifier_free_guidance else latents
        latent_model_input = self.scheduler.scale_model_input(latent_model_input, t)

        # predict the noise residual
        noise_pred = self.unet(
            latent_model_input,
            t,
            encoder_hidden_states=prompt_embeds,
            class_labels=image_embeds,
            cross_attention_kwargs=cross_attention_kwargs,
            return_dict=False,
        )[0]

        # perform guidance
        if do_classifier_free_guidance:
            noise_pred_uncond, noise_pred_text = noise_pred.chunk(2)
            noise_pred = noise_pred_uncond + guidance_scale * (noise_pred_text - noise_pred_uncond)

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

        if callback is not None and i % callback_steps == 0:
            step_idx = i // getattr(self.scheduler, "order", 1)
            callback(step_idx, t, latents)

    # 9. Post-processing
    if not output_type == "latent":
        image = self.vae.decode(latents / self.vae.config.scaling_factor, return_dict=False)[0]
    else:
        image = latents

    image = self.image_processor.postprocess(image, output_type=output_type)

    if not return_dict:
        return (image,)

    return ImagePipelineOutput(images=image)

mindone.diffusers.StableUnCLIPImg2ImgPipeline.encode_prompt(prompt, num_images_per_prompt, do_classifier_free_guidance, negative_prompt=None, prompt_embeds=None, negative_prompt_embeds=None, lora_scale=None, clip_skip=None)

Encodes the prompt into text encoder hidden states.

PARAMETER	DESCRIPTION
prompt	prompt to be encoded TYPE: `str` or `List[str]`, *optional*
num_images_per_prompt	number of images that should be generated per prompt TYPE: `int`
do_classifier_free_guidance	whether to use classifier free guidance or not TYPE: `bool`
negative_prompt	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 less than 1). TYPE: `str` or `List[str]`, *optional* DEFAULT: None
prompt_embeds	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. TYPE: `ms.Tensor`, *optional* DEFAULT: None
negative_prompt_embeds	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. TYPE: `ms.Tensor`, *optional* DEFAULT: None
lora_scale	A LoRA scale that will be applied to all LoRA layers of the text encoder if LoRA layers are loaded. TYPE: `float`, *optional* DEFAULT: None
clip_skip	Number of layers to be skipped from CLIP while computing the prompt embeddings. A value of 1 means that the output of the pre-final layer will be used for computing the prompt embeddings. TYPE: `int`, *optional* DEFAULT: None

Source code in mindone/diffusers/pipelines/stable_diffusion/pipeline_stable_unclip_img2img.py

def encode_prompt(
    self,
    prompt,
    num_images_per_prompt,
    do_classifier_free_guidance,
    negative_prompt=None,
    prompt_embeds: Optional[ms.Tensor] = None,
    negative_prompt_embeds: Optional[ms.Tensor] = None,
    lora_scale: Optional[float] = None,
    clip_skip: Optional[int] = None,
):
    r"""
    Encodes the prompt into text encoder hidden states.

    Args:
        prompt (`str` or `List[str]`, *optional*):
            prompt to be encoded
        num_images_per_prompt (`int`):
            number of images that should be generated per prompt
        do_classifier_free_guidance (`bool`):
            whether to use classifier free guidance or not
        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
            less than `1`).
        prompt_embeds (`ms.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.
        negative_prompt_embeds (`ms.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.
        lora_scale (`float`, *optional*):
            A LoRA scale that will be applied to all LoRA layers of the text encoder if LoRA layers are loaded.
        clip_skip (`int`, *optional*):
            Number of layers to be skipped from CLIP while computing the prompt embeddings. A value of 1 means that
            the output of the pre-final layer will be used for computing the prompt embeddings.
    """
    # 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, StableDiffusionLoraLoaderMixin):
        self._lora_scale = lora_scale

        # dynamically adjust the LoRA scale
        scale_lora_layers(self.text_encoder, lora_scale)

    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]

    if prompt_embeds is None:
        # textual inversion: process multi-vector tokens if necessary
        if isinstance(self, TextualInversionLoaderMixin):
            prompt = self.maybe_convert_prompt(prompt, self.tokenizer)

        text_inputs = self.tokenizer(
            prompt,
            padding="max_length",
            max_length=self.tokenizer.model_max_length,
            truncation=True,
            return_tensors="np",
        )
        text_input_ids = text_inputs.input_ids
        untruncated_ids = self.tokenizer(prompt, padding="longest", return_tensors="np").input_ids

        if untruncated_ids.shape[-1] >= text_input_ids.shape[-1] and not np.array_equal(
            text_input_ids, untruncated_ids
        ):
            removed_text = self.tokenizer.batch_decode(untruncated_ids[:, self.tokenizer.model_max_length - 1 : -1])
            logger.warning(
                "The following part of your input was truncated because CLIP can only handle sequences up to"
                f" {self.tokenizer.model_max_length} tokens: {removed_text}"
            )

        if hasattr(self.text_encoder.config, "use_attention_mask") and self.text_encoder.config.use_attention_mask:
            attention_mask = ms.Tensor(text_inputs.attention_mask)
        else:
            attention_mask = None

        if clip_skip is None:
            prompt_embeds = self.text_encoder(ms.Tensor(text_input_ids), attention_mask=attention_mask)
            prompt_embeds = prompt_embeds[0]
        else:
            prompt_embeds = self.text_encoder(
                ms.Tensor(text_input_ids), attention_mask=attention_mask, output_hidden_states=True
            )
            # Access the `hidden_states` first, that contains a tuple of
            # all the hidden states from the encoder layers. Then index into
            # the tuple to access the hidden states from the desired layer.
            prompt_embeds = prompt_embeds[-1][-(clip_skip + 1)]
            # We also need to apply the final LayerNorm here to not mess with the
            # representations. The `last_hidden_states` that we typically use for
            # obtaining the final prompt representations passes through the LayerNorm
            # layer.
            prompt_embeds = self.text_encoder.text_model.final_layer_norm(prompt_embeds)

    if self.text_encoder is not None:
        prompt_embeds_dtype = self.text_encoder.dtype
    elif self.unet is not None:
        prompt_embeds_dtype = self.unet.dtype
    else:
        prompt_embeds_dtype = prompt_embeds.dtype

    prompt_embeds = prompt_embeds.to(dtype=prompt_embeds_dtype)

    bs_embed, seq_len, _ = prompt_embeds.shape
    # duplicate text embeddings for each generation per prompt, using mps friendly method
    prompt_embeds = prompt_embeds.tile((1, num_images_per_prompt, 1))
    prompt_embeds = prompt_embeds.view(bs_embed * num_images_per_prompt, seq_len, -1)

    # get unconditional embeddings for classifier free guidance
    if do_classifier_free_guidance and negative_prompt_embeds is None:
        uncond_tokens: List[str]
        if negative_prompt is None:
            uncond_tokens = [""] * batch_size
        elif 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 isinstance(negative_prompt, str):
            uncond_tokens = [negative_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`."
            )
        else:
            uncond_tokens = negative_prompt

        # textual inversion: process multi-vector tokens if necessary
        if isinstance(self, TextualInversionLoaderMixin):
            uncond_tokens = self.maybe_convert_prompt(uncond_tokens, self.tokenizer)

        max_length = prompt_embeds.shape[1]
        uncond_input = self.tokenizer(
            uncond_tokens,
            padding="max_length",
            max_length=max_length,
            truncation=True,
            return_tensors="np",
        )

        if hasattr(self.text_encoder.config, "use_attention_mask") and self.text_encoder.config.use_attention_mask:
            attention_mask = ms.Tensor(uncond_input.attention_mask)
        else:
            attention_mask = None

        negative_prompt_embeds = self.text_encoder(
            ms.Tensor(uncond_input.input_ids),
            attention_mask=attention_mask,
        )
        negative_prompt_embeds = negative_prompt_embeds[0]

    if do_classifier_free_guidance:
        # duplicate unconditional embeddings for each generation per prompt, using mps friendly method
        seq_len = negative_prompt_embeds.shape[1]

        negative_prompt_embeds = negative_prompt_embeds.to(dtype=prompt_embeds_dtype)

        negative_prompt_embeds = negative_prompt_embeds.tile((1, num_images_per_prompt, 1))
        negative_prompt_embeds = negative_prompt_embeds.view(batch_size * num_images_per_prompt, seq_len, -1)

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

    return prompt_embeds, negative_prompt_embeds

mindone.diffusers.StableUnCLIPImg2ImgPipeline.noise_image_embeddings(image_embeds, noise_level, noise=None, generator=None)

Add noise to the image embeddings. The amount of noise is controlled by a noise_level input. A higher noise_level increases the variance in the final un-noised images.

The noise is applied in two ways: 1. A noise schedule is applied directly to the embeddings. 2. A vector of sinusoidal time embeddings are appended to the output.

In both cases, the amount of noise is controlled by the same noise_level.

The embeddings are normalized before the noise is applied and un-normalized after the noise is applied.

Source code in mindone/diffusers/pipelines/stable_diffusion/pipeline_stable_unclip_img2img.py

def noise_image_embeddings(
    self,
    image_embeds: ms.Tensor,
    noise_level: int,
    noise: Optional[ms.Tensor] = None,
    generator: Optional[np.random.Generator] = None,
):
    """
    Add noise to the image embeddings. The amount of noise is controlled by a `noise_level` input. A higher
    `noise_level` increases the variance in the final un-noised images.

    The noise is applied in two ways:
    1. A noise schedule is applied directly to the embeddings.
    2. A vector of sinusoidal time embeddings are appended to the output.

    In both cases, the amount of noise is controlled by the same `noise_level`.

    The embeddings are normalized before the noise is applied and un-normalized after the noise is applied.
    """
    if noise is None:
        noise = randn_tensor(image_embeds.shape, generator=generator, dtype=image_embeds.dtype)

    # noise_level = ms.tensor([noise_level] * image_embeds.shape[0])  # shape = [1, 1], we need shape = [1]
    noise_level = ms.tensor(noise_level.tile((image_embeds.shape[0],)))

    image_embeds = self.image_normalizer.scale(image_embeds)

    image_embeds = self.image_noising_scheduler.add_noise(image_embeds, timesteps=noise_level, noise=noise)

    image_embeds = self.image_normalizer.unscale(image_embeds)

    noise_level = get_timestep_embedding(
        timesteps=noise_level, embedding_dim=image_embeds.shape[-1], flip_sin_to_cos=True, downscale_freq_shift=0
    )

    # `get_timestep_embeddings` does not contain any weights and will always return f32 tensors,
    # but we might actually be running in fp16. so we need to cast here.
    # there might be better ways to encapsulate this.
    noise_level = noise_level.to(image_embeds.dtype)

    image_embeds = ops.cat((image_embeds, noise_level), 1)

    return image_embeds

mindone.diffusers.pipelines.ImagePipelineOutput dataclass

Bases: BaseOutput

Output class for image pipelines.

Source code in mindone/diffusers/pipelines/pipeline_utils.py

@dataclass
class ImagePipelineOutput(BaseOutput):
    """
    Output class for image pipelines.

    Args:
        images (`List[PIL.Image.Image]` or `np.ndarray`)
            List of denoised PIL images of length `batch_size` or NumPy array of shape `(batch_size, height, width,
            num_channels)`.
    """

    images: Union[List[PIL.Image.Image], np.ndarray]