stylegan2/inversion.py

import math
import os

import torch
from torch import optim
from torch.nn import functional as FF
from torchvision import transforms
from PIL import Image
from tqdm import tqdm
import dataclasses

from .lpips import util


def noise_regularize(noises):
    loss = 0

    for noise in noises:
        size = noise.shape[2]

        while True:
            loss = (
                loss
                + (noise * torch.roll(noise, shifts=1, dims=3)).mean().pow(2)
                + (noise * torch.roll(noise, shifts=1, dims=2)).mean().pow(2)
            )

            if size <= 8:
                break

            noise = noise.reshape([-1, 1, size // 2, 2, size // 2, 2])
            noise = noise.mean([3, 5])
            size //= 2

    return loss


def noise_normalize_(noises):
    for noise in noises:
        mean = noise.mean()
        std = noise.std()

        noise.data.add_(-mean).div_(std)


def get_lr(t, initial_lr, rampdown=0.25, rampup=0.05):
    lr_ramp = min(1, (1 - t) / rampdown)
    lr_ramp = 0.5 - 0.5 * math.cos(lr_ramp * math.pi)
    lr_ramp = lr_ramp * min(1, t / rampup)

    return initial_lr * lr_ramp


def latent_noise(latent, strength):
    noise = torch.randn_like(latent) * strength

    return latent + noise


def make_image(tensor):
    return (
        tensor.detach()
        .clamp_(min=-1, max=1)
        .add(1)
        .div_(2)
        .mul(255)
        .type(torch.uint8)
        .permute(0, 2, 3, 1)
        .to("cpu")
        .numpy()
    )


@dataclasses.dataclass
class InverseConfig:
    lr_warmup = 0.05
    lr_decay = 0.25
    lr = 0.1
    noise = 0.05
    noise_decay = 0.75
    step = 1000
    noise_regularize = 1e5
    mse = 0
    w_plus = False,


def inverse_image(
    g_ema,
    image,
    image_size=256,
    config=InverseConfig()
):
    device = "cuda"
    args = config

    n_mean_latent = 10000

    resize = min(image_size, 256)

    transform = transforms.Compose(
        [
            transforms.Resize(resize),
            transforms.CenterCrop(resize),
            transforms.ToTensor(),
            transforms.Normalize([0.5, 0.5, 0.5], [0.5, 0.5, 0.5]),
        ]
    )

    imgs = []
    img = transform(image)
    imgs.append(img)

    imgs = torch.stack(imgs, 0).to(device)

    with torch.no_grad():
        noise_sample = torch.randn(n_mean_latent, 512, device=device)
        latent_out = g_ema.style(noise_sample)

        latent_mean = latent_out.mean(0)
        latent_std = ((latent_out - latent_mean).pow(2).sum() / n_mean_latent) ** 0.5

    percept = util.PerceptualLoss(
        model="net-lin", net="vgg", use_gpu=device.startswith("cuda")
    )

    noises_single = g_ema.make_noise()
    noises = []
    for noise in noises_single:
        noises.append(noise.repeat(imgs.shape[0], 1, 1, 1).normal_())

    latent_in = latent_mean.detach().clone().unsqueeze(0).repeat(imgs.shape[0], 1)

    if args.w_plus:
        latent_in = latent_in.unsqueeze(1).repeat(1, g_ema.n_latent, 1)

    latent_in.requires_grad = True

    for noise in noises:
        noise.requires_grad = True

    optimizer = optim.Adam([latent_in] + noises, lr=args.lr)

    pbar = tqdm(range(args.step))
    latent_path = []

    for i in pbar:
        t = i / args.step
        lr = get_lr(t, args.lr)
        optimizer.param_groups[0]["lr"] = lr
        noise_strength = latent_std * args.noise * max(0, 1 - t / args.noise_decay) ** 2
        latent_n = latent_noise(latent_in, noise_strength.item())

        latent, noise = g_ema.prepare([latent_n], input_is_latent=True, noise=noises)
        img_gen, F = g_ema.generate(latent, noise)

        batch, channel, height, width = img_gen.shape

        if height > 256:
            factor = height // 256

            img_gen = img_gen.reshape(
                batch, channel, height // factor, factor, width // factor, factor
            )
            img_gen = img_gen.mean([3, 5])

        p_loss = percept(img_gen, imgs).sum()
        n_loss = noise_regularize(noises)
        mse_loss = FF.mse_loss(img_gen, imgs)

        loss = p_loss + args.noise_regularize * n_loss + args.mse * mse_loss

        optimizer.zero_grad()
        loss.backward()
        optimizer.step()

        noise_normalize_(noises)

        if (i + 1) % 100 == 0:
            latent_path.append(latent_in.detach().clone())

        pbar.set_description(
            (
                f"perceptual: {p_loss.item():.4f}; noise regularize: {n_loss.item():.4f};"
                f" mse: {mse_loss.item():.4f}; lr: {lr:.4f}"
            )
        )

    latent, noise = g_ema.prepare([latent_path[-1]], input_is_latent=True, noise=noises)
    img_gen, F = g_ema.generate(latent, noise)

    img_ar = make_image(img_gen)

    i = 0

    noise_single = []
    for noise in noises:
        noise_single.append(noise[i: i + 1])

    result = {
        "latent": latent,
        "noise": noise_single,
        'F': F,
        "sample": img_gen,
    }

    pil_img = Image.fromarray(img_ar[i])
    pil_img.save('project.png')

    return result