post release v0.39.0 (#14116)

* post release v0.39.0

* style.

* remove deprecated code path.

* remove deprecated method.

Author: sayakpaul

docs/source/en/api/pipelines/animatediff.md

@@ -68,7 +68,7 @@ scheduler = DDIMScheduler.from_pretrained(
 pipe.scheduler = scheduler
 
 # enable memory savings
-pipe.enable_vae_slicing()
+pipe.vae.enable_slicing()
 pipe.enable_model_cpu_offload()
 
 output = pipe(
@@ -395,8 +395,8 @@ pipe = AnimateDiffSDXLPipeline.from_pretrained(
 ).to("cuda")
 
 # enable memory savings
-pipe.enable_vae_slicing()
-pipe.enable_vae_tiling()
+pipe.vae.enable_slicing()
+pipe.vae.enable_tiling()
 
 output = pipe(
     prompt="a panda surfing in the ocean, realistic, high quality",
@@ -441,7 +441,7 @@ scheduler = DDIMScheduler.from_pretrained(
 pipe.scheduler = scheduler
 
 # enable memory savings
-pipe.enable_vae_slicing()
+pipe.vae.enable_slicing()
 pipe.enable_model_cpu_offload()
 
 # helper function to load videos
@@ -640,7 +640,7 @@ scheduler = DDIMScheduler.from_pretrained(
 pipe.scheduler = scheduler
 
 # enable memory savings
-pipe.enable_vae_slicing()
+pipe.vae.enable_slicing()
 pipe.enable_model_cpu_offload()
 
 output = pipe(
@@ -714,7 +714,7 @@ scheduler = DDIMScheduler.from_pretrained(
 pipe.scheduler = scheduler
 
 # enable memory savings
-pipe.enable_vae_slicing()
+pipe.vae.enable_slicing()
 pipe.enable_model_cpu_offload()
 
 output = pipe(
@@ -772,8 +772,8 @@ pipe.scheduler = DDIMScheduler.from_pretrained(
 )
 
 # enable memory savings
-pipe.enable_vae_slicing()
-pipe.enable_vae_tiling()
+pipe.vae.enable_slicing()
+pipe.vae.enable_tiling()
 
 # enable FreeInit
 # Refer to the enable_free_init documentation for a full list of configurable parameters
@@ -840,7 +840,7 @@ pipe.scheduler = LCMScheduler.from_config(pipe.scheduler.config, beta_schedule="
 
 pipe.load_lora_weights("wangfuyun/AnimateLCM", weight_name="sd15_lora_beta.safetensors", adapter_name="lcm-lora")
 
-pipe.enable_vae_slicing()
+pipe.vae.enable_slicing()
 pipe.enable_model_cpu_offload()
 
 output = pipe(
@@ -882,7 +882,7 @@ pipe.load_lora_weights("wangfuyun/AnimateLCM", weight_name="sd15_lora_beta.safet
 pipe.load_lora_weights("guoyww/animatediff-motion-lora-tilt-up", adapter_name="tilt-up")
 
 pipe.set_adapters(["lcm-lora", "tilt-up"], [1.0, 0.8])
-pipe.enable_vae_slicing()
+pipe.vae.enable_slicing()
 pipe.enable_model_cpu_offload()
 
 output = pipe(

docs/source/en/api/pipelines/controlnet.md

@@ -37,8 +37,6 @@ The original codebase can be found at [lllyasviel/ControlNet](https://github.com
 	- __call__
 	- enable_attention_slicing
 	- disable_attention_slicing
-	- enable_vae_slicing
-	- disable_vae_slicing
 	- enable_xformers_memory_efficient_attention
 	- disable_xformers_memory_efficient_attention
 	- load_textual_inversion
@@ -49,8 +47,6 @@ The original codebase can be found at [lllyasviel/ControlNet](https://github.com
 	- __call__
 	- enable_attention_slicing
 	- disable_attention_slicing
-	- enable_vae_slicing
-	- disable_vae_slicing
 	- enable_xformers_memory_efficient_attention
 	- disable_xformers_memory_efficient_attention
 	- load_textual_inversion
@@ -61,8 +57,6 @@ The original codebase can be found at [lllyasviel/ControlNet](https://github.com
 	- __call__
 	- enable_attention_slicing
 	- disable_attention_slicing
-	- enable_vae_slicing
-	- disable_vae_slicing
 	- enable_xformers_memory_efficient_attention
 	- disable_xformers_memory_efficient_attention
 	- load_textual_inversion

docs/source/en/api/pipelines/latent_consistency_models.md

@@ -643,10 +643,6 @@ export_to_gif(frames, "animation.gif")
     - __call__
     - enable_freeu
     - disable_freeu
-    - enable_vae_slicing
-    - disable_vae_slicing
-    - enable_vae_tiling
-    - disable_vae_tiling
 
 ## LatentConsistencyModelImg2ImgPipeline
 
@@ -655,10 +651,6 @@ export_to_gif(frames, "animation.gif")
     - __call__
     - enable_freeu
     - disable_freeu
-    - enable_vae_slicing
-    - disable_vae_slicing
-    - enable_vae_tiling
-    - disable_vae_tiling
 
 ## StableDiffusionPipelineOutput
 

docs/source/en/api/pipelines/mochi.md

@@ -86,7 +86,7 @@ pipe = MochiPipeline.from_pretrained("genmo/mochi-1-preview")
 
 # Enable memory savings
 pipe.enable_model_cpu_offload()
-pipe.enable_vae_tiling()
+pipe.vae.enable_tiling()
 
 prompt = "Close-up of a chameleon's eye, with its scaly skin changing color. Ultra high resolution 4k."
 
@@ -109,7 +109,7 @@ pipe = MochiPipeline.from_pretrained("genmo/mochi-1-preview", variant="bf16", to
 
 # Enable memory savings
 pipe.enable_model_cpu_offload()
-pipe.enable_vae_tiling()
+pipe.vae.enable_tiling()
 
 prompt = "Close-up of a chameleon's eye, with its scaly skin changing color. Ultra high resolution 4k."
 frames = pipe(prompt, num_frames=85).frames[0]
@@ -138,7 +138,7 @@ from diffusers.utils import export_to_video
 from diffusers.video_processor import VideoProcessor
 
 pipe = MochiPipeline.from_pretrained("genmo/mochi-1-preview", force_zeros_for_empty_prompt=True)
-pipe.enable_vae_tiling()
+pipe.vae.enable_tiling()
 pipe.enable_model_cpu_offload()
 
 prompt =  "An aerial shot of a parade of elephants walking across the African savannah. The camera showcases the herd and the surrounding landscape."
@@ -205,7 +205,7 @@ transformer = MochiTransformer3DModel.from_pretrained(
 
 pipe = MochiPipeline.from_pretrained(model_id,  transformer=transformer)
 pipe.enable_model_cpu_offload()
-pipe.enable_vae_tiling()
+pipe.vae.enable_tiling()
 
 with torch.autocast(device_type="cuda", dtype=torch.bfloat16, cache_enabled=False):
     frames = pipe(
@@ -245,7 +245,7 @@ transformer = MochiTransformer3DModel.from_pretrained(ckpt_path, torch_dtype=tor
 
 pipe = MochiPipeline.from_pretrained(model_id,  transformer=transformer)
 pipe.enable_model_cpu_offload()
-pipe.enable_vae_tiling()
+pipe.vae.enable_tiling()
 
 with torch.autocast(device_type="cuda", dtype=torch.bfloat16, cache_enabled=False):
     frames = pipe(

docs/source/en/api/pipelines/stable_diffusion/adapter.md

@@ -29,8 +29,6 @@ This model was contributed by the community contributor [HimariO](https://github
     - __call__
     - enable_attention_slicing
     - disable_attention_slicing
-    - enable_vae_slicing
-    - disable_vae_slicing
     - enable_xformers_memory_efficient_attention
     - disable_xformers_memory_efficient_attention
 
@@ -41,7 +39,5 @@ This model was contributed by the community contributor [HimariO](https://github
     - __call__
     - enable_attention_slicing
     - disable_attention_slicing
-    - enable_vae_slicing
-    - disable_vae_slicing
     - enable_xformers_memory_efficient_attention
     - disable_xformers_memory_efficient_attention

docs/source/en/api/pipelines/stable_diffusion/text2img.md

@@ -34,12 +34,8 @@ The abstract from the paper is:
 	- __call__
 	- enable_attention_slicing
 	- disable_attention_slicing
-	- enable_vae_slicing
-	- disable_vae_slicing
 	- enable_xformers_memory_efficient_attention
 	- disable_xformers_memory_efficient_attention
-	- enable_vae_tiling
-	- disable_vae_tiling
 	- load_textual_inversion
 	- from_single_file
 	- load_lora_weights

docs/source/en/api/pipelines/stable_unclip.md

@@ -106,8 +106,6 @@ image
 	- __call__
 	- enable_attention_slicing
 	- disable_attention_slicing
-	- enable_vae_slicing
-	- disable_vae_slicing
 	- enable_xformers_memory_efficient_attention
 	- disable_xformers_memory_efficient_attention
 
@@ -118,8 +116,6 @@ image
 	- __call__
 	- enable_attention_slicing
 	- disable_attention_slicing
-	- enable_vae_slicing
-	- disable_vae_slicing
 	- enable_xformers_memory_efficient_attention
 	- disable_xformers_memory_efficient_attention
 

docs/source/en/optimization/memory.md

@@ -173,7 +173,7 @@ VAE slicing saves memory by splitting large batches of inputs into a single batc
 
 For example, if you're generating 4 images at once, decoding would increase peak activation memory by 4x. VAE slicing reduces this by only decoding 1 image at a time instead of all 4 images at once.
 
-Call [`~StableDiffusionPipeline.enable_vae_slicing`] to enable sliced VAE. You can expect a small increase in performance when decoding multi-image batches and no performance impact for single-image batches.
+Call [`~AutoencoderKL.enable_slicing`] to enable sliced VAE. You can expect a small increase in performance when decoding multi-image batches and no performance impact for single-image batches.
 
 ```py
 import torch
@@ -183,7 +183,7 @@ pipeline = StableDiffusionXLPipeline.from_pretrained(
     "stabilityai/stable-diffusion-xl-base-1.0",
     torch_dtype=torch.float16,
 ).to("cuda")
-pipeline.enable_vae_slicing()
+pipeline.vae.enable_slicing()
 pipeline(["An astronaut riding a horse on Mars"]*32).images[0]
 print(f"Max memory reserved: {torch.cuda.max_memory_allocated() / 1024**3:.2f} GB")
 ```
@@ -195,7 +195,7 @@ print(f"Max memory reserved: {torch.cuda.max_memory_allocated() / 1024**3:.2f} G
 
 VAE tiling saves memory by dividing an image into smaller overlapping tiles instead of processing the entire image at once. This also reduces peak memory usage because the GPU is only processing a tile at a time.
 
-Call [`~StableDiffusionPipeline.enable_vae_tiling`] to enable VAE tiling. The generated image may have some tone variation from tile-to-tile because they're decoded separately, but there shouldn't be any obvious seams between the tiles. Tiling is disabled for resolutions lower than a pre-specified (but configurable) limit. For example, this limit is 512x512 for the VAE in [`StableDiffusionPipeline`].
+Call [`~AutoencoderKL.enable_tiling`] to enable VAE tiling. The generated image may have some tone variation from tile-to-tile because they're decoded separately, but there shouldn't be any obvious seams between the tiles. Tiling is disabled for resolutions lower than a pre-specified (but configurable) limit. For example, this limit is 512x512 for the VAE in [`StableDiffusionPipeline`].
 
 ```py
 import torch
@@ -205,7 +205,7 @@ from diffusers.utils import load_image
 pipeline = AutoPipelineForImage2Image.from_pretrained(
     "stabilityai/stable-diffusion-xl-base-1.0", torch_dtype=torch.float16
 ).to("cuda")
-pipeline.enable_vae_tiling()
+pipeline.vae.enable_tiling()
 
 init_image = load_image("https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/diffusers/img2img-sdxl-init.png")
 prompt = "Astronaut in a jungle, cold color palette, muted colors, detailed, 8k"

docs/source/en/using-diffusers/ip_adapter.md

@@ -180,7 +180,7 @@ scheduler = DDIMScheduler.from_pretrained(
     steps_offset=1,
 )
 pipeline.scheduler = scheduler
-pipeline.enable_vae_slicing()
+pipeline.vae.enable_slicing()
 pipeline.load_ip_adapter("h94/IP-Adapter", subfolder="models", weight_name="ip-adapter_sd15.bin")
 pipeline.enable_model_cpu_offload()
 

docs/source/ko/optimization/fp16.md

@@ -112,7 +112,7 @@ image = pipe(prompt).images[0]
 
 이를 [`~StableDiffusionPipeline.enable_attention_slicing`] 또는 [`~StableDiffusionPipeline.enable_xformers_memory_efficient_attention`]과 결합하여 메모리 사용을 추가로 최소화할 수 있습니다.
 
-VAE 디코드를 한 번에 하나씩 수행하려면 추론 전에 파이프라인에서 [`~StableDiffusionPipeline.enable_vae_slicing`]을 호출합니다. 예를 들어:
+VAE 디코드를 한 번에 하나씩 수행하려면 추론 전에 파이프라인에서 [`~AutoencoderKL.enable_slicing`]을 호출합니다. 예를 들어:
 
 ```Python
 import torch
@@ -126,7 +126,7 @@ pipe = StableDiffusionPipeline.from_pretrained(
 pipe = pipe.to("cuda")
 
 prompt = "a photo of an astronaut riding a horse on mars"
-pipe.enable_vae_slicing()
+pipe.vae.enable_slicing()
 images = pipe([prompt] * 32).images
 ```