bit diffusion

Bitdiffusion – Pytorch

Implementierung von Bit Diffusion, dem Versuch von Hintons Gruppe zur diskreten Entrauschungsdiffusion, in Pytorch

Beim Text scheint es, als hätten sie das Ziel verfehlt, aber die Forschungsrichtung scheint immer noch vielversprechend. Ich denke, ein sauberes Repository wird der Forschungsgemeinschaft viele Vorteile für diejenigen bringen, die von hier aus weitermachen.

Installieren

$ pip install bit-diffusion

Verwendung

 from bit_diffusion import Unet , Trainer , BitDiffusion

model = Unet (
    dim = 32 ,
    channels = 3 ,
    dim_mults = ( 1 , 2 , 4 , 8 ),
). cuda ()

bit_diffusion = BitDiffusion (
    model ,
    image_size = 128 ,
    timesteps = 100 ,
    time_difference = 0.1 ,       # they found in the paper that at lower number of timesteps, a time difference during sampling of greater than 0 helps FID. as timesteps increases, this time difference can be set to 0 as it does not help
    use_ddim = True              # use ddim
). cuda ()

trainer = Trainer (
    bit_diffusion ,
    '/path/to/your/data' ,             # path to your folder of images
    results_folder = './results' ,     # where to save results
    num_samples = 16 ,                 # number of samples
    train_batch_size = 4 ,             # training batch size
    gradient_accumulate_every = 4 ,    # gradient accumulation
    train_lr = 1e-4 ,                  # learning rate
    save_and_sample_every = 1000 ,     # how often to save and sample
    train_num_steps = 700000 ,         # total training steps
    ema_decay = 0.995 ,                # exponential moving average decay
)

trainer . train ()

Die Ergebnisse werden regelmäßig im Ordner ./results gespeichert

Wenn Sie außerhalb des Trainer mit der Unet und BitDiffusion -Klasse experimentieren möchten

 import torch
from bit_diffusion import Unet , BitDiffusion

model = Unet (
    dim = 64 ,
    dim_mults = ( 1 , 2 , 4 , 8 )
)

bit_diffusion = BitDiffusion (
    model ,
    image_size = 128 ,
    timesteps = 1000
)

training_images = torch . randn ( 8 , 3 , 128 , 128 ) # images are normalized from 0 to 1
loss = bit_diffusion ( training_images )
loss . backward ()
# after a lot of training

sampled_images = bit_diffusion . sample ( batch_size = 4 )
sampled_images . shape # (4, 3, 128, 128) 

Zitate

 @article { Chen2022AnalogBG ,
    title   = { Analog Bits: Generating Discrete Data using Diffusion Models with Self-Conditioning } ,
    author  = { Ting Chen and Ruixiang Zhang and Geoffrey E. Hinton } ,
    journal = { ArXiv } ,
    year    = { 2022 } ,
    volume  = { abs/2208.04202 }
}