mmdit

MMDiT

Implementación de una sola capa del MMDiT, propuesta por Esser et al. en Difusión Estable 3, en Pytorch

Además de una reproducción directa, también se generalizará a > 2 modalidades, ya que puedo imaginar un MMDiT para imágenes, audio y texto.

También ofrecerá una variante improvisada de atención personal que selecciona de forma adaptativa los pesos a usar mediante la activación aprendida. Esta idea surgió de las convoluciones adaptativas aplicadas por Kang et al. para GigaGAN.

Instalar

$ pip install mmdit

Uso

 import torch
from mmdit import MMDiTBlock

# define mm dit block

block = MMDiTBlock (
    dim_joint_attn = 512 ,
    dim_cond = 256 ,
    dim_text = 768 ,
    dim_image = 512 ,
    qk_rmsnorm = True
)

# mock inputs

time_cond = torch . randn ( 2 , 256 )

text_tokens = torch . randn ( 2 , 512 , 768 )
text_mask = torch . ones (( 2 , 512 )). bool ()

image_tokens = torch . randn ( 2 , 1024 , 512 )

# single block forward

text_tokens_next , image_tokens_next = block (
    time_cond = time_cond ,
    text_tokens = text_tokens ,
    text_mask = text_mask ,
    image_tokens = image_tokens
)

Se puede utilizar una versión generalizada como tal.

 import torch
from mmdit . mmdit_generalized_pytorch import MMDiT

mmdit = MMDiT (
    depth = 2 , 
    dim_modalities = ( 768 , 512 , 384 ),
    dim_joint_attn = 512 ,
    dim_cond = 256 ,
    qk_rmsnorm = True
)

# mock inputs

time_cond = torch . randn ( 2 , 256 )

text_tokens = torch . randn ( 2 , 512 , 768 )
text_mask = torch . ones (( 2 , 512 )). bool ()

video_tokens = torch . randn ( 2 , 1024 , 512 )

audio_tokens = torch . randn ( 2 , 256 , 384 )

# forward

text_tokens , video_tokens , audio_tokens = mmdit (
    modality_tokens = ( text_tokens , video_tokens , audio_tokens ),
    modality_masks = ( text_mask , None , None ),
    time_cond = time_cond ,
)

Citas

 @article { Esser2024ScalingRF ,
    title   = { Scaling Rectified Flow Transformers for High-Resolution Image Synthesis } ,
    author  = { Patrick Esser and Sumith Kulal and A. Blattmann and Rahim Entezari and Jonas Muller and Harry Saini and Yam Levi and Dominik Lorenz and Axel Sauer and Frederic Boesel and Dustin Podell and Tim Dockhorn and Zion English and Kyle Lacey and Alex Goodwin and Yannik Marek and Robin Rombach } ,
    journal = { ArXiv } ,
    year    = { 2024 } ,
    volume  = { abs/2403.03206 } ,
    url     = { https://api.semanticscholar.org/CorpusID:268247980 }
}

 @inproceedings { Darcet2023VisionTN ,
    title   = { Vision Transformers Need Registers } ,
    author  = { Timoth'ee Darcet and Maxime Oquab and Julien Mairal and Piotr Bojanowski } ,
    year    = { 2023 } ,
    url     = { https://api.semanticscholar.org/CorpusID:263134283 }
}

 @article { Zhu2024HyperConnections ,
    title   = { Hyper-Connections } ,
    author  = { Defa Zhu and Hongzhi Huang and Zihao Huang and Yutao Zeng and Yunyao Mao and Banggu Wu and Qiyang Min and Xun Zhou } ,
    journal = { ArXiv } ,
    year    = { 2024 } ,
    volume  = { abs/2409.19606 } ,
    url     = { https://api.semanticscholar.org/CorpusID:272987528 }
}