Skip to content

Attend-and-Excite

Attend-and-Excite for Stable Diffusion was proposed in Attend-and-Excite: Attention-Based Semantic Guidance for Text-to-Image Diffusion Models and provides textual attention control over image generation.

The abstract from the paper is:

Recent text-to-image generative models have demonstrated an unparalleled ability to generate diverse and creative imagery guided by a target text prompt. While revolutionary, current state-of-the-art diffusion models may still fail in generating images that fully convey the semantics in the given text prompt. We analyze the publicly available Stable Diffusion model and assess the existence of catastrophic neglect, where the model fails to generate one or more of the subjects from the input prompt. Moreover, we find that in some cases the model also fails to correctly bind attributes (e.g., colors) to their corresponding subjects. To help mitigate these failure cases, we introduce the concept of Generative Semantic Nursing (GSN), where we seek to intervene in the generative process on the fly during inference time to improve the faithfulness of the generated images. Using an attention-based formulation of GSN, dubbed Attend-and-Excite, we guide the model to refine the cross-attention units to attend to all subject tokens in the text prompt and strengthen - or excite - their activations, encouraging the model to generate all subjects described in the text prompt. We compare our approach to alternative approaches and demonstrate that it conveys the desired concepts more faithfully across a range of text prompts.

You can find additional information about Attend-and-Excite on the project page, the original codebase, or try it out in a demo.

Tip

Make sure to check out the Schedulers guide to learn how to explore the tradeoff between scheduler speed and quality, and see the reuse components across pipelines section to learn how to efficiently load the same components into multiple pipelines.

mindone.diffusers.StableDiffusionAttendAndExcitePipeline

Bases: DiffusionPipeline, StableDiffusionMixin, TextualInversionLoaderMixin

Pipeline for text-to-image generation using Stable Diffusion and Attend-and-Excite.

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
PARAMETER DESCRIPTION
vae

Variational Auto-Encoder (VAE) model to encode and decode images to and from latent representations.

TYPE: [`AutoencoderKL`]

text_encoder

Frozen text-encoder (clip-vit-large-patch14).

TYPE: [`~transformers.CLIPTextModel`]

tokenizer

A CLIPTokenizer to tokenize text.

TYPE: [`~transformers.CLIPTokenizer`]

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. Can be one of [DDIMScheduler], [LMSDiscreteScheduler], or [PNDMScheduler].

TYPE: [`SchedulerMixin`]

safety_checker

Classification module that estimates whether generated images could be considered offensive or harmful. Please refer to the model card for more details about a model's potential harms.

TYPE: [`StableDiffusionSafetyChecker`]

feature_extractor

A CLIPImageProcessor to extract features from generated images; used as inputs to the safety_checker.

TYPE: [`~transformers.CLIPImageProcessor`]

Source code in mindone/diffusers/pipelines/stable_diffusion_attend_and_excite/pipeline_stable_diffusion_attend_and_excite.py
 169
 170
 171
 172
 173
 174
 175
 176
 177
 178
 179
 180
 181
 182
 183
 184
 185
 186
 187
 188
 189
 190
 191
 192
 193
 194
 195
 196
 197
 198
 199
 200
 201
 202
 203
 204
 205
 206
 207
 208
 209
 210
 211
 212
 213
 214
 215
 216
 217
 218
 219
 220
 221
 222
 223
 224
 225
 226
 227
 228
 229
 230
 231
 232
 233
 234
 235
 236
 237
 238
 239
 240
 241
 242
 243
 244
 245
 246
 247
 248
 249
 250
 251
 252
 253
 254
 255
 256
 257
 258
 259
 260
 261
 262
 263
 264
 265
 266
 267
 268
 269
 270
 271
 272
 273
 274
 275
 276
 277
 278
 279
 280
 281
 282
 283
 284
 285
 286
 287
 288
 289
 290
 291
 292
 293
 294
 295
 296
 297
 298
 299
 300
 301
 302
 303
 304
 305
 306
 307
 308
 309
 310
 311
 312
 313
 314
 315
 316
 317
 318
 319
 320
 321
 322
 323
 324
 325
 326
 327
 328
 329
 330
 331
 332
 333
 334
 335
 336
 337
 338
 339
 340
 341
 342
 343
 344
 345
 346
 347
 348
 349
 350
 351
 352
 353
 354
 355
 356
 357
 358
 359
 360
 361
 362
 363
 364
 365
 366
 367
 368
 369
 370
 371
 372
 373
 374
 375
 376
 377
 378
 379
 380
 381
 382
 383
 384
 385
 386
 387
 388
 389
 390
 391
 392
 393
 394
 395
 396
 397
 398
 399
 400
 401
 402
 403
 404
 405
 406
 407
 408
 409
 410
 411
 412
 413
 414
 415
 416
 417
 418
 419
 420
 421
 422
 423
 424
 425
 426
 427
 428
 429
 430
 431
 432
 433
 434
 435
 436
 437
 438
 439
 440
 441
 442
 443
 444
 445
 446
 447
 448
 449
 450
 451
 452
 453
 454
 455
 456
 457
 458
 459
 460
 461
 462
 463
 464
 465
 466
 467
 468
 469
 470
 471
 472
 473
 474
 475
 476
 477
 478
 479
 480
 481
 482
 483
 484
 485
 486
 487
 488
 489
 490
 491
 492
 493
 494
 495
 496
 497
 498
 499
 500
 501
 502
 503
 504
 505
 506
 507
 508
 509
 510
 511
 512
 513
 514
 515
 516
 517
 518
 519
 520
 521
 522
 523
 524
 525
 526
 527
 528
 529
 530
 531
 532
 533
 534
 535
 536
 537
 538
 539
 540
 541
 542
 543
 544
 545
 546
 547
 548
 549
 550
 551
 552
 553
 554
 555
 556
 557
 558
 559
 560
 561
 562
 563
 564
 565
 566
 567
 568
 569
 570
 571
 572
 573
 574
 575
 576
 577
 578
 579
 580
 581
 582
 583
 584
 585
 586
 587
 588
 589
 590
 591
 592
 593
 594
 595
 596
 597
 598
 599
 600
 601
 602
 603
 604
 605
 606
 607
 608
 609
 610
 611
 612
 613
 614
 615
 616
 617
 618
 619
 620
 621
 622
 623
 624
 625
 626
 627
 628
 629
 630
 631
 632
 633
 634
 635
 636
 637
 638
 639
 640
 641
 642
 643
 644
 645
 646
 647
 648
 649
 650
 651
 652
 653
 654
 655
 656
 657
 658
 659
 660
 661
 662
 663
 664
 665
 666
 667
 668
 669
 670
 671
 672
 673
 674
 675
 676
 677
 678
 679
 680
 681
 682
 683
 684
 685
 686
 687
 688
 689
 690
 691
 692
 693
 694
 695
 696
 697
 698
 699
 700
 701
 702
 703
 704
 705
 706
 707
 708
 709
 710
 711
 712
 713
 714
 715
 716
 717
 718
 719
 720
 721
 722
 723
 724
 725
 726
 727
 728
 729
 730
 731
 732
 733
 734
 735
 736
 737
 738
 739
 740
 741
 742
 743
 744
 745
 746
 747
 748
 749
 750
 751
 752
 753
 754
 755
 756
 757
 758
 759
 760
 761
 762
 763
 764
 765
 766
 767
 768
 769
 770
 771
 772
 773
 774
 775
 776
 777
 778
 779
 780
 781
 782
 783
 784
 785
 786
 787
 788
 789
 790
 791
 792
 793
 794
 795
 796
 797
 798
 799
 800
 801
 802
 803
 804
 805
 806
 807
 808
 809
 810
 811
 812
 813
 814
 815
 816
 817
 818
 819
 820
 821
 822
 823
 824
 825
 826
 827
 828
 829
 830
 831
 832
 833
 834
 835
 836
 837
 838
 839
 840
 841
 842
 843
 844
 845
 846
 847
 848
 849
 850
 851
 852
 853
 854
 855
 856
 857
 858
 859
 860
 861
 862
 863
 864
 865
 866
 867
 868
 869
 870
 871
 872
 873
 874
 875
 876
 877
 878
 879
 880
 881
 882
 883
 884
 885
 886
 887
 888
 889
 890
 891
 892
 893
 894
 895
 896
 897
 898
 899
 900
 901
 902
 903
 904
 905
 906
 907
 908
 909
 910
 911
 912
 913
 914
 915
 916
 917
 918
 919
 920
 921
 922
 923
 924
 925
 926
 927
 928
 929
 930
 931
 932
 933
 934
 935
 936
 937
 938
 939
 940
 941
 942
 943
 944
 945
 946
 947
 948
 949
 950
 951
 952
 953
 954
 955
 956
 957
 958
 959
 960
 961
 962
 963
 964
 965
 966
 967
 968
 969
 970
 971
 972
 973
 974
 975
 976
 977
 978
 979
 980
 981
 982
 983
 984
 985
 986
 987
 988
 989
 990
 991
 992
 993
 994
 995
 996
 997
 998
 999
1000
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
class StableDiffusionAttendAndExcitePipeline(DiffusionPipeline, StableDiffusionMixin, TextualInversionLoaderMixin):
    r"""
    Pipeline for text-to-image generation using Stable Diffusion and Attend-and-Excite.

    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

    Args:
        vae ([`AutoencoderKL`]):
            Variational Auto-Encoder (VAE) model to encode and decode images to and from latent representations.
        text_encoder ([`~transformers.CLIPTextModel`]):
            Frozen text-encoder ([clip-vit-large-patch14](https://huggingface.co/openai/clip-vit-large-patch14)).
        tokenizer ([`~transformers.CLIPTokenizer`]):
            A `CLIPTokenizer` to tokenize text.
        unet ([`UNet2DConditionModel`]):
            A `UNet2DConditionModel` to denoise the encoded image latents.
        scheduler ([`SchedulerMixin`]):
            A scheduler to be used in combination with `unet` to denoise the encoded image latents. Can be one of
            [`DDIMScheduler`], [`LMSDiscreteScheduler`], or [`PNDMScheduler`].
        safety_checker ([`StableDiffusionSafetyChecker`]):
            Classification module that estimates whether generated images could be considered offensive or harmful.
            Please refer to the [model card](https://huggingface.co/runwayml/stable-diffusion-v1-5) for more details
            about a model's potential harms.
        feature_extractor ([`~transformers.CLIPImageProcessor`]):
            A `CLIPImageProcessor` to extract features from generated images; used as inputs to the `safety_checker`.
    """

    model_cpu_offload_seq = "text_encoder->unet->vae"
    _optional_components = ["safety_checker", "feature_extractor"]
    _exclude_from_cpu_offload = ["safety_checker"]

    def __init__(
        self,
        vae: AutoencoderKL,
        text_encoder: CLIPTextModel,
        tokenizer: CLIPTokenizer,
        unet: UNet2DConditionModel,
        scheduler: KarrasDiffusionSchedulers,
        safety_checker: StableDiffusionSafetyChecker,
        feature_extractor: CLIPImageProcessor,
        requires_safety_checker: bool = True,
    ):
        super().__init__()

        if safety_checker is None and requires_safety_checker:
            logger.warning(
                f"You have disabled the safety checker for {self.__class__} by passing `safety_checker=None`. Ensure"
                " that you abide to the conditions of the Stable Diffusion license and do not expose unfiltered"
                " results in services or applications open to the public. Both the diffusers team and Hugging Face"
                " strongly recommend to keep the safety filter enabled in all public facing circumstances, disabling"
                " it only for use-cases that involve analyzing network behavior or auditing its results. For more"
                " information, please have a look at https://github.com/huggingface/diffusers/pull/254 ."
            )

        if safety_checker is not None and feature_extractor is None:
            raise ValueError(
                "Make sure to define a feature extractor when loading {self.__class__} if you want to use the safety"
                " checker. If you do not want to use the safety checker, you can pass `'safety_checker=None'` instead."
            )

        self.register_modules(
            vae=vae,
            text_encoder=text_encoder,
            tokenizer=tokenizer,
            unet=unet,
            scheduler=scheduler,
            safety_checker=safety_checker,
            feature_extractor=feature_extractor,
        )
        self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1)
        self.image_processor = VaeImageProcessor(vae_scale_factor=self.vae_scale_factor)
        self.register_to_config(requires_safety_checker=requires_safety_checker)

    # 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 = mint.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.run_safety_checker
    def run_safety_checker(self, image, dtype):
        if self.safety_checker is None:
            has_nsfw_concept = None
        else:
            if ops.is_tensor(image):
                feature_extractor_input = self.image_processor.postprocess(image, output_type="pil")
            else:
                feature_extractor_input = self.image_processor.numpy_to_pil(image)
            safety_checker_input = self.feature_extractor(feature_extractor_input, return_tensors="np")
            image, has_nsfw_concept = self.safety_checker(
                images=image, clip_input=ms.tensor(safety_checker_input.pixel_values).to(dtype)
            )

            # Warning for safety checker operations here as it couldn't been done in construct()
            if mint.any(has_nsfw_concept):
                logger.warning(
                    "Potential NSFW content was detected in one or more images. A black image will be returned instead."
                    " Try again with a different prompt and/or seed."
                )
        return image, has_nsfw_concept

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

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

        if prompt_embeds is not None and negative_prompt_embeds is not None:
            if prompt_embeds.shape != negative_prompt_embeds.shape:
                raise ValueError(
                    "`prompt_embeds` and `negative_prompt_embeds` must have the same shape when passed directly, but"
                    f" got: `prompt_embeds` {prompt_embeds.shape} != `negative_prompt_embeds`"
                    f" {negative_prompt_embeds.shape}."
                )

        indices_is_list_ints = isinstance(indices, list) and isinstance(indices[0], int)
        indices_is_list_list_ints = (
            isinstance(indices, list) and isinstance(indices[0], list) and isinstance(indices[0][0], int)
        )

        if not indices_is_list_ints and not indices_is_list_list_ints:
            raise TypeError("`indices` must be a list of ints or a list of a list of ints")

        if indices_is_list_ints:
            indices_batch_size = 1
        elif indices_is_list_list_ints:
            indices_batch_size = len(indices)

        if prompt is not None and isinstance(prompt, str):
            prompt_batch_size = 1
        elif prompt is not None and isinstance(prompt, list):
            prompt_batch_size = len(prompt)
        elif prompt_embeds is not None:
            prompt_batch_size = prompt_embeds.shape[0]

        if indices_batch_size != prompt_batch_size:
            raise ValueError(
                f"indices batch size must be same as prompt batch size. indices batch size: {indices_batch_size}, prompt batch size: {prompt_batch_size}"
            )

    # 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

    @staticmethod
    def _compute_max_attention_per_index(
        attention_maps: ms.Tensor,
        indices: List[int],
    ) -> List[ms.Tensor]:
        """Computes the maximum attention value for each of the tokens we wish to alter."""
        attention_for_text = attention_maps[:, :, 1:-1]
        attention_for_text *= 100
        attention_for_text = mint.nn.functional.softmax(attention_for_text, dim=-1)

        # Shift indices since we removed the first token
        indices = [index - 1 for index in indices]

        # Extract the maximum values
        max_indices_list = []
        for i in indices:
            image = attention_for_text[:, :, i]
            smoothing = GaussianSmoothing()
            input = mint.nn.functional.pad(image.unsqueeze(0).unsqueeze(0), (1, 1, 1, 1), mode="reflect")
            image = smoothing(input).squeeze(0).squeeze(0)
            max_indices_list.append(image.max())
        return max_indices_list

    def _aggregate_and_get_max_attention_per_token(
        self,
        indices: List[int],
    ):
        """Aggregates the attention for each token and computes the max activation value for each token to alter."""
        attention_maps = self.attention_store.aggregate_attention(
            from_where=("up", "down", "mid"),
        )
        max_attention_per_index = self._compute_max_attention_per_index(
            attention_maps=attention_maps,
            indices=indices,
        )
        return max_attention_per_index

    @staticmethod
    def _compute_loss(max_attention_per_index: List[ms.Tensor]) -> ms.Tensor:
        """Computes the attend-and-excite loss using the maximum attention value for each token."""
        losses = [max(0, 1.0 - curr_max) for curr_max in max_attention_per_index]
        loss = max(losses)
        return loss

    def _update_latent(self, latents: ms.Tensor, grad_cond: ms.Tensor, step_size: float) -> ms.Tensor:
        """Update the latent according to the computed loss."""
        latents = latents - ms.tensor(step_size, dtype=latents.dtype) * grad_cond
        return latents

    def _perform_iterative_refinement_step(
        self,
        latents: ms.Tensor,
        indices: List[int],
        loss: ms.Tensor,
        threshold: float,
        text_embeddings: ms.Tensor,
        step_size: float,
        t: int,
        max_refinement_steps: int = 20,
    ):
        """
        Performs the iterative latent refinement introduced in the paper. Here, we continuously update the latent code
        according to our loss objective until the given threshold is reached for all tokens.
        """
        iteration = 0
        target_loss = max(0, 1.0 - threshold)
        while loss > target_loss:
            iteration += 1

            latents = ops.stop_gradient(latents.clone())

            grad_fn = ms.value_and_grad(self.iterative_refinement_step, grad_position=0)
            loss, grad_cond = grad_fn(latents, t, text_embeddings, indices)

            if loss != 0:
                latents = self._update_latent(latents, grad_cond, step_size)

            logger.info(f"\t Try {iteration}. loss: {loss}")

            if iteration >= max_refinement_steps:
                logger.info(f"\t Exceeded max number of iterations ({max_refinement_steps})! ")
                break

        # Run one more time but don't compute gradients and update the latents.
        # We just need to compute the new loss - the grad update will occur below
        latents = ops.stop_gradient(latents.clone())
        _ = self.unet(latents, t, encoder_hidden_states=text_embeddings)[0]
        # self.unet.zero_grad()

        # Get max activation value for each subject token
        max_attention_per_index = self._aggregate_and_get_max_attention_per_token(
            indices=indices,
        )
        loss = self._compute_loss(max_attention_per_index)
        logger.info(f"\t Finished with loss of: {loss}")
        return loss, latents, max_attention_per_index

    def register_attention_control(self):
        attn_procs = {}
        cross_att_count = 0
        for name in self.unet.attn_processors.keys():
            if name.startswith("mid_block"):
                place_in_unet = "mid"
            elif name.startswith("up_blocks"):
                place_in_unet = "up"
            elif name.startswith("down_blocks"):
                place_in_unet = "down"
            else:
                continue

            cross_att_count += 1
            attn_procs[name] = AttendExciteAttnProcessor(attnstore=self.attention_store, place_in_unet=place_in_unet)

        self.unet.set_attn_processor(attn_procs)
        self.attention_store.num_att_layers = cross_att_count

    def get_indices(self, prompt: str) -> Dict[str, int]:
        """Utility function to list the indices of the tokens you wish to alte"""
        ids = self.tokenizer(prompt).input_ids
        indices = {i: tok for tok, i in zip(self.tokenizer.convert_ids_to_tokens(ids), range(len(ids)))}
        return indices

    def iterative_refinement_step(self, latents, t, text_embeddings, indices):
        self.unet(latents, t, encoder_hidden_states=text_embeddings)[0]
        # self.unet.zero_grad()

        # Get max activation value for each subject token
        max_attention_per_index = self._aggregate_and_get_max_attention_per_token(
            indices=indices,
        )

        loss = self._compute_loss(max_attention_per_index)

        return loss

    def train_step(self, latent, index, text_embedding, t, cross_attention_kwargs, thresholds, i, step_size):
        self.unet(
            latent,
            t,
            encoder_hidden_states=text_embedding,
            cross_attention_kwargs=cross_attention_kwargs,
        )[0]
        # self.unet.zero_grad()

        # Get max activation value for each subject token
        max_attention_per_index = self._aggregate_and_get_max_attention_per_token(
            indices=index,
        )

        loss = self._compute_loss(max_attention_per_index=max_attention_per_index)

        # If this is an iterative refinement step, verify we have reached the desired threshold for all
        if i in thresholds.keys() and loss > 1.0 - thresholds[i]:
            loss, latent, max_attention_per_index = self._perform_iterative_refinement_step(
                latents=latent,
                indices=index,
                loss=loss,
                threshold=thresholds[i],
                text_embeddings=text_embedding,
                step_size=step_size[i],
                t=t,
            )

        return loss

    def __call__(
        self,
        prompt: Union[str, List[str]],
        token_indices: Union[List[int], List[List[int]]],
        height: Optional[int] = None,
        width: Optional[int] = None,
        num_inference_steps: int = 50,
        guidance_scale: float = 7.5,
        negative_prompt: Optional[Union[str, List[str]]] = None,
        num_images_per_prompt: int = 1,
        eta: float = 0.0,
        generator: Optional[Union[np.random.Generator, List[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,
        max_iter_to_alter: int = 25,
        thresholds: dict = {0: 0.05, 10: 0.5, 20: 0.8},
        scale_factor: int = 20,
        attn_res: Optional[Tuple[int]] = (16, 16),
        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 image generation. If not defined, you need to pass `prompt_embeds`.
            token_indices (`List[int]`):
                The token indices to alter with attend-and-excite.
            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 50):
                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 7.5):
                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.stable_diffusion.StableDiffusionPipelineOutput`] 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).
            max_iter_to_alter (`int`, *optional*, defaults to `25`):
                Number of denoising steps to apply attend-and-excite. The `max_iter_to_alter` denoising steps are when
                attend-and-excite is applied. For example, if `max_iter_to_alter` is `25` and there are a total of `30`
                denoising steps, the first `25` denoising steps applies attend-and-excite and the last `5` will not.
            thresholds (`dict`, *optional*, defaults to `{0: 0.05, 10: 0.5, 20: 0.8}`):
                Dictionary defining the iterations and desired thresholds to apply iterative latent refinement in.
            scale_factor (`int`, *optional*, default to 20):
                Scale factor to control the step size of each attend-and-excite update.
            attn_res (`tuple`, *optional*, default computed from width and height):
                The 2D resolution of the semantic attention map.
            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.stable_diffusion.StableDiffusionPipelineOutput`] or `tuple`:
                If `return_dict` is `True`, [`~pipelines.stable_diffusion.StableDiffusionPipelineOutput`] is returned,
                otherwise a `tuple` is returned where the first element is a list with the generated images and the
                second element is a list of `bool`s indicating whether the corresponding generated image contains
                "not-safe-for-work" (nsfw) content.
        """

        # 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,
            token_indices,
            height,
            width,
            callback_steps,
            negative_prompt,
            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]

        # 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
        prompt_embeds, negative_prompt_embeds = self.encode_prompt(
            prompt,
            num_images_per_prompt,
            do_classifier_free_guidance,
            negative_prompt,
            prompt_embeds=prompt_embeds,
            negative_prompt_embeds=negative_prompt_embeds,
            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 = mint.cat([negative_prompt_embeds, prompt_embeds])

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

        # 5. Prepare latent variables
        num_channels_latents = self.unet.config.in_channels
        latents = self.prepare_latents(
            batch_size * num_images_per_prompt,
            num_channels_latents,
            height,
            width,
            prompt_embeds.dtype,
            generator,
            latents,
        )

        # 6. 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)

        if attn_res is None:
            attn_res = int(np.ceil(width / 32)), int(np.ceil(height / 32))
        self.attention_store = AttentionStore(attn_res)
        original_attn_proc = self.unet.attn_processors
        self.register_attention_control()

        # default config for step size from original repo
        scale_range = np.linspace(1.0, 0.5, len(self.scheduler.timesteps))
        step_size = scale_factor * np.sqrt(scale_range)

        text_embeddings = (
            prompt_embeds[batch_size * num_images_per_prompt :] if do_classifier_free_guidance else prompt_embeds
        )

        if isinstance(token_indices[0], int):
            token_indices = [token_indices]

        indices = []

        for ind in token_indices:
            indices = indices + [ind] * num_images_per_prompt

        # 7. Denoising loop
        num_warmup_steps = len(timesteps) - num_inference_steps * self.scheduler.order
        with self.progress_bar(total=num_inference_steps) as progress_bar:
            for i, t in enumerate(timesteps):
                # Attend and excite process
                latents = ops.stop_gradient(latents.clone())
                updated_latents = []
                for latent, index, text_embedding in zip(latents, indices, text_embeddings):
                    for name in self.unet.attn_processors.keys():
                        self.unet.attn_processors[name].requires_grad = True
                    # Forward pass of denoising with text conditioning
                    latent = latent.unsqueeze(0)
                    text_embedding = text_embedding.unsqueeze(0)
                    grad_fn = ms.value_and_grad(self.train_step, grad_position=0)
                    loss, grad_cond = grad_fn(
                        latent, index, text_embedding, t, cross_attention_kwargs, thresholds, i, step_size
                    )

                    # Perform gradient update
                    if i < max_iter_to_alter:
                        if loss != 0:
                            latent = self._update_latent(
                                latents=latent,
                                grad_cond=grad_cond,
                                step_size=step_size[i],
                            )
                        logger.info(f"Iteration {i} | Loss: {loss.numpy():0.4f}")

                    updated_latents.append(latent)

                latents = mint.cat(updated_latents, dim=0)

                # expand the latents if we are doing classifier free guidance
                latent_model_input = mint.cat([latents] * 2) if do_classifier_free_guidance else latents
                latent_model_input = self.scheduler.scale_model_input(latent_model_input, t)

                for name in self.unet.attn_processors.keys():
                    self.unet.attn_processors[name].requires_grad = False

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

                # call the callback, if provided
                if i == len(timesteps) - 1 or ((i + 1) > num_warmup_steps and (i + 1) % self.scheduler.order == 0):
                    progress_bar.update()
                    if callback is not None and i % callback_steps == 0:
                        step_idx = i // getattr(self.scheduler, "order", 1)
                        callback(step_idx, t, latents)

        # 8. Post-processing
        if not output_type == "latent":
            image = self.vae.decode(latents / self.vae.config.scaling_factor, return_dict=False)[0]
            image, has_nsfw_concept = self.run_safety_checker(image, prompt_embeds.dtype)
        else:
            image = latents
            has_nsfw_concept = None

        if has_nsfw_concept is None:
            do_denormalize = [True] * image.shape[0]
        else:
            do_denormalize = [not has_nsfw for has_nsfw in has_nsfw_concept]

        image = self.image_processor.postprocess(image, output_type=output_type, do_denormalize=do_denormalize)
        # make sure to set the original attention processors back
        self.unet.set_attn_processor(original_attn_proc)

        if not return_dict:
            return (image, has_nsfw_concept)

        return StableDiffusionPipelineOutput(images=image, nsfw_content_detected=has_nsfw_concept)

mindone.diffusers.StableDiffusionAttendAndExcitePipeline.__call__(prompt, token_indices, height=None, width=None, num_inference_steps=50, guidance_scale=7.5, 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, max_iter_to_alter=25, thresholds={0: 0.05, 10: 0.5, 20: 0.8}, scale_factor=20, attn_res=(16, 16), 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*

token_indices

The token indices to alter with attend-and-excite.

TYPE: `List[int]`

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 50 DEFAULT: 50

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 7.5 DEFAULT: 7.5

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.stable_diffusion.StableDiffusionPipelineOutput] 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

max_iter_to_alter

Number of denoising steps to apply attend-and-excite. The max_iter_to_alter denoising steps are when attend-and-excite is applied. For example, if max_iter_to_alter is 25 and there are a total of 30 denoising steps, the first 25 denoising steps applies attend-and-excite and the last 5 will not.

TYPE: `int`, *optional*, defaults to `25` DEFAULT: 25

thresholds

0.05, 10: 0.5, 20: 0.8}`): Dictionary defining the iterations and desired thresholds to apply iterative latent refinement in.

TYPE: `dict`, *optional*, defaults to `{0 DEFAULT: {0: 0.05, 10: 0.5, 20: 0.8}

scale_factor

Scale factor to control the step size of each attend-and-excite update.

TYPE: `int`, *optional*, default to 20 DEFAULT: 20

attn_res

The 2D resolution of the semantic attention map.

TYPE: `tuple`, *optional*, default computed from width and height DEFAULT: (16, 16)

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.stable_diffusion.StableDiffusionPipelineOutput] or tuple: If return_dict is True, [~pipelines.stable_diffusion.StableDiffusionPipelineOutput] is returned, otherwise a tuple is returned where the first element is a list with the generated images and the second element is a list of bools indicating whether the corresponding generated image contains "not-safe-for-work" (nsfw) content.

Source code in mindone/diffusers/pipelines/stable_diffusion_attend_and_excite/pipeline_stable_diffusion_attend_and_excite.py
 766
 767
 768
 769
 770
 771
 772
 773
 774
 775
 776
 777
 778
 779
 780
 781
 782
 783
 784
 785
 786
 787
 788
 789
 790
 791
 792
 793
 794
 795
 796
 797
 798
 799
 800
 801
 802
 803
 804
 805
 806
 807
 808
 809
 810
 811
 812
 813
 814
 815
 816
 817
 818
 819
 820
 821
 822
 823
 824
 825
 826
 827
 828
 829
 830
 831
 832
 833
 834
 835
 836
 837
 838
 839
 840
 841
 842
 843
 844
 845
 846
 847
 848
 849
 850
 851
 852
 853
 854
 855
 856
 857
 858
 859
 860
 861
 862
 863
 864
 865
 866
 867
 868
 869
 870
 871
 872
 873
 874
 875
 876
 877
 878
 879
 880
 881
 882
 883
 884
 885
 886
 887
 888
 889
 890
 891
 892
 893
 894
 895
 896
 897
 898
 899
 900
 901
 902
 903
 904
 905
 906
 907
 908
 909
 910
 911
 912
 913
 914
 915
 916
 917
 918
 919
 920
 921
 922
 923
 924
 925
 926
 927
 928
 929
 930
 931
 932
 933
 934
 935
 936
 937
 938
 939
 940
 941
 942
 943
 944
 945
 946
 947
 948
 949
 950
 951
 952
 953
 954
 955
 956
 957
 958
 959
 960
 961
 962
 963
 964
 965
 966
 967
 968
 969
 970
 971
 972
 973
 974
 975
 976
 977
 978
 979
 980
 981
 982
 983
 984
 985
 986
 987
 988
 989
 990
 991
 992
 993
 994
 995
 996
 997
 998
 999
1000
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
def __call__(
    self,
    prompt: Union[str, List[str]],
    token_indices: Union[List[int], List[List[int]]],
    height: Optional[int] = None,
    width: Optional[int] = None,
    num_inference_steps: int = 50,
    guidance_scale: float = 7.5,
    negative_prompt: Optional[Union[str, List[str]]] = None,
    num_images_per_prompt: int = 1,
    eta: float = 0.0,
    generator: Optional[Union[np.random.Generator, List[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,
    max_iter_to_alter: int = 25,
    thresholds: dict = {0: 0.05, 10: 0.5, 20: 0.8},
    scale_factor: int = 20,
    attn_res: Optional[Tuple[int]] = (16, 16),
    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 image generation. If not defined, you need to pass `prompt_embeds`.
        token_indices (`List[int]`):
            The token indices to alter with attend-and-excite.
        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 50):
            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 7.5):
            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.stable_diffusion.StableDiffusionPipelineOutput`] 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).
        max_iter_to_alter (`int`, *optional*, defaults to `25`):
            Number of denoising steps to apply attend-and-excite. The `max_iter_to_alter` denoising steps are when
            attend-and-excite is applied. For example, if `max_iter_to_alter` is `25` and there are a total of `30`
            denoising steps, the first `25` denoising steps applies attend-and-excite and the last `5` will not.
        thresholds (`dict`, *optional*, defaults to `{0: 0.05, 10: 0.5, 20: 0.8}`):
            Dictionary defining the iterations and desired thresholds to apply iterative latent refinement in.
        scale_factor (`int`, *optional*, default to 20):
            Scale factor to control the step size of each attend-and-excite update.
        attn_res (`tuple`, *optional*, default computed from width and height):
            The 2D resolution of the semantic attention map.
        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.stable_diffusion.StableDiffusionPipelineOutput`] or `tuple`:
            If `return_dict` is `True`, [`~pipelines.stable_diffusion.StableDiffusionPipelineOutput`] is returned,
            otherwise a `tuple` is returned where the first element is a list with the generated images and the
            second element is a list of `bool`s indicating whether the corresponding generated image contains
            "not-safe-for-work" (nsfw) content.
    """

    # 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,
        token_indices,
        height,
        width,
        callback_steps,
        negative_prompt,
        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]

    # 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
    prompt_embeds, negative_prompt_embeds = self.encode_prompt(
        prompt,
        num_images_per_prompt,
        do_classifier_free_guidance,
        negative_prompt,
        prompt_embeds=prompt_embeds,
        negative_prompt_embeds=negative_prompt_embeds,
        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 = mint.cat([negative_prompt_embeds, prompt_embeds])

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

    # 5. Prepare latent variables
    num_channels_latents = self.unet.config.in_channels
    latents = self.prepare_latents(
        batch_size * num_images_per_prompt,
        num_channels_latents,
        height,
        width,
        prompt_embeds.dtype,
        generator,
        latents,
    )

    # 6. 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)

    if attn_res is None:
        attn_res = int(np.ceil(width / 32)), int(np.ceil(height / 32))
    self.attention_store = AttentionStore(attn_res)
    original_attn_proc = self.unet.attn_processors
    self.register_attention_control()

    # default config for step size from original repo
    scale_range = np.linspace(1.0, 0.5, len(self.scheduler.timesteps))
    step_size = scale_factor * np.sqrt(scale_range)

    text_embeddings = (
        prompt_embeds[batch_size * num_images_per_prompt :] if do_classifier_free_guidance else prompt_embeds
    )

    if isinstance(token_indices[0], int):
        token_indices = [token_indices]

    indices = []

    for ind in token_indices:
        indices = indices + [ind] * num_images_per_prompt

    # 7. Denoising loop
    num_warmup_steps = len(timesteps) - num_inference_steps * self.scheduler.order
    with self.progress_bar(total=num_inference_steps) as progress_bar:
        for i, t in enumerate(timesteps):
            # Attend and excite process
            latents = ops.stop_gradient(latents.clone())
            updated_latents = []
            for latent, index, text_embedding in zip(latents, indices, text_embeddings):
                for name in self.unet.attn_processors.keys():
                    self.unet.attn_processors[name].requires_grad = True
                # Forward pass of denoising with text conditioning
                latent = latent.unsqueeze(0)
                text_embedding = text_embedding.unsqueeze(0)
                grad_fn = ms.value_and_grad(self.train_step, grad_position=0)
                loss, grad_cond = grad_fn(
                    latent, index, text_embedding, t, cross_attention_kwargs, thresholds, i, step_size
                )

                # Perform gradient update
                if i < max_iter_to_alter:
                    if loss != 0:
                        latent = self._update_latent(
                            latents=latent,
                            grad_cond=grad_cond,
                            step_size=step_size[i],
                        )
                    logger.info(f"Iteration {i} | Loss: {loss.numpy():0.4f}")

                updated_latents.append(latent)

            latents = mint.cat(updated_latents, dim=0)

            # expand the latents if we are doing classifier free guidance
            latent_model_input = mint.cat([latents] * 2) if do_classifier_free_guidance else latents
            latent_model_input = self.scheduler.scale_model_input(latent_model_input, t)

            for name in self.unet.attn_processors.keys():
                self.unet.attn_processors[name].requires_grad = False

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

            # call the callback, if provided
            if i == len(timesteps) - 1 or ((i + 1) > num_warmup_steps and (i + 1) % self.scheduler.order == 0):
                progress_bar.update()
                if callback is not None and i % callback_steps == 0:
                    step_idx = i // getattr(self.scheduler, "order", 1)
                    callback(step_idx, t, latents)

    # 8. Post-processing
    if not output_type == "latent":
        image = self.vae.decode(latents / self.vae.config.scaling_factor, return_dict=False)[0]
        image, has_nsfw_concept = self.run_safety_checker(image, prompt_embeds.dtype)
    else:
        image = latents
        has_nsfw_concept = None

    if has_nsfw_concept is None:
        do_denormalize = [True] * image.shape[0]
    else:
        do_denormalize = [not has_nsfw for has_nsfw in has_nsfw_concept]

    image = self.image_processor.postprocess(image, output_type=output_type, do_denormalize=do_denormalize)
    # make sure to set the original attention processors back
    self.unet.set_attn_processor(original_attn_proc)

    if not return_dict:
        return (image, has_nsfw_concept)

    return StableDiffusionPipelineOutput(images=image, nsfw_content_detected=has_nsfw_concept)

mindone.diffusers.StableDiffusionAttendAndExcitePipeline.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_attend_and_excite/pipeline_stable_diffusion_attend_and_excite.py
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
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.StableDiffusionAttendAndExcitePipeline.get_indices(prompt)

Utility function to list the indices of the tokens you wish to alte

Source code in mindone/diffusers/pipelines/stable_diffusion_attend_and_excite/pipeline_stable_diffusion_attend_and_excite.py
717
718
719
720
721
def get_indices(self, prompt: str) -> Dict[str, int]:
    """Utility function to list the indices of the tokens you wish to alte"""
    ids = self.tokenizer(prompt).input_ids
    indices = {i: tok for tok, i in zip(self.tokenizer.convert_ids_to_tokens(ids), range(len(ids)))}
    return indices

mindone.diffusers.pipelines.stable_diffusion.StableDiffusionPipelineOutput dataclass

Bases: BaseOutput

Output class for Stable Diffusion pipelines.

Source code in mindone/diffusers/pipelines/stable_diffusion/pipeline_output.py
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
@dataclass
class StableDiffusionPipelineOutput(BaseOutput):
    """
    Output class for Stable Diffusion 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)`.
        nsfw_content_detected (`List[bool]`)
            List indicating whether the corresponding generated image contains "not-safe-for-work" (nsfw) content or
            `None` if safety checking could not be performed.
    """

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