From c25967ce2e0eb018476cd1414845946cd01ac6ab Mon Sep 17 00:00:00 2001 From: Patrick von Platen Date: Tue, 21 Nov 2023 11:17:29 +0000 Subject: [PATCH] Make vae robust --- .../pipelines/alt_diffusion/pipeline_alt_diffusion.py | 2 +- .../pipelines/alt_diffusion/pipeline_alt_diffusion_img2img.py | 2 +- src/diffusers/pipelines/animatediff/pipeline_animatediff.py | 2 +- src/diffusers/pipelines/audioldm/pipeline_audioldm.py | 2 +- src/diffusers/pipelines/audioldm2/pipeline_audioldm2.py | 2 +- src/diffusers/pipelines/controlnet/pipeline_controlnet.py | 2 +- .../pipelines/controlnet/pipeline_controlnet_img2img.py | 2 +- .../pipelines/controlnet/pipeline_controlnet_inpaint.py | 2 +- .../pipelines/controlnet/pipeline_controlnet_inpaint_sd_xl.py | 2 +- .../pipelines/controlnet/pipeline_controlnet_sd_xl.py | 4 +++- .../pipelines/controlnet/pipeline_controlnet_sd_xl_img2img.py | 2 +- .../pipelines/controlnet/pipeline_flax_controlnet.py | 2 +- .../pipeline_latent_consistency_img2img.py | 2 +- .../pipeline_latent_consistency_text2img.py | 2 +- src/diffusers/pipelines/musicldm/pipeline_musicldm.py | 2 +- .../pipelines/paint_by_example/pipeline_paint_by_example.py | 2 +- src/diffusers/pipelines/pixart_alpha/pipeline_pixart_alpha.py | 2 +- .../pipeline_semantic_stable_diffusion.py | 2 +- .../pipelines/stable_diffusion/pipeline_cycle_diffusion.py | 2 +- .../pipelines/stable_diffusion/pipeline_stable_diffusion.py | 2 +- .../pipeline_stable_diffusion_attend_and_excite.py | 2 +- .../stable_diffusion/pipeline_stable_diffusion_depth2img.py | 2 +- .../stable_diffusion/pipeline_stable_diffusion_diffedit.py | 2 +- .../stable_diffusion/pipeline_stable_diffusion_gligen.py | 2 +- .../pipeline_stable_diffusion_gligen_text_image.py | 2 +- .../pipeline_stable_diffusion_image_variation.py | 2 +- .../stable_diffusion/pipeline_stable_diffusion_img2img.py | 2 +- .../stable_diffusion/pipeline_stable_diffusion_inpaint.py | 2 +- .../pipeline_stable_diffusion_inpaint_legacy.py | 2 +- .../pipeline_stable_diffusion_instruct_pix2pix.py | 2 +- .../stable_diffusion/pipeline_stable_diffusion_k_diffusion.py | 2 +- .../pipeline_stable_diffusion_latent_upscale.py | 2 +- .../stable_diffusion/pipeline_stable_diffusion_ldm3d.py | 2 +- .../pipeline_stable_diffusion_model_editing.py | 2 +- .../stable_diffusion/pipeline_stable_diffusion_panorama.py | 2 +- .../stable_diffusion/pipeline_stable_diffusion_paradigms.py | 2 +- .../pipeline_stable_diffusion_pix2pix_zero.py | 2 +- .../stable_diffusion/pipeline_stable_diffusion_sag.py | 2 +- .../stable_diffusion/pipeline_stable_diffusion_upscale.py | 2 +- .../pipelines/stable_diffusion/pipeline_stable_unclip.py | 2 +- .../stable_diffusion/pipeline_stable_unclip_img2img.py | 2 +- .../stable_diffusion_safe/pipeline_stable_diffusion_safe.py | 2 +- .../stable_diffusion_xl/pipeline_flax_stable_diffusion_xl.py | 2 +- .../stable_diffusion_xl/pipeline_stable_diffusion_xl.py | 2 +- .../pipeline_stable_diffusion_xl_img2img.py | 2 +- .../pipeline_stable_diffusion_xl_inpaint.py | 2 +- .../pipeline_stable_diffusion_xl_instruct_pix2pix.py | 2 +- .../t2i_adapter/pipeline_stable_diffusion_adapter.py | 2 +- .../t2i_adapter/pipeline_stable_diffusion_xl_adapter.py | 2 +- .../text_to_video_synthesis/pipeline_text_to_video_synth.py | 2 +- .../pipeline_text_to_video_synth_img2img.py | 2 +- src/diffusers/pipelines/unidiffuser/pipeline_unidiffuser.py | 2 +- .../versatile_diffusion/pipeline_versatile_diffusion.py | 2 +- .../pipeline_versatile_diffusion_dual_guided.py | 2 +- .../pipeline_versatile_diffusion_image_variation.py | 2 +- .../pipeline_versatile_diffusion_text_to_image.py | 2 +- 56 files changed, 58 insertions(+), 56 deletions(-) diff --git a/src/diffusers/pipelines/alt_diffusion/pipeline_alt_diffusion.py b/src/diffusers/pipelines/alt_diffusion/pipeline_alt_diffusion.py index 335df9e6f4..863140b7db 100644 --- a/src/diffusers/pipelines/alt_diffusion/pipeline_alt_diffusion.py +++ b/src/diffusers/pipelines/alt_diffusion/pipeline_alt_diffusion.py @@ -198,7 +198,7 @@ class AltDiffusionPipeline(DiffusionPipeline, TextualInversionLoaderMixin, LoraL safety_checker=safety_checker, feature_extractor=feature_extractor, ) - self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1) + self.vae_scale_factor = 2 ** (len(getattr(self.vae.config, 'block_out_channels', self.vae.config.decoder_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) diff --git a/src/diffusers/pipelines/alt_diffusion/pipeline_alt_diffusion_img2img.py b/src/diffusers/pipelines/alt_diffusion/pipeline_alt_diffusion_img2img.py index 7f24bad90f..077e4ee855 100644 --- a/src/diffusers/pipelines/alt_diffusion/pipeline_alt_diffusion_img2img.py +++ b/src/diffusers/pipelines/alt_diffusion/pipeline_alt_diffusion_img2img.py @@ -236,7 +236,7 @@ class AltDiffusionImg2ImgPipeline( safety_checker=safety_checker, feature_extractor=feature_extractor, ) - self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1) + self.vae_scale_factor = 2 ** (len(getattr(self.vae.config, 'block_out_channels', self.vae.config.decoder_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) diff --git a/src/diffusers/pipelines/animatediff/pipeline_animatediff.py b/src/diffusers/pipelines/animatediff/pipeline_animatediff.py index b63acb9a5f..4e106c7fc5 100644 --- a/src/diffusers/pipelines/animatediff/pipeline_animatediff.py +++ b/src/diffusers/pipelines/animatediff/pipeline_animatediff.py @@ -128,7 +128,7 @@ class AnimateDiffPipeline(DiffusionPipeline, TextualInversionLoaderMixin, LoraLo motion_adapter=motion_adapter, scheduler=scheduler, ) - self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1) + self.vae_scale_factor = 2 ** (len(getattr(self.vae.config, 'block_out_channels', self.vae.config.decoder_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 with num_images_per_prompt -> num_videos_per_prompt diff --git a/src/diffusers/pipelines/audioldm/pipeline_audioldm.py b/src/diffusers/pipelines/audioldm/pipeline_audioldm.py index 3345fb6e75..cb15ff9474 100644 --- a/src/diffusers/pipelines/audioldm/pipeline_audioldm.py +++ b/src/diffusers/pipelines/audioldm/pipeline_audioldm.py @@ -93,7 +93,7 @@ class AudioLDMPipeline(DiffusionPipeline): scheduler=scheduler, vocoder=vocoder, ) - self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1) + self.vae_scale_factor = 2 ** (len(getattr(self.vae.config, 'block_out_channels', self.vae.config.decoder_block_out_channels)) - 1) # Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion.StableDiffusionPipeline.enable_vae_slicing def enable_vae_slicing(self): diff --git a/src/diffusers/pipelines/audioldm2/pipeline_audioldm2.py b/src/diffusers/pipelines/audioldm2/pipeline_audioldm2.py index b2dd9f7bb0..2aa7663c19 100644 --- a/src/diffusers/pipelines/audioldm2/pipeline_audioldm2.py +++ b/src/diffusers/pipelines/audioldm2/pipeline_audioldm2.py @@ -171,7 +171,7 @@ class AudioLDM2Pipeline(DiffusionPipeline): scheduler=scheduler, vocoder=vocoder, ) - self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1) + self.vae_scale_factor = 2 ** (len(getattr(self.vae.config, 'block_out_channels', self.vae.config.decoder_block_out_channels)) - 1) # Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion.StableDiffusionPipeline.enable_vae_slicing def enable_vae_slicing(self): diff --git a/src/diffusers/pipelines/controlnet/pipeline_controlnet.py b/src/diffusers/pipelines/controlnet/pipeline_controlnet.py index 04ca51b19f..5b4ac04f1c 100644 --- a/src/diffusers/pipelines/controlnet/pipeline_controlnet.py +++ b/src/diffusers/pipelines/controlnet/pipeline_controlnet.py @@ -173,7 +173,7 @@ class StableDiffusionControlNetPipeline( safety_checker=safety_checker, feature_extractor=feature_extractor, ) - self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1) + self.vae_scale_factor = 2 ** (len(getattr(self.vae.config, 'block_out_channels', self.vae.config.decoder_block_out_channels)) - 1) self.image_processor = VaeImageProcessor(vae_scale_factor=self.vae_scale_factor, do_convert_rgb=True) self.control_image_processor = VaeImageProcessor( vae_scale_factor=self.vae_scale_factor, do_convert_rgb=True, do_normalize=False diff --git a/src/diffusers/pipelines/controlnet/pipeline_controlnet_img2img.py b/src/diffusers/pipelines/controlnet/pipeline_controlnet_img2img.py index 8249c44b4d..975d8b3aaf 100644 --- a/src/diffusers/pipelines/controlnet/pipeline_controlnet_img2img.py +++ b/src/diffusers/pipelines/controlnet/pipeline_controlnet_img2img.py @@ -207,7 +207,7 @@ class StableDiffusionControlNetImg2ImgPipeline( safety_checker=safety_checker, feature_extractor=feature_extractor, ) - self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1) + self.vae_scale_factor = 2 ** (len(getattr(self.vae.config, 'block_out_channels', self.vae.config.decoder_block_out_channels)) - 1) self.image_processor = VaeImageProcessor(vae_scale_factor=self.vae_scale_factor, do_convert_rgb=True) self.control_image_processor = VaeImageProcessor( vae_scale_factor=self.vae_scale_factor, do_convert_rgb=True, do_normalize=False diff --git a/src/diffusers/pipelines/controlnet/pipeline_controlnet_inpaint.py b/src/diffusers/pipelines/controlnet/pipeline_controlnet_inpaint.py index 399cfdcf9c..ceeda0e9bd 100644 --- a/src/diffusers/pipelines/controlnet/pipeline_controlnet_inpaint.py +++ b/src/diffusers/pipelines/controlnet/pipeline_controlnet_inpaint.py @@ -329,7 +329,7 @@ class StableDiffusionControlNetInpaintPipeline( safety_checker=safety_checker, feature_extractor=feature_extractor, ) - self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1) + self.vae_scale_factor = 2 ** (len(getattr(self.vae.config, 'block_out_channels', self.vae.config.decoder_block_out_channels)) - 1) self.image_processor = VaeImageProcessor(vae_scale_factor=self.vae_scale_factor) self.mask_processor = VaeImageProcessor( vae_scale_factor=self.vae_scale_factor, do_normalize=False, do_binarize=True, do_convert_grayscale=True diff --git a/src/diffusers/pipelines/controlnet/pipeline_controlnet_inpaint_sd_xl.py b/src/diffusers/pipelines/controlnet/pipeline_controlnet_inpaint_sd_xl.py index f29d3bd515..feef813cf6 100644 --- a/src/diffusers/pipelines/controlnet/pipeline_controlnet_inpaint_sd_xl.py +++ b/src/diffusers/pipelines/controlnet/pipeline_controlnet_inpaint_sd_xl.py @@ -198,7 +198,7 @@ class StableDiffusionXLControlNetInpaintPipeline( ) self.register_to_config(force_zeros_for_empty_prompt=force_zeros_for_empty_prompt) self.register_to_config(requires_aesthetics_score=requires_aesthetics_score) - self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1) + self.vae_scale_factor = 2 ** (len(getattr(self.vae.config, 'block_out_channels', self.vae.config.decoder_block_out_channels)) - 1) self.image_processor = VaeImageProcessor(vae_scale_factor=self.vae_scale_factor) self.mask_processor = VaeImageProcessor( vae_scale_factor=self.vae_scale_factor, do_normalize=False, do_binarize=True, do_convert_grayscale=True diff --git a/src/diffusers/pipelines/controlnet/pipeline_controlnet_sd_xl.py b/src/diffusers/pipelines/controlnet/pipeline_controlnet_sd_xl.py index d6278c4f04..1b3aa5a0b5 100644 --- a/src/diffusers/pipelines/controlnet/pipeline_controlnet_sd_xl.py +++ b/src/diffusers/pipelines/controlnet/pipeline_controlnet_sd_xl.py @@ -171,7 +171,9 @@ class StableDiffusionXLControlNetPipeline( controlnet=controlnet, scheduler=scheduler, ) - self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1) + out_channels = getattr(self.vae.config, "block_out_channels", self.vae.config.decoder_block_out_channels) + self.vae_scale_factor = 2 ** (len(out_channels) - 1) + self.image_processor = VaeImageProcessor(vae_scale_factor=self.vae_scale_factor, do_convert_rgb=True) self.control_image_processor = VaeImageProcessor( vae_scale_factor=self.vae_scale_factor, do_convert_rgb=True, do_normalize=False diff --git a/src/diffusers/pipelines/controlnet/pipeline_controlnet_sd_xl_img2img.py b/src/diffusers/pipelines/controlnet/pipeline_controlnet_sd_xl_img2img.py index 06b4f6f55e..fb5bf7096e 100644 --- a/src/diffusers/pipelines/controlnet/pipeline_controlnet_sd_xl_img2img.py +++ b/src/diffusers/pipelines/controlnet/pipeline_controlnet_sd_xl_img2img.py @@ -224,7 +224,7 @@ class StableDiffusionXLControlNetImg2ImgPipeline( controlnet=controlnet, scheduler=scheduler, ) - self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1) + self.vae_scale_factor = 2 ** (len(getattr(self.vae.config, 'block_out_channels', self.vae.config.decoder_block_out_channels)) - 1) self.image_processor = VaeImageProcessor(vae_scale_factor=self.vae_scale_factor, do_convert_rgb=True) self.control_image_processor = VaeImageProcessor( vae_scale_factor=self.vae_scale_factor, do_convert_rgb=True, do_normalize=False diff --git a/src/diffusers/pipelines/controlnet/pipeline_flax_controlnet.py b/src/diffusers/pipelines/controlnet/pipeline_flax_controlnet.py index e1f508dc1e..f1f7922db0 100644 --- a/src/diffusers/pipelines/controlnet/pipeline_flax_controlnet.py +++ b/src/diffusers/pipelines/controlnet/pipeline_flax_controlnet.py @@ -175,7 +175,7 @@ class FlaxStableDiffusionControlNetPipeline(FlaxDiffusionPipeline): safety_checker=safety_checker, feature_extractor=feature_extractor, ) - self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1) + self.vae_scale_factor = 2 ** (len(getattr(self.vae.config, 'block_out_channels', self.vae.config.decoder_block_out_channels)) - 1) def prepare_text_inputs(self, prompt: Union[str, List[str]]): if not isinstance(prompt, (str, list)): diff --git a/src/diffusers/pipelines/latent_consistency_models/pipeline_latent_consistency_img2img.py b/src/diffusers/pipelines/latent_consistency_models/pipeline_latent_consistency_img2img.py index 46993fa9a0..e94dfa40a2 100644 --- a/src/diffusers/pipelines/latent_consistency_models/pipeline_latent_consistency_img2img.py +++ b/src/diffusers/pipelines/latent_consistency_models/pipeline_latent_consistency_img2img.py @@ -153,7 +153,7 @@ class LatentConsistencyModelImg2ImgPipeline( " information, please have a look at https://github.com/huggingface/diffusers/pull/254 ." ) - self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1) + self.vae_scale_factor = 2 ** (len(getattr(self.vae.config, 'block_out_channels', self.vae.config.decoder_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.enable_vae_slicing diff --git a/src/diffusers/pipelines/latent_consistency_models/pipeline_latent_consistency_text2img.py b/src/diffusers/pipelines/latent_consistency_models/pipeline_latent_consistency_text2img.py index 884110d162..cb3f8c5ce2 100644 --- a/src/diffusers/pipelines/latent_consistency_models/pipeline_latent_consistency_text2img.py +++ b/src/diffusers/pipelines/latent_consistency_models/pipeline_latent_consistency_text2img.py @@ -140,7 +140,7 @@ class LatentConsistencyModelPipeline( safety_checker=safety_checker, feature_extractor=feature_extractor, ) - self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1) + self.vae_scale_factor = 2 ** (len(getattr(self.vae.config, 'block_out_channels', self.vae.config.decoder_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) diff --git a/src/diffusers/pipelines/musicldm/pipeline_musicldm.py b/src/diffusers/pipelines/musicldm/pipeline_musicldm.py index 9e6b6fea13..5758079e7c 100644 --- a/src/diffusers/pipelines/musicldm/pipeline_musicldm.py +++ b/src/diffusers/pipelines/musicldm/pipeline_musicldm.py @@ -111,7 +111,7 @@ class MusicLDMPipeline(DiffusionPipeline): scheduler=scheduler, vocoder=vocoder, ) - self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1) + self.vae_scale_factor = 2 ** (len(getattr(self.vae.config, 'block_out_channels', self.vae.config.decoder_block_out_channels)) - 1) # Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion.StableDiffusionPipeline.enable_vae_slicing def enable_vae_slicing(self): diff --git a/src/diffusers/pipelines/paint_by_example/pipeline_paint_by_example.py b/src/diffusers/pipelines/paint_by_example/pipeline_paint_by_example.py index ffcbacc3b7..72aef1eff4 100644 --- a/src/diffusers/pipelines/paint_by_example/pipeline_paint_by_example.py +++ b/src/diffusers/pipelines/paint_by_example/pipeline_paint_by_example.py @@ -204,7 +204,7 @@ class PaintByExamplePipeline(DiffusionPipeline): safety_checker=safety_checker, feature_extractor=feature_extractor, ) - self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1) + self.vae_scale_factor = 2 ** (len(getattr(self.vae.config, 'block_out_channels', self.vae.config.decoder_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) diff --git a/src/diffusers/pipelines/pixart_alpha/pipeline_pixart_alpha.py b/src/diffusers/pipelines/pixart_alpha/pipeline_pixart_alpha.py index f4e61bdc94..936f10cc5d 100644 --- a/src/diffusers/pipelines/pixart_alpha/pipeline_pixart_alpha.py +++ b/src/diffusers/pipelines/pixart_alpha/pipeline_pixart_alpha.py @@ -141,7 +141,7 @@ class PixArtAlphaPipeline(DiffusionPipeline): tokenizer=tokenizer, text_encoder=text_encoder, vae=vae, transformer=transformer, scheduler=scheduler ) - self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1) + self.vae_scale_factor = 2 ** (len(getattr(self.vae.config, 'block_out_channels', self.vae.config.decoder_block_out_channels)) - 1) self.image_processor = VaeImageProcessor(vae_scale_factor=self.vae_scale_factor) # Adapted from https://github.com/PixArt-alpha/PixArt-alpha/blob/master/diffusion/model/utils.py diff --git a/src/diffusers/pipelines/semantic_stable_diffusion/pipeline_semantic_stable_diffusion.py b/src/diffusers/pipelines/semantic_stable_diffusion/pipeline_semantic_stable_diffusion.py index 19bd1f1615..fa8d2a7f44 100644 --- a/src/diffusers/pipelines/semantic_stable_diffusion/pipeline_semantic_stable_diffusion.py +++ b/src/diffusers/pipelines/semantic_stable_diffusion/pipeline_semantic_stable_diffusion.py @@ -87,7 +87,7 @@ class SemanticStableDiffusionPipeline(DiffusionPipeline): safety_checker=safety_checker, feature_extractor=feature_extractor, ) - self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1) + self.vae_scale_factor = 2 ** (len(getattr(self.vae.config, 'block_out_channels', self.vae.config.decoder_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) diff --git a/src/diffusers/pipelines/stable_diffusion/pipeline_cycle_diffusion.py b/src/diffusers/pipelines/stable_diffusion/pipeline_cycle_diffusion.py index 9bdb6d824f..7fb76a875f 100644 --- a/src/diffusers/pipelines/stable_diffusion/pipeline_cycle_diffusion.py +++ b/src/diffusers/pipelines/stable_diffusion/pipeline_cycle_diffusion.py @@ -223,7 +223,7 @@ class CycleDiffusionPipeline(DiffusionPipeline, TextualInversionLoaderMixin, Lor safety_checker=safety_checker, feature_extractor=feature_extractor, ) - self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1) + self.vae_scale_factor = 2 ** (len(getattr(self.vae.config, 'block_out_channels', self.vae.config.decoder_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) diff --git a/src/diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion.py b/src/diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion.py index 486841d9db..f6a44ce345 100644 --- a/src/diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion.py +++ b/src/diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion.py @@ -193,7 +193,7 @@ class StableDiffusionPipeline(DiffusionPipeline, TextualInversionLoaderMixin, Lo safety_checker=safety_checker, feature_extractor=feature_extractor, ) - self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1) + self.vae_scale_factor = 2 ** (len(getattr(self.vae.config, 'block_out_channels', self.vae.config.decoder_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) diff --git a/src/diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_attend_and_excite.py b/src/diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_attend_and_excite.py index 2e040306ab..65477e300c 100644 --- a/src/diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_attend_and_excite.py +++ b/src/diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_attend_and_excite.py @@ -238,7 +238,7 @@ class StableDiffusionAttendAndExcitePipeline(DiffusionPipeline, TextualInversion safety_checker=safety_checker, feature_extractor=feature_extractor, ) - self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1) + self.vae_scale_factor = 2 ** (len(getattr(self.vae.config, 'block_out_channels', self.vae.config.decoder_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) diff --git a/src/diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_depth2img.py b/src/diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_depth2img.py index f4debbf422..11043377ba 100644 --- a/src/diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_depth2img.py +++ b/src/diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_depth2img.py @@ -140,7 +140,7 @@ class StableDiffusionDepth2ImgPipeline(DiffusionPipeline, TextualInversionLoader depth_estimator=depth_estimator, feature_extractor=feature_extractor, ) - self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1) + self.vae_scale_factor = 2 ** (len(getattr(self.vae.config, 'block_out_channels', self.vae.config.decoder_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 diff --git a/src/diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_diffedit.py b/src/diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_diffedit.py index 9a9acba214..4a5dc356ff 100644 --- a/src/diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_diffedit.py +++ b/src/diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_diffedit.py @@ -366,7 +366,7 @@ class StableDiffusionDiffEditPipeline(DiffusionPipeline, TextualInversionLoaderM feature_extractor=feature_extractor, inverse_scheduler=inverse_scheduler, ) - self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1) + self.vae_scale_factor = 2 ** (len(getattr(self.vae.config, 'block_out_channels', self.vae.config.decoder_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) diff --git a/src/diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_gligen.py b/src/diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_gligen.py index ef88230b44..247b69b7ab 100644 --- a/src/diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_gligen.py +++ b/src/diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_gligen.py @@ -167,7 +167,7 @@ class StableDiffusionGLIGENPipeline(DiffusionPipeline): safety_checker=safety_checker, feature_extractor=feature_extractor, ) - self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1) + self.vae_scale_factor = 2 ** (len(getattr(self.vae.config, 'block_out_channels', self.vae.config.decoder_block_out_channels)) - 1) self.image_processor = VaeImageProcessor(vae_scale_factor=self.vae_scale_factor, do_convert_rgb=True) self.register_to_config(requires_safety_checker=requires_safety_checker) diff --git a/src/diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_gligen_text_image.py b/src/diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_gligen_text_image.py index 57a1d62c4f..2494ce4fb6 100644 --- a/src/diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_gligen_text_image.py +++ b/src/diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_gligen_text_image.py @@ -225,7 +225,7 @@ class StableDiffusionGLIGENTextImagePipeline(DiffusionPipeline): safety_checker=safety_checker, feature_extractor=feature_extractor, ) - self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1) + self.vae_scale_factor = 2 ** (len(getattr(self.vae.config, 'block_out_channels', self.vae.config.decoder_block_out_channels)) - 1) self.image_processor = VaeImageProcessor(vae_scale_factor=self.vae_scale_factor, do_convert_rgb=True) self.register_to_config(requires_safety_checker=requires_safety_checker) diff --git a/src/diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_image_variation.py b/src/diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_image_variation.py index e8f48a1630..209ef2532f 100644 --- a/src/diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_image_variation.py +++ b/src/diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_image_variation.py @@ -125,7 +125,7 @@ class StableDiffusionImageVariationPipeline(DiffusionPipeline): safety_checker=safety_checker, feature_extractor=feature_extractor, ) - self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1) + self.vae_scale_factor = 2 ** (len(getattr(self.vae.config, 'block_out_channels', self.vae.config.decoder_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) diff --git a/src/diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_img2img.py b/src/diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_img2img.py index d53ff8d3dc..e3e443d01b 100644 --- a/src/diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_img2img.py +++ b/src/diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_img2img.py @@ -230,7 +230,7 @@ class StableDiffusionImg2ImgPipeline( safety_checker=safety_checker, feature_extractor=feature_extractor, ) - self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1) + self.vae_scale_factor = 2 ** (len(getattr(self.vae.config, 'block_out_channels', self.vae.config.decoder_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) diff --git a/src/diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_inpaint.py b/src/diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_inpaint.py index f37dc49c2a..e06c50f4e7 100644 --- a/src/diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_inpaint.py +++ b/src/diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_inpaint.py @@ -298,7 +298,7 @@ class StableDiffusionInpaintPipeline( safety_checker=safety_checker, feature_extractor=feature_extractor, ) - self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1) + self.vae_scale_factor = 2 ** (len(getattr(self.vae.config, 'block_out_channels', self.vae.config.decoder_block_out_channels)) - 1) self.image_processor = VaeImageProcessor(vae_scale_factor=self.vae_scale_factor) self.mask_processor = VaeImageProcessor( vae_scale_factor=self.vae_scale_factor, do_normalize=False, do_binarize=True, do_convert_grayscale=True diff --git a/src/diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_inpaint_legacy.py b/src/diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_inpaint_legacy.py index 55db2a0a7d..04f01d3a18 100644 --- a/src/diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_inpaint_legacy.py +++ b/src/diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_inpaint_legacy.py @@ -212,7 +212,7 @@ class StableDiffusionInpaintPipelineLegacy( safety_checker=safety_checker, feature_extractor=feature_extractor, ) - self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1) + self.vae_scale_factor = 2 ** (len(getattr(self.vae.config, 'block_out_channels', self.vae.config.decoder_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) diff --git a/src/diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_instruct_pix2pix.py b/src/diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_instruct_pix2pix.py index daf5478cd2..74ddcb2592 100644 --- a/src/diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_instruct_pix2pix.py +++ b/src/diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_instruct_pix2pix.py @@ -132,7 +132,7 @@ class StableDiffusionInstructPix2PixPipeline(DiffusionPipeline, TextualInversion safety_checker=safety_checker, feature_extractor=feature_extractor, ) - self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1) + self.vae_scale_factor = 2 ** (len(getattr(self.vae.config, 'block_out_channels', self.vae.config.decoder_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) diff --git a/src/diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_k_diffusion.py b/src/diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_k_diffusion.py index f6742eed2c..99409af9da 100755 --- a/src/diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_k_diffusion.py +++ b/src/diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_k_diffusion.py @@ -116,7 +116,7 @@ class StableDiffusionKDiffusionPipeline(DiffusionPipeline, TextualInversionLoade feature_extractor=feature_extractor, ) self.register_to_config(requires_safety_checker=requires_safety_checker) - self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1) + self.vae_scale_factor = 2 ** (len(getattr(self.vae.config, 'block_out_channels', self.vae.config.decoder_block_out_channels)) - 1) self.image_processor = VaeImageProcessor(vae_scale_factor=self.vae_scale_factor) model = ModelWrapper(unet, scheduler.alphas_cumprod) diff --git a/src/diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_latent_upscale.py b/src/diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_latent_upscale.py index 4cde54ac58..6d6f23778d 100644 --- a/src/diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_latent_upscale.py +++ b/src/diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_latent_upscale.py @@ -98,7 +98,7 @@ class StableDiffusionLatentUpscalePipeline(DiffusionPipeline, FromSingleFileMixi unet=unet, scheduler=scheduler, ) - self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1) + self.vae_scale_factor = 2 ** (len(getattr(self.vae.config, 'block_out_channels', self.vae.config.decoder_block_out_channels)) - 1) self.image_processor = VaeImageProcessor(vae_scale_factor=self.vae_scale_factor, resample="bicubic") def _encode_prompt(self, prompt, device, do_classifier_free_guidance, negative_prompt): diff --git a/src/diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_ldm3d.py b/src/diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_ldm3d.py index f7136a65da..119f654781 100644 --- a/src/diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_ldm3d.py +++ b/src/diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_ldm3d.py @@ -157,7 +157,7 @@ class StableDiffusionLDM3DPipeline( safety_checker=safety_checker, feature_extractor=feature_extractor, ) - self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1) + self.vae_scale_factor = 2 ** (len(getattr(self.vae.config, 'block_out_channels', self.vae.config.decoder_block_out_channels)) - 1) self.image_processor = VaeImageProcessorLDM3D(vae_scale_factor=self.vae_scale_factor) self.register_to_config(requires_safety_checker=requires_safety_checker) diff --git a/src/diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_model_editing.py b/src/diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_model_editing.py index 6ada244211..24c85bbd28 100644 --- a/src/diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_model_editing.py +++ b/src/diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_model_editing.py @@ -113,7 +113,7 @@ class StableDiffusionModelEditingPipeline(DiffusionPipeline, TextualInversionLoa safety_checker=safety_checker, feature_extractor=feature_extractor, ) - self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1) + self.vae_scale_factor = 2 ** (len(getattr(self.vae.config, 'block_out_channels', self.vae.config.decoder_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) diff --git a/src/diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_panorama.py b/src/diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_panorama.py index ce3e694e7e..80d29da641 100644 --- a/src/diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_panorama.py +++ b/src/diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_panorama.py @@ -127,7 +127,7 @@ class StableDiffusionPanoramaPipeline(DiffusionPipeline, TextualInversionLoaderM safety_checker=safety_checker, feature_extractor=feature_extractor, ) - self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1) + self.vae_scale_factor = 2 ** (len(getattr(self.vae.config, 'block_out_channels', self.vae.config.decoder_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) diff --git a/src/diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_paradigms.py b/src/diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_paradigms.py index 5e8c6ce58d..2ff32854ef 100644 --- a/src/diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_paradigms.py +++ b/src/diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_paradigms.py @@ -138,7 +138,7 @@ class StableDiffusionParadigmsPipeline( safety_checker=safety_checker, feature_extractor=feature_extractor, ) - self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1) + self.vae_scale_factor = 2 ** (len(getattr(self.vae.config, 'block_out_channels', self.vae.config.decoder_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) diff --git a/src/diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_pix2pix_zero.py b/src/diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_pix2pix_zero.py index 4877041399..221cfc2c61 100644 --- a/src/diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_pix2pix_zero.py +++ b/src/diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_pix2pix_zero.py @@ -364,7 +364,7 @@ class StableDiffusionPix2PixZeroPipeline(DiffusionPipeline): caption_generator=caption_generator, inverse_scheduler=inverse_scheduler, ) - self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1) + self.vae_scale_factor = 2 ** (len(getattr(self.vae.config, 'block_out_channels', self.vae.config.decoder_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) diff --git a/src/diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_sag.py b/src/diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_sag.py index 56eb38c653..e0b7430532 100644 --- a/src/diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_sag.py +++ b/src/diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_sag.py @@ -150,7 +150,7 @@ class StableDiffusionSAGPipeline(DiffusionPipeline, TextualInversionLoaderMixin) safety_checker=safety_checker, feature_extractor=feature_extractor, ) - self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1) + self.vae_scale_factor = 2 ** (len(getattr(self.vae.config, 'block_out_channels', self.vae.config.decoder_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) diff --git a/src/diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_upscale.py b/src/diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_upscale.py index da89505017..eb955069fe 100644 --- a/src/diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_upscale.py +++ b/src/diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_upscale.py @@ -140,7 +140,7 @@ class StableDiffusionUpscalePipeline( watermarker=watermarker, feature_extractor=feature_extractor, ) - self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1) + self.vae_scale_factor = 2 ** (len(getattr(self.vae.config, 'block_out_channels', self.vae.config.decoder_block_out_channels)) - 1) self.image_processor = VaeImageProcessor(vae_scale_factor=self.vae_scale_factor, resample="bicubic") self.register_to_config(max_noise_level=max_noise_level) diff --git a/src/diffusers/pipelines/stable_diffusion/pipeline_stable_unclip.py b/src/diffusers/pipelines/stable_diffusion/pipeline_stable_unclip.py index eb4542888c..77dd7b144c 100644 --- a/src/diffusers/pipelines/stable_diffusion/pipeline_stable_unclip.py +++ b/src/diffusers/pipelines/stable_diffusion/pipeline_stable_unclip.py @@ -147,7 +147,7 @@ class StableUnCLIPPipeline(DiffusionPipeline, TextualInversionLoaderMixin, LoraL vae=vae, ) - self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1) + self.vae_scale_factor = 2 ** (len(getattr(self.vae.config, 'block_out_channels', self.vae.config.decoder_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.enable_vae_slicing diff --git a/src/diffusers/pipelines/stable_diffusion/pipeline_stable_unclip_img2img.py b/src/diffusers/pipelines/stable_diffusion/pipeline_stable_unclip_img2img.py index 73638fdd15..ff234e609e 100644 --- a/src/diffusers/pipelines/stable_diffusion/pipeline_stable_unclip_img2img.py +++ b/src/diffusers/pipelines/stable_diffusion/pipeline_stable_unclip_img2img.py @@ -148,7 +148,7 @@ class StableUnCLIPImg2ImgPipeline(DiffusionPipeline, TextualInversionLoaderMixin vae=vae, ) - self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1) + self.vae_scale_factor = 2 ** (len(getattr(self.vae.config, 'block_out_channels', self.vae.config.decoder_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.enable_vae_slicing diff --git a/src/diffusers/pipelines/stable_diffusion_safe/pipeline_stable_diffusion_safe.py b/src/diffusers/pipelines/stable_diffusion_safe/pipeline_stable_diffusion_safe.py index eb24cbfd94..3faf558513 100644 --- a/src/diffusers/pipelines/stable_diffusion_safe/pipeline_stable_diffusion_safe.py +++ b/src/diffusers/pipelines/stable_diffusion_safe/pipeline_stable_diffusion_safe.py @@ -142,7 +142,7 @@ class StableDiffusionPipelineSafe(DiffusionPipeline): feature_extractor=feature_extractor, ) self._safety_text_concept = safety_concept - self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1) + self.vae_scale_factor = 2 ** (len(getattr(self.vae.config, 'block_out_channels', self.vae.config.decoder_block_out_channels)) - 1) self.register_to_config(requires_safety_checker=requires_safety_checker) @property diff --git a/src/diffusers/pipelines/stable_diffusion_xl/pipeline_flax_stable_diffusion_xl.py b/src/diffusers/pipelines/stable_diffusion_xl/pipeline_flax_stable_diffusion_xl.py index 8f043c7c66..b426acb3af 100644 --- a/src/diffusers/pipelines/stable_diffusion_xl/pipeline_flax_stable_diffusion_xl.py +++ b/src/diffusers/pipelines/stable_diffusion_xl/pipeline_flax_stable_diffusion_xl.py @@ -65,7 +65,7 @@ class FlaxStableDiffusionXLPipeline(FlaxDiffusionPipeline): unet=unet, scheduler=scheduler, ) - self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1) + self.vae_scale_factor = 2 ** (len(getattr(self.vae.config, 'block_out_channels', self.vae.config.decoder_block_out_channels)) - 1) def prepare_inputs(self, prompt: Union[str, List[str]]): if not isinstance(prompt, (str, list)): diff --git a/src/diffusers/pipelines/stable_diffusion_xl/pipeline_stable_diffusion_xl.py b/src/diffusers/pipelines/stable_diffusion_xl/pipeline_stable_diffusion_xl.py index c9e68976cc..8e8b82967f 100644 --- a/src/diffusers/pipelines/stable_diffusion_xl/pipeline_stable_diffusion_xl.py +++ b/src/diffusers/pipelines/stable_diffusion_xl/pipeline_stable_diffusion_xl.py @@ -176,7 +176,7 @@ class StableDiffusionXLPipeline( scheduler=scheduler, ) self.register_to_config(force_zeros_for_empty_prompt=force_zeros_for_empty_prompt) - self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1) + self.vae_scale_factor = 2 ** (len(getattr(self.vae.config, 'block_out_channels', self.vae.config.decoder_block_out_channels)) - 1) self.image_processor = VaeImageProcessor(vae_scale_factor=self.vae_scale_factor) self.default_sample_size = self.unet.config.sample_size diff --git a/src/diffusers/pipelines/stable_diffusion_xl/pipeline_stable_diffusion_xl_img2img.py b/src/diffusers/pipelines/stable_diffusion_xl/pipeline_stable_diffusion_xl_img2img.py index 40eabca82b..9584aebb1d 100644 --- a/src/diffusers/pipelines/stable_diffusion_xl/pipeline_stable_diffusion_xl_img2img.py +++ b/src/diffusers/pipelines/stable_diffusion_xl/pipeline_stable_diffusion_xl_img2img.py @@ -191,7 +191,7 @@ class StableDiffusionXLImg2ImgPipeline( ) self.register_to_config(force_zeros_for_empty_prompt=force_zeros_for_empty_prompt) self.register_to_config(requires_aesthetics_score=requires_aesthetics_score) - self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1) + self.vae_scale_factor = 2 ** (len(getattr(self.vae.config, 'block_out_channels', self.vae.config.decoder_block_out_channels)) - 1) self.image_processor = VaeImageProcessor(vae_scale_factor=self.vae_scale_factor) add_watermarker = add_watermarker if add_watermarker is not None else is_invisible_watermark_available() diff --git a/src/diffusers/pipelines/stable_diffusion_xl/pipeline_stable_diffusion_xl_inpaint.py b/src/diffusers/pipelines/stable_diffusion_xl/pipeline_stable_diffusion_xl_inpaint.py index 97cd16b5fa..c44f84efb8 100644 --- a/src/diffusers/pipelines/stable_diffusion_xl/pipeline_stable_diffusion_xl_inpaint.py +++ b/src/diffusers/pipelines/stable_diffusion_xl/pipeline_stable_diffusion_xl_inpaint.py @@ -339,7 +339,7 @@ class StableDiffusionXLInpaintPipeline( ) self.register_to_config(force_zeros_for_empty_prompt=force_zeros_for_empty_prompt) self.register_to_config(requires_aesthetics_score=requires_aesthetics_score) - self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1) + self.vae_scale_factor = 2 ** (len(getattr(self.vae.config, 'block_out_channels', self.vae.config.decoder_block_out_channels)) - 1) self.image_processor = VaeImageProcessor(vae_scale_factor=self.vae_scale_factor) self.mask_processor = VaeImageProcessor( vae_scale_factor=self.vae_scale_factor, do_normalize=False, do_binarize=True, do_convert_grayscale=True diff --git a/src/diffusers/pipelines/stable_diffusion_xl/pipeline_stable_diffusion_xl_instruct_pix2pix.py b/src/diffusers/pipelines/stable_diffusion_xl/pipeline_stable_diffusion_xl_instruct_pix2pix.py index f391105cc4..dec8b5c02f 100644 --- a/src/diffusers/pipelines/stable_diffusion_xl/pipeline_stable_diffusion_xl_instruct_pix2pix.py +++ b/src/diffusers/pipelines/stable_diffusion_xl/pipeline_stable_diffusion_xl_instruct_pix2pix.py @@ -178,7 +178,7 @@ class StableDiffusionXLInstructPix2PixPipeline( scheduler=scheduler, ) self.register_to_config(force_zeros_for_empty_prompt=force_zeros_for_empty_prompt) - self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1) + self.vae_scale_factor = 2 ** (len(getattr(self.vae.config, 'block_out_channels', self.vae.config.decoder_block_out_channels)) - 1) self.image_processor = VaeImageProcessor(vae_scale_factor=self.vae_scale_factor) self.default_sample_size = self.unet.config.sample_size diff --git a/src/diffusers/pipelines/t2i_adapter/pipeline_stable_diffusion_adapter.py b/src/diffusers/pipelines/t2i_adapter/pipeline_stable_diffusion_adapter.py index 103cd70952..872faffb0a 100644 --- a/src/diffusers/pipelines/t2i_adapter/pipeline_stable_diffusion_adapter.py +++ b/src/diffusers/pipelines/t2i_adapter/pipeline_stable_diffusion_adapter.py @@ -198,7 +198,7 @@ class StableDiffusionAdapterPipeline(DiffusionPipeline): safety_checker=safety_checker, feature_extractor=feature_extractor, ) - self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1) + self.vae_scale_factor = 2 ** (len(getattr(self.vae.config, 'block_out_channels', self.vae.config.decoder_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) diff --git a/src/diffusers/pipelines/t2i_adapter/pipeline_stable_diffusion_xl_adapter.py b/src/diffusers/pipelines/t2i_adapter/pipeline_stable_diffusion_xl_adapter.py index a163a27a8b..13959d4a6b 100644 --- a/src/diffusers/pipelines/t2i_adapter/pipeline_stable_diffusion_xl_adapter.py +++ b/src/diffusers/pipelines/t2i_adapter/pipeline_stable_diffusion_xl_adapter.py @@ -187,7 +187,7 @@ class StableDiffusionXLAdapterPipeline( scheduler=scheduler, ) self.register_to_config(force_zeros_for_empty_prompt=force_zeros_for_empty_prompt) - self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1) + self.vae_scale_factor = 2 ** (len(getattr(self.vae.config, 'block_out_channels', self.vae.config.decoder_block_out_channels)) - 1) self.image_processor = VaeImageProcessor(vae_scale_factor=self.vae_scale_factor) self.default_sample_size = self.unet.config.sample_size diff --git a/src/diffusers/pipelines/text_to_video_synthesis/pipeline_text_to_video_synth.py b/src/diffusers/pipelines/text_to_video_synthesis/pipeline_text_to_video_synth.py index f5af49daef..9029772925 100644 --- a/src/diffusers/pipelines/text_to_video_synthesis/pipeline_text_to_video_synth.py +++ b/src/diffusers/pipelines/text_to_video_synthesis/pipeline_text_to_video_synth.py @@ -115,7 +115,7 @@ class TextToVideoSDPipeline(DiffusionPipeline, TextualInversionLoaderMixin, Lora unet=unet, scheduler=scheduler, ) - self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1) + self.vae_scale_factor = 2 ** (len(getattr(self.vae.config, 'block_out_channels', self.vae.config.decoder_block_out_channels)) - 1) # Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion.StableDiffusionPipeline.enable_vae_slicing def enable_vae_slicing(self): diff --git a/src/diffusers/pipelines/text_to_video_synthesis/pipeline_text_to_video_synth_img2img.py b/src/diffusers/pipelines/text_to_video_synthesis/pipeline_text_to_video_synth_img2img.py index 5cc37449b5..a3b01f1a1a 100644 --- a/src/diffusers/pipelines/text_to_video_synthesis/pipeline_text_to_video_synth_img2img.py +++ b/src/diffusers/pipelines/text_to_video_synthesis/pipeline_text_to_video_synth_img2img.py @@ -177,7 +177,7 @@ class VideoToVideoSDPipeline(DiffusionPipeline, TextualInversionLoaderMixin, Lor unet=unet, scheduler=scheduler, ) - self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1) + self.vae_scale_factor = 2 ** (len(getattr(self.vae.config, 'block_out_channels', self.vae.config.decoder_block_out_channels)) - 1) # Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion.StableDiffusionPipeline.enable_vae_slicing def enable_vae_slicing(self): diff --git a/src/diffusers/pipelines/unidiffuser/pipeline_unidiffuser.py b/src/diffusers/pipelines/unidiffuser/pipeline_unidiffuser.py index 4f3e003de0..bec7a5a091 100644 --- a/src/diffusers/pipelines/unidiffuser/pipeline_unidiffuser.py +++ b/src/diffusers/pipelines/unidiffuser/pipeline_unidiffuser.py @@ -117,7 +117,7 @@ class UniDiffuserPipeline(DiffusionPipeline): scheduler=scheduler, ) - self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1) + self.vae_scale_factor = 2 ** (len(getattr(self.vae.config, 'block_out_channels', self.vae.config.decoder_block_out_channels)) - 1) self.image_processor = VaeImageProcessor(vae_scale_factor=self.vae_scale_factor) self.num_channels_latents = vae.config.latent_channels diff --git a/src/diffusers/pipelines/versatile_diffusion/pipeline_versatile_diffusion.py b/src/diffusers/pipelines/versatile_diffusion/pipeline_versatile_diffusion.py index 68c720ab2a..a1dbc5fb68 100644 --- a/src/diffusers/pipelines/versatile_diffusion/pipeline_versatile_diffusion.py +++ b/src/diffusers/pipelines/versatile_diffusion/pipeline_versatile_diffusion.py @@ -76,7 +76,7 @@ class VersatileDiffusionPipeline(DiffusionPipeline): vae=vae, scheduler=scheduler, ) - self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1) + self.vae_scale_factor = 2 ** (len(getattr(self.vae.config, 'block_out_channels', self.vae.config.decoder_block_out_channels)) - 1) @torch.no_grad() def image_variation( diff --git a/src/diffusers/pipelines/versatile_diffusion/pipeline_versatile_diffusion_dual_guided.py b/src/diffusers/pipelines/versatile_diffusion/pipeline_versatile_diffusion_dual_guided.py index a248c25a55..65bb4a898a 100644 --- a/src/diffusers/pipelines/versatile_diffusion/pipeline_versatile_diffusion_dual_guided.py +++ b/src/diffusers/pipelines/versatile_diffusion/pipeline_versatile_diffusion_dual_guided.py @@ -93,7 +93,7 @@ class VersatileDiffusionDualGuidedPipeline(DiffusionPipeline): vae=vae, scheduler=scheduler, ) - self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1) + self.vae_scale_factor = 2 ** (len(getattr(self.vae.config, 'block_out_channels', self.vae.config.decoder_block_out_channels)) - 1) self.image_processor = VaeImageProcessor(vae_scale_factor=self.vae_scale_factor) if self.text_unet is not None and ( diff --git a/src/diffusers/pipelines/versatile_diffusion/pipeline_versatile_diffusion_image_variation.py b/src/diffusers/pipelines/versatile_diffusion/pipeline_versatile_diffusion_image_variation.py index 4f9c0bd9f4..be99d78aef 100644 --- a/src/diffusers/pipelines/versatile_diffusion/pipeline_versatile_diffusion_image_variation.py +++ b/src/diffusers/pipelines/versatile_diffusion/pipeline_versatile_diffusion_image_variation.py @@ -76,7 +76,7 @@ class VersatileDiffusionImageVariationPipeline(DiffusionPipeline): vae=vae, scheduler=scheduler, ) - self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1) + self.vae_scale_factor = 2 ** (len(getattr(self.vae.config, 'block_out_channels', self.vae.config.decoder_block_out_channels)) - 1) self.image_processor = VaeImageProcessor(vae_scale_factor=self.vae_scale_factor) def _encode_prompt(self, prompt, device, num_images_per_prompt, do_classifier_free_guidance, negative_prompt): diff --git a/src/diffusers/pipelines/versatile_diffusion/pipeline_versatile_diffusion_text_to_image.py b/src/diffusers/pipelines/versatile_diffusion/pipeline_versatile_diffusion_text_to_image.py index b6b61d754b..c19850b549 100644 --- a/src/diffusers/pipelines/versatile_diffusion/pipeline_versatile_diffusion_text_to_image.py +++ b/src/diffusers/pipelines/versatile_diffusion/pipeline_versatile_diffusion_text_to_image.py @@ -81,7 +81,7 @@ class VersatileDiffusionTextToImagePipeline(DiffusionPipeline): vae=vae, scheduler=scheduler, ) - self.vae_scale_factor = 2 ** (len(self.vae.config.block_out_channels) - 1) + self.vae_scale_factor = 2 ** (len(getattr(self.vae.config, 'block_out_channels', self.vae.config.decoder_block_out_channels)) - 1) self.image_processor = VaeImageProcessor(vae_scale_factor=self.vae_scale_factor) if self.text_unet is not None: