siren pytorch

SIREN di Pytorch

Implementasi Pytorch dari SIREN - Representasi Neural Implisit dengan Fungsi Aktivasi Berkala

Memasang

$ pip install siren-pytorch

Penggunaan

Jaringan saraf berlapis-lapis berbasis SIREN

 import torch
from torch import nn
from siren_pytorch import SirenNet

net = SirenNet (
    dim_in = 2 ,                        # input dimension, ex. 2d coor
    dim_hidden = 256 ,                  # hidden dimension
    dim_out = 3 ,                       # output dimension, ex. rgb value
    num_layers = 5 ,                    # number of layers
    final_activation = nn . Sigmoid (),   # activation of final layer (nn.Identity() for direct output)
    w0_initial = 30.                   # different signals may require different omega_0 in the first layer - this is a hyperparameter
)

coor = torch . randn ( 1 , 2 )
net ( coor ) # (1, 3) <- rgb value

Satu lapisan SIREN

 import torch
from siren_pytorch import Siren

neuron = Siren (
    dim_in = 3 ,
    dim_out = 256
)

coor = torch . randn ( 1 , 3 )
neuron ( coor ) # (1, 256)

Aktivasi sinus (hanya pembungkus torch.sin )

 import torch
from siren_pytorch import Sine

act = Sine ( 1. )
coor = torch . randn ( 1 , 2 )
act ( coor )

Wrapper untuk melatih gambar tertentu dengan tinggi dan lebar tertentu dari SirenNet tertentu, dan kemudian menghasilkannya.

 import torch
from torch import nn
from siren_pytorch import SirenNet , SirenWrapper

net = SirenNet (
    dim_in = 2 ,                        # input dimension, ex. 2d coor
    dim_hidden = 256 ,                  # hidden dimension
    dim_out = 3 ,                       # output dimension, ex. rgb value
    num_layers = 5 ,                    # number of layers
    w0_initial = 30.                   # different signals may require different omega_0 in the first layer - this is a hyperparameter
)

wrapper = SirenWrapper (
    net ,
    image_width = 256 ,
    image_height = 256
)

img = torch . randn ( 1 , 3 , 256 , 256 )
loss = wrapper ( img )
loss . backward ()

# after much training ...
# simply invoke the wrapper without passing in anything

pred_img = wrapper () # (1, 3, 256, 256) 

Modulasi dengan Kode Laten

Sebuah makalah baru mengusulkan bahwa cara terbaik untuk mengkondisikan Sirene dengan kode laten adalah dengan meneruskan vektor laten melalui jaringan umpan maju modulator, di mana keadaan tersembunyi setiap lapisan dikalikan secara elemen dengan lapisan Sirene yang sesuai.

Anda dapat menggunakan ini hanya dengan menetapkan kata kunci tambahan latent_dim , pada SirenWrapper

 import torch
from torch import nn
from siren_pytorch import SirenNet , SirenWrapper

net = SirenNet (
    dim_in = 2 ,                        # input dimension, ex. 2d coor
    dim_hidden = 256 ,                  # hidden dimension
    dim_out = 3 ,                       # output dimension, ex. rgb value
    num_layers = 5 ,                    # number of layers
    w0_initial = 30.                   # different signals may require different omega_0 in the first layer - this is a hyperparameter
)

wrapper = SirenWrapper (
    net ,
    latent_dim = 512 ,
    image_width = 256 ,
    image_height = 256
)

latent = nn . Parameter ( torch . zeros ( 512 ). normal_ ( 0 , 1e-2 ))
img = torch . randn ( 1 , 3 , 256 , 256 )

loss = wrapper ( img , latent = latent )
loss . backward ()

# after much training ...
# simply invoke the wrapper without passing in anything

pred_img = wrapper ( latent = latent ) # (1, 3, 256, 256) 

Kutipan

 @misc { sitzmann2020implicit ,
    title   = { Implicit Neural Representations with Periodic Activation Functions } ,
    author  = { Vincent Sitzmann and Julien N. P. Martel and Alexander W. Bergman and David B. Lindell and Gordon Wetzstein } ,
    year    = { 2020 } ,
    eprint  = { 2006.09661 } ,
    archivePrefix = { arXiv } ,
    primaryClass = { cs.CV }
}

 @misc { mehta2021modulated ,
    title   = { Modulated Periodic Activations for Generalizable Local Functional Representations } , 
    author  = { Ishit Mehta and Michaël Gharbi and Connelly Barnes and Eli Shechtman and Ravi Ramamoorthi and Manmohan Chandraker } ,
    year    = { 2021 } ,
    eprint  = { 2104.03960 } ,
    archivePrefix = { arXiv } ,
    primaryClass = { cs.CV }
}