Add VidTok AutoEncoders (#11261)

* add_autoencoder_vidtok

* format standardization

* remove small functions

* making the code style more diffusers-like

* Apply style fixes

* Add dummy objects for AutoencoderVidTok

* Fix AutoencoderVidTok avg_pool3d BFloat16 CPU compatibility

* skip test_layerwise_casting_training test

* Apply style fixes

---------

Co-authored-by: annitang1997 <memory97@sjtu.edu.cn>
Co-authored-by: github-actions[bot] <github-actions[bot]@users.noreply.github.com>
Co-authored-by: dg845 <58458699+dg845@users.noreply.github.com>

Author: 4 people

src/diffusers/__init__.py

@@ -204,6 +204,7 @@
             "AutoencoderOobleck",
             "AutoencoderRAE",
             "AutoencoderTiny",
+            "AutoencoderVidTok",
             "AutoModel",
             "BriaFiboTransformer2DModel",
             "BriaTransformer2DModel",
@@ -978,6 +979,7 @@
             AutoencoderOobleck,
             AutoencoderRAE,
             AutoencoderTiny,
+            AutoencoderVidTok,
             AutoModel,
             BriaFiboTransformer2DModel,
             BriaTransformer2DModel,

src/diffusers/models/__init__.py

@@ -51,6 +51,7 @@
     _import_structure["autoencoders.autoencoder_oobleck"] = ["AutoencoderOobleck"]
     _import_structure["autoencoders.autoencoder_rae"] = ["AutoencoderRAE"]
     _import_structure["autoencoders.autoencoder_tiny"] = ["AutoencoderTiny"]
+    _import_structure["autoencoders.autoencoder_vidtok"] = ["AutoencoderVidTok"]
     _import_structure["autoencoders.consistency_decoder_vae"] = ["ConsistencyDecoderVAE"]
     _import_structure["autoencoders.vq_model"] = ["VQModel"]
     _import_structure["cache_utils"] = ["CacheMixin"]
@@ -171,6 +172,7 @@
             AutoencoderOobleck,
             AutoencoderRAE,
             AutoencoderTiny,
+            AutoencoderVidTok,
             ConsistencyDecoderVAE,
             VQModel,
         )

src/diffusers/models/autoencoders/__init__.py

@@ -20,5 +20,6 @@
 from .autoencoder_oobleck import AutoencoderOobleck
 from .autoencoder_rae import AutoencoderRAE
 from .autoencoder_tiny import AutoencoderTiny
+from .autoencoder_vidtok import AutoencoderVidTok
 from .consistency_decoder_vae import ConsistencyDecoderVAE
 from .vq_model import VQModel