repo
stringlengths
1
99
file
stringlengths
13
215
code
stringlengths
12
59.2M
file_length
int64
12
59.2M
avg_line_length
float64
3.82
1.48M
max_line_length
int64
12
2.51M
extension_type
stringclasses
1 value
DDoS
DDoS-master/models/srVAE/srVAE.py
from functools import partial import numpy as np import torch import torch.nn as nn from torchvision import transforms from .backbone.densenet16x32 import * from .priors.realnvp import RealNVP # --------- Utility functions --------- def get_shape(z_dim): """ Given the dimentionality of the latent space, re-shape it to an appropriate 3-D tensor. """ d = 8 if (z_dim%d==0) and (z_dim // (d*d) > 0): # cx8x8 H = W = d C = z_dim // (d*d) return (C, H, W) raise "Latent space can not mapped to a 3-D tensor. \ Please choose another dimentionality (power of 2)." # ----- Two Staged VAE ----- class srVAE(nn.Module): """ Super-Resolution Variational Auto-Encoder (srVAE). A Two Staged Visual Processing Variational AutoEncoder. Author: Ioannis Gatopoulos. """ def __init__(self, x_shape, y_shape=(3, 16, 16), u_dim=args.u_dim, z_dim=args.z_dim, prior=args.prior, device="cuda"): super().__init__() self.device = device self.x_shape = x_shape self.y_shape = (x_shape[0], y_shape[1], y_shape[2]) self.u_shape = get_shape(u_dim) self.z_shape = get_shape(z_dim) # q(y|x): deterministic "compressed" transformation self.compressed_transform = transforms.Compose([ transforms.ToPILImage(), transforms.Resize((self.y_shape[1], self.y_shape[2])), transforms.ToTensor() ]) # p(u) self.p_u = RealNVP(self.u_shape) # q(u | y) self.q_u = q_u(self.u_shape, self.y_shape) # p(z | y) self.p_z = p_z(self.z_shape, (self.y_shape, self.u_shape)) # q(z | x) self.q_z = q_z(self.z_shape, self.x_shape) # p(y | u) self.p_y = p_y(self.y_shape, self.u_shape) # p(x | y, z) self.p_x = p_x(self.x_shape, (self.y_shape, self.z_shape)) # likelihood distribution self.recon_loss = partial(dmol_loss) self.sample_distribution = partial(sample_from_dmol) def compressed_transoformation(self, input): y = [] for x in input: y.append(self.compressed_transform(x.cpu())) return torch.stack(y).to(self.device) def initialize(self, dataloader): """ Data dependent init for weight normalization (Automatically done during the first forward pass). """ with torch.no_grad(): x, _ = next(iter(dataloader)) x = x.to(self.device) output = self.forward(x) self.calculate_elbo(x, output) return @staticmethod def reparameterize(z_mean, z_log_var): """ z ~ N(z| z_mu, z_logvar) """ epsilon = torch.randn_like(z_mean) return z_mean + torch.exp(0.5*z_log_var)*epsilon @torch.no_grad() def generate(self, n_samples=20): # u ~ p(u) u = self.p_u.sample(self.u_shape, n_samples=n_samples, device=self.device).to(self.device) # p(y|u) y_logits = self.p_y(u) y_hat = self.sample_distribution(y_logits, nc=self.y_shape[0]) # z ~ p(z|y, u) z_p_mean, z_p_logvar = self.p_z((y_hat, u)) z_p = self.reparameterize(z_p_mean, z_p_logvar) # x ~ p(x|y,z) x_logits = self.p_x((y_hat, z_p)) x_hat = self.sample_distribution(x_logits, nc=self.x_shape[0]) return x_hat, y_hat @torch.no_grad() def reconstruct(self, x, **kwargs): outputs = self.forward(x) y_hat = self.sample_distribution(outputs.get('y_logits'), nc=self.y_shape[0]) x_hat = self.sample_distribution(outputs.get('x_logits'), nc=self.x_shape[0]) return outputs.get('y'), y_hat, x_hat @torch.no_grad() def super_resolution(self, y): # u ~ q(u| y) u_q_mean, u_q_logvar = self.q_u(y) u_q = self.reparameterize(u_q_mean, u_q_logvar) # z ~ p(z|y) z_p_mean, z_p_logvar = self.p_z((y, u_q)) z_p = self.reparameterize(z_p_mean, z_p_logvar) # x ~ p(x|y,z) x_logits = self.p_x((y, z_p)) x_hat = self.sample_distribution(x_logits) return x_hat def calculate_elbo(self, x, outputs, **kwargs): # unpack variables y, x_logits, y_logits = outputs.get('y'), outputs.get('x_logits'), outputs.get('y_logits') u_q, u_q_mean, u_q_logvar = outputs.get('u_q'), outputs.get('u_q_mean'), outputs.get('u_q_logvar') z_q, z_q_mean, z_q_logvar = outputs.get('z_q'), outputs.get('z_q_mean'), outputs.get('z_q_logvar') z_p_mean, z_p_logvar = outputs.get('z_p_mean'), outputs.get('z_p_logvar') # Reconstraction loss RE_x = self.recon_loss(x, x_logits, nc=self.x_shape[0]) RE_y = self.recon_loss(y, y_logits, nc=self.y_shape[0]) # Regularization loss log_p_u = self.p_u.log_p(u_q, dim=1) log_q_u = log_normal_diag(u_q, u_q_mean, u_q_logvar) KL_u = log_q_u - log_p_u log_p_z = log_normal_diag(z_q, z_p_mean, z_p_logvar) log_q_z = log_normal_diag(z_q, z_q_mean, z_q_logvar) KL_z = log_q_z - log_p_z # Total lower bound loss nelbo = - (RE_x + RE_y - KL_u - KL_z).mean() diagnostics = { "bpd" : (nelbo.item()) / (np.prod(x.shape[1:]) * np.log(2.)), "nelbo" : nelbo.item(), "RE" : - (RE_x + RE_y).mean().item(), "RE_x" : - RE_x.mean().item(), "RE_y" : - RE_y.mean().item(), "KL" : (KL_z + KL_u).mean().item(), "KL_u" : KL_u.mean().item(), "KL_z" : KL_z.mean().item(), } return nelbo, diagnostics def forward(self, x, **kwargs): """ Forward pass through the inference and the generative model. """ # y ~ f(x) (determinist) y = self.compressed_transoformation(x) # u ~ q(u| y) u_q_mean, u_q_logvar = self.q_u(y) u_q = self.reparameterize(u_q_mean, u_q_logvar) # z ~ q(z| x, y) z_q_mean, z_q_logvar = self.q_z(x) z_q = self.reparameterize(z_q_mean, z_q_logvar) # x ~ p(x| y, z) x_logits = self.p_x((y, z_q)) # y ~ p(y| u) y_logits = self.p_y(u_q) # z ~ p(z| x) z_p_mean, z_p_logvar = self.p_z((y, u_q)) return { 'u_q_mean' : u_q_mean, 'u_q_logvar' : u_q_logvar, 'u_q' : u_q, 'z_q_mean' : z_q_mean, 'z_q_logvar' : z_q_logvar, 'z_q' : z_q, 'z_p_mean' : z_p_mean, 'z_p_logvar' : z_p_logvar, 'y' : y, 'y_logits' : y_logits, 'x_logits' : x_logits } if __name__ == "__main__": pass
6,789
29.3125
122
py
DDoS
DDoS-master/models/srVAE/backbone/densenet16x32.py
import torch import torch.nn as nn import torch.nn.functional as F from src.modules.nn_layers import * from src.modules.distributions import n_embenddings from src.utils.args import args class q_u(nn.Module): """ Encoder q(u|y) """ def __init__(self, output_shape, input_shape): super().__init__() nc_in = input_shape[0] nc_out = 2 * output_shape[0] self.core_nn = nn.Sequential( DenselyEncoder( in_channels=nc_in, out_channels=nc_out, growth_rate=64, steps=3, scale_factor=1) ) def forward(self, input): mu, logvar = self.core_nn(input).chunk(2, 1) return mu, F.hardtanh(logvar, min_val=-7, max_val=7.) class p_y(nn.Module): """ Dencoder p(y|u) """ def __init__(self, output_shape, input_shape): super().__init__() nc_in = input_shape[0] nc_out = n_embenddings(output_shape[0]) self.core_nn = nn.Sequential( DenselyDecoder( in_channels=nc_in, out_channels=nc_out, growth_rate=128, steps=4, scale_factor=1) ) def forward(self, input): logits = self.core_nn(input) return logits class q_z(nn.Module): """ Encoder q(z|x) """ def __init__(self, output_shape, input_shape): super().__init__() nc_in = input_shape[0] nc_out = 2 * output_shape[0] self.core_nn = nn.Sequential( DenselyEncoder( in_channels=nc_in, out_channels=nc_out, growth_rate=16, steps=4, scale_factor=2) ) def forward(self, input): mu, logvar = self.core_nn(input).chunk(2, 1) return mu, F.hardtanh(logvar, min_val=-7, max_val=7.) class p_z(nn.Module): """ Encoder p(z| y, u) """ def __init__(self, output_shape, input_shape): super().__init__() nc_y_in, nc_u_in = input_shape[0][0], input_shape[1][0] nc_out = 2 * output_shape[0] self.y_nn = nn.Sequential( DenselyEncoder( in_channels=nc_y_in, out_channels=nc_out//2, growth_rate=32, steps=5, scale_factor=1), nn.ELU(inplace=True) ) self.u_nn = nn.Sequential( DenselyNetwork( in_channels=nc_u_in, out_channels=nc_out//2, growth_rate=64, steps=3, blocks=3, act=True) ) self.core_nn = nn.Sequential( DenselyNetwork( in_channels=nc_out, out_channels=nc_out, growth_rate=64, steps=3, blocks=3, act=None) ) def forward(self, input): y, u = input[0], input[1] y_out = self.y_nn(y) u_out = self.u_nn(u) joint = torch.cat((y_out, u_out), 1) mu, logvar = self.core_nn(joint).chunk(2, 1) return mu, F.hardtanh(logvar, min_val=-7, max_val=7.) class p_x(nn.Module): """ p(x| y, z) """ def __init__(self, output_shape, input_shape): super().__init__() nc_y_in, nc_z_in = input_shape[0][0], input_shape[1][0] nc_out = n_embenddings(output_shape[0]) self.z_nn = nn.Sequential( DenselyDecoder( in_channels=nc_z_in, out_channels=nc_out, growth_rate=64, steps=8, scale_factor=2) ) self.core_nn = nn.Sequential( DenselyNetwork( in_channels=nc_out + 3, out_channels=nc_out, growth_rate=64, steps=5, blocks=3, act=None) ) def forward(self, input): y, z = input[0], input[1] y_out = F.interpolate(y, size=[32, 32], align_corners=False, mode='bilinear') z_out = self.z_nn(z) joint = torch.cat((y_out, z_out), 1) logits = self.core_nn(joint) return logits if __name__ == "__main__": pass
4,289
23.94186
85
py
DDoS
DDoS-master/models/srVAE/priors/mog.py
import numpy as np import torch import torch.nn as nn from torch.autograd import Variable from .prior import Prior from src.modules.nn_layers import * from src.modules.distributions import * from src.utils import args # Modified vertion of: https://github.com/divymurli/VAEs class MixtureOfGaussians(Prior): def __init__(self, z_shape, num_mixtures=1000): super().__init__() self.z_shape = z_shape self.z_dim = np.prod(z_shape) self.k = num_mixtures # Mixture of Gaussians prior self.z_pre = torch.nn.Parameter(torch.randn(1, 2 * self.k, self.z_dim).to(args.device) / np.sqrt(self.k * self.z_dim)) # Uniform weighting self.pi = torch.nn.Parameter(torch.ones(self.k).to(args.device) / self.k, requires_grad=False) def sample_gaussian(self, m, v): """ Element-wise application reparameterization trick to sample from Gaussian """ sample = torch.randn(m.shape).to(args.device) z = m + (v**0.5)*sample return z def log_sum_exp(self, x, dim=0): """ Compute the log(sum(exp(x), dim)) in a numerically stable manner """ max_x = torch.max(x, dim)[0] new_x = x - max_x.unsqueeze(dim).expand_as(x) return max_x + (new_x.exp().sum(dim)).log() def log_mean_exp(self, x, dim): """ Compute the log(mean(exp(x), dim)) in a numerically stable manner """ return self.log_sum_exp(x, dim) - np.log(x.size(dim)) def log_normal(self, x, m, v): """ Computes the elem-wise log probability of a Gaussian and then sum over the last dim. Basically we're assuming all dims are batch dims except for the last dim. """ const = -0.5 * x.size(-1) * torch.log(2*torch.tensor(np.pi)) log_det = -0.5 * torch.sum(torch.log(v), dim = -1) log_exp = -0.5 * torch.sum((x - m)**2/v, dim = -1) log_prob = const + log_det + log_exp return log_prob def log_normal_mixture(self, z, m, v): """ Computes log probability of a uniformly-weighted Gaussian mixture. """ z = z.view(z.shape[0], 1, -1) log_probs = self.log_normal(z, m, v) log_prob = self.log_mean_exp(log_probs, 1) return log_prob def gaussian_parameters(self, h, dim=-1): m, h = torch.split(h, h.size(dim) // 2, dim=dim) v = F.softplus(h) + 1e-8 return m, v def sample(self, n_samples=1, **kwargs): idx = torch.distributions.categorical.Categorical(self.pi).sample((n_samples,)) m, v = self.gaussian_parameters(self.z_pre.squeeze(0), dim=0) m, v = m[idx], v[idx] z_samples = self.sample_gaussian(m, v) return z_samples.view(z_samples.shape[0], *self.z_shape) def log_p(self, z, **kwargs): return self.forward(z) def forward(self, z, dim=None, **kwargs): """ Computes the mixture of Gaussian prior """ m, v = self.gaussian_parameters(self.z_pre, dim=1) log_p_z = self.log_normal_mixture(z, m, v) return log_p_z def __str__(self): return "MixtureOfGaussians" if __name__ == "__main__": pass
3,267
32.010101
94
py
DDoS
DDoS-master/models/srVAE/priors/prior.py
import torch import torch.nn as nn class Prior(nn.Module): def __init__(self): super().__init__() def sample(self, **kwargs): raise NotImplementedError def log_p(self, input, **kwargs): return self.forward(z) def forward(self, input, **kwargs): raise NotImplementedError def __str__(self): raise NotImplementedError if __name__ == "__main__": pass
420
16.541667
39
py
DDoS
DDoS-master/models/srVAE/priors/standard_normal.py
import math import torch class StandardNormal: def __init__(self, z_shape): self.z_shape = z_shape def sample(self, n_samples=1, **kwargs): return torch.randn((n_samples, *self.z_shape)) def log_p(self, z, **kwargs): return self.forward(z) def forward(self, z, **kwargs): """ Outputs the log p(z). """ log_probs = z.pow(2) + math.log(math.pi * 2.) log_probs = -0.5 * log_probs.view(z.size(0), -1).sum(dim=1) return log_probs def __call__(self, z, **kwargs): return self.forward(z, **kwargs) def __str__(self): return "StandardNormal" if __name__ == "__main__": pass
681
21
67
py
DDoS
DDoS-master/models/srVAE/priors/realnvp/distributions/mog.py
import numpy as np import torch import torch.nn as nn from torch.autograd import Variable from src.modules.nn_layers import * from src.modules.distributions import * from src.utils import args class MixtureOfGaussians(nn.Module): def __init__(self, z_shape, num_mixtures=10): super().__init__() self.z_shape = z_shape self.z_dim = np.prod(z_shape) self.k = num_mixtures # Mixture of Gaussians prior self.z_pre = torch.nn.Parameter(torch.randn(1, 2 * self.k, self.z_dim).to(args.device) / np.sqrt(self.k * self.z_dim)) # Uniform weighting self.pi = torch.nn.Parameter(torch.ones(self.k).to(args.device) / self.k, requires_grad=False) def sample_gaussian(self, m, v): """ Element-wise application reparameterization trick to sample from Gaussian """ sample = torch.randn(m.shape).to(args.device) z = m + (v**0.5)*sample return z def log_sum_exp(self, x, dim=0): """ Compute the log(sum(exp(x), dim)) in a numerically stable manner """ max_x = torch.max(x, dim)[0] new_x = x - max_x.unsqueeze(dim).expand_as(x) return max_x + (new_x.exp().sum(dim)).log() def log_mean_exp(self, x, dim): """ Compute the log(mean(exp(x), dim)) in a numerically stable manner """ return self.log_sum_exp(x, dim) - np.log(x.size(dim)) def log_normal(self, x, m, v): """ Computes the elem-wise log probability of a Gaussian and then sum over the last dim. Basically we're assuming all dims are batch dims except for the last dim. """ const = -0.5 * x.size(-1) * torch.log(2*torch.tensor(np.pi)) log_det = -0.5 * torch.sum(torch.log(v), dim = -1) log_exp = -0.5 * torch.sum((x - m)**2/v, dim = -1) log_prob = const + log_det + log_exp return log_prob def log_normal_mixture(self, z, m, v): """ Computes log probability of a uniformly-weighted Gaussian mixture. """ z = z.view(z.shape[0], 1, -1) log_probs = self.log_normal(z, m, v) log_prob = self.log_mean_exp(log_probs, 1) return log_prob def gaussian_parameters(self, h, dim=-1): m, h = torch.split(h, h.size(dim) // 2, dim=dim) v = F.softplus(h) + 1e-8 return m, v def sample(self, n_samples=1, **kwargs): idx = torch.distributions.categorical.Categorical(self.pi).sample((n_samples,)) m, v = self.gaussian_parameters(self.z_pre.squeeze(0), dim=0) m, v = m[idx], v[idx] z_samples = self.sample_gaussian(m, v) return z_samples.view(z_samples.shape[0], *self.z_shape) def log_p(self, z, **kwargs): return self.forward(z) def forward(self, z, dim=None, **kwargs): """ Computes the mixture of Gaussian prior """ m, v = self.gaussian_parameters(self.z_pre, dim=1) log_p_z = self.log_normal_mixture(z, m, v) return log_p_z def __str__(self): return "MixtureOfGaussians" if __name__ == "__main__": pass
3,185
32.536842
94
py
DDoS
DDoS-master/models/srVAE/priors/realnvp/distributions/standard_normal.py
import math import torch import torch.nn as nn class StandardNormal: """ Isotropic Standard Normal distribution. """ def __init__(self, z_shape): self.z_shape = z_shape def sample(self, n_samples=1, **kwargs): return torch.randn((n_samples, *self.z_shape)) def log_p(self, z, **kwargs): return self.forward(z) def forward(self, z, **kwargs): """ Outputs the log p(z). """ log_probs = z.pow(2) + math.log(math.pi * 2.) log_probs = -0.5 * log_probs.view(z.size(0), -1).sum(dim=1) return log_probs def __call__(self, z, **kwargs): return self.forward(z, **kwargs) def __str__(self): return "StandardNormal" if __name__ == "__main__": pass
764
20.857143
67
py
DDoS
DDoS-master/models/srVAE/priors/realnvp/util/array_util.py
import torch import torch.nn.functional as F def squeeze_2x2(x, reverse=False, alt_order=False): """For each spatial position, a sub-volume of shape `1x1x(N^2 * C)`, reshape into a sub-volume of shape `NxNxC`, where `N = block_size`. Adapted from: https://github.com/tensorflow/models/blob/master/research/real_nvp/real_nvp_utils.py See Also: - TensorFlow nn.depth_to_space: https://www.tensorflow.org/api_docs/python/tf/nn/depth_to_space - Figure 3 of RealNVP paper: https://arxiv.org/abs/1605.08803 Args: x (torch.Tensor): Input tensor of shape (B, C, H, W). reverse (bool): Whether to do a reverse squeeze (unsqueeze). alt_order (bool): Whether to use alternate ordering. """ block_size = 2 if alt_order: n, c, h, w = x.size() if reverse: if c % 4 != 0: raise ValueError('Number of channels must be divisible by 4, got {}.'.format(c)) c //= 4 else: if h % 2 != 0: raise ValueError('Height must be divisible by 2, got {}.'.format(h)) if w % 2 != 0: raise ValueError('Width must be divisible by 4, got {}.'.format(w)) # Defines permutation of input channels (shape is (4, 1, 2, 2)). squeeze_matrix = torch.tensor([[[[1., 0.], [0., 0.]]], [[[0., 0.], [0., 1.]]], [[[0., 1.], [0., 0.]]], [[[0., 0.], [1., 0.]]]], dtype=x.dtype, device=x.device) perm_weight = torch.zeros((4 * c, c, 2, 2), dtype=x.dtype, device=x.device) for c_idx in range(c): slice_0 = slice(c_idx * 4, (c_idx + 1) * 4) slice_1 = slice(c_idx, c_idx + 1) perm_weight[slice_0, slice_1, :, :] = squeeze_matrix shuffle_channels = torch.tensor([c_idx * 4 for c_idx in range(c)] + [c_idx * 4 + 1 for c_idx in range(c)] + [c_idx * 4 + 2 for c_idx in range(c)] + [c_idx * 4 + 3 for c_idx in range(c)]) perm_weight = perm_weight[shuffle_channels, :, :, :] if reverse: x = F.conv_transpose2d(x, perm_weight, stride=2) else: x = F.conv2d(x, perm_weight, stride=2) else: b, c, h, w = x.size() x = x.permute(0, 2, 3, 1) if reverse: if c % 4 != 0: raise ValueError('Number of channels {} is not divisible by 4'.format(c)) x = x.view(b, h, w, c // 4, 2, 2) x = x.permute(0, 1, 4, 2, 5, 3) x = x.contiguous().view(b, 2 * h, 2 * w, c // 4) else: if h % 2 != 0 or w % 2 != 0: raise ValueError('Expected even spatial dims HxW, got {}x{}'.format(h, w)) x = x.view(b, h // 2, 2, w // 2, 2, c) x = x.permute(0, 1, 3, 5, 2, 4) x = x.contiguous().view(b, h // 2, w // 2, c * 4) x = x.permute(0, 3, 1, 2) return x def checkerboard_mask(height, width, reverse=False, dtype=torch.float32, device=None, requires_grad=False): """Get a checkerboard mask, such that no two entries adjacent entries have the same value. In non-reversed mask, top-left entry is 0. Args: height (int): Number of rows in the mask. width (int): Number of columns in the mask. reverse (bool): If True, reverse the mask (i.e., make top-left entry 1). Useful for alternating masks in RealNVP. dtype (torch.dtype): Data type of the tensor. device (torch.device): Device on which to construct the tensor. requires_grad (bool): Whether the tensor requires gradient. Returns: mask (torch.tensor): Checkerboard mask of shape (1, 1, height, width). """ checkerboard = [[((i % 2) + j) % 2 for j in range(width)] for i in range(height)] mask = torch.tensor(checkerboard, dtype=dtype, device=device, requires_grad=requires_grad) if reverse: mask = 1 - mask # Reshape to (1, 1, height, width) for broadcasting with tensors of shape (B, C, H, W) mask = mask.view(1, 1, height, width) return mask
4,369
40.226415
103
py
DDoS
DDoS-master/models/srVAE/priors/realnvp/util/norm_util.py
import functools import torch import torch.nn as nn def get_norm_layer(norm_type='instance'): if norm_type == 'batch': return functools.partial(nn.BatchNorm2d, affine=True) elif norm_type == 'instance': return functools.partial(nn.InstanceNorm2d, affine=False) else: raise NotImplementedError('Invalid normalization type: {}'.format(norm_type)) def get_param_groups(net, weight_decay, norm_suffix='weight_g', verbose=False): """Get two parameter groups from `net`: One named "normalized" which will override the optimizer with `weight_decay`, and one named "unnormalized" which will inherit all hyperparameters from the optimizer. Args: net (torch.nn.Module): Network to get parameters from weight_decay (float): Weight decay to apply to normalized weights. norm_suffix (str): Suffix to select weights that should be normalized. For WeightNorm, using 'weight_g' normalizes the scale variables. verbose (bool): Print out number of normalized and unnormalized parameters. """ norm_params = [] unnorm_params = [] for n, p in net.named_parameters(): if n.endswith(norm_suffix): norm_params.append(p) else: unnorm_params.append(p) param_groups = [{'name': 'normalized', 'params': norm_params, 'weight_decay': weight_decay}, {'name': 'unnormalized', 'params': unnorm_params}] if verbose: print('{} normalized parameters'.format(len(norm_params))) print('{} unnormalized parameters'.format(len(unnorm_params))) return param_groups class WNConv2d(nn.Module): """Weight-normalized 2d convolution. Args: in_channels (int): Number of channels in the input. out_channels (int): Number of channels in the output. kernel_size (int): Side length of each convolutional kernel. padding (int): Padding to add on edges of input. bias (bool): Use bias in the convolution operation. """ def __init__(self, in_channels, out_channels, kernel_size, padding, bias=True): super(WNConv2d, self).__init__() self.conv = nn.utils.weight_norm( nn.Conv2d(in_channels, out_channels, kernel_size, padding=padding, bias=bias)) def forward(self, x): x = self.conv(x) return x class BatchNormStats2d(nn.Module): """Compute BatchNorm2d normalization statistics: `mean` and `var`. Useful for keeping track of sum of log-determinant of Jacobians in flow models. Args: num_features (int): Number of features in the input (i.e., `C` in `(N, C, H, W)`). eps (float): Added to the denominator for numerical stability. decay (float): The value used for the running_mean and running_var computation. Different from conventional momentum, see `nn.BatchNorm2d` for more. """ def __init__(self, num_features, eps=1e-5, decay=0.1): super(BatchNormStats2d, self).__init__() self.eps = eps self.register_buffer('running_mean', torch.zeros(num_features)) self.register_buffer('running_var', torch.ones(num_features)) self.decay = decay def forward(self, x, training): # Get mean and variance per channel if training: channels = x.transpose(0, 1).contiguous().view(x.size(1), -1) used_mean, used_var = channels.mean(-1), channels.var(-1) curr_mean, curr_var = used_mean, used_var # Update variables self.running_mean = self.running_mean - self.decay * (self.running_mean - curr_mean) self.running_var = self.running_var - self.decay * (self.running_var - curr_var) else: used_mean = self.running_mean used_var = self.running_var used_var += self.eps # Reshape to (N, C, H, W) used_mean = used_mean.view(1, x.size(1), 1, 1).expand_as(x) used_var = used_var.view(1, x.size(1), 1, 1).expand_as(x) return used_mean, used_var
4,052
37.971154
96
py
DDoS
DDoS-master/models/srVAE/priors/realnvp/model/real_nvp.py
import torch import torch.nn as nn import torch.nn.functional as F import numpy as np from .coupling_layer import CouplingLayer, MaskType from ..util import squeeze_2x2 from ..distributions import StandardNormal # Modified vertion of: https://github.com/chrischute/real-nvp class RealNVP(nn.Module): """RealNVP Model Codebase from Chris Chute: https://github.com/chrischute/real-nvp Based on the paper: "Density estimation using Real NVP" by Laurent Dinh, Jascha Sohl-Dickstein, and Samy Bengio (https://arxiv.org/abs/1605.08803). Args: num_scales (int): Number of scales in the RealNVP model. in_channels (int): Number of channels in the input. mid_channels (int): Number of channels in the intermediate layers. num_blocks (int): Number of residual blocks in the s and t network of `Coupling` layers. """ def __init__(self, input_shape, mid_channels=64, num_blocks=5, num_scales=2, prior='std_normal'): super().__init__() self.flows = _RealNVP(0, num_scales, input_shape[0], mid_channels, num_blocks) # self.nbits = 8. if prior=='std_normal': self.prior = StandardNormal(input_shape) elif prior=='mog': self.prior = MixtureOfGaussians(input_shape) @torch.no_grad() def sample(self, z_shape, n_samples, device, **kwargs): """Sample from RealNVP model. Args: z_shape (tuple): n_samples (int): Number of samples to generate. device (torch.device): Device to use. """ z = self.prior.sample(n_samples).to(device) x, _ = self.forward(z, reverse=True) return x def log_p(self, x, **kwargs): """ returns the log likelihood. """ z, sldj = self.forward(x, reverse=False) ll = (self.prior.log_p(z) + sldj) # prior_ll = -0.5 * (z ** 2 + np.log(2 * np.pi)) # prior_ll = prior_ll.flatten(1).sum(-1) - np.log(2**self.nbits) * np.prod(z.size()[1:]) # ll = prior_ll + sldj # ll = ll.mean() return ll def forward(self, x, reverse=False): sldj = None if not reverse: sldj = 0 # we do not quintize ! # quintize ! # x = (x * (2**self.nbits - 1) + torch.rand_like(x)) / (2**self.nbits) x, sldj = self.flows(x, sldj, reverse) return x, sldj class _RealNVP(nn.Module): """Recursive builder for a `RealNVP` model. Each `_RealNVPBuilder` corresponds to a single scale in `RealNVP`, and the constructor is recursively called to build a full `RealNVP` model. Args: scale_idx (int): Index of current scale. num_scales (int): Number of scales in the RealNVP model. in_channels (int): Number of channels in the input. mid_channels (int): Number of channels in the intermediate layers. num_blocks (int): Number of residual blocks in the s and t network of `Coupling` layers. """ def __init__(self, scale_idx, num_scales, in_channels, mid_channels, num_blocks): super(_RealNVP, self).__init__() self.is_last_block = scale_idx == num_scales - 1 self.in_couplings = nn.ModuleList([ CouplingLayer(in_channels, mid_channels, num_blocks, MaskType.CHECKERBOARD, reverse_mask=False), CouplingLayer(in_channels, mid_channels, num_blocks, MaskType.CHECKERBOARD, reverse_mask=True), CouplingLayer(in_channels, mid_channels, num_blocks, MaskType.CHECKERBOARD, reverse_mask=False) ]) if self.is_last_block: self.in_couplings.append( CouplingLayer(in_channels, mid_channels, num_blocks, MaskType.CHECKERBOARD, reverse_mask=True)) else: self.out_couplings = nn.ModuleList([ CouplingLayer(4 * in_channels, 2 * mid_channels, num_blocks, MaskType.CHANNEL_WISE, reverse_mask=False), CouplingLayer(4 * in_channels, 2 * mid_channels, num_blocks, MaskType.CHANNEL_WISE, reverse_mask=True), CouplingLayer(4 * in_channels, 2 * mid_channels, num_blocks, MaskType.CHANNEL_WISE, reverse_mask=False) ]) self.next_block = _RealNVP(scale_idx + 1, num_scales, 2 * in_channels, 2 * mid_channels, num_blocks) def forward(self, x, sldj, reverse=False): if reverse: if not self.is_last_block: # Re-squeeze -> split -> next block x = squeeze_2x2(x, reverse=False, alt_order=True) x, x_split = x.chunk(2, dim=1) x, sldj = self.next_block(x, sldj, reverse) x = torch.cat((x, x_split), dim=1) x = squeeze_2x2(x, reverse=True, alt_order=True) # Squeeze -> 3x coupling (channel-wise) x = squeeze_2x2(x, reverse=False) for coupling in reversed(self.out_couplings): x, sldj = coupling(x, sldj, reverse) x = squeeze_2x2(x, reverse=True) for coupling in reversed(self.in_couplings): x, sldj = coupling(x, sldj, reverse) else: for coupling in self.in_couplings: x, sldj = coupling(x, sldj, reverse) if not self.is_last_block: # Squeeze -> 3x coupling (channel-wise) x = squeeze_2x2(x, reverse=False) for coupling in self.out_couplings: x, sldj = coupling(x, sldj, reverse) x = squeeze_2x2(x, reverse=True) # Re-squeeze -> split -> next block x = squeeze_2x2(x, reverse=False, alt_order=True) x, x_split = x.chunk(2, dim=1) x, sldj = self.next_block(x, sldj, reverse) x = torch.cat((x, x_split), dim=1) x = squeeze_2x2(x, reverse=True, alt_order=True) return x, sldj
5,949
37.636364
120
py
DDoS
DDoS-master/models/srVAE/priors/realnvp/model/coupling_layer.py
import torch import torch.nn as nn from enum import IntEnum from ..util import checkerboard_mask from src.modules.nn_layers import * class MaskType(IntEnum): CHECKERBOARD = 0 CHANNEL_WISE = 1 class CouplingLayer(nn.Module): """Coupling layer in RealNVP. Args: in_channels (int): Number of channels in the input. mid_channels (int): Number of channels in the `s` and `t` network. num_blocks (int): Number of residual blocks in the `s` and `t` network. mask_type (MaskType): One of `MaskType.CHECKERBOARD` or `MaskType.CHANNEL_WISE`. reverse_mask (bool): Whether to reverse the mask. Useful for alternating masks. """ def __init__(self, in_channels, mid_channels, num_blocks, mask_type, reverse_mask): super(CouplingLayer, self).__init__() # Save mask info self.mask_type = mask_type self.reverse_mask = reverse_mask if self.mask_type == MaskType.CHANNEL_WISE: in_channels //= 2 # Build scale and translate network growth_rate, steps = 64, 5 self.st_net = nn.Sequential( DenseNetLayer(inplanes=in_channels, growth_rate=growth_rate, steps=steps), Conv2d(in_channels + growth_rate*steps, 2*in_channels, kernel_size=3, stride=1, padding=1) ) # Learnable scale for s self.rescale = nn.utils.weight_norm(Rescale(in_channels)) def forward(self, x, sldj=None, reverse=True): if self.mask_type == MaskType.CHECKERBOARD: # Checkerboard mask b = checkerboard_mask(x.size(2), x.size(3), self.reverse_mask, device=x.device) x_b = x * b st = self.st_net(x_b) s, t = st.chunk(2, dim=1) s = self.rescale(torch.tanh(s)) s = s * (1 - b) t = t * (1 - b) # Scale and translate if reverse: inv_exp_s = s.mul(-1).exp() if torch.isnan(inv_exp_s).any(): raise RuntimeError('Scale factor has NaN entries') x = x * inv_exp_s - t else: exp_s = s.exp() if torch.isnan(exp_s).any(): raise RuntimeError('Scale factor has NaN entries') x = (x + t) * exp_s # Add log-determinant of the Jacobian sldj += s.view(s.size(0), -1).sum(-1) else: # Channel-wise mask if self.reverse_mask: x_id, x_change = x.chunk(2, dim=1) else: x_change, x_id = x.chunk(2, dim=1) st = self.st_net(x_id) s, t = st.chunk(2, dim=1) s = self.rescale(torch.tanh(s)) # Scale and translate if reverse: inv_exp_s = s.mul(-1).exp() if torch.isnan(inv_exp_s).any(): raise RuntimeError('Scale factor has NaN entries') x_change = x_change * inv_exp_s - t else: exp_s = s.exp() if torch.isnan(exp_s).any(): raise RuntimeError('Scale factor has NaN entries') x_change = (x_change + t) * exp_s # Add log-determinant of the Jacobian sldj += s.view(s.size(0), -1).sum(-1) if self.reverse_mask: x = torch.cat((x_id, x_change), dim=1) else: x = torch.cat((x_change, x_id), dim=1) return x, sldj class Rescale(nn.Module): """Per-channel rescaling. Need a proper `nn.Module` so we can wrap it with `torch.nn.utils.weight_norm`. Args: num_channels (int): Number of channels in the input. """ def __init__(self, num_channels): super(Rescale, self).__init__() self.weight = nn.Parameter(torch.ones(num_channels, 1, 1)) def forward(self, x): x = self.weight * x return x
4,031
32.04918
91
py
DDoS
DDoS-master/models/ShuffleUNet/icnr.py
import torch import torch.nn as nn def ICNR(tensor, upscale_factor=2, inizializer=nn.init.kaiming_normal_): new_shape = [int(tensor.shape[0] / (upscale_factor ** 2))] + list(tensor.shape[1:]) subkernel = torch.zeros(new_shape) subkernel = inizializer(subkernel) subkernel = subkernel.transpose(0, 1) subkernel = subkernel.contiguous().view(subkernel.shape[0], subkernel.shape[1], -1) kernel = subkernel.repeat(1, 1, upscale_factor ** 2) transposed_shape = [tensor.shape[1]] + [tensor.shape[0]] + list(tensor.shape[2:]) kernel = kernel.contiguous().view(transposed_shape) kernel = kernel.transpose(0, 1) return kernel
708
31.227273
87
py
DDoS
DDoS-master/models/ShuffleUNet/pixel_shuffle.py
import torch.nn as nn from . import icnr def _pixel_shuffle(input, upscale_factor): r"""Rearranges elements in a Tensor of shape :math:`(N, C, d_{1}, d_{2}, ..., d_{n})` to a tensor of shape :math:`(N, C/(r^n), d_{1}*r, d_{2}*r, ..., d_{n}*r)`. Where :math:`n` is the dimensionality of the data. See :class:`~torch.nn.PixelShuffle` for details. Args: input (Variable): Input upscale_factor (int): factor to increase spatial resolution by Examples:: # 1D example #>>> input = torch.Tensor(1, 4, 8) #>>> output = F.pixel_shuffle(input, 2) #>>> print(output.size()) torch.Size([1, 2, 16]) # 2D example #>>> input = torch.Tensor(1, 9, 8, 8) #>>> output = F.pixel_shuffle(input, 3) #>>> print(output.size()) torch.Size([1, 1, 24, 24]) # 3D example #>>> input = torch.Tensor(1, 8, 16, 16, 16) #>>> output = F.pixel_shuffle(input, 2) #>>> print(output.size()) torch.Size([1, 1, 32, 32, 32]) """ input_size = list(input.size()) dimensionality = len(input_size) - 2 input_size[1] //= (upscale_factor ** dimensionality) output_size = [dim * upscale_factor for dim in input_size[2:]] input_view = input.contiguous().view( input_size[0], input_size[1], *(([upscale_factor] * dimensionality) + input_size[2:]) ) indicies = list(range(2, 2 + 2 * dimensionality)) indicies = indicies[1::2] + indicies[0::2] shuffle_out = input_view.permute(0, 1, *(indicies[::-1])).contiguous() return shuffle_out.view(input_size[0], input_size[1], *output_size) class PixelShuffle(nn.Module): def __init__(self, in_c, out_c, kernel, stride, bias=True, d=3): super(PixelShuffle, self).__init__() if d==3: self.conv = nn.Conv3d(in_c, out_c, kernel_size=kernel, stride=stride, bias=bias, padding=kernel//2) else: self.conv = nn.Conv2d(in_c, out_c, kernel_size=kernel, stride=stride, bias=bias, padding=kernel//2) self.icnr_weights = icnr.ICNR(self.conv.weight, 2) self.conv.weight.data.copy_(self.icnr_weights) def forward(self, x): x = self.conv(x) x = _pixel_shuffle(x, 2) return x
2,276
35.142857
111
py
DDoS
DDoS-master/models/ShuffleUNet/net.py
import sys import torch import torch.nn as nn from . import pixel_shuffle, pixel_unshuffle # -------------------------------------------------------------------------------------------------------------------------------------------------## class _double_conv(nn.Module): """ Double Convolution Block """ def __init__(self, in_channels, out_channels, k_size, stride, bias=True, conv_layer=nn.Conv3d): super(_double_conv, self).__init__() self.conv_1 = conv_layer(in_channels=in_channels, out_channels=out_channels, kernel_size=k_size, stride=stride, padding=k_size // 2, bias=bias) self.conv_2 = conv_layer(in_channels=out_channels, out_channels=out_channels, kernel_size=k_size, stride=stride, padding=k_size // 2, bias=bias) self.relu = nn.ReLU(inplace=True) def forward(self, x): x = self.conv_1(x) x = self.relu((x)) x = self.conv_2(x) x = self.relu((x)) return x class _conv_decomp(nn.Module): """ Convolutional Decomposition Block """ def __init__(self, in_channels, out_channels, k_size, stride, bias=True, conv_layer=nn.Conv3d): super(_conv_decomp, self).__init__() self.conv1 = conv_layer(in_channels=in_channels, out_channels=out_channels, kernel_size=k_size, stride=stride, padding=k_size // 2, bias=bias) self.conv2 = conv_layer(in_channels=in_channels, out_channels=out_channels, kernel_size=k_size, stride=stride, padding=k_size // 2, bias=bias) self.conv3 = conv_layer(in_channels=in_channels, out_channels=out_channels, kernel_size=k_size, stride=stride, padding=k_size // 2, bias=bias) self.conv4 = conv_layer(in_channels=in_channels, out_channels=out_channels, kernel_size=k_size, stride=stride, padding=k_size // 2, bias=bias) self.relu = nn.ReLU(inplace=True) def forward(self, x): x1 = self.conv1(x) x1 = self.relu((x1)) x2 = self.conv2(x) x2 = self.relu((x2)) x3 = self.conv3(x) x3 = self.relu((x3)) x4 = self.conv4(x) x4 = self.relu((x4)) return x1, x2, x3, x4 class _concat(nn.Module): """ Skip-Addition block """ def __init__(self): super(_concat, self).__init__() def forward(self, e1, e2, e3, e4, d1, d2, d3, d4): self.X1 = e1 + d1 self.X2 = e2 + d2 self.X3 = e3 + d3 self.X4 = e4 + d4 x = torch.cat([self.X1, self.X2, self.X3, self.X4], dim=1) return x # -------------------------------------------------------------------------------------------------------------------------------------------------## class ShuffleUNet(nn.Module): def __init__(self, d=3, in_ch=1, num_features=64, n_levels=3, out_ch=1, kernel_size=3, stride=1): super(ShuffleUNet, self).__init__() self.n_levels = n_levels num_features = num_features filters = [num_features] for _ in range(n_levels): filters.append(filters[-1]*2) if d==3: conv_layer = nn.Conv3d ps_fact = (2 ** 2) elif d==2: conv_layer = nn.Conv2d ps_fact = 2 else: sys.exit("Invalid d") # Input self.conv_inp = _double_conv(in_ch, filters[0], kernel_size, stride, conv_layer=conv_layer) #Contraction path self.wave_down = nn.ModuleList() self.pix_unshuff = nn.ModuleList() self.conv_enc = nn.ModuleList() for i in range(0, n_levels): self.wave_down.append(_conv_decomp(filters[i], filters[i], kernel_size, stride, conv_layer=conv_layer)) self.pix_unshuff.append(pixel_unshuffle.PixelUnshuffle(num_features * (2**i), num_features * (2**i), kernel_size, stride, d=d)) self.conv_enc.append(_double_conv(filters[i], filters[i+1], kernel_size, stride, conv_layer=conv_layer)) #Expansion path self.cat = _concat() self.pix_shuff = nn.ModuleList() self.wave_up = nn.ModuleList() self.convup = nn.ModuleList() for i in range(n_levels-1,-1,-1): self.pix_shuff.append(pixel_shuffle.PixelShuffle(num_features * (2**(i+1)), num_features * (2**(i+1)) * ps_fact, kernel_size, stride, d=d)) self.wave_up.append(_conv_decomp(filters[i], filters[i], kernel_size, stride, conv_layer=conv_layer)) self.convup.append(_double_conv(filters[i] * 5, filters[i], kernel_size, stride, conv_layer=conv_layer)) #FC self.out = conv_layer(filters[0], out_ch, kernel_size=1, stride=1, padding=0, bias=True) #Weight init for m in self.modules(): if isinstance(m, conv_layer): weight = nn.init.kaiming_normal_(m.weight, nonlinearity='relu') m.weight.data.copy_(weight) if m.bias is not None: m.bias.data.zero_() def forward(self, x): encs = [self.conv_inp(x)] waves = [] for i in range(self.n_levels): waves.append(self.wave_down[i](encs[-1])) _tmp = self.pix_unshuff[i](waves[-1][-1]) encs.append(self.conv_enc[i](_tmp)) dec = encs.pop() for i in range(self.n_levels): _tmp = self.pix_shuff[i](dec) _tmp_waves = self.wave_up[i](_tmp) + waves.pop() _tmp_cat = self.cat(*_tmp_waves) dec = self.convup[i](torch.cat([encs.pop(), _tmp_cat], dim=1)) return self.out(dec)
5,659
36.733333
151
py
DDoS
DDoS-master/models/ShuffleUNet/pixel_unshuffle.py
import torch.nn as nn from . import icnr class _double_conv_3d(nn.Module): """ Convolution Block """ def __init__(self, in_channels, out_channels, k_size, stride, bias=True): super(_double_conv_3d, self).__init__() self.conv = nn.Sequential( nn.Conv3d(in_channels=in_channels, out_channels=out_channels, kernel_size=k_size, stride=stride, padding=k_size//2, bias=bias), nn.BatchNorm3d(num_features=out_channels), nn.ReLU(inplace=True), nn.Conv3d(in_channels=out_channels, out_channels=out_channels, kernel_size=k_size, stride=stride, padding=k_size//2, bias=bias), nn.BatchNorm3d(num_features=out_channels), nn.ReLU(inplace=True) ) def forward(self, x): x = self.conv(x) return x class _double_conv_2d(nn.Module): """ Convolution Block """ def __init__(self, in_channels, out_channels, k_size, stride, bias=True): super(_double_conv_2d, self).__init__() self.conv = nn.Sequential( nn.Conv2d(in_channels=in_channels, out_channels=out_channels, kernel_size=k_size, stride=stride, padding=k_size//2, bias=bias), nn.BatchNorm2d(num_features=out_channels), nn.ReLU(inplace=True), nn.Conv2d(in_channels=out_channels, out_channels=out_channels, kernel_size=k_size, stride=stride, padding=k_size//2, bias=bias), nn.BatchNorm2d(num_features=out_channels), nn.ReLU(inplace=True) ) def forward(self, x): x = self.conv(x) return x def _pixel_unshuffle_3d(input, upscale_factor): r"""Rearranges elements in a Tensor of shape :math:(C, rH, rW) to a tensor of shape :math:(*, r^2C, H, W). written by: Zhaoyi Yan, https://github.com/Zhaoyi-Yan and Kai Zhang, https://github.com/cszn/FFDNet 01/01/2019 """ batch_size, channels, depth, in_height, in_width = input.size() depth_final = depth // upscale_factor out_height = in_height // upscale_factor out_width = in_width // upscale_factor input_view = input.contiguous().view( batch_size, channels, depth_final, upscale_factor, out_height, upscale_factor, out_width, upscale_factor) channels *= upscale_factor ** 3 unshuffle_out = input_view.permute(0, 1, 3, 5, 7, 2, 4, 6).contiguous() return unshuffle_out.view(batch_size, channels, depth_final, out_height, out_width) def _pixel_unshuffle_2d(input, upscale_factor): r"""Rearranges elements in a Tensor of shape :math:(C, rH, rW) to a tensor of shape :math:(*, r^2C, H, W). written by: Zhaoyi Yan, https://github.com/Zhaoyi-Yan and Kai Zhang, https://github.com/cszn/FFDNet 01/01/2019 """ batch_size, channels, in_height, in_width = input.size() out_height = in_height // upscale_factor out_width = in_width // upscale_factor input_view = input.contiguous().view( batch_size, channels, out_height, upscale_factor, out_width, upscale_factor) channels *= upscale_factor ** 2 unshuffle_out = input_view.permute(0, 1, 3, 5, 2, 4).contiguous() return unshuffle_out.view(batch_size, channels, out_height, out_width) class PixelUnshuffle(nn.Module): def __init__(self, in_c, out_c, kernel, stride, bias=True, d=3): super(PixelUnshuffle, self).__init__() if d == 3: self.conv = nn.Conv3d(in_c, out_c, kernel_size=kernel, stride=stride, bias=bias, padding=kernel//2) self.down_conv = _double_conv_3d(out_c*8, out_c, kernel, stride, bias) self.pu = _pixel_unshuffle_3d else: self.conv = nn.Conv2d(in_c, out_c, kernel_size=kernel, stride=stride, bias=bias, padding=kernel//2) self.down_conv = _double_conv_2d(out_c*4, out_c, kernel, stride, bias) self.pu = _pixel_unshuffle_2d self.icnr_weights = icnr.ICNR(self.conv.weight, 2) self.conv.weight.data.copy_(self.icnr_weights) def forward(self, x): x = self.conv(x) x = self.down_conv(self.pu(x, 2)) return x
4,185
37.054545
111
py
DDoS
DDoS-master/visualisation/num4trilinear.py
from glob import glob import torch from tqdm import tqdm import os import nibabel as nib import numpy as np import pandas as pd import torch.nn.functional as F from utils.utilities import calc_metircs fully_root = "/mnt/MEMoRIAL/MEMoRIAL_SharedStorage_M1.2+4+7/Chompunuch/PhD/Data/3DDynTest/MickAbdomen3DDyn/DynProtocol3/Filtered/hrTestDynConST" under_root = "/mnt/MEMoRIAL/MEMoRIAL_SharedStorage_M1.2+4+7/Chompunuch/PhD/Data/3DDynTest/MickAbdomen3DDyn/DynProtocol3/Filtered/usTestDynConST" interp = "trilinear" files = sorted(glob(f"{under_root}/**/*.nii.gz", recursive=True)) metrics = [] for f in tqdm(files): if ("Center" not in f and "Centre" not in f) or "TP00" in f or "WoPad" not in f: continue fully_parts = f.replace(under_root, fully_root).split(os.path.sep) undersampling = fully_parts[-3] del fully_parts[-3] f_fully = os.path.sep.join(fully_parts) vol_under = np.array(nib.load(f).get_fdata()) vol_fully = np.array(nib.load(f_fully).get_fdata()) vol_under /= vol_under.max() vol_fully /= vol_fully.max() vol_under = F.interpolate(torch.from_numpy(vol_under).unsqueeze(0).unsqueeze(0), size=vol_fully.shape, mode=interp, align_corners=False).squeeze().numpy() inp_metrics, inp_ssimMAP = calc_metircs(vol_fully, vol_under, tag="ZPad") inp_metrics["file"] = fully_parts[-2] + "_" + fully_parts[-1] inp_metrics["subject"] = fully_parts[-6] + "_" + fully_parts[-5] inp_metrics["undersampling"] = undersampling + "WoPad" inp_metrics["model"] = interp.capitalize() inp_metrics["DiffZPad"] = np.std(vol_fully - vol_under) metrics.append(inp_metrics) df = pd.DataFrame.from_dict(metrics) df.to_csv(f"{os.path.dirname(under_root)}/noprevnorm_metrics_{interp}.csv")
1,754
38
158
py
DDoS
DDoS-master/utils/elastic_transform.py
#!/usr/bin/env python ''' Purpose : ''' from numbers import Number from typing import Optional, Tuple, Union import numpy as np import torch import torch as th import torch.nn as nn import torch.nn.functional as F from torch.nn.parameter import Parameter __author__ = "Kartik Prabhu, Mahantesh Pattadkal, and Soumick Chatterjee" __copyright__ = "Copyright 2022, Faculty of Computer Science, Otto von Guericke University Magdeburg, Germany" __credits__ = ["Kartik Prabhu", "Mahantesh Pattadkal", "Soumick Chatterjee"] __license__ = "GPL" __version__ = "1.0.0" __maintainer__ = "Soumick Chatterjee" __email__ = "soumick.chatterjee@ovgu.de" __status__ = "Production" from torchio.utils import to_tuple try: from torchio import RandomElasticDeformation except: from torchio.transforms.augmentation import RandomElasticDeformation from torch.cuda.amp import autocast from airlab import utils as tu from airlab.transformation.pairwise import _KernelTransformation from airlab.transformation.utils import compute_grid from airlab.utils import kernelFunction as utils SPLINE_ORDER = 3 """ Warp image with displacement * input(tensor) : input of shape (N, C, H_\text{in}, W_\text{in})(N,C,H,W) (4-D case)(N,C,D,H ,W) (5-D case) * grid(tensor): flow-field of shape (N, H_\text{out}, W_\text{out}, 2)(N,H ,W,2) (4-D case) or (N, D, H, W, 3)(N,D,H,W,3) (5-D case) * mult = true if batched input """ def warp_image(image, displacement, multi=False): image_size = image.size() #[B, D, H, W] batch_size = image_size[0] if multi: image_size = image_size[2:]#[D, H, W] grid = compute_grid(image_size, dtype=image.dtype, device=image.device) grid = displacement + grid grid = torch.cat([grid] * batch_size, dim=0) # batch number of times # warp image if multi: warped_image = F.grid_sample(image, grid) #[B, C, D, H, W] else: warped_image = F.grid_sample(image.unsqueeze(0).unsqueeze(0), grid) #[B, C, D, H, W], unsqueeze to give batch and channel dimension return warped_image #[B, C, D, H, W] """ Base class for kernel transformations """ class _ParameterizedKernelTransformation(_KernelTransformation): def __init__(self, image_size, rnd_grid_params=None, diffeomorphic=False, dtype=th.float32, device='cpu'): super(_ParameterizedKernelTransformation, self).__init__(image_size, diffeomorphic, dtype, device) self.rnd_grid_params = rnd_grid_params def get_coarse_field(self, grid_shape, max_displacement, num_locked_borders, ): coarse_field = th.rand(self._dim, *grid_shape) # [0, 1) coarse_field -= 0.5 # [-0.5, 0.5) coarse_field *= 2 # [-1, 1] for dimension in range(3): # [-max_displacement, max_displacement) coarse_field[dimension, ...] *= max_displacement[dimension] # Set displacement to 0 at the borders for i in range(num_locked_borders): coarse_field[:, i, :] = 0 coarse_field[:, -1 - i, :] = 0 coarse_field[:, :, i] = 0 coarse_field[:, :, -1 - i] = 0 return coarse_field.unsqueeze(0) def _initialize(self): cp_grid = np.ceil(np.divide(self._image_size, self._stride)).astype(dtype=int) # new image size after convolution inner_image_size = np.multiply(self._stride, cp_grid) - (self._stride - 1) # add one control point at each side cp_grid = cp_grid + 2 # image size with additional control points new_image_size = np.multiply(self._stride, cp_grid) - (self._stride - 1) # center image between control points image_size_diff = inner_image_size - self._image_size image_size_diff_floor = np.floor((np.abs(image_size_diff)/2))*np.sign(image_size_diff) self._crop_start = image_size_diff_floor + np.remainder(image_size_diff, 2)*np.sign(image_size_diff) self._crop_end = image_size_diff_floor # create transformation parameters if self.rnd_grid_params is None: cp_grid = [1, self._dim] + cp_grid.tolist() self.trans_parameters = Parameter(th.Tensor(*cp_grid)) self.trans_parameters.data.fill_(0) else: self.trans_parameters = Parameter(self.get_coarse_field(cp_grid, self.rnd_grid_params['max_displacement'], self.rnd_grid_params['num_locked_borders'])) # copy to gpu if needed self.to(dtype=self._dtype, device=self._device) # convert to integer self._padding = self._padding.astype(dtype=int).tolist() self._stride = self._stride.astype(dtype=int).tolist() self._crop_start = self._crop_start.astype(dtype=int) self._crop_end = self._crop_end.astype(dtype=int) size = [1, 1] + new_image_size.astype(dtype=int).tolist() self._displacement_tmp = th.empty(*size, dtype=self._dtype, device=self._device) size = [1, 1] + self._image_size.astype(dtype=int).tolist() self._displacement = th.empty(*size, dtype=self._dtype, device=self._device) """ bspline kernel transformation """ class ParameterizedBsplineTransformation(_ParameterizedKernelTransformation): def __init__(self, image_size, sigma, rnd_grid_params=None, diffeomorphic=False, order=2, dtype=th.float32, device='cpu'): super(ParameterizedBsplineTransformation, self).__init__(image_size, rnd_grid_params, diffeomorphic, dtype, device) self._stride = np.array(sigma) # compute bspline kernel self._kernel = utils.bspline_kernel(sigma, dim=self._dim, order=order, asTensor=True, dtype=dtype) self._padding = (np.array(self._kernel.size()) - 1) / 2 self._kernel.unsqueeze_(0).unsqueeze_(0) self._kernel = self._kernel.expand(self._dim, *((np.ones(self._dim + 1, dtype=int)*-1).tolist())) self._kernel = self._kernel.to(dtype=dtype, device=self._device) self._initialize() class RandomElasticDeformation(nn.Module): def __init__( self, num_control_points: Union[int, Tuple[int, int, int]] = 7, max_displacement: Union[float, Tuple[float, float, float]] = 7.5, locked_borders: int = 2, ): super().__init__() self.num_control_points = to_tuple(num_control_points, length=3) self.parse_control_points(self.num_control_points) self.max_displacement = to_tuple(max_displacement, length=3) self.parse_max_displacement(self.max_displacement) self.num_locked_borders = locked_borders if locked_borders not in (0, 1, 2): raise ValueError('locked_borders must be 0, 1, or 2') if locked_borders == 2 and 4 in self.num_control_points: message = ( 'Setting locked_borders to 2 and using less than 5 control' 'points results in an identity transform. Lock fewer borders' ' or use more control points.' ) raise ValueError(message) self.bspline_params = {'max_displacement':self.max_displacement, 'num_locked_borders':self.num_locked_borders} @staticmethod def parse_control_points( num_control_points: Tuple[int, int, int], ) -> None: for axis, number in enumerate(num_control_points): if not isinstance(number, int) or number < 4: message = ( f'The number of control points for axis {axis} must be' f' an integer greater than 3, not {number}' ) raise ValueError(message) @staticmethod def parse_max_displacement( max_displacement: Tuple[float, float, float], ) -> None: for axis, number in enumerate(max_displacement): if not isinstance(number, Number) or number < 0: message = ( 'The maximum displacement at each control point' f' for axis {axis} must be' f' a number greater or equal to 0, not {number}' ) raise ValueError(message) """ Images: shape of [N,D,H,W] or [N,H,W] """ def forward(self, images): bspline_transform = ParameterizedBsplineTransformation(images.size()[2:], #ignore batch and channel dim sigma=self.num_control_points, rnd_grid_params=self.bspline_params, diffeomorphic=True, order=SPLINE_ORDER, device=images.device) displacement = bspline_transform.get_displacement() inv_displacement = bspline_transform.get_inverse_displacement() warped_images = warp_image(images, displacement, multi=True) return warped_images, displacement, inv_displacement
9,147
40.022422
163
py
DDoS
DDoS-master/utils/datasets_dyn.py
# from __future__ import self.logger.debug_function, division import fnmatch import glob import os import sys from random import randint, random, seed import nibabel import numpy as np import pandas as pd import torch import torch.nn as nn import torch.nn.functional as F import torch.utils.data import torchvision.transforms as transforms from torch.utils.data import Dataset from utils.customutils import createCenterRatioMask, performUndersampling __author__ = "Soumick Chatterjee" __copyright__ = "Copyright 2022, Faculty of Computer Science, Otto von Guericke University Magdeburg, Germany" __credits__ = ["Soumick Chatterjee", "Chompunuch Sarasaen"] __license__ = "GPL" __version__ = "1.0.0" __maintainer__ = "Soumick Chatterjee" __email__ = "soumick.chatterjee@ovgu.de" __status__ = "Production" torch.manual_seed(2020) np.random.seed(2020) seed(2020) class SRDataset(Dataset): def __init__(self,logger, patch_size, dir_path, label_dir_path, stride_depth=16, stride_length=32, stride_width=32, Size=4000, fly_under_percent=None, patch_size_us=None, return_coords=False, pad_patch=True, pre_interpolate=None, norm_data=True, pre_load=False, dyn=True, noncumulative=False): self.patch_size = patch_size #-1 = full vol self.stride_depth = stride_depth self.stride_length = stride_length self.stride_width = stride_width self.size = Size self.logger = logger self.fly_under_percent = fly_under_percent #if None, then use already undersampled data. Gets priority over patch_size_us. They are both mutually exclusive self.return_coords = return_coords self.pad_patch = pad_patch self.pre_interpolate = pre_interpolate if patch_size == patch_size_us: patch_size_us = None if patch_size!=-1 and patch_size_us is not None: stride_length_us = stride_length // (patch_size//patch_size_us) stride_width_us = stride_width // (patch_size//patch_size_us) self.stride_length_us = stride_length_us self.stride_width_us = stride_width_us elif patch_size==-1: patch_size_us = None if self.fly_under_percent is not None: patch_size_us = None self.patch_size_us = patch_size_us #If already downsampled data is supplied, then this can be used. Calculate already based on the downsampling size. self.norm_data = norm_data self.pre_load = pre_load self.dyn = dyn self.noncumulative = noncumulative self.pre_loaded_lbl = {} self.pre_loaded_img = {} if not self.norm_data: print("No Norm") #TODO remove # Constants self.IMAGE_FILE_NAME = "imageFilename" self.IMAGE_FILE_SHAPE = "imageFileShape" self.IMAGE_FILE_MAXVAL = "imageFileMaxVal" self.LABEL_FILE_NAME = "labelFilename" self.LABEL_FILE_SHAPE = "labelFileShape" self.LABEL_FILE_MAXVAL = "labelFileMaxVal" self.LABEL_PREV_FILE_NAME = "labelPrevFilename" self.LABEL_PREV_FILE_SHAPE = "labelPrevFileShape" self.LABEL_PREV_FILE_MAXVAL = "labelPrevFileMaxVal" self.STARTINDEX_DEPTH = "startIndex_depth" self.STARTINDEX_LENGTH = "startIndex_length" self.STARTINDEX_WIDTH = "startIndex_width" self.STARTINDEX_DEPTH_US = "startIndex_depth_us" self.STARTINDEX_LENGTH_US = "startIndex_length_us" self.STARTINDEX_WIDTH_US = "startIndex_width_us" self.trans = transforms.ToTensor() # used to convert tiffimagefile to tensor dataDict = { self.IMAGE_FILE_NAME: [], self.IMAGE_FILE_SHAPE: [], self.IMAGE_FILE_MAXVAL:[], self.LABEL_FILE_NAME: [], self.LABEL_FILE_SHAPE: [], self.LABEL_FILE_MAXVAL:[], self.STARTINDEX_DEPTH: [],self.STARTINDEX_LENGTH: [],self.STARTINDEX_WIDTH: [], self.STARTINDEX_DEPTH_US: [],self.STARTINDEX_LENGTH_US: [],self.STARTINDEX_WIDTH_US: []} column_names = [ self.IMAGE_FILE_NAME, self.IMAGE_FILE_SHAPE, self.IMAGE_FILE_MAXVAL, self.LABEL_FILE_NAME, self.LABEL_FILE_SHAPE, self.LABEL_FILE_MAXVAL, self.STARTINDEX_DEPTH, self.STARTINDEX_LENGTH,self.STARTINDEX_WIDTH, self.STARTINDEX_DEPTH_US, self.STARTINDEX_LENGTH_US,self.STARTINDEX_WIDTH_US] self.data = pd.DataFrame(columns=column_names) files_us = glob.glob(dir_path+'/**/*.nii', recursive = True) files_us += glob.glob(dir_path+'/**/*.nii.gz', recursive = True) for imageFileName in files_us: labelFileName = imageFileName.replace(dir_path[:-1], label_dir_path[:-1]) #[:-1] is needed to remove the trailing slash for shitty windows if imageFileName == labelFileName: sys.exit('Input and Output save file') if not(os.path.isfile(imageFileName) and os.path.isfile(labelFileName)): #trick to include the other file extension if labelFileName.endswith('.nii.nii.gz'): labelFileName = labelFileName.replace('.nii.nii.gz', '.nii.gz') elif labelFileName.endswith('.nii.gz'): labelFileName = labelFileName.replace('.nii.gz', '.nii') else: labelFileName = labelFileName.replace('.nii', '.nii.gz') #check again, after replacing the file extension if not(os.path.isfile(imageFileName) and os.path.isfile(labelFileName)): self.logger.debug("skipping file as label for the corresponding image doesn't exist :"+ str(imageFileName)) continue imageFile = nibabel.load(imageFileName) # shape (Length X Width X Depth X Channels) header_shape_us = imageFile.header.get_data_shape() imageFile_data = imageFile.get_data() imageFile_max = imageFile_data.max() labelFile = nibabel.load(labelFileName) # shape (Length X Width X Depth X Channels) - changed to label file name as input image can have different (lower) size header_shape = labelFile.header.get_data_shape() labelFile_data = labelFile.get_data() labelFile_max = labelFile_data.max() self.logger.debug(header_shape) n_depth,n_length,n_width = header_shape[2],header_shape[0],header_shape[1] # gives depth which is no. of slices n_depth_us,n_length_us,n_width_us = header_shape_us[2],header_shape_us[0],header_shape_us[1] # gives depth which is no. of slices if self.pre_load: self.pre_loaded_img[imageFileName] = imageFile_data self.pre_loaded_lbl[labelFileName] = labelFile_data if patch_size!=1 and (n_depth<patch_size or n_length<patch_size or n_width<patch_size): self.logger.debug("skipping file because of its size being less than the patch size :"+ str(imageFileName)) continue ############ Following the fully sampled size if patch_size != -1: depth_i =0 ranger_depth = int((n_depth-patch_size)/stride_depth)+1 for depth_index in range(ranger_depth if n_depth%patch_size==0 else ranger_depth+1): # iterate through the whole image voxel, and extract patch length_i = 0 # self.logger.debug("depth") # self.logger.debug(depth_i) ranger_length = int((n_length-patch_size)/stride_length)+1 for length_index in range(ranger_length if n_length%patch_size==0 else ranger_length+1): width_i = 0 # self.logger.debug("length") # self.logger.debug(length_i) ranger_width = int((n_width - patch_size)/stride_width)+1 for width_index in range(ranger_width if n_width%patch_size==0 else ranger_width+1): # self.logger.debug("width") # self.logger.debug(width_i) dataDict[self.IMAGE_FILE_NAME].append(imageFileName) dataDict[self.IMAGE_FILE_SHAPE].append(header_shape_us) dataDict[self.IMAGE_FILE_MAXVAL].append(imageFile_max) dataDict[self.LABEL_FILE_NAME].append(labelFileName) dataDict[self.LABEL_FILE_SHAPE].append(header_shape) dataDict[self.LABEL_FILE_MAXVAL].append(labelFile_max) dataDict[self.STARTINDEX_DEPTH].append(depth_i) dataDict[self.STARTINDEX_LENGTH].append(length_i) dataDict[self.STARTINDEX_WIDTH].append(width_i) if patch_size_us is None: #data is zero padded dataDict[self.STARTINDEX_DEPTH_US].append(depth_i) dataDict[self.STARTINDEX_LENGTH_US].append(length_i) dataDict[self.STARTINDEX_WIDTH_US].append(width_i) width_i += stride_width length_i += stride_length depth_i += stride_depth else: dataDict[self.IMAGE_FILE_NAME].append(imageFileName) dataDict[self.IMAGE_FILE_SHAPE].append(header_shape_us) dataDict[self.IMAGE_FILE_MAXVAL].append(imageFile_max) dataDict[self.LABEL_FILE_NAME].append(labelFileName) dataDict[self.LABEL_FILE_SHAPE].append(header_shape) dataDict[self.LABEL_FILE_MAXVAL].append(labelFile_max) dataDict[self.STARTINDEX_DEPTH].append(0) dataDict[self.STARTINDEX_LENGTH].append(0) dataDict[self.STARTINDEX_WIDTH].append(0) dataDict[self.STARTINDEX_DEPTH_US].append(0) dataDict[self.STARTINDEX_LENGTH_US].append(0) dataDict[self.STARTINDEX_WIDTH_US].append(0) ############ Following the undersampled size, only if patch_size_us has been provied if patch_size_us is not None: depth_i =0 ranger_depth = int((n_depth_us-patch_size_us)/stride_depth)+1 for depth_index in range(ranger_depth if n_depth_us%patch_size_us==0 else ranger_depth+1): # iterate through the whole image voxel, and extract patch length_i = 0 # self.logger.debug("depth") # self.logger.debug(depth_i) ranger_length = int((n_length_us-patch_size_us)/stride_length_us)+1 for length_index in range(ranger_length if n_length_us%patch_size_us==0 else ranger_length+1): width_i = 0 # self.logger.debug("length") # self.logger.debug(length_i) ranger_width = int((n_width_us - patch_size_us)/stride_width_us)+1 for width_index in range(ranger_width if n_width_us%patch_size_us==0 else ranger_width+1): # self.logger.debug("width") # self.logger.debug(width_i) dataDict[self.STARTINDEX_DEPTH_US].append(depth_i) dataDict[self.STARTINDEX_LENGTH_US].append(length_i) dataDict[self.STARTINDEX_WIDTH_US].append(width_i) width_i += stride_width_us length_i += stride_length_us depth_i += stride_depth self.data = pd.DataFrame.from_dict(dataDict) self.logger.debug(len(self.data)) if self.dyn: inp_dicts, files_inp = self._process_TPs(files_us) files_gt = glob.glob(label_dir_path+'/**/*.nii', recursive = True) files_gt += glob.glob(label_dir_path+'/**/*.nii.gz', recursive = True) gt_dicts, _ = self._process_TPs(files_gt) tp_dicts = [] for filename in files_inp: inp_files = [d for d in inp_dicts if filename in d['filename']] gt_files = [d for d in gt_dicts if filename in d['filename']] tps = list(set(dic["tp"] for dic in inp_files)) tp_prev = tps.pop(0) for tp in tps: # inp_tp_prev = [d for d in inp_files if tp_prev == d['tp']] gt_tp_prev = [d for d in gt_files if tp_prev == d['tp']] inp_tp = [d for d in inp_files if tp == d['tp']] # gt_tp = [d for d in gt_files if tp == d['tp']] tp_prev = tp if not self.noncumulative else tp_prev gt_tp_prev_datum = self.data[self.data[self.LABEL_FILE_NAME] == gt_tp_prev[0]['path']] tp_dict = { self.LABEL_PREV_FILE_NAME: gt_tp_prev[0]['path'], self.LABEL_PREV_FILE_MAXVAL: gt_tp_prev_datum[self.LABEL_FILE_MAXVAL].iloc[0], self.LABEL_PREV_FILE_SHAPE: gt_tp_prev_datum[self.LABEL_FILE_SHAPE].iloc[0], # "inp_tp_prev": inp_tp_prev[0]['path'], # "gt": gt_tp[0]['path'], "inp_tpkey": inp_tp[0]['path'], "subject_filename": filename, "tpID":tp } tp_dicts.append(tp_dict) self.tp_data = pd.DataFrame.from_dict(tp_dicts) self.data = pd.merge(self.tp_data, self.data, how="left", left_on="inp_tpkey", right_on=self.IMAGE_FILE_NAME) if Size is not None and len(self.data) > Size: self.logger.debug('Dataset is larger tham supplied size. Choosing s subset randomly of size '+str(Size)) self.data = self.data.sample(n = Size, replace = False, random_state=2020) if patch_size!=-1 and fly_under_percent is not None: self.mask = createCenterRatioMask(np.zeros((patch_size,patch_size,patch_size)), fly_under_percent) def _process_TPs(self, files): f_dicts = [] for f in files: f_info = {"path": f} f_parts = os.path.normpath(f).split(os.sep) tp = fnmatch.filter(f_parts, "TP*")[0] f_info["filename"] = "_".join(f_parts[f_parts.index(tp)+1:]) f_info["tp"] = int(tp[2:]) f_dicts.append(f_info) f_dicts = sorted(f_dicts, key=lambda k: k['tp']) filenames = list(set(dic["filename"] for dic in f_dicts)) return f_dicts, filenames def __len__(self): return len(self.data) def __getitem__(self, index): imageFile_max = self.data.iloc[index][self.IMAGE_FILE_MAXVAL] labelFile_max = self.data.iloc[index][self.LABEL_FILE_MAXVAL] if self.pre_load: groundTruthImages = self.pre_loaded_lbl[self.data.iloc[index][self.LABEL_FILE_NAME]] groundTruthImages_handler = groundTruthImages else: groundTruthImages = nibabel.load(self.data.iloc[index][self.LABEL_FILE_NAME]) groundTruthImages_handler = groundTruthImages.dataobj startIndex_depth = self.data.iloc[index][self.STARTINDEX_DEPTH] startIndex_length = self.data.iloc[index][self.STARTINDEX_LENGTH] startIndex_width = self.data.iloc[index][self.STARTINDEX_WIDTH] start_coords = [(startIndex_depth, startIndex_length, startIndex_width)] if self.patch_size_us is not None: startIndex_depth_us = self.data.iloc[index][self.STARTINDEX_DEPTH_US] startIndex_length_us = self.data.iloc[index][self.STARTINDEX_LENGTH_US] startIndex_width_us = self.data.iloc[index][self.STARTINDEX_WIDTH_US] start_coords = start_coords + [(startIndex_depth_us, startIndex_length_us, startIndex_width_us)] if self.patch_size != -1: if len(groundTruthImages.shape) == 4: #don't know why, but an additional dim is noticed in some of the fully-sampled NIFTIs target_voxel = groundTruthImages_handler[startIndex_length:startIndex_length+self.patch_size, startIndex_width:startIndex_width+self.patch_size, 0, startIndex_depth:startIndex_depth+self.patch_size]#.squeeze() else: target_voxel = groundTruthImages_handler[startIndex_length:startIndex_length+self.patch_size, startIndex_width:startIndex_width+self.patch_size, startIndex_depth:startIndex_depth+self.patch_size]#.squeeze() else: if len(groundTruthImages.shape) == 4: #don't know why, but an additional dim is noticed in some of the fully-sampled NIFTIs target_voxel = groundTruthImages_handler[:, :, 0, :]#.squeeze() else: target_voxel = groundTruthImages_handler[...]#.squeeze() if self.fly_under_percent is not None: if self.patch_size != -1: voxel = abs(performUndersampling(np.array(target_voxel).copy(), mask=self.mask, zeropad=False)) voxel = voxel[...,::2] #2 for 25% - harcoded. TODO fix it else: mask = createCenterRatioMask(target_voxel, self.fly_under_percent) voxel = abs(performUndersampling(np.array(target_voxel).copy(), mask=mask, zeropad=False)) voxel = voxel[...,::2] #2 for 25% - harcoded. TODO fix it else: if self.pre_load: images = self.pre_loaded_img[self.data.iloc[index][self.IMAGE_FILE_NAME]] images_handler = images else: images = nibabel.load(self.data.iloc[index][self.IMAGE_FILE_NAME]) images_handler = images.dataobj images = nibabel.load(self.data.iloc[index][self.IMAGE_FILE_NAME]) if self.patch_size_us is not None: voxel = images_handler[startIndex_length_us:startIndex_length_us+self.patch_size_us, startIndex_width_us:startIndex_width_us+self.patch_size_us, startIndex_depth_us:startIndex_depth_us+self.patch_size]#.squeeze() else: if self.patch_size != -1 and self.pre_interpolate is None: voxel = images_handler[startIndex_length:startIndex_length+self.patch_size, startIndex_width:startIndex_width+self.patch_size, startIndex_depth:startIndex_depth+self.patch_size]#.squeeze() else: voxel = images_handler[...] target_slices = np.moveaxis(np.array(target_voxel), -1, 0).astype( np.float32) # get slices in range, convert to array, change axis of depth (because nibabel gives LXWXD, but we need in DXLXW) slices = np.moveaxis(np.array(voxel),-1, 0).astype(np.float32) #get slices in range, convert to array, change axis of depth (because nibabel gives LXWXD, but we need in DXLXW) patch = torch.from_numpy(slices) # patch = patch/torch.max(patch)# normalisation if self.pre_interpolate: patch = F.interpolate(patch.unsqueeze(0).unsqueeze(0), size=tuple(np.roll(groundTruthImages.shape, 1)), mode=self.pre_interpolate, align_corners=False).squeeze() if self.patch_size != -1: patch = patch[startIndex_depth:startIndex_depth+self.patch_size, startIndex_length:startIndex_length+self.patch_size, startIndex_width:startIndex_width+self.patch_size] if self.norm_data: patch = patch/imageFile_max# normalisation targetPatch = torch.from_numpy(target_slices) # targetPatch = targetPatch/torch.max(targetPatch) if self.norm_data: targetPatch = targetPatch/labelFile_max if self.dyn: if self.pre_load: prevTPImages = self.pre_loaded_lbl[self.data.iloc[index][self.LABEL_PREV_FILE_NAME]] prevTPImages_handler = prevTPImages else: prevTPImages = nibabel.load(self.data.iloc[index][self.LABEL_PREV_FILE_NAME]) prevTPImages_handler = prevTPImages.dataobj if self.patch_size != -1: if len(prevTPImages.shape) == 4: #don't know why, but an additional dim is noticed in some of the fully-sampled NIFTIs prevTP_voxel = prevTPImages_handler[startIndex_length:startIndex_length+self.patch_size, startIndex_width:startIndex_width+self.patch_size, 0, startIndex_depth:startIndex_depth+self.patch_size]#.squeeze() else: prevTP_voxel = prevTPImages_handler[startIndex_length:startIndex_length+self.patch_size, startIndex_width:startIndex_width+self.patch_size, startIndex_depth:startIndex_depth+self.patch_size]#.squeeze() else: if len(prevTPImages.shape) == 4: #don't know why, but an additional dim is noticed in some of the fully-sampled NIFTIs prevTP_voxel = prevTPImages_handler[:, :, 0, :]#.squeeze() else: prevTP_voxel = prevTPImages_handler[...]#.squeeze() prevTP_slices = np.moveaxis(np.array(prevTP_voxel), -1, 0).astype(np.float32) prevTPPatch = torch.from_numpy(prevTP_slices) # prevTPPatch = prevTPPatch/torch.max(prevTPPatch) if self.norm_data: prevTPPatch = prevTPPatch/self.data.iloc[index][self.LABEL_PREV_FILE_MAXVAL] #to deal the patches which has smaller size if self.pad_patch: pad = () for dim in range(len(targetPatch.shape)): pad_needed = self.patch_size - targetPatch.shape[dim] pad_dim = (pad_needed//2, pad_needed-(pad_needed//2)) pad += pad_dim if self.patch_size_us is None and self.fly_under_percent is None: pad_us = pad else: pad_us = () if self.patch_size_us is None and self.fly_under_percent is not None: real_patch_us = int(self.patch_size * (self.fly_under_percent*2)) #TODO: works for 25%, but not sure about others. Need to fix the logic else: real_patch_us = self.patch_size_us for dim in range(len(patch.shape)): pad_needed = real_patch_us - patch.shape[dim] pad_dim = (pad_needed//2, pad_needed-(pad_needed//2)) pad_us += pad_dim patch = F.pad(patch, pad_us[::-1]) #tuple has to be reveresed before using it for padding. As the tuple contains in DHW manner, and input is needed as WHD mannger targetPatch = F.pad(targetPatch, pad[::-1]) if self.dyn: prevTPPatch = F.pad(prevTPPatch, pad[::-1]) else: pad = None if self.dyn: patch = torch.stack([prevTPPatch, patch]) else: patch = patch.unsqueeze(0) targetPatch = targetPatch.unsqueeze(0) if self.return_coords is True: lblfilename = self.data.iloc[index][self.LABEL_FILE_NAME] return patch, targetPatch, np.array(start_coords), os.path.basename(os.path.dirname(lblfilename)) +"_"+os.path.basename(lblfilename), np.array([(self.data.iloc[index][self.LABEL_FILE_SHAPE]), (self.data.iloc[index][self.IMAGE_FILE_SHAPE])]), np.array(pad[::-1]) if pad is not None else -1 else: return patch, targetPatch # DATASET_FOLDER = "/nfs1/schatter/Chimp/data_3D_sr/" # DATASET_FOLDER = r"S:\MEMoRIAL_SharedStorage_M1.2+4+7\Data\Skyra\unet_3D_sr" # US_Folder = 'Center25Mask' # patch_size=64 # import logging # logger = logging.getLogger('x') # traindataset = SRDataset(logger, patch_size, DATASET_FOLDER + '/usVal/' + US_Folder + '/', DATASET_FOLDER + '/hrVal/', stride_depth =64, # stride_length=64, stride_width=64,Size =10, patch_size_us=None, return_coords=True) # train_loader = torch.utils.data.DataLoader(traindataset, batch_size=8, shuffle=True) # for epoch in range(3): # for batch_index, (local_batch, local_labels) in enumerate(train_loader): # self.logger.debug(str(epoch) + " "+ str(batch_index))
24,444
54.938215
300
py
DDoS
DDoS-master/utils/data.py
import fnmatch import os import random from glob import glob import numpy as np import torch import torchio as tio from torchio.data.io import read_image from .motion import MotionCorrupter __author__ = "Soumick Chatterjee" __copyright__ = "Copyright 2022, Faculty of Computer Science, Otto von Guericke University Magdeburg, Germany" __credits__ = ["Soumick Chatterjee"] __license__ = "GPL" __version__ = "1.0.0" __maintainer__ = "Soumick Chatterjee" __email__ = "soumick.chatterjee@ovgu.de" __status__ = "Production" def create_trainDS(path, p=1, **kwargs): files = glob(path+"/**/*.nii", recursive=True) + glob(path+"/**/*.nii.gz", recursive=True) subjects = [] for file in files: subjects.append(tio.Subject( im=tio.ScalarImage(file), filename=os.path.basename(file), )) moco = MotionCorrupter(**kwargs) transforms = [ tio.Lambda(moco.perform, p = p) ] transform = tio.Compose(transforms) subjects_dataset = tio.SubjectsDataset(subjects, transform=transform) return subjects_dataset def create_trainDS_precorrupt(path_gt, path_corrupt, p=1, norm_mode=0): files = glob(path_gt+"/**/*.nii", recursive=True) + glob(path_gt+"/**/*.nii.gz", recursive=True) subjects = [] for file in files: subjects.append(tio.Subject( im=tio.ScalarImage(file), filename=os.path.basename(file), )) transforms = [ ReadCorrupted(path_corrupt=path_corrupt, p=p, norm_mode=norm_mode) ] transform = tio.Compose(transforms) subjects_dataset = tio.SubjectsDataset(subjects, transform=transform) return subjects_dataset def createTIODS(path_gt, path_corrupt, is_infer=False, p=1, transforms = [], **kwargs): files_gt = glob(path_gt+"/**/*.nii", recursive=True) + glob(path_gt+"/**/*.nii.gz", recursive=True) if path_corrupt: files_inp = glob(path_corrupt+"/**/*.nii", recursive=True) + glob(path_corrupt+"/**/*.nii.gz", recursive=True) corruptFly = False else: files_inp = files_gt.copy() corruptFly = True subjects = [] for file in files_inp: filename = os.path.basename(file) gt_files = [f for f in files_gt if filename in f] if len(gt_files) > 0: gt_path = gt_files[0] files_gt.remove(gt_path) subjects.append(tio.Subject( gt=tio.ScalarImage(gt_path), inp=tio.ScalarImage(file), filename=filename, tag="CorruptNGT", )) if corruptFly: moco = MotionCorrupter(**kwargs) transforms.append(tio.Lambda(moco.perform, p = p)) transform = tio.Compose(transforms) subjects_dataset = tio.SubjectsDataset(subjects, transform=transform) return subjects_dataset def __process_TPs(files): f_dicts = [] for f in files: f_info = {"path": f} f_parts = os.path.normpath(f).split(os.sep) tp = fnmatch.filter(f_parts, "TP*")[0] f_info["filename"] = "_".join(f_parts[f_parts.index(tp)+1:]) f_info["tp"] = int(tp[2:]) f_dicts.append(f_info) f_dicts = sorted(f_dicts, key=lambda k: k['tp']) filenames = list(set(dic["filename"] for dic in f_dicts)) return f_dicts, filenames class ProcessTIOSubsTPs(): def __init__(self): pass def __call__(self, subject): gt_tp_prev = subject['gt_tp_prev'][tio.DATA] inp_tp = subject['inp'][tio.DATA] subject["inp"][tio.DATA] = torch.cat([gt_tp_prev, inp_tp], dim=0) return subject def createTIODynDS(path_gt, path_corrupt, is_infer=False, p=1, transforms = [], **kwargs): files_gt = glob(path_gt+"/**/*.nii", recursive=True) + glob(path_gt+"/**/*.nii.gz", recursive=True) if path_corrupt: files_inp = glob(path_corrupt+"/**/*.nii", recursive=True) + glob(path_corrupt+"/**/*.nii.gz", recursive=True) corruptFly = False else: files_inp = files_gt.copy() corruptFly = True subjects = [] inp_dicts, files_inp = __process_TPs(files_inp) gt_dicts, _ = __process_TPs(files_gt) for filename in files_inp: inp_files = [d for d in inp_dicts if filename in d['filename']] gt_files = [d for d in gt_dicts if filename in d['filename']] tps = list(set(dic["tp"] for dic in inp_files)) tp_prev = tps.pop(0) for tp in tps: inp_tp_prev = [d for d in inp_files if tp_prev == d['tp']] gt_tp_prev = [d for d in gt_files if tp_prev == d['tp']] inp_tp = [d for d in inp_files if tp == d['tp']] gt_tp = [d for d in gt_files if tp == d['tp']] tp_prev = tp if len(gt_tp_prev) > 0 and len(gt_tp) > 0: subjects.append(tio.Subject( gt_tp_prev=tio.ScalarImage(gt_tp_prev[0]['path']), inp_tp_prev=tio.ScalarImage(inp_tp_prev[0]['path']), gt=tio.ScalarImage(gt_tp[0]['path']), inp=tio.ScalarImage(inp_tp[0]['path']), filename=filename, tp=tp, tag="CorruptNGT", )) else: print("Warning: Not Implemented if GT is missing. Skipping Sub-TP.") continue if corruptFly: moco = MotionCorrupter(**kwargs) transforms.append(tio.Lambda(moco.perform, p = p)) transforms.append(ProcessTIOSubsTPs()) transform = tio.Compose(transforms) subjects_dataset = tio.SubjectsDataset(subjects, transform=transform) return subjects_dataset def create_patchDS(train_subs, val_subs, patch_size, patch_qlen, patch_per_vol, inference_strides): train_queue = None val_queue = None if train_subs is not None: sampler = tio.data.UniformSampler(patch_size) train_queue = tio.Queue( subjects_dataset=train_subs, max_length=patch_qlen, samples_per_volume=patch_per_vol, sampler=sampler, num_workers=0, start_background=True ) if val_subs is not None: overlap = np.subtract(patch_size, inference_strides) grid_samplers = [] for i in range(len(val_subs)): grid_sampler = tio.inference.GridSampler(val_subs[i], patch_size, overlap) grid_samplers.append(grid_sampler) val_queue = torch.utils.data.ConcatDataset(grid_samplers) return train_queue, val_queue class ReadCorrupted(tio.transforms.Transform): def __init__(self, path_corrupt, p=1, norm_mode=0): super().__init__(p=p) self.path_corrupt=path_corrupt self.norm_mode = norm_mode def apply_transform(self, subject): corrupted_query = subject.filename.split(".")[0]+"*" files = glob(self.path_corrupt+"/**/"+corrupted_query, recursive=True) corrupt_path = files[random.randint(0, len(files)-1)] transformed, _ = read_image(corrupt_path) vol = subject['im'][tio.DATA].float() transformed = transformed.float() if self.norm_mode==1: vol = vol/vol.max() transformed = transformed/transformed.max() elif self.norm_mode==2: vol = (vol-vol.min())/(vol.max()-vol.min()) transformed = (transformed-transformed.min())/(transformed.max()-transformed.min()) subject['im'][tio.DATA] = torch.cat([vol,transformed], 0) return subject
7,985
38.93
118
py
DDoS
DDoS-master/utils/interpnorm_vols.py
import os import random from glob import glob import nibabel as nib import numpy as np import torch import torch.nn.functional as F from tqdm import tqdm __author__ = "Soumick Chatterjee" __copyright__ = "Copyright 2022, Faculty of Computer Science, Otto von Guericke University Magdeburg, Germany" __credits__ = ["Soumick Chatterjee", "Chompunuch Sarasaen"] __license__ = "GPL" __version__ = "1.0.0" __maintainer__ = "Soumick Chatterjee" __email__ = "soumick.chatterjee@ovgu.de" __status__ = "Production" path_woZPad = "/project/schatter/Chimp/Data/usCHAOSWoT2/Center6p25MaskWoPad" path_GT = "/project/schatter/Chimp/Data/hrCHAOS" outpath_interpNorm = "/project/schatter/Chimp/Data/usCHAOSWoT2/Center6p25MaskWoPad_Tri_Norm" outpath_GTNorm = "/project/schatter/Chimp/Data/hrCHAOS_Norm" files = glob(path_woZPad+"/**/*.nii", recursive=True) + glob(path_woZPad+"/**/*.nii.gz", recursive=True) files_gt = glob(path_GT+"/**/*.nii", recursive=True) + glob(path_GT+"/**/*.nii.gz", recursive=True) def SaveNIFTI(data, path): os.makedirs(os.path.split(path)[0], exist_ok=True) nib.save(nib.Nifti1Image(data, np.eye(4)), path) for file in tqdm(files): filename = os.path.basename(file) gt_files = [f for f in files_gt if filename in f] gt_path = gt_files[0] gt = nib.load(gt_path).dataobj[...] gt_max = gt.max() gt = (gt.astype(np.float32))/gt_max SaveNIFTI(gt, gt_path.replace(path_GT, outpath_GTNorm)) images = nib.load(file).dataobj[...] img_max = images.max() images = torch.from_numpy(images.astype(np.float32)) images = F.interpolate(images.unsqueeze(0).unsqueeze(0), mode="trilinear", size=gt.shape).squeeze() images = (images/img_max).numpy() SaveNIFTI(images, file.replace(path_woZPad, outpath_interpNorm))
1,778
35.306122
110
py
DDoS
DDoS-master/utils/utilities.py
import os from copy import deepcopy from statistics import median import random import nibabel as nib import numpy as np import torch import torch.nn.functional as F import torchcomplex.nn.functional as cF import torchio as tio import torchvision.utils as vutils from scipy import ndimage import wandb from pynufft import NUFFT from sewar.full_ref import ssim as SSIM2DCalc from sewar.full_ref import uqi as UQICalc from skimage.metrics import (normalized_root_mse, peak_signal_noise_ratio, structural_similarity) from torchcomplex.utils.signaltools import resample from tricorder.math.transforms.fourier import fftNc_np, ifftNc_np from utils.elastic_transform import RandomElasticDeformation, warp_image __author__ = "Soumick Chatterjee" __copyright__ = "Copyright 2022, Faculty of Computer Science, Otto von Guericke University Magdeburg, Germany" __credits__ = ["Soumick Chatterjee", "Chompunuch Sarasaen"] __license__ = "GPL" __version__ = "1.0.0" __maintainer__ = "Soumick Chatterjee" __email__ = "soumick.chatterjee@ovgu.de" __status__ = "Production" class Interpolator(): def __init__(self, mode=None): if mode in ["sinc", "nearest", "linear", "bilinear", "bicubic", "trilinear", "area"]: self.mode = mode else: self.mode = None def perform_sinc(self, images, out_shape): axes = np.argwhere(np.equal(images.shape[2:], out_shape) == False).squeeze(1) #2 dims for batch and channel out_shape = [out_shape[i] for i in axes] return resample(images, out_shape, axis=axes+2) #2 dims for batch and channel def __call__(self, images, out_shape): if self.mode is None: return images elif images.is_complex(): return cF.interpolate(images, size=out_shape, mode=self.mode) elif self.mode == "sinc": return self.perform_sinc(images, out_shape) else: return F.interpolate(images, size=out_shape, mode=self.mode) def tensorboard_images(writer, inputs, outputs, targets, epoch, section='train'): writer.add_image('{}/output'.format(section), vutils.make_grid(outputs[0, 0, ...], normalize=True, scale_each=True), epoch) if inputs is not None: writer.add_image('{}/input'.format(section), vutils.make_grid(inputs[0, 0, ...], normalize=True, scale_each=True), epoch) if targets is not None: writer.add_image('{}/target'.format(section), vutils.make_grid(targets[0, 0, ...], normalize=True, scale_each=True), epoch) def SaveNIFTI(data, file_path): """Save a NIFTI file using given file path from an array Using: NiBabel""" if(np.iscomplex(data).any()): data = abs(data) nii = nib.Nifti1Image(data, np.eye(4)) nib.save(nii, file_path) def sharpTP(vol, alpha=0.5): filteredVOL = ndimage.gaussian_filter(vol, 1) return vol + alpha * (vol - filteredVOL) def applyDCS(output, fully, under_mask=None, missing_mask=None, mat=None, isCartesian=True, norm=True): if norm: fully /= fully.max() output /= output.max() if isCartesian: fullyK = fftNc_np(fully, axes=(0,1)) underK = fullyK*under_mask outK = fftNc_np(output, axes=(0,1)) missingK = outK*missing_mask finalK = underK+missingK return abs(ifftNc_np(finalK, axes=(0,1))) else: om = mat['om'] invom = mat['invom'] fullom = mat['fullom'] dcfFullRes = mat['dcfFullRes'].squeeze() imageSize = fully.shape[0] baseresolution = imageSize*2 interpolationSize4NUFFT = 6 NufftObjOM = NUFFT() NufftObjInvOM = NUFFT() NufftObjFullOM = NUFFT() Nd = (baseresolution, baseresolution) # image size Kd = (baseresolution*2, baseresolution*2) # k-space size - TODO: multiply back by 2 Jd = (interpolationSize4NUFFT, interpolationSize4NUFFT) # interpolation size NufftObjOM.plan(om, Nd, Kd, Jd) NufftObjInvOM.plan(invom, Nd, Kd, Jd) NufftObjFullOM.plan(fullom, Nd, Kd, Jd) for slc in range(fully.shape[-1]): oversam_fully = np.zeros((baseresolution,baseresolution), dtype=fully.dtype) oversam_fully[imageSize//2:imageSize+imageSize//2,imageSize//2:imageSize+imageSize//2] = fully[...,slc] oversam_output = np.zeros((baseresolution,baseresolution), dtype=output.dtype) oversam_output[imageSize//2:imageSize+imageSize//2,imageSize//2:imageSize+imageSize//2] = output[...,slc] yUnder = NufftObjOM.forward(oversam_fully) yMissing = NufftObjInvOM.forward(oversam_output) yCorrected = np.concatenate((yUnder,yMissing)) yCorrected = np.multiply(dcfFullRes,yCorrected) oversam_output_corrected = NufftObjFullOM.adjoint(yCorrected) output_corrected = oversam_output_corrected[imageSize//2:imageSize+imageSize//2,imageSize//2:imageSize+imageSize//2] output[...,slc] = abs(output_corrected).astype(fully.dtype) return output def process_DDoS_SRPrev(SRPrev, start_coords, patch_size, pad, lr_imgs): for i in range(lr_imgs.shape[0]): (startIndex_depth, startIndex_length, startIndex_width) = start_coords[i][0].numpy() if patch_size != -1: prevTP_voxel = SRPrev[startIndex_length:startIndex_length+patch_size, startIndex_width:startIndex_width+patch_size, startIndex_depth:startIndex_depth+patch_size]#.squeeze() else: prevTP_voxel = SRPrev[...]#.squeeze() prevTP_slices = np.moveaxis(np.array(prevTP_voxel), -1, 0).astype(np.float32) prevTPPatch = torch.from_numpy(prevTP_slices) prevTPPatch = prevTPPatch/SRPrev.max() lr_imgs[i,0] = F.pad(prevTPPatch, tuple(pad[i])) return lr_imgs def process_valBatch(batch): inp = [] gt = [] gt_flag = [] for i in range(len(batch['tag'])): gt_flag.append(True) batch_tag = batch['tag'][i] if batch_tag == "CorruptNGT" or batch_tag == "GTOnly": inp.append(batch['inp'][tio.DATA][i,...]) gt.append(batch['gt'][tio.DATA][i,...]) elif batch_tag == "FlyCorrupt": gt.append(batch['im'][tio.DATA][i,0,...].unsqueeze(1)) if batch['im'][tio.DATA].shape[1] == 2: inp.append(batch['im'][tio.DATA][i,1,...].unsqueeze(1)) else: #Use motion free image inp.append(deepcopy(gt[-1])) elif batch_tag == "CorruptOnly": inp.append(batch['inp'][tio.DATA][i,...]) gt.append(batch['inp'][tio.DATA][i,...]) gt_flag[-1] = False inp = torch.stack(inp,dim=0) gt = torch.stack(gt,dim=0) return inp, gt, gt_flag def getSSIM(gt, out, gt_flag=None, data_range=1): if gt_flag is None: gt_flag = np.ones(gt.shape[0]) vals = [] for i in range(gt.shape[0]): if not bool(gt_flag[i]): continue for j in range(gt.shape[1]): vals.append(structural_similarity(gt[i,j,...], out[i,j,...], data_range=data_range)) return median(vals) def calc_metircs(gt, out, tag): ssim, ssimMAP = structural_similarity(gt, out, data_range=1, full=True) nrmse = normalized_root_mse(gt, out) psnr = peak_signal_noise_ratio(gt, out, data_range=1) uqi = UQICalc(gt, out) metrics = { "SSIM"+tag: ssim, "NRMSE"+tag: nrmse, "PSNR"+tag: psnr, "UQI"+tag: uqi } return metrics, ssimMAP def MinMax(data): return (data-data.min())/(data.max()-data.min()) def convert_image(img, source, target): """ Convert an image from a source format to a target format. :param img: image :param source: source format, one of 'pil' (PIL image), '[0, 1]' or '[-1, 1]' (pixel value ranges) :param target: target format, one of 'pil' (PIL image), '[0, 255]', '[0, 1]', '[-1, 1]' (pixel value ranges), 'imagenet-norm' (pixel values standardized by imagenet mean and std.), 'y-channel' (luminance channel Y in the YCbCr color format, used to calculate PSNR and SSIM) :return: converted image """ assert source in {'pil', '[0, 1]', '[-1, 1]'}, "Cannot convert from source format %s!" % source assert target in {'pil', '[0, 255]', '[0, 1]', '[-1, 1]', 'imagenet-norm', 'y-channel'}, "Cannot convert to target format %s!" % target # Convert from source to [0, 1] if source == 'pil': img = FT.to_tensor(img) elif source == '[0, 1]': pass # already in [0, 1] elif source == '[-1, 1]': img = (img + 1.) / 2. # Convert from [0, 1] to target if target == 'pil': img = FT.to_pil_image(img) elif target == '[0, 255]': img = 255. * img elif target == '[0, 1]': pass # already in [0, 1] elif target == '[-1, 1]': img = 2. * img - 1. elif target == 'imagenet-norm': if img.ndimension() == 3: img = (img - imagenet_mean) / imagenet_std elif img.ndimension() == 4: img = (img - imagenet_mean_cuda) / imagenet_std_cuda elif target == 'y-channel': # Based on definitions at https://github.com/xinntao/BasicSR/wiki/Color-conversion-in-SR # torch.dot() does not work the same way as numpy.dot() # So, use torch.matmul() to find the dot product between the last dimension of an 4-D tensor and a 1-D tensor img = torch.matmul(255. * img.permute(0, 2, 3, 1)[:, 4:-4, 4:-4, :], rgb_weights) / 255. + 16. return img class ResSaver(): def __init__(self, out_path, save_inp=False, save_out=True, analyse_out=True, do_norm=False): self.out_path = out_path self.save_inp = save_inp self.do_norm = do_norm self.save_out = save_out self.analyse_out = analyse_out def CalcNSave(self, out, inp, gt, outfolder, already_numpy=False): outpath = os.path.join(self.out_path, outfolder) os.makedirs(outpath, exist_ok=True) if not already_numpy: inp = inp.numpy() out = out.numpy() if self.save_out: SaveNIFTI(out, os.path.join(outpath, "out.nii.gz")) if self.save_inp: SaveNIFTI(inp, os.path.join(outpath, "inp.nii.gz")) if gt is not None: if not already_numpy: gt = gt.numpy() if self.do_norm: inp = convert_image(inp, source='[-1, 1]', target='[0, 1]') #inp/inp.max() gt = convert_image(gt, source='[-1, 1]', target='[0, 1]') #gt/gt.max() if self.analyse_out: out = convert_image(out, source='[-1, 1]', target='[0, 1]') #out/out.max() out_metrics, out_ssimMAP = calc_metircs(gt, out, tag="Out") SaveNIFTI(out_ssimMAP, os.path.join(outpath, "ssimMAPOut.nii.gz")) else: out_metrics = {} inp_metrics, inp_ssimMAP = calc_metircs(gt, inp, tag="Inp") SaveNIFTI(inp_ssimMAP, os.path.join(outpath, "ssimMAPInp.nii.gz")) metrics = {**out_metrics, **inp_metrics} return metrics #The WnB functions are here, but not been tested (even not finished) def WnB_ArtefactLog_DS(run, datasets, meta={}, names = ["training", "validation", "test"], description="Train-Val(-Test) Split"): raw_data = wandb.Artifact("DSSplit", type="dataset", description=description, metadata={"sizes": [len(dataset) for dataset in datasets], **meta}) for name, dataset in zip(names, datasets): with raw_data.new_file(name + ".npz", mode="wb") as file: np.savez(file, ds=dataset) run.log_artifact(raw_data) def WnB_ReadArtefact_DS(run, tag="latest", names = ["training", "validation", "test"]): raw_data_artifact = run.use_artifact('DSSplit:'+tag) raw_dataset = raw_data_artifact.download() datasets = [] for split in names: raw_split = np.load(os.path.join(raw_dataset, split + ".npz"))['ds'] datasets.append(raw_split) return datasets def WnB_ArtefactLog_Model(run, model, config, description="MyModel"): model_artifact = wandb.Artifact("Model", type="model", description=description, metadata=dict(config)) model.save("initialized_model.keras") model_artifact.add_file("initialized_model.keras") wandb.save("initialized_model.keras") run.log_artifact(model_artifact) def WnB_ReadArtefact_Model(run, tag="latest"): model_artifact = run.use_artifact('Model:'+tag) model_dir = model_artifact.download() model_path = os.path.join(model_dir, "initialized_model.keras") model = keras.models.load_model(model_path) model_config = model_artifact.metadata return model, model_config def deformOTF(input_batch): elastic = RandomElasticDeformation( num_control_points=random.choice([5, 6, 7]), max_displacement=random.uniform(0.7, 2.0), locked_borders=2 ) elastic.cuda() input_batch_transformed, _, _ = elastic(input_batch) input_batch_transformed = torch.nan_to_num(input_batch_transformed) return input_batch_transformed #/ torch.amax(input_batch_transformed, dim=[1,2,3,4])
13,876
38.991354
184
py
DDoS
DDoS-master/utils/datasets.py
# from __future__ import self.logger.debug_function, division import glob import os import sys from random import randint, random, seed import nibabel import numpy as np import pandas as pd import torch import torch.nn as nn import torch.nn.functional as F import torch.utils.data import torchvision.transforms as transforms from torch.utils.data import Dataset from utils.customutils import createCenterRatioMask, performUndersampling __author__ = "Soumick Chatterjee" __copyright__ = "Copyright 2022, Faculty of Computer Science, Otto von Guericke University Magdeburg, Germany" __credits__ = ["Soumick Chatterjee", "Chompunuch Sarasaen"] __license__ = "GPL" __version__ = "1.0.0" __maintainer__ = "Soumick Chatterjee" __email__ = "soumick.chatterjee@ovgu.de" __status__ = "Production" torch.manual_seed(2020) np.random.seed(2020) seed(2020) class SRDataset(Dataset): def __init__(self,logger, patch_size, dir_path, label_dir_path, stride_depth=16, stride_length=32, stride_width=32, Size=4000, fly_under_percent=None, patch_size_us=None, return_coords=False, pad_patch=True, pre_interpolate=None, norm_data=True, pre_load=False): self.patch_size = patch_size #-1 = full vol self.stride_depth = stride_depth self.stride_length = stride_length self.stride_width = stride_width self.size = Size self.logger = logger self.fly_under_percent = fly_under_percent #if None, then use already undersampled data. Gets priority over patch_size_us. They are both mutually exclusive self.return_coords = return_coords self.pad_patch = pad_patch self.pre_interpolate = pre_interpolate if patch_size == patch_size_us: patch_size_us = None if patch_size!=-1 and patch_size_us is not None: stride_length_us = stride_length // (patch_size//patch_size_us) stride_width_us = stride_width // (patch_size//patch_size_us) self.stride_length_us = stride_length_us self.stride_width_us = stride_width_us elif patch_size==-1: patch_size_us = None if self.fly_under_percent is not None: patch_size_us = None self.patch_size_us = patch_size_us #If already downsampled data is supplied, then this can be used. Calculate already based on the downsampling size. self.norm_data = norm_data self.pre_load = pre_load self.pre_loaded_lbl = {} self.pre_loaded_img = {} if not self.norm_data: print("No Norm") #TODO remove # Constants self.IMAGE_FILE_NAME = "imageFilename" self.IMAGE_FILE_SHAPE = "imageFileShape" self.IMAGE_FILE_MAXVAL = "imageFileMaxVal" self.LABEL_FILE_NAME = "labelFilename" self.LABEL_FILE_SHAPE = "labelFileShape" self.LABEL_FILE_MAXVAL = "labelFileMaxVal" self.STARTINDEX_DEPTH = "startIndex_depth" self.STARTINDEX_LENGTH = "startIndex_length" self.STARTINDEX_WIDTH = "startIndex_width" self.STARTINDEX_DEPTH_US = "startIndex_depth_us" self.STARTINDEX_LENGTH_US = "startIndex_length_us" self.STARTINDEX_WIDTH_US = "startIndex_width_us" self.trans = transforms.ToTensor() # used to convert tiffimagefile to tensor dataDict = { self.IMAGE_FILE_NAME: [], self.IMAGE_FILE_SHAPE: [], self.IMAGE_FILE_MAXVAL:[], self.LABEL_FILE_NAME: [], self.LABEL_FILE_SHAPE: [], self.LABEL_FILE_MAXVAL:[], self.STARTINDEX_DEPTH: [],self.STARTINDEX_LENGTH: [],self.STARTINDEX_WIDTH: [], self.STARTINDEX_DEPTH_US: [],self.STARTINDEX_LENGTH_US: [],self.STARTINDEX_WIDTH_US: []} column_names = [ self.IMAGE_FILE_NAME, self.IMAGE_FILE_SHAPE, self.IMAGE_FILE_MAXVAL, self.LABEL_FILE_NAME, self.LABEL_FILE_SHAPE, self.LABEL_FILE_MAXVAL, self.STARTINDEX_DEPTH, self.STARTINDEX_LENGTH,self.STARTINDEX_WIDTH, self.STARTINDEX_DEPTH_US, self.STARTINDEX_LENGTH_US,self.STARTINDEX_WIDTH_US] self.data = pd.DataFrame(columns=column_names) files_us = glob.glob(dir_path+'/**/*.nii', recursive = True) files_us += glob.glob(dir_path+'/**/*.nii.gz', recursive = True) for imageFileName in files_us: labelFileName = imageFileName.replace(dir_path[:-1], label_dir_path[:-1]) #[:-1] is needed to remove the trailing slash for shitty windows if imageFileName == labelFileName: sys.exit('Input and Output save file') if not(os.path.isfile(imageFileName) and os.path.isfile(labelFileName)): #trick to include the other file extension if labelFileName.endswith('.nii.nii.gz'): labelFileName = labelFileName.replace('.nii.nii.gz', '.nii.gz') elif labelFileName.endswith('.nii.gz'): labelFileName = labelFileName.replace('.nii.gz', '.nii') else: labelFileName = labelFileName.replace('.nii', '.nii.gz') #check again, after replacing the file extension if not(os.path.isfile(imageFileName) and os.path.isfile(labelFileName)): self.logger.debug("skipping file as label for the corresponding image doesn't exist :"+ str(imageFileName)) continue imageFile = nibabel.load(imageFileName) # shape (Length X Width X Depth X Channels) header_shape_us = imageFile.header.get_data_shape() imageFile_data = imageFile.get_data() imageFile_max = imageFile_data.max() labelFile = nibabel.load(labelFileName) # shape (Length X Width X Depth X Channels) - changed to label file name as input image can have different (lower) size header_shape = labelFile.header.get_data_shape() labelFile_data = labelFile.get_data() labelFile_max = labelFile_data.max() self.logger.debug(header_shape) n_depth,n_length,n_width = header_shape[2],header_shape[0],header_shape[1] # gives depth which is no. of slices n_depth_us,n_length_us,n_width_us = header_shape_us[2],header_shape_us[0],header_shape_us[1] # gives depth which is no. of slices if self.pre_load: self.pre_loaded_img[imageFileName] = imageFile_data self.pre_loaded_lbl[labelFileName] = labelFile_data if patch_size!=1 and (n_depth<patch_size or n_length<patch_size or n_width<patch_size): self.logger.debug("skipping file because of its size being less than the patch size :"+ str(imageFileName)) continue ############ Following the fully sampled size if patch_size != -1: depth_i =0 ranger_depth = int((n_depth-patch_size)/stride_depth)+1 for depth_index in range(ranger_depth if n_depth%patch_size==0 else ranger_depth+1): # iterate through the whole image voxel, and extract patch length_i = 0 # self.logger.debug("depth") # self.logger.debug(depth_i) ranger_length = int((n_length-patch_size)/stride_length)+1 for length_index in range(ranger_length if n_length%patch_size==0 else ranger_length+1): width_i = 0 # self.logger.debug("length") # self.logger.debug(length_i) ranger_width = int((n_width - patch_size)/stride_width)+1 for width_index in range(ranger_width if n_width%patch_size==0 else ranger_width+1): # self.logger.debug("width") # self.logger.debug(width_i) dataDict[self.IMAGE_FILE_NAME].append(imageFileName) dataDict[self.IMAGE_FILE_SHAPE].append(header_shape_us) dataDict[self.IMAGE_FILE_MAXVAL].append(imageFile_max) dataDict[self.LABEL_FILE_NAME].append(labelFileName) dataDict[self.LABEL_FILE_SHAPE].append(header_shape) dataDict[self.LABEL_FILE_MAXVAL].append(labelFile_max) dataDict[self.STARTINDEX_DEPTH].append(depth_i) dataDict[self.STARTINDEX_LENGTH].append(length_i) dataDict[self.STARTINDEX_WIDTH].append(width_i) if patch_size_us is None: #data is zero padded dataDict[self.STARTINDEX_DEPTH_US].append(depth_i) dataDict[self.STARTINDEX_LENGTH_US].append(length_i) dataDict[self.STARTINDEX_WIDTH_US].append(width_i) width_i += stride_width length_i += stride_length depth_i += stride_depth else: dataDict[self.IMAGE_FILE_NAME].append(imageFileName) dataDict[self.IMAGE_FILE_SHAPE].append(header_shape_us) dataDict[self.IMAGE_FILE_MAXVAL].append(imageFile_max) dataDict[self.LABEL_FILE_NAME].append(labelFileName) dataDict[self.LABEL_FILE_SHAPE].append(header_shape) dataDict[self.LABEL_FILE_MAXVAL].append(labelFile_max) dataDict[self.STARTINDEX_DEPTH].append(0) dataDict[self.STARTINDEX_LENGTH].append(0) dataDict[self.STARTINDEX_WIDTH].append(0) dataDict[self.STARTINDEX_DEPTH_US].append(0) dataDict[self.STARTINDEX_LENGTH_US].append(0) dataDict[self.STARTINDEX_WIDTH_US].append(0) ############ Following the undersampled size, only if patch_size_us has been provied if patch_size_us is not None: depth_i =0 ranger_depth = int((n_depth_us-patch_size_us)/stride_depth)+1 for depth_index in range(ranger_depth if n_depth_us%patch_size_us==0 else ranger_depth+1): # iterate through the whole image voxel, and extract patch length_i = 0 # self.logger.debug("depth") # self.logger.debug(depth_i) ranger_length = int((n_length_us-patch_size_us)/stride_length_us)+1 for length_index in range(ranger_length if n_length_us%patch_size_us==0 else ranger_length+1): width_i = 0 # self.logger.debug("length") # self.logger.debug(length_i) ranger_width = int((n_width_us - patch_size_us)/stride_width_us)+1 for width_index in range(ranger_width if n_width_us%patch_size_us==0 else ranger_width+1): # self.logger.debug("width") # self.logger.debug(width_i) dataDict[self.STARTINDEX_DEPTH_US].append(depth_i) dataDict[self.STARTINDEX_LENGTH_US].append(length_i) dataDict[self.STARTINDEX_WIDTH_US].append(width_i) width_i += stride_width_us length_i += stride_length_us depth_i += stride_depth self.data = pd.DataFrame.from_dict(dataDict) self.logger.debug(len(self.data)) if Size is not None and len(self.data) > Size: self.logger.debug('Dataset is larger tham supplied size. Choosing s subset randomly of size '+str(Size)) self.data = self.data.sample(n = Size, replace = False, random_state=2020) if patch_size!=-1 and fly_under_percent is not None: self.mask = createCenterRatioMask(np.zeros((patch_size,patch_size,patch_size)), fly_under_percent) def __len__(self): return len(self.data) def __getitem__(self, index): ''' imageFilename: 0 imageFileShape: 1 imageFileMaxVal: 2 labelFilename: 3 labelFileShape: 4 labelFileMaxVal: 5 startIndex_depth : 6 startIndex_length : 7 startIndex_width : 8 startIndex_depth_us : 9 startIndex_length_us : 10 startIndex_width_us : 11 ''' imageFile_max = self.data.iloc[index, 2] labelFile_max = self.data.iloc[index, 5] if self.pre_load: groundTruthImages = self.pre_loaded_lbl[self.data.iloc[index, 3]] groundTruthImages_handler = groundTruthImages else: groundTruthImages = nibabel.load(self.data.iloc[index, 3]) groundTruthImages_handler = groundTruthImages.dataobj startIndex_depth = self.data.iloc[index, 6] startIndex_length = self.data.iloc[index, 7] startIndex_width = self.data.iloc[index, 8] start_coords = [(startIndex_depth, startIndex_length, startIndex_width)] if self.patch_size_us is not None: startIndex_depth_us = self.data.iloc[index, 9] startIndex_length_us = self.data.iloc[index, 10] startIndex_width_us = self.data.iloc[index, 11] start_coords = start_coords + [(startIndex_depth_us, startIndex_length_us, startIndex_width_us)] if self.patch_size != -1: if len(groundTruthImages.shape) == 4: #don't know why, but an additional dim is noticed in some of the fully-sampled NIFTIs target_voxel = groundTruthImages_handler[startIndex_length:startIndex_length+self.patch_size, startIndex_width:startIndex_width+self.patch_size, 0, startIndex_depth:startIndex_depth+self.patch_size]#.squeeze() else: target_voxel = groundTruthImages_handler[startIndex_length:startIndex_length+self.patch_size, startIndex_width:startIndex_width+self.patch_size, startIndex_depth:startIndex_depth+self.patch_size]#.squeeze() else: if len(groundTruthImages.shape) == 4: #don't know why, but an additional dim is noticed in some of the fully-sampled NIFTIs target_voxel = groundTruthImages_handler[:, :, 0, :]#.squeeze() else: target_voxel = groundTruthImages_handler[...]#.squeeze() if self.fly_under_percent is not None: if self.patch_size != -1: voxel = abs(performUndersampling(np.array(target_voxel).copy(), mask=self.mask, zeropad=False)) voxel = voxel[...,::2] #2 for 25% - harcoded. TODO fix it else: mask = createCenterRatioMask(target_voxel, self.fly_under_percent) voxel = abs(performUndersampling(np.array(target_voxel).copy(), mask=mask, zeropad=False)) voxel = voxel[...,::2] #2 for 25% - harcoded. TODO fix it else: if self.pre_load: images = self.pre_loaded_img[self.data.iloc[index, 0]] images_handler = images else: images = nibabel.load(self.data.iloc[index, 0]) images_handler = images.dataobj images = nibabel.load(self.data.iloc[index, 0]) if self.patch_size_us is not None: voxel = images_handler[startIndex_length_us:startIndex_length_us+self.patch_size_us, startIndex_width_us:startIndex_width_us+self.patch_size_us, startIndex_depth_us:startIndex_depth_us+self.patch_size]#.squeeze() else: if self.patch_size != -1 and self.pre_interpolate is None: voxel = images_handler[startIndex_length:startIndex_length+self.patch_size, startIndex_width:startIndex_width+self.patch_size, startIndex_depth:startIndex_depth+self.patch_size]#.squeeze() else: voxel = images_handler[...] target_slices = np.moveaxis(np.array(target_voxel), -1, 0).astype( np.float32) # get slices in range, convert to array, change axis of depth (because nibabel gives LXWXD, but we need in DXLXW) slices = np.moveaxis(np.array(voxel),-1, 0).astype(np.float32) #get slices in range, convert to array, change axis of depth (because nibabel gives LXWXD, but we need in DXLXW) patch = torch.from_numpy(slices) # patch = patch/torch.max(patch)# normalisation if self.pre_interpolate: patch = F.interpolate(patch.unsqueeze(0).unsqueeze(0), size=tuple(np.roll(groundTruthImages.shape, 1)), mode=self.pre_interpolate, align_corners=False).squeeze() if self.patch_size != -1: patch = patch[startIndex_depth:startIndex_depth+self.patch_size, startIndex_length:startIndex_length+self.patch_size, startIndex_width:startIndex_width+self.patch_size] if self.norm_data: patch = patch/imageFile_max# normalisation targetPatch = torch.from_numpy(target_slices) # targetPatch = targetPatch/torch.max(targetPatch) if self.norm_data: targetPatch = targetPatch/labelFile_max #to deal the patches which has smaller size if self.pad_patch: pad = () for dim in range(len(targetPatch.shape)): pad_needed = self.patch_size - targetPatch.shape[dim] pad_dim = (pad_needed//2, pad_needed-(pad_needed//2)) pad += pad_dim if self.patch_size_us is None and self.fly_under_percent is None: pad_us = pad else: pad_us = () if self.patch_size_us is None and self.fly_under_percent is not None: real_patch_us = int(self.patch_size * (self.fly_under_percent*2)) #TODO: works for 25%, but not sure about others. Need to fix the logic else: real_patch_us = self.patch_size_us for dim in range(len(patch.shape)): pad_needed = real_patch_us - patch.shape[dim] pad_dim = (pad_needed//2, pad_needed-(pad_needed//2)) pad_us += pad_dim patch = F.pad(patch, pad_us[::-1]) #tuple has to be reveresed before using it for padding. As the tuple contains in DHW manner, and input is needed as WHD mannger targetPatch = F.pad(targetPatch, pad[::-1]) else: pad = None if self.return_coords is True: return patch, targetPatch, np.array(start_coords), os.path.basename(self.data.iloc[index, 3]), np.array([(self.data.iloc[index, 4]), (self.data.iloc[index, 1])]), np.array(pad[::-1]) if pad is not None else -1 else: return patch, targetPatch # DATASET_FOLDER = "/nfs1/schatter/Chimp/data_3D_sr/" # DATASET_FOLDER = r"S:\MEMoRIAL_SharedStorage_M1.2+4+7\Data\Skyra\unet_3D_sr" # US_Folder = 'Center25Mask' # patch_size=64 # import logging # logger = logging.getLogger('x') # traindataset = SRDataset(logger, patch_size, DATASET_FOLDER + '/usVal/' + US_Folder + '/', DATASET_FOLDER + '/hrVal/', stride_depth =64, # stride_length=64, stride_width=64,Size =10, patch_size_us=None, return_coords=True) # train_loader = torch.utils.data.DataLoader(traindataset, batch_size=8, shuffle=True) # for epoch in range(3): # for batch_index, (local_batch, local_labels) in enumerate(train_loader): # self.logger.debug(str(epoch) + " "+ str(batch_index))
19,535
53.266667
261
py
DDoS
DDoS-master/utils/motion.py
import math import multiprocessing.dummy as multiprocessing import random from collections import defaultdict from typing import List import numpy as np import SimpleITK as sitk import torch import torchio as tio from scipy.ndimage import affine_transform from torchio.transforms import Motion, RandomMotion from torchio.transforms.interpolation import Interpolation # import multiprocessing __author__ = "Soumick Chatterjee, Alessandro Sciarra" __copyright__ = "Copyright 2022, Faculty of Computer Science, Otto von Guericke University Magdeburg, Germany" __credits__ = ["Soumick Chatterjee", "Alessandro Sciarra"] __license__ = "GPL" __version__ = "1.0.0" __maintainer__ = "Soumick Chatterjee" __email__ = "soumick.chatterjee@ovgu.de" __status__ = "Production" class CustomMotion(Motion): def __init__(self, noise_dir=2, **kargs): super(CustomMotion, self).__init__(**kargs) self.noise_dir = noise_dir def add_artifact( self, image: sitk.Image, transforms: List[sitk.Euler3DTransform], times: np.ndarray, interpolation: Interpolation, ): images = self.resample_images(image, transforms, interpolation) arrays = [sitk.GetArrayViewFromImage(im) for im in images] arrays = [array.transpose() for array in arrays] # ITK to NumPy spectra = [self.fourier_transform(array) for array in arrays] self.sort_spectra(spectra, times) result_spectrum = np.empty_like(spectra[0]) noise_dir = self.noise_dir if noise_dir == -1: noise_dir = random.randint(0,1) last_index = result_spectrum.shape[noise_dir] #it can be 0, 1 or 2 indices = (last_index * times).astype(int).tolist() indices.append(last_index) ini = 0 for spectrum, fin in zip(spectra, indices): if noise_dir == 0: result_spectrum[..., ini:fin,:,:] = spectrum[..., ini:fin,:,:] #depending upon last_index value, move ini:fin left or right [at the end :,: for 0, : for 1, none for 2] elif noise_dir == 1: result_spectrum[..., ini:fin,:] = spectrum[..., ini:fin,:] else: #original result_spectrum[..., ini:fin] = spectrum[..., ini:fin] ini = fin result_image = np.real(self.inv_fourier_transform(result_spectrum)) return result_image.astype(np.float32) class CustomRandomMotion(RandomMotion): def __init__(self, noise_dir=2, **kwargs): super(CustomRandomMotion, self).__init__(**kwargs) self.noise_dir = noise_dir def apply_transform(self, subject): arguments = defaultdict(dict) for name, image in self.get_images_dict(subject).items(): params = self.get_params( self.degrees_range, self.translation_range, self.num_transforms, is_2d=image.is_2d(), ) times_params, degrees_params, translation_params = params arguments['times'][name] = times_params arguments['degrees'][name] = degrees_params arguments['translation'][name] = translation_params arguments['image_interpolation'][name] = self.image_interpolation transform = CustomMotion(noise_dir=self.noise_dir,**self.add_include_exclude(arguments)) transformed = transform(subject) return transformed class RealityMotion(): def __init__(self, n_threads = 4, mu = 0, sigma = 0.1, random_sigma=True): self.n_threads = n_threads self.mu = mu self.sigma = sigma self.sigma_limit = sigma self.random_sigma = random_sigma def __perform_singlePE(self, idx): rot_x = np.random.normal(self.mu, self.sigma, 1)*random.randint(-1,1) rot_y = np.random.normal(self.mu, self.sigma, 1)*random.randint(-1,1) rot_z = np.random.normal(self.mu, self.sigma, 1)*random.randint(-1,1) tran_x = int(np.random.normal(self.mu, self.sigma, 1)*random.randint(-1,1)) tran_y = int(np.random.normal(self.mu, self.sigma, 1)*random.randint(-1,1)) tran_z = int(np.random.normal(self.mu, self.sigma, 1)*random.randint(-1,1)) temp_vol = self.__rot_tran_3d(self.in_vol, rot_x, rot_y, rot_z, tran_x, tran_y, tran_z) temp_k = np.fft.fftn(temp_vol) for slc in range(self.in_vol.shape[2]): self.out_k[idx,:,slc]=temp_k[idx,:,slc] def __call__(self, vol): if self.random_sigma: self.sigma = random.uniform(0, self.sigma_limit) shape = vol.shape device = vol.device self.in_vol = vol.squeeze().cpu().numpy() self.in_vol = self.in_vol/self.in_vol.max() self.out_k = np.zeros((self.in_vol.shape)) + 0j if self.n_threads > 0: pool = multiprocessing.Pool(self.n_threads) pool.map(self.__perform_singlePE, range(self.in_vol.shape[0])) else: for idx in range(self.in_vol.shape[0]): self.__perform_singlePE(idx) vol = np.abs(np.fft.ifftn(self.out_k)) vol = torch.from_numpy(vol).view(shape).to(device) del self.in_vol, self.out_k return vol def __x_rotmat(self, theta): cos_t = math.cos(theta) sin_t = math.sin(theta) return np.array([[1, 0, 0], [0, cos_t, -sin_t], [0, sin_t, cos_t]]) def __y_rotmat(self, theta): cos_t = math.cos(theta) sin_t = math.sin(theta) return np.array([[cos_t, 0, sin_t], [0, 1, 0], [-sin_t, 0, cos_t]]) def __z_rotmat(self, theta): cos_t = math.cos(theta) sin_t = math.sin(theta) return np.array([[cos_t, -sin_t, 0], [sin_t, cos_t, 0], [0, 0, 1]]) def __rot_tran_3d(self, J, rot_x, rot_y, rot_z, tran_x, tran_y, tran_z): M = self.__x_rotmat(rot_x) * self.__y_rotmat(rot_y) * self.__z_rotmat(rot_z) translation = ([tran_x, tran_y, tran_z]) K = affine_transform(J, M, translation, order=1) return K/(K.max()+1e-16) class MotionCorrupter(): def __init__(self, mode=0, degrees=10, translation=10, num_transforms=2, image_interpolation='linear', norm_mode=0, noise_dir=2, mu=0, sigma=0.1, random_sigma=False, n_threads=4): self.mode = mode #0: TorchIO's version, 1: Custom direction specific motion self.degrees = degrees self.translation = translation self.num_transforms = num_transforms self.image_interpolation = image_interpolation self.norm_mode = norm_mode #0: No Norm, 1: Divide by Max, 2: MinMax self.noise_dir = noise_dir #0, 1 or 2 - which direction the motion is generated, only for custom random self.mu = mu #Only for Reality Motion self.sigma = sigma #Only for Reality Motion self.random_sigma = random_sigma #Only for Reality Motion - to randomise the sigma value, treating the provided sigma as upper limit and 0 as lower self.n_threads = n_threads #Only for Reality Motion - to apply motion for each thread encoding line parallel, max thread controlled by this. Set to 0 to perform serially. if mode==0: #TorchIO's version self.corrupter = tio.transforms.RandomMotion(degrees=degrees, translation=translation, num_transforms=num_transforms, image_interpolation=image_interpolation) elif mode==1: #Custom Motion self.corrupter = CustomRandomMotion(degrees=degrees, translation=translation, num_transforms=num_transforms, image_interpolation=image_interpolation, noise_dir=noise_dir) elif mode==2: #Reality motion. self.corrupter = RealityMotion(n_threads=n_threads, mu=mu, sigma=sigma, random_sigma=random_sigma) def perform(self, vol): vol = vol.float() transformed = self.corrupter(vol) if self.norm_mode==1: vol = vol/vol.max() transformed = transformed/transformed.max() elif self.norm_mode==2: vol = (vol-vol.min())/(vol.max()-vol.min()) transformed = (transformed-transformed.min())/(transformed.max()-transformed.min()) return torch.cat([vol,transformed], 0)
8,325
44.497268
183
py
DDoS
DDoS-master/utils/padding.py
#parital source: https://github.com/c22n/unet-pytorch from typing import Tuple, Union import torch import torch.nn as nn import torch.nn.functional as F from torch.autograd import Function, Variable from torch.nn.modules.utils import _ntuple __author__ = "Soumick Chatterjee" __copyright__ = "Copyright 2022, Soumick Chatterjee & OvGU:ESF:MEMoRIAL" __credits__ = ["Soumick Chatterjee"] __license__ = "GPL" __version__ = "1.0.0" __email__ = "soumick.chatterjee@ovgu.de" __status__ = "Production" def flip(x: Variable, dim: int) -> Variable: """Flip torch Variable along given dimension axis.""" xsize = x.size() dim = x.dim() + dim if dim < 0 else dim x = x.contiguous().view(-1, *xsize[dim:]) x = x.view(x.size(0), x.size(1), -1)[:, getattr(torch.arange(x.size(1)-1, -1, -1), ('cpu', 'cuda')[x.is_cuda])().long(), :] return x.view(xsize) class ReflectionPad3d(nn.Module): """Wrapper for ReflectionPadNd function in 3 dimensions.""" def __init__(self, padding: Union[int, Tuple[int]]): super(ReflectionPad3d, self).__init__() self.padding = _ntuple(6)(padding) def forward(self, input: Variable) -> Variable: return ReflectionPadNd.apply(input, self.padding) def __repr__(self) -> str: return self.__class__.__name__ + '(' \ + str(self.padding) + ')' class ReflectionPadNd(Function): """Padding for same convolutional layer.""" # @staticmethod # def symbolic(g, input: Variable, padding: Union[int, Tuple[int]]): # paddings = prepare_onnx_paddings(len(input.type().sizes()), pad) # return g.op("Pad", input, pads_i=paddings, mode_s="reflect") @staticmethod def forward(ctx: Function, input: Variable, pad: Tuple[int]) -> Variable: ctx.pad = pad ctx.input_size = input.size() ctx.l_inp = len(input.size()) ctx.pad_tup = tuple([(a, b) for a, b in zip(pad[:-1:2], pad[1::2])] [::-1]) ctx.l_pad = len(ctx.pad_tup) ctx.l_diff = ctx.l_inp - ctx.l_pad assert ctx.l_inp >= ctx.l_pad new_dim = tuple([sum((d,) + ctx.pad_tup[i]) for i, d in enumerate(input.size()[-ctx.l_pad:])]) assert all([d > 0 for d in new_dim]), 'input is too small' # Create output tensor by concatenating with reflected chunks. output = input.new(input.size()[:(ctx.l_diff)] + new_dim).zero_() c_input = input for i, p in zip(range(ctx.l_inp)[-ctx.l_pad:], ctx.pad_tup): if p[0] > 0: chunk1 = flip(c_input.narrow(i, 0, pad[0]), i) c_input = torch.cat((chunk1, c_input), i) if p[1] > 0: chunk2 = flip(c_input.narrow(i, c_input.shape[i]-p[1], p[1]), i) c_input = torch.cat((c_input, chunk2), i) output.copy_(c_input) return output @staticmethod def backward(ctx: Function, grad_output: Variable) -> Variable: grad_input = Variable(grad_output.data.new(ctx.input_size).zero_()) grad_input_slices = [slice(0, x,) for x in ctx.input_size] cg_output = grad_output for i_s, p in zip(range(ctx.l_inp)[-ctx.l_pad:], ctx.pad_tup): if p[0] > 0: cg_output = cg_output.narrow(i_s, p[0], cg_output.size(i_s) - p[0]) if p[1] > 0: cg_output = cg_output.narrow(i_s, 0, cg_output.size(i_s) - p[1]) gis = tuple(grad_input_slices) grad_input[gis] = cg_output return grad_input, None, None
3,703
36.04
80
py
DDoS
DDoS-master/utils/pLoss/Resnet2D.py
#!/usr/bin/env python """ Original file Resnet2Dv2b14 of NCC1701 """ import torch import torch.nn as nn import torch.nn.functional as F #from utils.TorchAct.pelu import PELU_oneparam as PELU __author__ = "Soumick Chatterjee" __copyright__ = "Copyright 2018, Soumick Chatterjee & OvGU:ESF:MEMoRIAL" __credits__ = ["Soumick Chatterjee"] __license__ = "GPL" __version__ = "1.0.0" __email__ = "soumick.chatterjee@ovgu.de" __status__ = "Production" class ResidualBlock(nn.Module): def __init__(self, in_features, relu, norm): super(ResidualBlock, self).__init__() conv_block = [ nn.ReflectionPad2d(1), nn.Conv2d(in_features, in_features, 3), norm(in_features), relu(), nn.Dropout2d(p=0.2), nn.ReflectionPad2d(1), nn.Conv2d(in_features, in_features, 3), norm(in_features) ] self.conv_block = nn.Sequential(*conv_block) def forward(self, x): return x + self.conv_block(x) class ResNet(nn.Module): def __init__(self, in_channels=1, out_channels=1, res_blocks=14, starting_n_features=64, updown_blocks=2, is_relu_leaky=True, final_out_sigmoid=True, do_batchnorm=True): #should use 14 as that gives number of trainable parameters close to number of possible pixel values in a image 256x256 super(ResNet, self).__init__() if is_relu_leaky: relu = nn.PReLU else: relu = nn.ReLU if do_batchnorm: norm = nn.BatchNorm2d else: norm = nn.InstanceNorm2d # Initial convolution block model = [ nn.ReflectionPad2d(3), nn.Conv2d(in_channels, starting_n_features, 7), norm(starting_n_features), relu() ] # Downsampling in_features = starting_n_features out_features = in_features*2 for _ in range(updown_blocks): model += [ nn.Conv2d(in_features, out_features, 3, stride=2, padding=1), norm(out_features), relu() ] in_features = out_features out_features = in_features*2 # Residual blocks for _ in range(res_blocks): model += [ResidualBlock(in_features, relu, norm)] # Upsampling out_features = in_features//2 for _ in range(updown_blocks): model += [ nn.ConvTranspose2d(in_features, out_features, 3, stride=2, padding=1, output_padding=1), norm(out_features), relu() ] in_features = out_features out_features = in_features//2 # Output layer model += [ nn.ReflectionPad2d(3), nn.Conv2d(starting_n_features, out_channels, 7)] #final activation if final_out_sigmoid: model += [ nn.Sigmoid(), ] else: model += [ relu(), ] self.model = nn.Sequential(*model) def forward(self, input): return self.model(input)
3,150
31.484536
294
py
DDoS
DDoS-master/utils/pLoss/VesselSeg_UNet3d_DeepSup.py
# -*- coding: utf-8 -*- """ """ # from __future__ import print_function, division import torch import torch.nn as nn import torch.utils.data #from Utils.wta import KWinnersTakeAll __author__ = "Kartik Prabhu, Mahantesh Pattadkal, and Soumick Chatterjee" __copyright__ = "Copyright 2022, Faculty of Computer Science, Otto von Guericke University Magdeburg, Germany" __credits__ = ["Kartik Prabhu", "Mahantesh Pattadkal", "Soumick Chatterjee"] __license__ = "GPL" __version__ = "1.0.0" __maintainer__ = "Soumick Chatterjee" __email__ = "soumick.chatterjee@ovgu.de" __status__ = "Production" class conv_block(nn.Module): """ Convolution Block """ def __init__(self, in_channels, out_channels, k_size=3, stride=1, padding=1, bias=True): super(conv_block, self).__init__() self.conv = nn.Sequential( nn.Conv3d(in_channels=in_channels, out_channels=out_channels, kernel_size=k_size, stride=stride, padding=padding, bias=bias), nn.BatchNorm3d(num_features=out_channels), nn.ReLU(inplace=True), nn.Conv3d(in_channels=out_channels, out_channels=out_channels, kernel_size=k_size, stride=stride, padding=padding, bias=bias), nn.BatchNorm3d(num_features=out_channels), nn.ReLU(inplace=True) ) def forward(self, x): x = self.conv(x) return x class conv_block_v2(nn.Module): """ Convolution Block With WTA """ def __init__(self, in_channels, out_channels, k_size=3, stride=1, padding=1, bias=True): super(conv_block_v2, self).__init__() self.conv = nn.Sequential( nn.Conv3d(in_channels=in_channels, out_channels=out_channels, kernel_size=k_size, stride=stride, padding=padding, bias=bias), nn.BatchNorm3d(num_features=out_channels), nn.ReLU(inplace=True), nn.Conv3d(in_channels=out_channels, out_channels=out_channels, kernel_size=k_size, stride=stride, padding=padding, bias=bias), nn.BatchNorm3d(num_features=out_channels), nn.ReLU(inplace=True), #KWinnersTakeAll(0.02) ) def forward(self, x): x = self.conv(x) return x class up_conv(nn.Module): """ Up Convolution Block """ # def __init__(self, in_ch, out_ch): def __init__(self, in_channels, out_channels, k_size=3, stride=1, padding=1, bias=True): super(up_conv, self).__init__() self.up = nn.Sequential( nn.Upsample(scale_factor=2), nn.Conv3d(in_channels=in_channels, out_channels=out_channels, kernel_size=k_size, stride=stride, padding=padding, bias=bias), nn.BatchNorm3d(num_features=out_channels), nn.ReLU(inplace=True)) def forward(self, x): x = self.up(x) return x class U_Net_DeepSup(nn.Module): """ UNet - Basic Implementation Input _ [batch * channel(# of channels of each image) * depth(# of frames) * height * width]. Paper : https://arxiv.org/abs/1505.04597 """ def __init__(self, in_ch=1, out_ch=1): super(U_Net_DeepSup, self).__init__() n1 = 64 #TODO: original paper starts with 64 filters = [n1, n1 * 2, n1 * 4, n1 * 8, n1 * 16] # 64,128,256,512,1024 self.Maxpool1 = nn.MaxPool3d(kernel_size=2, stride=2) self.Maxpool2 = nn.MaxPool3d(kernel_size=2, stride=2) self.Maxpool3 = nn.MaxPool3d(kernel_size=2, stride=2) self.Maxpool4 = nn.MaxPool3d(kernel_size=2, stride=2) self.Conv1 = conv_block(in_ch, filters[0]) self.Conv2 = conv_block(filters[0], filters[1]) self.Conv3 = conv_block(filters[1], filters[2]) self.Conv4 = conv_block(filters[2], filters[3]) self.Conv5 = conv_block(filters[3], filters[4]) #1x1x1 Convolution for Deep Supervision self.Conv_d3 = conv_block(filters[1], 1) self.Conv_d4 = conv_block(filters[2], 1) self.Up5 = up_conv(filters[4], filters[3]) self.Up_conv5 = conv_block(filters[4], filters[3]) self.Up4 = up_conv(filters[3], filters[2]) self.Up_conv4 = conv_block(filters[3], filters[2]) self.Up3 = up_conv(filters[2], filters[1]) self.Up_conv3 = conv_block(filters[2], filters[1]) self.Up2 = up_conv(filters[1], filters[0]) self.Up_conv2 = conv_block(filters[1], filters[0]) self.Conv = nn.Conv3d(filters[0], out_ch, kernel_size=1, stride=1, padding=0) # self.active = torch.nn.Sigmoid() def forward(self, x): # print("unet") # print(x.shape) # print(padded.shape) e1 = self.Conv1(x) # print("conv1:") # print(e1.shape) e2 = self.Maxpool1(e1) e2 = self.Conv2(e2) # print("conv2:") # print(e2.shape) e3 = self.Maxpool2(e2) e3 = self.Conv3(e3) # print("conv3:") # print(e3.shape) e4 = self.Maxpool3(e3) e4 = self.Conv4(e4) # print("conv4:") # print(e4.shape) e5 = self.Maxpool4(e4) e5 = self.Conv5(e5) # print("conv5:") # print(e5.shape) d5 = self.Up5(e5) # print("d5:") # print(d5.shape) # print("e4:") # print(e4.shape) d5 = torch.cat((e4, d5), dim=1) d5 = self.Up_conv5(d5) # print("upconv5:") # print(d5.size) d4 = self.Up4(d5) # print("d4:") # print(d4.shape) d4 = torch.cat((e3, d4), dim=1) d4 = self.Up_conv4(d4) d4_out = self.Conv_d4(d4) # print("upconv4:") # print(d4.shape) d3 = self.Up3(d4) d3 = torch.cat((e2, d3), dim=1) d3 = self.Up_conv3(d3) d3_out = self.Conv_d3(d3) # print("upconv3:") # print(d3.shape) d2 = self.Up2(d3) d2 = torch.cat((e1, d2), dim=1) d2 = self.Up_conv2(d2) # print("upconv2:") # print(d2.shape) out = self.Conv(d2) # print("out:") # print(out.shape) # d1 = self.active(out) return out, d3_out , d4_out class U_Net_DeepSup_level4(nn.Module): """ UNet - Basic Implementation Input _ [batch * channel(# of channels of each image) * depth(# of frames) * height * width]. Paper : https://arxiv.org/abs/1505.04597 """ def __init__(self, in_ch=1, out_ch=1): super(U_Net_DeepSup_level4, self).__init__() n1 = 64 #TODO: original paper starts with 64 filters = [n1, n1 * 2, n1 * 4, n1 * 8, n1 * 16] # 64,128,256,512,1024 self.Maxpool1 = nn.MaxPool3d(kernel_size=2, stride=2) self.Maxpool2 = nn.MaxPool3d(kernel_size=2, stride=2) self.Maxpool3 = nn.MaxPool3d(kernel_size=2, stride=2) self.Maxpool4 = nn.MaxPool3d(kernel_size=2, stride=2) self.Conv1 = conv_block(in_ch, filters[0]) self.Conv2 = conv_block(filters[0], filters[1]) self.Conv3 = conv_block(filters[1], filters[2]) self.Conv4 = conv_block(filters[2], filters[3]) self.Conv5 = conv_block(filters[3], filters[4]) #1x1x1 Convolution for Deep Supervision self.Conv_d3 = conv_block(filters[1], 1) self.Conv_d4 = conv_block(filters[2], 1) self.Conv_d5 = conv_block(filters[3], 1) self.Up5 = up_conv(filters[4], filters[3]) self.Up_conv5 = conv_block(filters[4], filters[3]) self.Up4 = up_conv(filters[3], filters[2]) self.Up_conv4 = conv_block(filters[3], filters[2]) self.Up3 = up_conv(filters[2], filters[1]) self.Up_conv3 = conv_block(filters[2], filters[1]) self.Up2 = up_conv(filters[1], filters[0]) self.Up_conv2 = conv_block(filters[1], filters[0]) self.Conv = nn.Conv3d(filters[0], out_ch, kernel_size=1, stride=1, padding=0) # self.active = torch.nn.Sigmoid() def forward(self, x): # print("unet") # print(x.shape) # print(padded.shape) e1 = self.Conv1(x) # print("conv1:") # print(e1.shape) e2 = self.Maxpool1(e1) e2 = self.Conv2(e2) # print("conv2:") # print(e2.shape) e3 = self.Maxpool2(e2) e3 = self.Conv3(e3) # print("conv3:") # print(e3.shape) e4 = self.Maxpool3(e3) e4 = self.Conv4(e4) # print("conv4:") # print(e4.shape) e5 = self.Maxpool4(e4) e5 = self.Conv5(e5) # print("conv5:") # print(e5.shape) d5 = self.Up5(e5) # print("d5:") # print(d5.shape) # print("e4:") # print(e4.shape) d5 = torch.cat((e4, d5), dim=1) d5 = self.Up_conv5(d5) # print("upconv5:") # print(d5.size) d5_out = self.Conv_d5(d5) d4 = self.Up4(d5) # print("d4:") # print(d4.shape) d4 = torch.cat((e3, d4), dim=1) d4 = self.Up_conv4(d4) d4_out = self.Conv_d4(d4) # print("upconv4:") # print(d4.shape) d3 = self.Up3(d4) d3 = torch.cat((e2, d3), dim=1) d3 = self.Up_conv3(d3) d3_out = self.Conv_d3(d3) # print("upconv3:") # print(d3.shape) d2 = self.Up2(d3) d2 = torch.cat((e1, d2), dim=1) d2 = self.Up_conv2(d2) # print("upconv2:") # print(d2.shape) out = self.Conv(d2) # print("out:") # print(out.shape) # d1 = self.active(out) return [out, d3_out , d4_out , d5_out] class U_Net_DeepSup_level4_wta(nn.Module): """ UNet - Basic Implementation Input _ [batch * channel(# of channels of each image) * depth(# of frames) * height * width]. Paper : https://arxiv.org/abs/1505.04597 """ def __init__(self, in_ch=1, out_ch=1): super(U_Net_DeepSup_level4_wta, self).__init__() n1 = 64 # TODO: original paper starts with 64 filters = [n1, n1 * 2, n1 * 4, n1 * 8, n1 * 16] # 64,128,256,512,1024 self.Maxpool1 = nn.MaxPool3d(kernel_size=2, stride=2) self.Maxpool2 = nn.MaxPool3d(kernel_size=2, stride=2) self.Maxpool3 = nn.MaxPool3d(kernel_size=2, stride=2) self.Maxpool4 = nn.MaxPool3d(kernel_size=2, stride=2) self.Conv1 = conv_block_v2(in_ch, filters[0]) self.Conv2 = conv_block_v2(filters[0], filters[1]) self.Conv3 = conv_block_v2(filters[1], filters[2]) self.Conv4 = conv_block_v2(filters[2], filters[3]) self.Conv5 = conv_block_v2(filters[3], filters[4]) # 1x1x1 Convolution for Deep Supervision self.Conv_d3 = conv_block_v2(filters[1], 1) self.Conv_d4 = conv_block_v2(filters[2], 1) self.Conv_d5 = conv_block_v2(filters[3], 1) self.Up5 = up_conv(filters[4], filters[3]) self.Up_conv5 = conv_block(filters[4], filters[3]) self.Up4 = up_conv(filters[3], filters[2]) self.Up_conv4 = conv_block(filters[3], filters[2]) self.Up3 = up_conv(filters[2], filters[1]) self.Up_conv3 = conv_block(filters[2], filters[1]) self.Up2 = up_conv(filters[1], filters[0]) self.Up_conv2 = conv_block(filters[1], filters[0]) self.Conv = nn.Conv3d(filters[0], out_ch, kernel_size=1, stride=1, padding=0) # self.active = torch.nn.Sigmoid() def forward(self, x): # print("unet") # print(x.shape) # print(padded.shape) e1 = self.Conv1(x) # print("conv1:") # print(e1.shape) e2 = self.Maxpool1(e1) e2 = self.Conv2(e2) # print("conv2:") # print(e2.shape) e3 = self.Maxpool2(e2) e3 = self.Conv3(e3) # print("conv3:") # print(e3.shape) e4 = self.Maxpool3(e3) e4 = self.Conv4(e4) # print("conv4:") # print(e4.shape) e5 = self.Maxpool4(e4) e5 = self.Conv5(e5) # print("conv5:") # print(e5.shape) d5 = self.Up5(e5) # print("d5:") # print(d5.shape) # print("e4:") # print(e4.shape) d5 = torch.cat((e4, d5), dim=1) d5 = self.Up_conv5(d5) # print("upconv5:") # print(d5.size) d5_out = self.Conv_d5(d5) d4 = self.Up4(d5) # print("d4:") # print(d4.shape) d4 = torch.cat((e3, d4), dim=1) d4 = self.Up_conv4(d4) d4_out = self.Conv_d4(d4) # print("upconv4:") # print(d4.shape) d3 = self.Up3(d4) d3 = torch.cat((e2, d3), dim=1) d3 = self.Up_conv3(d3) d3_out = self.Conv_d3(d3) # print("upconv3:") # print(d3.shape) d2 = self.Up2(d3) d2 = torch.cat((e1, d2), dim=1) d2 = self.Up_conv2(d2) # print("upconv2:") # print(d2.shape) out = self.Conv(d2) # print("out:") # print(out.shape) # d1 = self.active(out) return [out, d3_out, d4_out, d5_out]
13,151
29.09611
110
py
DDoS
DDoS-master/utils/pLoss/perceptual_loss.py
import math import torch import torch.nn as nn import torchvision # from utils.utils import * # from pytorch_msssim import SSIM from .Resnet2D import ResNet from .simpleunet import UNet from .VesselSeg_UNet3d_DeepSup import U_Net_DeepSup __author__ = "Soumick Chatterjee" __copyright__ = "Copyright 2022, Faculty of Computer Science, Otto von Guericke University Magdeburg, Germany" __credits__ = ["Soumick Chatterjee"] __license__ = "GPL" __version__ = "1.0.0" __maintainer__ = "Soumick Chatterjee" __email__ = "soumick.chatterjee@ovgu.de" __status__ = "Production" class PerceptualLoss(torch.nn.Module): #currently configured for 1 channel only, with datarange as 1 for SSIM def __init__(self, device="cuda:0", loss_model="densenet161", n_level=math.inf, resize=None, loss_type="L1", mean=[], std=[]): super(PerceptualLoss, self).__init__() blocks = [] if loss_model == "resnet2D": #TODO: not finished model = ResNet(in_channels=1, out_channels=1).to(device) chk = torch.load(r"./utils/pLoss/ResNet14_IXIT2_Base_d1p75_t0_n10_dir01_5depth_L1Loss_best.pth.tar", map_location=device) model.load_state_dict(chk['state_dict']) elif loss_model == "unet2D": model = UNet(in_channels=1, out_channels=1, depth=5, wf=6, padding=True, batch_norm=False, up_mode='upsample', droprate=0.0, is3D=False, returnBlocks=False, downPath=True, upPath=True).to(device) chk = torch.load(r"./utils/pLoss/SimpleU_IXIT2_Base_d1p75_t0_n10_dir01_5depth_L1Loss_best.pth.tar", map_location=device) model.load_state_dict(chk['state_dict']) blocks.append(model.down_path[0].block.eval()) if n_level >= 2: blocks.append( nn.Sequential( nn.AvgPool2d(2), model.down_path[1].block.eval() ) ) if n_level >= 3: blocks.append( nn.Sequential( nn.AvgPool2d(2), model.down_path[2].block.eval() ) ) if n_level >= 4: blocks.append( nn.Sequential( nn.AvgPool2d(2), model.down_path[3].block.eval() ) ) elif loss_model == "unet3Dds": model = U_Net_DeepSup().to(device) chk = torch.load(r"./utils/pLoss/VesselSeg_UNet3d_DeepSup.pth", map_location=device) model.load_state_dict(chk['state_dict']) blocks.append(model.Conv1.conv.eval()) if n_level >= 2: blocks.append( nn.Sequential( model.Maxpool1.eval(), model.Conv2.conv.eval() ) ) if n_level >= 3: blocks.append( nn.Sequential( model.Maxpool2.eval(), model.Conv3.conv.eval() ) ) if n_level >= 4: blocks.append( nn.Sequential( model.Maxpool3.eval(), model.Conv4.conv.eval() ) ) if n_level >= 5: blocks.append( nn.Sequential( model.Maxpool4.eval(), model.Conv5.conv.eval() ) ) elif loss_model == "resnext1012D": model = torchvision.models.resnext101_32x8d() model.conv1 = nn.Conv2d(1, model.conv1.out_channels, kernel_size=model.conv1.kernel_size, stride=model.conv1.stride, padding=model.conv1.padding, bias=False if model.conv1.bias is None else True) model.fc = nn.Linear(in_features=model.fc.in_features, out_features=33, bias=False if model.fc.bias is None else True) model.to(device) # chk = torch.load(r"./utils/pLoss/ResNet14_IXIT2_Base_d1p75_t0_n10_dir01_5depth_L1Loss_best.pth.tar", map_location=device) # model.load_state_dict(chk['state_dict']) blocks.append( nn.Sequential( model.conv1.eval(), model.bn1.eval(), model.relu.eval(), ) ) if n_level >= 2: blocks.append( nn.Sequential( model.maxpool.eval(), model.layer1.eval() ) ) if n_level >= 3: blocks.append(model.layer2.eval()) if n_level >= 4: blocks.append(model.layer3.eval()) if n_level >= 5: blocks.append(model.layer4.eval()) elif loss_model == "densenet161": model = torchvision.models.densenet161() model.features.conv0 = nn.Conv2d(1, model.features.conv0.out_channels, kernel_size=model.features.conv0.kernel_size, stride=model.features.conv0.stride, padding=model.features.conv0.padding, bias=False if model.features.conv0.bias is None else True) model.classifier = nn.Linear(in_features=model.classifier.in_features, out_features=33, bias=False if model.classifier.bias is None else True) model.to(device) # chk = torch.load(r"./utils/pLoss/ResNet14_IXIT2_Base_d1p75_t0_n10_dir01_5depth_L1Loss_best.pth.tar", map_location=device) # model.load_state_dict(chk['state_dict']) model = model.features blocks.append( nn.Sequential( model.conv0.eval(), model.norm0.eval(), model.relu0.eval(), ) ) if n_level >= 2: blocks.append( nn.Sequential( model.pool0.eval(), model.denseblock1.eval() ) ) if n_level >= 3: blocks.append(model.denseblock2.eval()) if n_level >= 4: blocks.append(model.denseblock3.eval()) if n_level >= 5: blocks.append(model.denseblock4.eval()) for bl in blocks: for params in bl.parameters(): params.requires_grad = False self.blocks = nn.ModuleList(blocks) self.transform = nn.functional.interpolate if (mean is not None and len(mean) > 1) and (std is not None and len(std) > 1) and (len(mean) == len(std)): self.mean = nn.Parameter(torch.tensor(mean).view(1,len(mean),1,1)) self.std = nn.Parameter(torch.tensor(std).view(1,len(std),1,1)) else: self.mean = None self.std = None self.resize = resize if loss_type == "L1": self.loss_func = torch.nn.functional.l1_loss elif loss_type == "MultiSSIM": self.loss_func = MultiSSIM(reduction='mean').to(device) elif loss_type == "SSIM3D": self.loss_func = SSIM(data_range=1, size_average=True, channel=1, spatial_dims=3).to(device) elif loss_type == "SSIM2D": self.loss_func = SSIM(data_range=1, size_average=True, channel=1, spatial_dims=2).to(device) def forward(self, input, target): if self.mean is not None: input = (input-self.mean) / self.std target = (target-self.mean) / self.std if self.resize: input = self.transform(input, mode='trilinear' if len(input.shape) == 5 else 'bilinear', size=self.resize, align_corners=False) target = self.transform(target, mode='trilinear' if len(input.shape) == 5 else 'bilinear', size=self.resize, align_corners=False) loss = 0.0 x = input y = target for block in self.blocks: x = block(x) y = block(y) loss += self.loss_func(x, y) return loss if __name__ == '__main__': x = PerceptualLoss(resize=None).cuda() a = torch.rand(2,1,24,24).cuda() b = torch.rand(2,1,24,24).cuda() l = x(a,b) sdsd
8,538
42.345178
154
py
DDoS
DDoS-master/utils/pLoss/simpleunet.py
import torch import torch.nn.functional as F from torch import nn __author__ = "Soumick Chatterjee" __copyright__ = "Copyright 2022, Faculty of Computer Science, Otto von Guericke University Magdeburg, Germany" __credits__ = ["Soumick Chatterjee", "Chompunuch Sarasaen"] __license__ = "GPL" __version__ = "1.0.0" __maintainer__ = "Soumick Chatterjee" __email__ = "soumick.chatterjee@ovgu.de" __status__ = "Production" class UNetConvBlock(nn.Module): def __init__(self, in_size, out_size, padding, batch_norm): super(UNetConvBlock, self).__init__() block = [] block.append(layer_conv(in_size, out_size, kernel_size=3, padding=int(padding))) block.append(nn.ReLU()) if batch_norm: block.append(layer_batchnorm(out_size)) block.append(layer_conv(out_size, out_size, kernel_size=3, padding=int(padding))) block.append(nn.ReLU()) if batch_norm: block.append(layer_batchnorm(out_size)) self.block = nn.Sequential(*block) def forward(self, x): out = self.block(x) return out class UNetUpBlock(nn.Module): def __init__(self, in_size, out_size, up_mode, padding, batch_norm): super(UNetUpBlock, self).__init__() if up_mode == 'upconv': self.up = layer_convtrans(in_size, out_size, kernel_size=2, stride=2) elif up_mode == 'upsample': self.up = nn.Sequential(nn.Upsample(mode=interp_mode, scale_factor=2), layer_conv(in_size, out_size, kernel_size=1)) self.conv_block = UNetConvBlock(in_size, out_size, padding, batch_norm) def center_crop(self, layer, target_size): _, _, layer_depth, layer_height, layer_width = layer.size() diff_z = (layer_depth - target_size[0]) // 2 diff_y = (layer_height - target_size[1]) // 2 diff_x = (layer_width - target_size[2]) // 2 return layer[:, :, diff_z:(diff_z + target_size[0]), diff_y:(diff_y + target_size[1]), diff_x:(diff_x + target_size[2])] # _, _, layer_height, layer_width = layer.size() #for 2D data # diff_y = (layer_height - target_size[0]) // 2 # diff_x = (layer_width - target_size[1]) // 2 # return layer[:, :, diff_y:(diff_y + target_size[0]), diff_x:(diff_x + target_size[1])] def forward(self, x, bridge): up = self.up(x) # bridge = self.center_crop(bridge, up.shape[2:]) #sending shape ignoring 2 digit, so target size start with 0,1,2 up = F.interpolate(up, size=bridge.shape[2:], mode=interp_mode) out = torch.cat([up, bridge], 1) out = self.conv_block(out) return out class UNet(nn.Module): """ Implementation of U-Net: Convolutional Networks for Biomedical Image Segmentation (Ronneberger et al., 2015) https://arxiv.org/abs/1505.04597 Using the default arguments will yield the exact version used in the original paper Args: in_channels (int): number of input channels out_channels (int): number of output channels depth (int): depth of the network wf (int): number of filters in the first layer is 2**wf padding (bool): if True, apply padding such that the input shape is the same as the output. This may introduce artifacts batch_norm (bool): Use BatchNorm after layers with an activation function up_mode (str): one of 'upconv' or 'upsample'. 'upconv' will use transposed convolutions for learned upsampling. 'upsample' will use bilinear upsampling. droprate (float): Rate of dropout. If undesired, then 0.0 is3D (bool): If a 3D or 2D version of U-net returnBlocks (bool) : If True, it will return the blocks created during downPath. If downPath is False, then it will be ignored downPath and upPath (bool): If only the downpath or uppath of the U-Net is needed, make the other one False Forward call: x (Tensor): Input Tensor blocks (list of Tensors): If only upPath is set to True, then this will be used during the forward of the uppath. If not desired, then supply blank list """ def __init__(self, in_channels=1, out_channels=1, depth=3, wf=6, padding=True, batch_norm=False, up_mode='upconv', droprate=0.0, is3D=False, returnBlocks=False, downPath=True, upPath=True): super(UNet, self).__init__() layers = {} if is3D: layers["layer_conv"] = nn.Conv3d layers["layer_convtrans"] = nn.ConvTranspose3d layers["layer_batchnorm"] = nn.BatchNorm3d layers["layer_drop"] = nn.Dropout3d layers["func_avgpool"] = F.avg_pool3d layers["interp_mode"] = 'trilinear' else: layers["layer_conv"] = nn.Conv2d layers["layer_convtrans"] = nn.ConvTranspose2d layers["layer_batchnorm"] = nn.BatchNorm2d layers["layer_drop"] = nn.Dropout2d layers["func_avgpool"] = F.avg_pool2d layers["interp_mode"] = 'bilinear' globals().update(layers) self.returnBlocks = returnBlocks self.do_down = downPath self.do_up = upPath self.padding = padding self.depth = depth self.dropout = layer_drop(p=droprate) prev_channels = in_channels self.down_path = nn.ModuleList() for i in range(depth): if self.do_down: self.down_path.append(UNetConvBlock(prev_channels, 2**(wf+i), padding, batch_norm)) prev_channels = 2**(wf+i) self.latent_channels = prev_channels self.up_path = nn.ModuleList() for i in reversed(range(depth - 1)): if self.do_up: self.up_path.append(UNetUpBlock(prev_channels, 2**(wf+i), up_mode, padding, batch_norm)) prev_channels = 2**(wf+i) if self.do_up: self.last = layer_conv(prev_channels, out_channels, kernel_size=1) def forward(self, x, blocks=()): if self.do_down: for i, down in enumerate(self.down_path): x = down(x) if i != len(self.down_path)-1: blocks += (x,) x = func_avgpool(x, 2) x = self.dropout(x) if self.do_up: for i, up in enumerate(self.up_path): x = up(x, blocks[-i-1]) x = self.last(x) if self.returnBlocks and self.do_down: return x, blocks else: return x if __name__ == '__main__': print('#### Test Case ###') from torch.autograd import Variable x = Variable(torch.rand(2,1,64,64)).cuda() model = UNet().cuda() y = model(x) print(y.shape)
7,121
39.237288
160
py
HIBPool
HIBPool-main/GIB.py
#!/usr/bin/env python # coding: utf-8 # In[ ]: from __future__ import print_function import numpy as np import pprint as pp from copy import deepcopy import pickle from numbers import Number from collections import OrderedDict import itertools import torch import torch.nn as nn from torch.autograd import Variable from torch.nn import functional as F import torch.optim as optim from torch.optim.lr_scheduler import ReduceLROnPlateau, LambdaLR from torch.distributions import constraints from torch.distributions.normal import Normal from torch.distributions.multivariate_normal import MultivariateNormal from torch.distributions.distribution import Distribution from torch.distributions.utils import broadcast_all import sys, os sys.path.append(os.path.join(os.path.dirname(__file__), '..')) from pytorch_net.modules import get_Layer, load_layer_dict, Simple_2_Symbolic from pytorch_net.util import forward, Loss_Fun, get_activation, get_criterion, get_criteria_value, get_optimizer, get_full_struct_param, plot_matrices, get_model_DL, PrecisionFloorLoss, get_list_DL, init_weight from pytorch_net.util import Early_Stopping, Performance_Monitor, record_data, to_np_array, to_Variable, make_dir, formalize_value, RampupLR, Transform_Label, view_item, load_model, save_model, to_cpu_recur, filter_kwargs # ## Training functionality: # In[ ]: def train( model, X = None, y = None, train_loader = None, validation_data = None, validation_loader = None, criterion = nn.MSELoss(), inspect_interval = 10, isplot = False, is_cuda = None, **kwargs ): """Training function for generic models. "model" can be a single model or a ordered list of models""" def get_regularization(model, loss_epoch, **kwargs): """Compute regularization.""" reg_dict = kwargs["reg_dict"] if "reg_dict" in kwargs else None reg = to_Variable([0], is_cuda = is_cuda) if reg_dict is not None: for reg_type, reg_coeff in reg_dict.items(): # Setting up regularization strength: if isinstance(reg_coeff, Number): reg_coeff_ele = reg_coeff else: if loss_epoch < len(reg_coeff): reg_coeff_ele = reg_coeff[loss_epoch] else: reg_coeff_ele = reg_coeff[-1] # Accumulate regularization: reg = reg + model.get_regularization(source=[reg_type], mode=reg_mode, **kwargs) * reg_coeff_ele return reg if is_cuda is None: if X is None and y is None: assert train_loader is not None is_cuda = train_loader.dataset.tensors[0].is_cuda else: is_cuda = X.is_cuda # Optimization kwargs: epochs = kwargs["epochs"] if "epochs" in kwargs else 10000 lr = kwargs["lr"] if "lr" in kwargs else 5e-3 lr_rampup_steps = kwargs["lr_rampup"] if "lr_rampup" in kwargs else 200 optim_type = kwargs["optim_type"] if "optim_type" in kwargs else "adam" optim_kwargs = kwargs["optim_kwargs"] if "optim_kwargs" in kwargs else {} scheduler_type = kwargs["scheduler_type"] if "scheduler_type" in kwargs else "ReduceLROnPlateau" gradient_noise = kwargs["gradient_noise"] if "gradient_noise" in kwargs else None data_loader_apply = kwargs["data_loader_apply"] if "data_loader_apply" in kwargs else None # Inspection kwargs: inspect_step = kwargs["inspect_step"] if "inspect_step" in kwargs else None # Whether to inspect each step inspect_items = kwargs["inspect_items"] if "inspect_items" in kwargs else None inspect_items_train = get_inspect_items_train(inspect_items) inspect_functions = kwargs["inspect_functions"] if "inspect_functions" in kwargs else None if inspect_functions is not None: for inspect_function_key in inspect_functions: if inspect_function_key not in inspect_items: inspect_items.append(inspect_function_key) inspect_items_interval = kwargs["inspect_items_interval"] if "inspect_items_interval" in kwargs else 1000 inspect_image_interval = kwargs["inspect_image_interval"] if "inspect_image_interval" in kwargs else None inspect_loss_precision = kwargs["inspect_loss_precision"] if "inspect_loss_precision" in kwargs else 4 callback = kwargs["callback"] if "callback" in kwargs else None # Saving kwargs: record_keys = kwargs["record_keys"] if "record_keys" in kwargs else ["loss"] filename = kwargs["filename"] if "filename" in kwargs else None if filename is not None: make_dir(filename) save_interval = kwargs["save_interval"] if "save_interval" in kwargs else None save_step = kwargs["save_step"] if "save_step" in kwargs else None logdir = kwargs["logdir"] if "logdir" in kwargs else None data_record = {key: [] for key in record_keys} info_to_save = kwargs["info_to_save"] if "info_to_save" in kwargs else None if info_to_save is not None: data_record.update(info_to_save) patience = kwargs["patience"] if "patience" in kwargs else 20 if patience is not None: early_stopping_epsilon = kwargs["early_stopping_epsilon"] if "early_stopping_epsilon" in kwargs else 0 early_stopping_monitor = kwargs["early_stopping_monitor"] if "early_stopping_monitor" in kwargs else "loss" early_stopping = Early_Stopping(patience = patience, epsilon = early_stopping_epsilon, mode = "max" if early_stopping_monitor in ["accuracy"] else "min") if logdir is not None: from pytorch_net.logger import Logger batch_idx = 0 logger = Logger(logdir) logimages = kwargs["logimages"] if "logimages" in kwargs else None reg_mode = kwargs["reg_mode"] if "reg_mode" in kwargs else "L1" if validation_loader is not None: assert validation_data is None X_valid, y_valid = None, None elif validation_data is not None: X_valid, y_valid = validation_data else: X_valid, y_valid = X, y # Setting up dynamic label noise: label_noise_matrix = kwargs["label_noise_matrix"] if "label_noise_matrix" in kwargs else None transform_label = Transform_Label(label_noise_matrix = label_noise_matrix, is_cuda=is_cuda) # Setting up cotrain optimizer: co_kwargs = kwargs["co_kwargs"] if "co_kwargs" in kwargs else None if co_kwargs is not None: co_optimizer = co_kwargs["co_optimizer"] co_model = co_kwargs["co_model"] co_criterion = co_kwargs["co_criterion"] if "co_criterion" in co_kwargs else None co_multi_step = co_kwargs["co_multi_step"] if "co_multi_step" in co_kwargs else 1 # Get original loss: if len(inspect_items_train) > 0: loss_value_train = get_loss(model, train_loader, X, y, criterion=criterion, loss_epoch=-1, transform_label=transform_label, **kwargs) info_dict_train = prepare_inspection(model, train_loader, X, y, transform_label=transform_label, **kwargs) if "loss" in record_keys: record_data(data_record, [loss_value_train], ["loss_tr"]) loss_original = get_loss(model, validation_loader, X_valid, y_valid, criterion=criterion, loss_epoch=-1, transform_label=transform_label, **kwargs) if "loss" in record_keys: record_data(data_record, [-1, loss_original], ["iter", "loss"]) if "reg" in record_keys and "reg_dict" in kwargs and len(kwargs["reg_dict"]) > 0: reg_value = get_regularization(model, loss_epoch=0, **kwargs) record_data(data_record, [reg_value], ["reg"]) if "param" in record_keys: record_data(data_record, [model.get_weights_bias(W_source="core", b_source="core")], ["param"]) if "param_grad" in record_keys: record_data(data_record, [model.get_weights_bias(W_source="core", b_source="core", is_grad=True)], ["param_grad"]) if co_kwargs is not None: co_loss_original = get_loss(co_model, validation_loader, X_valid, y_valid, criterion=criterion, loss_epoch=-1, transform_label=transform_label, **co_kwargs) if "co_loss" in record_keys: record_data(data_record, [co_loss_original], ["co_loss"]) if filename is not None and save_interval is not None: record_data(data_record, [{}], ["model_dict"]) # Setting up optimizer: parameters = model.parameters() num_params = len(list(model.parameters())) if num_params == 0: print("No parameters to optimize!") loss_value = get_loss(model, validation_loader, X_valid, y_valid, criterion = criterion, loss_epoch = -1, transform_label=transform_label, **kwargs) if "loss" in record_keys: record_data(data_record, [0, loss_value], ["iter", "loss"]) if "param" in record_keys: record_data(data_record, [model.get_weights_bias(W_source = "core", b_source = "core")], ["param"]) if "param_grad" in record_keys: record_data(data_record, [model.get_weights_bias(W_source = "core", b_source = "core", is_grad = True)], ["param_grad"]) if co_kwargs is not None: co_loss_value = get_loss(co_model, validation_loader, X_valid, y_valid, criterion = criterion, loss_epoch = -1, transform_label=transform_label, **co_kwargs) record_data(data_record, [co_loss_value], ["co_loss"]) return loss_original, loss_value, data_record optimizer = get_optimizer(optim_type, lr, parameters, **optim_kwargs) if "optimizer" not in kwargs or ("optimizer" in kwargs and kwargs["optimizer"] is None) else kwargs["optimizer"] # Initialize inspect_items: if inspect_items is not None: print("{}:".format(-1), end = "") print("\tlr: {0:.3e}\t loss:{1:.{2}f}".format(optimizer.param_groups[0]["lr"], loss_original, inspect_loss_precision), end = "") info_dict = prepare_inspection(model, validation_loader, X_valid, y_valid, transform_label=transform_label, **kwargs) if len(inspect_items_train) > 0: print("\tloss_tr: {0:.{1}f}".format(loss_value_train, inspect_loss_precision), end = "") info_dict_train = update_key_train(info_dict_train, inspect_items_train) info_dict.update(info_dict_train) if "reg" in record_keys and "reg_dict" in kwargs and len(kwargs["reg_dict"]) > 0: print("\treg:{0:.{1}f}".format(to_np_array(reg_value), inspect_loss_precision), end="") if len(info_dict) > 0: for item in inspect_items: if item in info_dict: print(" \t{0}: {1:.{2}f}".format(item, info_dict[item], inspect_loss_precision), end = "") if item in record_keys and item not in ["loss", "reg"]: record_data(data_record, [to_np_array(info_dict[item])], [item]) if co_kwargs is not None: co_info_dict = prepare_inspection(co_model, validation_loader, X_valid, y_valid, transform_label=transform_label, **co_kwargs) if "co_loss" in inspect_items: co_loss_value = get_loss(co_model, validation_loader, X_valid, y_valid, criterion=criterion, loss_epoch=-1, transform_label=transform_label, **co_kwargs) print("\tco_loss: {}".format(formalize_value(co_loss_value, inspect_loss_precision)), end="") if len(co_info_dict) > 0: for item in inspect_items: if item in co_info_dict: print(" \t{0}: {1}".format(item, formalize_value(co_info_dict[item], inspect_loss_precision)), end="") if item in record_keys and item != "loss": record_data(data_record, [to_np_array(co_info_dict[item])], [item]) print("\n") # Setting up gradient noise: if gradient_noise is not None: from pytorch_net.util import Gradient_Noise_Scale_Gen scale_gen = Gradient_Noise_Scale_Gen(epochs=epochs, gamma=gradient_noise["gamma"], # decay rate eta=gradient_noise["eta"], # starting variance gradient_noise_interval_epoch=1, ) gradient_noise_scale = scale_gen.generate_scale(verbose=True) # Set up learning rate scheduler: if scheduler_type is not None: if scheduler_type == "ReduceLROnPlateau": scheduler_patience = kwargs["scheduler_patience"] if "scheduler_patience" in kwargs else 40 scheduler_factor = kwargs["scheduler_factor"] if "scheduler_factor" in kwargs else 0.1 scheduler_verbose = kwargs["scheduler_verbose"] if "scheduler_verbose" in kwargs else False scheduler = ReduceLROnPlateau(optimizer, factor=scheduler_factor, patience=scheduler_patience, verbose=scheduler_verbose) elif scheduler_type == "LambdaLR": scheduler_lr_lambda = kwargs["scheduler_lr_lambda"] if "scheduler_lr_lambda" in kwargs else (lambda epoch: 0.97 ** (epoch // 2)) scheduler = LambdaLR(optimizer, lr_lambda=scheduler_lr_lambda) else: raise # Ramping or learning rate for the first lr_rampup_steps steps: if lr_rampup_steps is not None and train_loader is not None: scheduler_rampup = RampupLR(optimizer, num_steps=lr_rampup_steps) if hasattr(train_loader, "dataset"): data_size = len(train_loader.dataset) else: data_size = kwargs["data_size"] # Initialize logdir: if logdir is not None: if logimages is not None: for tag, image_fun in logimages["image_fun"].items(): image = image_fun(model, logimages["X"], logimages["y"]) logger.log_images(tag, image, -1) # Training: to_stop = False for i in range(epochs + 1): model.train() # Updating gradient noise: if gradient_noise is not None: hook_handle_list = [] if i % scale_gen.gradient_noise_interval_epoch == 0: for h in hook_handle_list: h.remove() hook_handle_list = [] scale_idx = int(i / scale_gen.gradient_noise_interval_epoch) if scale_idx >= len(gradient_noise_scale): current_gradient_noise_scale = gradient_noise_scale[-1] else: current_gradient_noise_scale = gradient_noise_scale[scale_idx] for param_group in optimizer.param_groups: for param in param_group["params"]: if param.requires_grad: h = param.register_hook(lambda grad: grad + Variable(torch.normal(mean=torch.zeros(grad.size()), std=current_gradient_noise_scale * torch.ones(grad.size())))) hook_handle_list.append(h) if X is not None and y is not None: if optim_type != "LBFGS": optimizer.zero_grad() reg = get_regularization(model, loss_epoch=i, **kwargs) loss = model.get_loss(X, transform_label(y), criterion=criterion, loss_epoch=i, **kwargs) + reg loss.backward() optimizer.step() else: # "LBFGS" is a second-order optimization algorithm that requires a slightly different procedure: def closure(): optimizer.zero_grad() reg = get_regularization(model, loss_epoch=i, **kwargs) loss = model.get_loss(X, transform_label(y), criterion=criterion, loss_epoch=i, **kwargs) + reg loss.backward() return loss optimizer.step(closure) # Cotrain step: if co_kwargs is not None: if "co_warmup_epochs" not in co_kwargs or "co_warmup_epochs" in co_kwargs and i >= co_kwargs["co_warmup_epochs"]: for _ in range(co_multi_step): co_optimizer.zero_grad() co_reg = get_regularization(co_model, loss_epoch=i, **co_kwargs) co_loss = co_model.get_loss(X, transform_label(y), criterion=co_criterion, loss_epoch=i, **co_kwargs) + co_reg co_loss.backward() co_optimizer.step() else: if inspect_step is not None: info_dict_step = {key: [] for key in inspect_items} if "loader_process" in kwargs and kwargs["loader_process"] is not None: train_loader = kwargs["loader_process"]("train") for k, data_batch in enumerate(train_loader): if isinstance(data_batch, tuple) or isinstance(data_batch, list): X_batch, y_batch = data_batch if data_loader_apply is not None: X_batch, y_batch = data_loader_apply(X_batch, y_batch) else: X_batch, y_batch = data_loader_apply(data_batch) if optim_type != "LBFGS": optimizer.zero_grad() reg = get_regularization(model, loss_epoch=i, **kwargs) loss = model.get_loss(X_batch, transform_label(y_batch), criterion=criterion, loss_epoch=i, loss_step=k, **kwargs) + reg loss.backward() if logdir is not None: batch_idx += 1 if len(model.info_dict) > 0: for item in inspect_items: if item in model.info_dict: logger.log_scalar(item, model.info_dict[item], batch_idx) optimizer.step() else: def closure(): optimizer.zero_grad() reg = get_regularization(model, loss_epoch=i, **kwargs) loss = model.get_loss(X_batch, transform_label(y_batch), criterion=criterion, loss_epoch=i, loss_step=k, **kwargs) + reg loss.backward() return loss if logdir is not None: batch_idx += 1 if len(model.info_dict) > 0: for item in inspect_items: if item in model.info_dict: logger.log_scalar(item, model.info_dict[item], batch_idx) optimizer.step(closure) # Rampup scheduler: if lr_rampup_steps is not None and i * data_size // len(X_batch) + k < lr_rampup_steps: scheduler_rampup.step() # Cotrain step: if co_kwargs is not None: if "co_warmup_epochs" not in co_kwargs or "co_warmup_epochs" in co_kwargs and i >= co_kwargs["co_warmup_epochs"]: for _ in range(co_multi_step): co_optimizer.zero_grad() co_reg = get_regularization(co_model, loss_epoch=i, **co_kwargs) co_loss = co_model.get_loss(X_batch, transform_label(y_batch), criterion=co_criterion, loss_epoch=i, loss_step=k, **co_kwargs) + co_reg co_loss.backward() if logdir is not None: if len(co_model.info_dict) > 0: for item in inspect_items: if item in co_model.info_dict: logger.log_scalar(item, co_model.info_dict[item], batch_idx) co_optimizer.step() # Inspect at each step: if inspect_step is not None: if k % inspect_step == 0: print("s{}:".format(k), end = "") info_dict = prepare_inspection(model, validation_loader, X_valid, y_valid, transform_label=transform_label, **kwargs) if "loss" in inspect_items: info_dict_step["loss"].append(loss.item()) print("\tloss: {0:.{1}f}".format(loss.item(), inspect_loss_precision), end="") if len(info_dict) > 0: for item in inspect_items: if item in info_dict: info_dict_step[item].append(info_dict[item]) print(" \t{0}: {1}".format(item, formalize_value(info_dict[item], inspect_loss_precision)), end = "") if co_kwargs is not None: if "co_warmup_epochs" not in co_kwargs or "co_warmup_epochs" in co_kwargs and i >= co_kwargs["co_warmup_epochs"]: co_info_dict = prepare_inspection(co_model, validation_loader, X_valid, y_valid, transform_label=transform_label, **co_kwargs) if "co_loss" in inspect_items: print("\tco_loss: {0:.{1}f}".format(co_loss.item(), inspect_loss_precision), end="") info_dict_step["co_loss"].append(co_loss.item()) if len(co_info_dict) > 0: for item in inspect_items: if item in co_info_dict and item != "co_loss": info_dict_step[item].append(co_info_dict[item]) print(" \t{0}: {1}".format(item, formalize_value(co_info_dict[item], inspect_loss_precision)), end="") print() if k % save_step == 0: if filename is not None: pickle.dump(model.model_dict, open(filename[:-2] + "_model.p", "wb")) if logdir is not None: # Log values and gradients of the parameters (histogram summary) # for tag, value in model.named_parameters(): # tag = tag.replace('.', '/') # logger.log_histogram(tag, to_np_array(value), i) # logger.log_histogram(tag + '/grad', to_np_array(value.grad), i) if logimages is not None: for tag, image_fun in logimages["image_fun"].items(): image = image_fun(model, logimages["X"], logimages["y"]) logger.log_images(tag, image, i) if i % inspect_interval == 0: model.eval() if inspect_items is not None and i % inspect_items_interval == 0 and len(inspect_items_train) > 0: loss_value_train = get_loss(model, train_loader, X, y, criterion = criterion, loss_epoch = i, transform_label=transform_label, **kwargs) info_dict_train = prepare_inspection(model, train_loader, X, y, transform_label=transform_label, **kwargs) loss_value = get_loss(model, validation_loader, X_valid, y_valid, criterion = criterion, loss_epoch = i, transform_label=transform_label, **kwargs) reg_value = get_regularization(model, loss_epoch = i, **kwargs) if scheduler_type is not None: if lr_rampup_steps is None or train_loader is None or (lr_rampup_steps is not None and i * data_size // len(X_batch) + k >= lr_rampup_steps): if scheduler_type == "ReduceLROnPlateau": scheduler.step(loss_value) else: scheduler.step() if callback is not None: assert callable(callback) callback(model = model, X = X_valid, y = y_valid, iteration = i, loss = loss_value, ) if patience is not None: if early_stopping_monitor == "loss": to_stop = early_stopping.monitor(loss_value) else: info_dict = prepare_inspection(model, validation_loader, X_valid, y_valid, transform_label=transform_label, **kwargs) to_stop = early_stopping.monitor(info_dict[early_stopping_monitor]) if inspect_items is not None: if i % inspect_items_interval == 0: # Get loss: print("{}:".format(i), end = "") print("\tlr: {0:.3e}\tloss: {1:.{2}f}".format(optimizer.param_groups[0]["lr"], loss_value, inspect_loss_precision), end = "") info_dict = prepare_inspection(model, validation_loader, X_valid, y_valid, transform_label=transform_label, **kwargs) if len(inspect_items_train) > 0: print("\tloss_tr: {0:.{1}f}".format(loss_value_train, inspect_loss_precision), end = "") info_dict_train = update_key_train(info_dict_train, inspect_items_train) info_dict.update(info_dict_train) if "reg" in inspect_items and "reg_dict" in kwargs and len(kwargs["reg_dict"]) > 0: print("\treg:{0:.{1}f}".format(to_np_array(reg_value), inspect_loss_precision), end="") # Print and record: if len(info_dict) > 0: for item in inspect_items: if item + "_val" in info_dict: print(" \t{0}: {1}".format(item, formalize_value(info_dict[item + "_val"], inspect_loss_precision)), end = "") if item in record_keys and item not in ["loss", "reg"]: record_data(data_record, [to_np_array(info_dict[item + "_val"])], [item]) # logger: if logdir is not None: for item in inspect_items: if item + "_val" in info_dict: logger.log_scalar(item + "_val", info_dict[item + "_val"], i) # Co_model: if co_kwargs is not None: co_loss_value = get_loss(co_model, validation_loader, X_valid, y_valid, criterion = criterion, loss_epoch = i, transform_label=transform_label, **co_kwargs) co_info_dict = prepare_inspection(co_model, validation_loader, X_valid, y_valid, transform_label=transform_label, **co_kwargs) if "co_loss" in inspect_items: print("\tco_loss: {0:.{1}f}".format(co_loss_value, inspect_loss_precision), end="") if len(co_info_dict) > 0: for item in inspect_items: if item + "_val" in co_info_dict: print(" \t{0}: {1}".format(item, formalize_value(co_info_dict[item + "_val"], inspect_loss_precision)), end="") if item in record_keys and item != "co_loss": record_data(data_record, [to_np_array(co_info_dict[item + "_val"])], [item]) if "co_loss" in record_keys: record_data(data_record, [co_loss_value], ["co_loss"]) # Training metrics: if inspect_step is not None: for item in info_dict_step: if len(info_dict_step[item]) > 0: print(" \t{0}_s: {1}".format(item, formalize_value(np.mean(info_dict_step[item]), inspect_loss_precision)), end = "") if item in record_keys and item != "loss": record_data(data_record, [np.mean(info_dict_step[item])], ["{}_s".format(item)]) # Record loss: if "loss" in record_keys: record_data(data_record, [i, loss_value], ["iter", "loss"]) if "reg" in record_keys and "reg_dict" in kwargs and len(kwargs["reg_dict"]) > 0: record_data(data_record, [reg_value], ["reg"]) if "param" in record_keys: record_data(data_record, [model.get_weights_bias(W_source = "core", b_source = "core")], ["param"]) if "param_grad" in record_keys: record_data(data_record, [model.get_weights_bias(W_source = "core", b_source = "core", is_grad = True)], ["param_grad"]) print("\n") try: sys.stdout.flush() except: pass if isplot: if inspect_image_interval is not None and hasattr(model, "plot"): if i % inspect_image_interval == 0: if gradient_noise is not None: print("gradient_noise: {0:.9f}".format(current_gradient_noise_scale)) plot_model(model, data_loader = validation_loader, X = X_valid, y = y_valid, transform_label=transform_label, data_loader_apply=data_loader_apply) if co_kwargs is not None and "inspect_image_interval" in co_kwargs and co_kwargs["inspect_image_interval"] and hasattr(co_model, "plot"): if i % co_kwargs["inspect_image_interval"] == 0: plot_model(co_model, data_loader = validation_loader, X = X_valid, y = y_valid, transform_label=transform_label, data_loader_apply=data_loader_apply) if save_interval is not None: if i % save_interval == 0: record_data(data_record, [model.model_dict], ["model_dict"]) if co_kwargs is not None: record_data(data_record, [co_model.model_dict], ["co_model_dict"]) if filename is not None: pickle.dump(data_record, open(filename, "wb")) if to_stop: break loss_value = get_loss(model, validation_loader, X_valid, y_valid, criterion=criterion, loss_epoch=epochs, transform_label=transform_label, **kwargs) if isplot: import matplotlib.pylab as plt for key, item in data_record.items(): if isinstance(item, Number) or len(data_record["iter"]) != len(item): continue if key not in ["iter", "model_dict"]: if key in ["accuracy"]: plt.figure(figsize = (8,6)) plt.plot(data_record["iter"], data_record[key]) plt.xlabel("epoch") plt.ylabel(key) plt.title(key) plt.show() else: plt.figure(figsize = (8,6)) plt.semilogy(data_record["iter"], data_record[key]) plt.xlabel("epoch") plt.ylabel(key) plt.title(key) plt.show() return loss_original, loss_value, data_record def train_simple(model, X, y, validation_data = None, inspect_interval = 5, **kwargs): """minimal version of training. "model" can be a single model or a ordered list of models""" def get_regularization(model, **kwargs): reg_dict = kwargs["reg_dict"] if "reg_dict" in kwargs else None reg = to_Variable([0], is_cuda = X.is_cuda) for model_ele in model: if reg_dict is not None: for reg_type, reg_coeff in reg_dict.items(): reg = reg + model_ele.get_regularization(source = [reg_type], mode = "L1", **kwargs) * reg_coeff return reg if not(isinstance(model, list) or isinstance(model, tuple)): model = [model] epochs = kwargs["epochs"] if "epochs" in kwargs else 2000 lr = kwargs["lr"] if "lr" in kwargs else 5e-3 optim_type = kwargs["optim_type"] if "optim_type" in kwargs else "adam" optim_kwargs = kwargs["optim_kwargs"] if "optim_kwargs" in kwargs else {} loss_type = kwargs["loss_type"] if "loss_type" in kwargs else "mse" early_stopping_epsilon = kwargs["early_stopping_epsilon"] if "early_stopping_epsilon" in kwargs else 0 patience = kwargs["patience"] if "patience" in kwargs else 40 record_keys = kwargs["record_keys"] if "record_keys" in kwargs else ["loss", "mse", "data_DL", "model_DL"] scheduler_type = kwargs["scheduler_type"] if "scheduler_type" in kwargs else "ReduceLROnPlateau" loss_precision_floor = kwargs["loss_precision_floor"] if "loss_precision_floor" in kwargs else PrecisionFloorLoss autoencoder = kwargs["autoencoder"] if "autoencoder" in kwargs else None data_record = {key: [] for key in record_keys} isplot = kwargs["isplot"] if "isplot" in kwargs else False if patience is not None: early_stopping = Early_Stopping(patience = patience, epsilon = early_stopping_epsilon) if validation_data is not None: X_valid, y_valid = validation_data else: X_valid, y_valid = X, y # Get original loss: criterion = get_criterion(loss_type, loss_precision_floor = loss_precision_floor) DL_criterion = Loss_Fun(core = "DLs", loss_precision_floor = loss_precision_floor, DL_sum = True) DL_criterion_absolute = Loss_Fun(core = "DLs", loss_precision_floor = PrecisionFloorLoss, DL_sum = True) pred_valid = forward(model, X_valid, **kwargs) loss_original = to_np_array(criterion(pred_valid, y_valid)) if "loss" in record_keys: record_data(data_record, [-1, loss_original], ["iter","loss"]) if "mse" in record_keys: record_data(data_record, [to_np_array(nn.MSELoss()(pred_valid, y_valid))], ["mse"]) if "data_DL" in record_keys: record_data(data_record, [to_np_array(DL_criterion(pred_valid, y_valid))], ["data_DL"]) if "data_DL_absolute" in record_keys: record_data(data_record, [to_np_array(DL_criterion_absolute(pred_valid, y_valid))], ["data_DL_absolute"]) if "model_DL" in record_keys: record_data(data_record, [get_model_DL(model)], ["model_DL"]) if "param" in record_keys: record_data(data_record, [model[0].get_weights_bias(W_source = "core", b_source = "core")], ["param"]) if "param_grad" in record_keys: record_data(data_record, [model[0].get_weights_bias(W_source = "core", b_source = "core", is_grad = True)], ["param_grad"]) if "param_collapse_layers" in record_keys: record_data(data_record, [simplify(deepcopy(model[0]), X, y, "collapse_layers", verbose = 0)[0] .get_weights_bias(W_source = "core", b_source = "core")], ["param"]) # Setting up optimizer: parameters = itertools.chain(*[model_ele.parameters() for model_ele in model]) num_params = np.sum([[len(list(model_ele.parameters())) for model_ele in model]]) if num_params == 0: print("No parameters to optimize!") pred_valid = forward(model, X_valid, **kwargs) loss_value = to_np_array(criterion(pred_valid, y_valid)) if "loss" in record_keys: record_data(data_record, [0, loss_value], ["iter", "loss"]) if "mse" in record_keys: record_data(data_record, [to_np_array(nn.MSELoss()(pred_valid, y_valid))], ["mse"]) if "data_DL" in record_keys: record_data(data_record, [to_np_array(DL_criterion(pred_valid, y_valid))], ["data_DL"]) if "data_DL_absolute" in record_keys: record_data(data_record, [to_np_array(DL_criterion_absolute(pred_valid, y_valid))], ["data_DL_absolute"]) if "model_DL" in record_keys: record_data(data_record, [get_model_DL(model)], ["model_DL"]) if "param" in record_keys: record_data(data_record, [model[0].get_weights_bias(W_source = "core", b_source = "core")], ["param"]) if "param_grad" in record_keys: record_data(data_record, [model[0].get_weights_bias(W_source = "core", b_source = "core", is_grad = True)], ["param_grad"]) if "param_collapse_layers" in record_keys: record_data(data_record, [simplify(deepcopy(model[0]), X, y, "collapse_layers", verbose = 0)[0] .get_weights_bias(W_source = "core", b_source = "core")], ["param"]) return loss_original, loss_value, data_record optimizer = get_optimizer(optim_type, lr, parameters, **optim_kwargs) # Set up learning rate scheduler: if scheduler_type is not None: if scheduler_type == "ReduceLROnPlateau": scheduler_patience = kwargs["scheduler_patience"] if "scheduler_patience" in kwargs else 10 scheduler_factor = kwargs["scheduler_factor"] if "scheduler_factor" in kwargs else 0.1 scheduler = ReduceLROnPlateau(optimizer, factor = scheduler_factor, patience = scheduler_patience) elif scheduler_type == "LambdaLR": scheduler_lr_lambda = kwargs["scheduler_lr_lambda"] if "scheduler_lr_lambda" in kwargs else (lambda epoch: 1 / (1 + 0.01 * epoch)) scheduler = LambdaLR(optimizer, lr_lambda = scheduler_lr_lambda) else: raise # Training: to_stop = False for i in range(epochs + 1): if optim_type != "LBFGS": optimizer.zero_grad() pred = forward(model, X, **kwargs) reg = get_regularization(model, **kwargs) loss = criterion(pred, y) + reg loss.backward() optimizer.step() else: # "LBFGS" is a second-order optimization algorithm that requires a slightly different procedure: def closure(): optimizer.zero_grad() pred = forward(model, X, **kwargs) reg = get_regularization(model, **kwargs) loss = criterion(pred, y) + reg loss.backward() return loss optimizer.step(closure) if i % inspect_interval == 0: pred_valid = forward(model, X_valid, **kwargs) loss_value = to_np_array(criterion(pred_valid, y_valid)) if scheduler_type is not None: if scheduler_type == "ReduceLROnPlateau": scheduler.step(loss_value) else: scheduler.step() if "loss" in record_keys: record_data(data_record, [i, loss_value], ["iter", "loss"]) if "mse" in record_keys: record_data(data_record, [to_np_array(nn.MSELoss()(pred_valid, y_valid))], ["mse"]) if "data_DL" in record_keys: record_data(data_record, [to_np_array(DL_criterion(pred_valid, y_valid))], ["data_DL"]) if "data_DL_absolute" in record_keys: record_data(data_record, [to_np_array(DL_criterion_absolute(pred_valid, y_valid))], ["data_DL_absolute"]) if "model_DL" in record_keys: record_data(data_record, [get_model_DL(model)], ["model_DL"]) if "param" in record_keys: record_data(data_record, [model[0].get_weights_bias(W_source = "core", b_source = "core")], ["param"]) if "param_grad" in record_keys: record_data(data_record, [model[0].get_weights_bias(W_source = "core", b_source = "core", is_grad = True)], ["param_grad"]) if "param_collapse_layers" in record_keys: record_data(data_record, [simplify(deepcopy(model[0]), X, y, "collapse_layers", verbose = 0)[0] .get_weights_bias(W_source = "core", b_source = "core")], ["param"]) if patience is not None: to_stop = early_stopping.monitor(loss_value) if to_stop: break pred_valid = forward(model, X_valid, **kwargs) loss_value = to_np_array(criterion(pred_valid, y_valid)) if isplot: import matplotlib.pylab as plt if "mse" in data_record: plt.semilogy(data_record["iter"], data_record["mse"]) plt.xlabel("epochs") plt.title("MSE") plt.show() if "loss" in data_record: plt.plot(data_record["iter"], data_record["loss"]) plt.xlabel("epochs") plt.title("Loss") plt.show() return loss_original, loss_value, data_record def load_model_dict_net(model_dict, is_cuda = False): net_type = model_dict["type"] if net_type.startswith("MLP"): return MLP(input_size = model_dict["input_size"], struct_param = model_dict["struct_param"] if "struct_param" in model_dict else None, W_init_list = model_dict["weights"] if "weights" in model_dict else None, b_init_list = model_dict["bias"] if "bias" in model_dict else None, settings = model_dict["settings"] if "settings" in model_dict else {}, is_cuda = is_cuda, ) elif net_type == "Labelmix_MLP": model = Labelmix_MLP(input_size=model_dict["input_size"], struct_param=model_dict["struct_param"], idx_label=model_dict["idx_label"] if "idx_label" in model_dict else None, is_cuda=is_cuda, ) if "state_dict" in model_dict: model.load_state_dict(model_dict["state_dict"]) return model elif net_type == "Multi_MLP": return Multi_MLP(input_size = model_dict["input_size"], struct_param = model_dict["struct_param"], W_init_list = model_dict["weights"] if "weights" in model_dict else None, b_init_list = model_dict["bias"] if "bias" in model_dict else None, settings = model_dict["settings"] if "settings" in model_dict else {}, is_cuda = is_cuda, ) elif net_type == "Branching_Net": return Branching_Net(net_base_model_dict = model_dict["net_base_model_dict"], net_1_model_dict = model_dict["net_1_model_dict"], net_2_model_dict = model_dict["net_2_model_dict"], is_cuda = is_cuda, ) elif net_type == "Fan_in_MLP": return Fan_in_MLP(model_dict_branch1=model_dict["model_dict_branch1"], model_dict_branch2=model_dict["model_dict_branch2"], model_dict_joint=model_dict["model_dict_joint"], is_cuda=is_cuda, ) elif net_type == "Net_reparam": return Net_reparam(model_dict=model_dict["model"], reparam_mode=model_dict["reparam_mode"], is_cuda=is_cuda, ) elif net_type == "Wide_ResNet": model = Wide_ResNet(depth=model_dict["depth"], widen_factor=model_dict["widen_factor"], input_channels=model_dict["input_channels"], output_size=model_dict["output_size"], dropout_rate=model_dict["dropout_rate"], is_cuda=is_cuda, ) if "state_dict" in model_dict: model.load_state_dict(model_dict["state_dict"]) return model elif net_type.startswith("ConvNet"): return ConvNet(input_channels = model_dict["input_channels"], struct_param = model_dict["struct_param"], W_init_list = model_dict["weights"] if "weights" in model_dict else None, b_init_list = model_dict["bias"] if "bias" in model_dict else None, settings = model_dict["settings"] if "settings" in model_dict else {}, return_indices = model_dict["return_indices"] if "return_indices" in model_dict else False, is_cuda = is_cuda, ) elif net_type == "Conv_Autoencoder": model = Conv_Autoencoder(input_channels_encoder = model_dict["input_channels_encoder"], input_channels_decoder = model_dict["input_channels_decoder"], struct_param_encoder = model_dict["struct_param_encoder"], struct_param_decoder = model_dict["struct_param_decoder"], settings = model_dict["settings"], is_cuda = is_cuda, ) if "encoder" in model_dict: model.encoder.load_model_dict(model_dict["encoder"]) if "decoder" in model_dict: model.decoder.load_model_dict(model_dict["decoder"]) return model elif model_dict["type"] == "Conv_Model": is_generative = model_dict["is_generative"] if "is_generative" in model_dict else False return Conv_Model(encoder_model_dict = model_dict["encoder_model_dict"] if not is_generative else None, core_model_dict = model_dict["core_model_dict"], decoder_model_dict = model_dict["decoder_model_dict"], latent_size = model_dict["latent_size"], is_generative = model_dict["is_generative"] if is_generative else False, is_res_block = model_dict["is_res_block"] if "is_res_block" in model_dict else False, is_cuda = is_cuda, ) else: raise Exception("net_type {} not recognized!".format(net_type)) def load_model_dict(model_dict, is_cuda = False): net_type = model_dict["type"] if net_type not in ["Model_Ensemble", "LSTM", "Model_with_Uncertainty", "Mixture_Model", "Mixture_Gaussian"]: return load_model_dict_net(model_dict, is_cuda = is_cuda) elif net_type == "Model_Ensemble": if model_dict["model_type"] == "MLP": model_ensemble = Model_Ensemble( num_models = model_dict["num_models"], input_size = model_dict["input_size"], model_type = model_dict["model_type"], output_size = model_dict["output_size"], is_cuda = is_cuda, # Here we just create some placeholder network. The model will be overwritten in the next steps: struct_param = [[1, "Simple_Layer", {}]], ) elif model_dict["model_type"] == "LSTM": model_ensemble = Model_Ensemble( num_models = model_dict["num_models"], input_size = model_dict["input_size"], model_type = model_dict["model_type"], output_size = model_dict["output_size"], is_cuda = is_cuda, # Here we just create some placeholder network. The model will be overwritten in the next steps: hidden_size = 3, output_struct_param = [[1, "Simple_Layer", {}]], ) else: raise for k in range(model_ensemble.num_models): setattr(model_ensemble, "model_{}".format(k), load_model_dict(model_dict["model_{}".format(k)], is_cuda = is_cuda)) return model_ensemble elif net_type == "Model_with_Uncertainty": return Model_with_Uncertainty(model_pred = load_model_dict(model_dict["model_pred"], is_cuda = is_cuda), model_logstd = load_model_dict(model_dict["model_logstd"], is_cuda = is_cuda)) elif net_type == "Mixture_Model": return Mixture_Model(model_dict_list=model_dict["model_dict_list"], weight_logits_model_dict=model_dict["weight_logits_model_dict"], num_components=model_dict["num_components"], is_cuda=is_cuda, ) elif net_type == "Mixture_Gaussian": return load_model_dict_Mixture_Gaussian(model_dict, is_cuda = is_cuda) else: raise Exception("net_type {} not recognized!".format(net_type)) ## Helper functions: def get_accuracy(pred, target): """Get accuracy from prediction and target""" assert len(pred.shape) == len(target.shape) == 1 assert len(pred) == len(target) pred, target = to_np_array(pred, target) accuracy = ((pred == target).sum().astype(float) / len(pred)) return accuracy def flatten(*tensors): """Flatten the tensor except the first dimension""" new_tensors = [] for tensor in tensors: new_tensors.append(tensor.view(tensor.size(0), -1)) if len(new_tensors) == 1: new_tensors = new_tensors[0] return new_tensors def fill_triangular(vec, dim, mode = "lower"): """Fill an lower or upper triangular matrices with given vectors""" num_examples, size = vec.shape assert size == dim * (dim + 1) // 2 matrix = torch.zeros(num_examples, dim, dim).to(vec.device) idx = (torch.tril(torch.ones(dim, dim)) == 1).unsqueeze(0) idx = idx.repeat(num_examples,1,1) if mode == "lower": matrix[idx] = vec.contiguous().view(-1) elif mode == "upper": matrix[idx] = vec.contiguous().view(-1) else: raise Exception("mode {} not recognized!".format(mode)) return matrix def matrix_diag_transform(matrix, fun): """Return the matrices whose diagonal elements have been executed by the function 'fun'.""" num_examples = len(matrix) idx = torch.eye(matrix.size(-1)).bool().unsqueeze(0) idx = idx.repeat(num_examples, 1, 1) new_matrix = matrix.clone() new_matrix[idx] = fun(matrix.diagonal(dim1 = 1, dim2 = 2).contiguous().view(-1)) return new_matrix def Zip(*data, **kwargs): """Recursive unzipping of data structure Example: Zip(*[(('a',2), 1), (('b',3), 2), (('c',3), 3), (('d',2), 4)]) ==> [[['a', 'b', 'c', 'd'], [2, 3, 3, 2]], [1, 2, 3, 4]] Each subtree in the original data must be in the form of a tuple. In the **kwargs, you can set the function that is applied to each fully unzipped subtree. """ import collections function = kwargs["function"] if "function" in kwargs else None if len(data) == 1: return data[0] data = [list(element) for element in zip(*data)] for i, element in enumerate(data): if isinstance(element[0], tuple): data[i] = Zip(*element, **kwargs) elif isinstance(element, list): if function is not None: data[i] = function(element) return data def get_loss(model, data_loader=None, X=None, y=None, criterion=None, transform_label=None, **kwargs): """Get loss using the whole data or data_loader. Return the average validation loss with np.ndarray format""" max_validation_iter = kwargs["max_validation_iter"] if "max_validation_iter" in kwargs else None if transform_label is None: transform_label = Transform_Label() if "loader_process" in kwargs and kwargs["loader_process"] is not None: data_loader = kwargs["loader_process"]("test") if data_loader is not None: assert X is None and y is None loss_record = 0 count = 0 # Taking the average of all metrics: for j, data_batch in enumerate(data_loader): if isinstance(data_batch, tuple) or isinstance(data_batch, list): X_batch, y_batch = data_batch if "data_loader_apply" in kwargs and kwargs["data_loader_apply"] is not None: X_batch, y_batch = kwargs["data_loader_apply"](X_batch, y_batch) else: X_batch, y_batch = kwargs["data_loader_apply"](data_batch) loss_ele = to_np_array(model.get_loss(X_batch, transform_label(y_batch), criterion = criterion, **kwargs)) if j == 0: all_info_dict = {key: 0 for key in model.info_dict.keys()} loss_record = loss_record + loss_ele count += 1 for key in model.info_dict: all_info_dict[key] = all_info_dict[key] + model.info_dict[key] if max_validation_iter is not None and count > max_validation_iter: break for key in model.info_dict: all_info_dict[key] = all_info_dict[key] / count loss = loss_record / count model.info_dict = deepcopy(all_info_dict) else: assert X is not None and y is not None loss = to_np_array(model.get_loss(X, transform_label(y), criterion = criterion, **kwargs)) return loss def plot_model(model, data_loader=None, X=None, y=None, transform_label=None, **kwargs): data_loader_apply = kwargs["data_loader_apply"] if "data_loader_apply" in kwargs else None max_validation_iter = kwargs["max_validation_iter"] if "max_validation_iter" in kwargs else None if transform_label is None: transform_label = Transform_Label() if "loader_process" in kwargs and kwargs["loader_process"] is not None: data_loader = kwargs["loader_process"]("test") if data_loader is not None: assert X is None and y is None X_all = [] y_all = [] for i, data_batch in enumerate(data_loader): if isinstance(data_batch, tuple) or isinstance(data_batch, list): X_batch, y_batch = data_batch if data_loader_apply is not None: X_batch, y_batch = data_loader_apply(X_batch, y_batch) else: X_batch, y_batch = data_loader_apply(data_batch) X_all.append(X_batch) y_all.append(y_batch) if max_validation_iter is not None and i >= max_validation_iter: break if not isinstance(X_all[0], torch.Tensor): X_all = Zip(*X_all, function = torch.cat) else: X_all = torch.cat(X_all, 0) y_all = torch.cat(y_all) model.plot(X_all, transform_label(y_all)) else: assert X is not None and y is not None model.plot(X, transform_label(y)) def prepare_inspection(model, data_loader=None, X=None, y=None, transform_label=None, **kwargs): inspect_functions = kwargs["inspect_functions"] if "inspect_functions" in kwargs else None max_validation_iter = kwargs["max_validation_iter"] if "max_validation_iter" in kwargs else None verbose = kwargs["verbose"] if "verbose" in kwargs else False if transform_label is None: transform_label = Transform_Label() if "loader_process" in kwargs and kwargs["loader_process"] is not None: data_loader = kwargs["loader_process"]("test") if data_loader is None: assert X is not None and y is not None all_dict_summary = model.prepare_inspection(X, transform_label(y), **kwargs) if inspect_functions is not None: for inspect_function_key, inspect_function in inspect_functions.items(): all_dict_summary[inspect_function_key] = inspect_function(model, X, y, **kwargs) else: assert X is None and y is None all_dict = {} for j, data_batch in enumerate(data_loader): if verbose is True: print("valid step: {}".format(j)) if isinstance(data_batch, tuple) or isinstance(data_batch, list): X_batch, y_batch = data_batch if "data_loader_apply" in kwargs and kwargs["data_loader_apply"] is not None: X_batch, y_batch = kwargs["data_loader_apply"](X_batch, y_batch) else: X_batch, y_batch = kwargs["data_loader_apply"](data_batch) info_dict = model.prepare_inspection(X_batch, transform_label(y_batch), valid_step=j, **kwargs) for key, item in info_dict.items(): if key not in all_dict: all_dict[key] = [item] else: all_dict[key].append(item) if inspect_functions is not None: for inspect_function_key, inspect_function in inspect_functions.items(): inspect_function_result = inspect_function(model, X_batch, transform_label(y_batch), **kwargs) if inspect_function_key not in all_dict: all_dict[inspect_function_key] = [inspect_function_result] else: all_dict[inspect_function_key].append(inspect_function_result) if max_validation_iter is not None and j >= max_validation_iter: break all_dict_summary = {} for key, item in all_dict.items(): all_dict_summary[key + "_val"] = np.mean(all_dict[key]) return all_dict_summary def get_inspect_items_train(inspect_items): if inspect_items is None: return [] inspect_items_train = [] for item in inspect_items: if item.endswith("_tr"): inspect_items_train.append("_".join(item.split("_")[:-1])) return inspect_items_train def update_key_train(info_dict_train, inspect_items_train): info_dict_train_new = {} for key, item in info_dict_train.items(): if key in inspect_items_train: info_dict_train_new[key + "_tr"] = item return deepcopy(info_dict_train_new) # ## Simplifying functionality: # In[ ]: def simplify(model, X=None, y=None, mode="full", isplot=False, target_name=None, validation_data=None, **kwargs): """Simplify a neural network model in various ways. "model" can be a single model or a ordered list of models""" verbose = kwargs["verbose"] if "verbose" in kwargs else 1 if validation_data is None: X_valid, y_valid = X, y else: X_valid, y_valid = validation_data simplify_criteria = kwargs["simplify_criteria"] if "simplify_criteria" in kwargs else ("DLs", 0.05, 3, "relative") # the first argument choose from "DL", "loss" simplify_epsilon = simplify_criteria[1] simplify_patience = simplify_criteria[2] simplify_compare_mode = simplify_criteria[3] performance_monitor = Performance_Monitor(patience = simplify_patience, epsilon = simplify_epsilon, compare_mode = simplify_compare_mode) record_keys = kwargs["record_keys"] if "record_keys" in kwargs else ["mse"] loss_precision_floor = kwargs["loss_precision_floor"] if "loss_precision_floor" in kwargs else PrecisionFloorLoss if X is not None: if y is None: y = Variable(forward(model, X, **kwargs).data, requires_grad = False) if not (isinstance(model, list) or isinstance(model, tuple)): model = [model] is_list = False else: is_list = True if mode == "full": mode = ["collapse_layers", "snap"] if not isinstance(mode, list): mode = [mode] # Obtain the original loss and setup criterion: loss_type = kwargs["loss_type"] if "loss_type" in kwargs else "mse" criterion = get_criterion(loss_type, loss_precision_floor = loss_precision_floor) DL_criterion = Loss_Fun(core = "DLs", loss_precision_floor = loss_precision_floor, DL_sum = True) loss_dict = OrderedDict() for mode_ele in mode: if verbose >= 1: print("\n" + "=" * 48 + "\nSimplifying mode: {}".format(mode_ele), end = "") if mode_ele == "snap": snap_mode = kwargs["snap_mode"] if "snap_mode" in kwargs else "integer" print(" {}".format(snap_mode), end = "") if target_name is not None: print(" for {}".format(target_name)) else: print() print("=" * 48) # Record the loss before simplification: if X is not None: pred_valid = forward(model, X_valid, **kwargs) loss_original = to_np_array(criterion(pred_valid, y_valid)) loss_list = [loss_original] if verbose >= 1: print("original_loss: {}".format(loss_original)) mse_record_whole = [to_np_array(nn.MSELoss()(pred_valid, y_valid))] data_DL_whole = [to_np_array(DL_criterion(pred_valid, y_valid))] model_DL_whole = [get_model_DL(model)] event_list = ["before simplification"] iter_end_whole = [1] is_accept_whole = [] if "param" in record_keys: param_record_whole = [model[0].get_weights_bias(W_source = "core", b_source = "core")] if "param_grad" in record_keys: param_grad_record_whole = [model[0].get_weights_bias(W_source = "core", b_source = "core", is_grad = True)] # Begin simplification: if mode_ele == "collapse_layers": all_collapse_dict = {} for model_id, model_ele in enumerate(model): # Obtain activations for each layer: activation_list = [] for k in range(len(model_ele.struct_param)): if "activation" in model_ele.struct_param[k][2]: activation_list.append(model_ele.struct_param[k][2]["activation"]) elif "activation" in model_ele.settings: activation_list.append(model_ele.settings["activation"]) else: activation_list.append("default") # Build the collapse_list that stipulates which layers to collapse: collapse_dict = {} current_start = None current_layer_type = None for k, activation in enumerate(activation_list): if activation == "linear" and k != len(activation_list) - 1: if k not in collapse_dict and current_start is None: # Create a new bunch: if model_ele.struct_param[k + 1][1] == model_ele.struct_param[k][1]: # The current layer must have the same layer_type as the next layer current_start = k collapse_dict[current_start] = [k] current_layer_type = model_ele.struct_param[k][1] else: # Adding to current bunch: if model_ele.struct_param[k + 1][1] == model_ele.struct_param[k][1] == current_layer_type: collapse_dict[current_start].append(k) else: collapse_dict[current_start].append(k) current_start = None else: if current_start is not None: collapse_dict[current_start].append(k) current_start = None # Build new layer: new_layer_info = {} for current_start, layer_ids in collapse_dict.items(): for i, layer_id in enumerate(layer_ids): layer = getattr(model_ele, "layer_{}".format(layer_id)) if i == 0: W_accum = layer.W_core b_accum = layer.b_core else: W_accum = torch.matmul(W_accum, layer.W_core) b_accum = torch.matmul(b_accum, layer.W_core) + layer.b_core if model_ele.is_cuda: W_accum = W_accum.cpu() b_accum = b_accum.cpu() last_layer_id = collapse_dict[current_start][-1] new_layer_info[current_start] = {"W_init": W_accum.data.numpy(), "b_init": b_accum.data.numpy(), "layer_struct_param": [b_accum.size(0), model_ele.struct_param[last_layer_id][1], deepcopy(model_ele.struct_param[last_layer_id][2])], } new_layer_info[current_start].pop("snap_dict", None) if verbose >= 1: print("model_id {}, layers collapsed: {}".format(model_id, collapse_dict)) # Rebuild the Net: if len(collapse_dict) > 0: all_collapse_dict[model_id] = {"collapse_dict": collapse_dict, "new_layer_info": new_layer_info, "collapse_layer_ids": [idx for item in collapse_dict.values() for idx in item], } # Rebuild the list of models: if len(all_collapse_dict) > 0: model_new = [] for model_id, model_ele in enumerate(model): if model_id in all_collapse_dict: W_list, b_list = model_ele.get_weights_bias(W_source = "core", b_source = "core") W_init_list = [] b_init_list = [] struct_param = [] for k in range(len(model_ele.struct_param)): if k not in all_collapse_dict[model_id]["collapse_layer_ids"]: struct_param.append(model_ele.struct_param[k]) W_init_list.append(W_list[k]) b_init_list.append(b_list[k]) else: if k in all_collapse_dict[model_id]["collapse_dict"].keys(): struct_param.append(all_collapse_dict[model_id]["new_layer_info"][k]["layer_struct_param"]) W_init_list.append(all_collapse_dict[model_id]["new_layer_info"][k]["W_init"]) b_init_list.append(all_collapse_dict[model_id]["new_layer_info"][k]["b_init"]) model_ele_new = MLP(input_size = model_ele.input_size, struct_param = struct_param, W_init_list = W_init_list, b_init_list = b_init_list, settings = model_ele.settings, is_cuda = model_ele.is_cuda, ) else: model_ele_new = model_ele model_new.append(model_ele_new) model = model_new # Calculate the loss again: pred_valid = forward(model, X_valid, **kwargs) loss_new = to_np_array(criterion(pred_valid, y_valid)) if verbose >= 1: print("after collapsing linear layers in all models, new loss {}".format(loss_new)) loss_list.append(loss_new) mse_record_whole.append(to_np_array(nn.MSELoss()(pred_valid, y_valid))) data_DL_whole.append(to_np_array(DL_criterion(pred_valid, y_valid))) model_DL_whole.append(get_model_DL(model)) if "param" in record_keys: param_record_whole.append(model[0].get_weights_bias(W_source = "core", b_source = "core")) if "param_grad" in record_keys: param_grad_record_whole.append(model[0].get_weights_bias(W_source = "core", b_source = "core", is_grad = True)) iter_end_whole.append(1) event_list.append({mode_ele: all_collapse_dict}) elif mode_ele in ["local", "snap"]: # 'local': greedily try reducing the input dimension by removing input dimension from the beginning; # 'snap': greedily snap each float parameter into an integer or rational number. Set argument 'snap_mode' == 'integer' or 'rational'. if mode_ele == "snap": target_params = [[(model_id, layer_id), "snap"] for model_id, model_ele in enumerate(model) for layer_id in range(len(model_ele.struct_param))] elif mode_ele == "local": for model_id, model_ele in enumerate(model): if len(model_ele.struct_param) > 0: first_model_id = model_id break first_layer = getattr(model[first_model_id], "layer_0") target_params = [[(first_model_id, 0), [[(("weight", (i, j)), 0.) for j in range(first_layer.output_size)] for i in range(first_layer.input_size)]]] else: raise excluded_idx_dict = {item[0]: [] for item in target_params} target_layer_ids_exclude = [] for (model_id, layer_id), target_list in target_params: layer = getattr(model[model_id], "layer_{}".format(layer_id)) if isinstance(target_list, list): max_passes = len(target_list) elif target_list == "snap": max_passes = (layer.input_size + 1) * layer.output_size if "max_passes" in kwargs: max_passes = min(max_passes, kwargs["max_passes"]) else: raise Exception("target_list {} not recognizable!".format(target_list)) if verbose >= 2: print("\n****starting model:****") model[model_id].get_weights_bias(W_source = "core", b_source = "core", verbose = True) print("********\n" ) performance_monitor.reset() criteria_value, criteria_result = get_criteria_value(model, X, y, criteria_type = simplify_criteria[0], criterion = criterion, **kwargs) to_stop, pivot_dict, log, _, pivot_id = performance_monitor.monitor(criteria_value, model_dict = model[model_id].model_dict, criteria_result = criteria_result) for i in range(max_passes): # Perform tentative simplification if isinstance(target_list, list): info = layer.simplify(mode = "snap", excluded_idx = excluded_idx_dict[(model_id, layer_id)], snap_targets = target_list[i], **kwargs) else: info = layer.simplify(mode = "snap", excluded_idx = excluded_idx_dict[(model_id, layer_id)], **kwargs) if len(info) == 0: target_layer_ids_exclude.append((model_id, layer_id)) print("Pass {0}, (model {1}, layer {2}) has no parameters to snap. Revert to pivot model. Go to next layer".format(i, model_id, layer_id)) break excluded_idx_dict[(model_id, layer_id)] = excluded_idx_dict[(model_id, layer_id)] + info _, loss_new, data_record = train_simple(model, X, y, optim_type = "adam", validation_data = validation_data, **kwargs) if verbose >= 2: print("=" * 8) model[model_id].get_weights_bias(W_source = "core", b_source = "core", verbose = True) criteria_value, criteria_result = get_criteria_value(model, X, y, criteria_type = simplify_criteria[0], criterion = criterion, **kwargs) to_stop, pivot_dict, log, is_accept, pivot_id = performance_monitor.monitor(criteria_value, model_dict = model[model_id].model_dict, criteria_result = criteria_result) is_accept_whole.append(is_accept) if is_accept: print('[Accepted] as pivot model!') print() # Check if the criterion after simplification and refit is worse. If it is worse than the simplify_epsilon, revert: if to_stop: target_layer_ids_exclude.append((model_id, layer_id)) if verbose >= 1: print("Pass {0}, loss: {1}\tDL: {2}. New snap {3} is do not improve by {4} = {5} for {6} steps. Revert the simplification to pivot model. Go to next layer.".format( i, view_item(log, ("criteria_result", "loss")), view_item(log, ("criteria_result", "DL")), info, simplify_criteria[0], simplify_epsilon, simplify_patience)) break mse_record_whole += data_record["mse"] data_DL_whole += data_record["data_DL"] model_DL_whole += data_record["model_DL"] if "param" in record_keys: param_record_whole += data_record["param"] if "param_grad" in record_keys: param_grad_record_whole += data_record["param_grad"] iter_end_whole.append(len(data_record["mse"])) model[model_id].reset_layer(layer_id, layer) loss_list.append(loss_new) event_list.append({mode_ele: ((model_id, layer_id), info)}) if verbose >= 1: print("Pass {0}, snap (model {1}, layer {2}), snap {3}. \tloss: {4}\tDL: {5}".format( i, model_id, layer_id, info, view_item(log, ("criteria_result", "loss")), view_item(log, ("criteria_result", "DL")))) # Update the whole model's struct_param and snap_dict: model[model_id].load_model_dict(pivot_dict["model_dict"]) model[model_id].synchronize_settings() if verbose >= 2: print("\n****pivot model at {}th transformation:****".format(pivot_id)) model[model_id].get_weights_bias(W_source = "core", b_source = "core", verbose = True) print("********\n" ) elif mode_ele == "pair_snap": model_new = [] for model_id, model_ele in enumerate(model): for layer_id, layer_struct_param in enumerate(model_ele.struct_param): if layer_struct_param[1] == "Symbolic_Layer": layer = getattr(model_ele, "layer_{}".format(layer_id)) max_passes = len(layer.get_param_dict()) - 1 if "max_passes" in kwargs: max_passes = min(max_passes, kwargs["max_passes"]) if verbose > 1: print("original:") print("symbolic_expression: ", layer.symbolic_expression) print("numerical_expression: ", layer.numerical_expression) print() performance_monitor.reset() criteria_value, criteria_result = get_criteria_value(model, X, y, criteria_type = simplify_criteria[0], criterion = criterion, **kwargs) to_stop, pivot_dict, log, _, pivot_id = performance_monitor.monitor(criteria_value, model_dict = model[model_id].model_dict, criteria_result = criteria_result) for i in range(max_passes): info = layer.simplify(mode = "pair_snap", **kwargs) if len(info) == 0: target_layer_ids_exclude.append((model_id, layer_id)) print("Pass {0}, (model {1}, layer {2}) has no parameters to pair_snap. Revert to pivot model. Go to next layer".format(i, model_id, layer_id)) break _, loss, data_record = train_simple(model, X, y, optim_type = "adam", epochs = 1000, validation_data = validation_data, **kwargs) criteria_value, criteria_result = get_criteria_value(model, X, y, criteria_type = simplify_criteria[0], criterion = criterion, **kwargs) to_stop, pivot_dict, log, is_accept, pivot_id = performance_monitor.monitor(criteria_value, model_dict = model[model_id].model_dict, criteria_result = criteria_result) is_accept_whole.append(is_accept) if to_stop: if verbose >= 1: print("\nPass {0}, loss: {1}\tDL: {2}. New snap {3} is do not improve by {4} = {5} for {6} steps. Revert the simplification to pivot model. Go to next layer.".format( i, view_item(log, ("criteria_result", "loss")), view_item(log, ("criteria_result", "DL")), info, simplify_criteria[0], simplify_epsilon, simplify_patience)) break mse_record_whole += data_record["mse"] data_DL_whole += data_record["data_DL"] model_DL_whole += data_record["model_DL"] if "param" in record_keys: param_record_whole += data_record["param"] if "param_grad" in record_keys: param_grad_record_whole += data_record["param_grad"] iter_end_whole.append(len(data_record["mse"])) model[model_id].reset_layer(layer_id, layer) loss_list.append(loss) event_list.append({mode_ele: ((model_id, layer_id), info)}) if verbose >= 1: print("\nPass {0}, snap (model {1}, layer {2}), snap {3}. \tloss: {4}\tDL: {5}".format( i, model_id, layer_id, info, view_item(log, ("criteria_result", "loss")), view_item(log, ("criteria_result", "DL")))) print("symbolic_expression: ", layer.symbolic_expression) print("numerical_expression: ", layer.numerical_expression) print() model[model_id].load_model_dict(pivot_dict["model_dict"]) print("final: \nsymbolic_expression: ", getattr(model[model_id], "layer_{0}".format(layer_id)).symbolic_expression) print("numerical_expression: ", getattr(model[model_id], "layer_{0}".format(layer_id)).numerical_expression) print() elif mode_ele[:11] == "to_symbolic": from sympy import Symbol force_simplification = kwargs["force_simplification"] if "force_simplification" in kwargs else False is_multi_model = True if len(model) > 1 else False for model_id, model_ele in enumerate(model): for layer_id, layer_struct_param in enumerate(model_ele.struct_param): prefix = "L{}_".format(layer_id) if layer_struct_param[1] == "Simple_Layer": # Obtain loss before simplification: layer = getattr(model_ele, "layer_{}".format(layer_id)) if X is not None: criteria_prev, criteria_result_prev = get_criteria_value(model, X, y, criteria_type = simplify_criteria[0], criterion = criterion, **kwargs) if mode_ele.split("_")[-1] == "separable": new_layer = Simple_2_Symbolic(layer, settings = model_ele.settings, mode = "separable", prefix = prefix) else: new_layer = Simple_2_Symbolic(layer, settings = model_ele.settings, prefix = prefix) model[model_id].reset_layer(layer_id, new_layer) if "snap_dict" in model_ele.settings and layer_id in model_ele.settings["snap_dict"]: subs_targets = [] for (pos, true_idx), item in model_ele.settings["snap_dict"][layer_id].items(): if pos == "weight": subs_targets.append((Symbol("W{0}{1}".format(true_idx[0], true_idx[1])), item["new_value"])) elif pos == "bias": subs_targets.append((Symbol("b{}".format(true_idx)), item["new_value"])) else: raise Exception("pos {} not recognized!".format(pos)) new_expression = [expression.subs(subs_targets) for expression in new_layer.symbolic_expression] new_layer.set_symbolic_expression(new_expression) model_ele.settings["snap_dict"].pop(layer_id) model_ele.struct_param[layer_id][2].update(new_layer.struct_param[2]) # Calculate the loss again: if X is not None: criteria_new, criteria_result_new = get_criteria_value(model, X, y, criteria_type = simplify_criteria[0], criterion = criterion, **kwargs) if verbose >= 1: print("Prev_loss: {0}, new loss: {1}\tprev_DL: {2:.9f}, new DL: {3:.9f}".format( criteria_result_prev["loss"], criteria_result_new["loss"], criteria_result_prev["DL"], criteria_result_new["DL"])) print() if criteria_new > criteria_prev * (1 + 0.05): print("to_symbolic DL increase more than 5%! ", end = "") if not force_simplification: print("Reset layer.") model[model_id].reset_layer(layer_id, layer) else: print("Nevertheless, force simplification.") loss_list.append(criteria_result_new["loss"]) print("{0} succeed. Prev_loss: {1}\tnew_loss: {2}\tprev_DL: {3:.9f}, new_DL: {4:.9f}".format( mode_ele, criteria_result_prev["loss"], criteria_result_new["loss"], criteria_result_prev["DL"], criteria_result_new["DL"])) else: print("{0} succeed.".format(mode_ele)) event_list.append({mode_ele: (model_id, layer_id)}) elif layer_struct_param[1] == "Sneuron_Layer": # Obtain loss before simplification: layer = getattr(model_ele, "layer_{0}".format(layer_id)) criteria_prev, criteria_result_prev = get_criteria_value(model, X, y, criteria_type = simplify_criteria[0], criterion = criterion, **kwargs) new_layer = Sneuron_2_Symbolic(layer, prefix = prefix) model[model_id].reset_layer(layer_id, new_layer) # Calculate the loss again: criteria_new, criteria_result_new = get_criteria_value(model, X, y, criteria_type = simplify_criteria[0], criterion = criterion, **kwargs) if verbose >= 1: print("Prev_loss: {0}, new loss: {1}\tprev_DL: {2:.9f}, new DL: {3:.9f}".format( criteria_result_prev["loss"], criteria_result_new["loss"], criteria_result_prev["DL"], criteria_result_new["DL"])) print() if criteria_new > criteria_prev * (1 + 0.05): print("to_symbolic DL increase more than 5%! ", end = "") if not force_simplification: print("Reset layer.") model[model_id].reset_layer(layer_id, layer) else: print("Nevertheless, force simplification.") loss_list.append(criteria_result_new["loss"]) event_list.append({mode_ele: (model_id, layer_id)}) print("{0} succeed. Prev_loss: {1}\tnew_loss: {2}\tprev_DL: {3:.9f}, new_DL: {4:.9f}".format( mode_ele, criteria_result_prev["loss"], criteria_result_new["loss"], criteria_result_prev["DL"], criteria_result_new["DL"])) if X is not None: mse_record_whole.append(to_np_array(nn.MSELoss()(pred_valid, y_valid))) data_DL_whole.append(to_np_array(DL_criterion(pred_valid, y_valid))) model_DL_whole.append(get_model_DL(model)) if "param" in record_keys: param_record_whole.append(model[0].get_weights_bias(W_source = "core", b_source = "core")) if "param_grad" in record_keys: param_grad_record_whole.append(model[0].get_weights_bias(W_source = "core", b_source = "core", is_grad = True)) iter_end_whole.append(1) elif mode_ele == "symbolic_simplification": """Collapse multi-layer symbolic expression""" from sympy import Symbol, Poly, expand, prod force_simplification = kwargs["force_simplification"] if "force_simplification" in kwargs else False numerical_threshold = kwargs["numerical_threshold"] if "numerical_threshold" in kwargs else None is_numerical = kwargs["is_numerical"] if "is_numerical" in kwargs else False max_poly_degree = kwargs["max_poly_degree"] if "max_poly_degree" in kwargs else None show_before_truncate = kwargs["show_before_truncate"] if "show_before_truncate" in kwargs else False for model_id, model_ele in enumerate(model): is_all_symbolic = True for layer_id, layer_struct_param in enumerate(model_ele.struct_param): if layer_struct_param[1] != "Symbolic_Layer": is_all_symbolic = False if is_all_symbolic: criteria_prev, criteria_result_prev = get_criteria_value(model, X, y, criteria_type = simplify_criteria[0], criterion = criterion, **kwargs) variables = OrderedDict() for i in range(model[0].layer_0.input_size): variables["x{0}".format(i)] = Symbol("x{0}".format(i)) expression = list(variables.values()) param_dict_all = {} # Collapse multiple layers: for layer_id, layer_struct_param in enumerate(model_ele.struct_param): layer = getattr(model_ele, "layer_{0}".format(layer_id)) layer_expression = deepcopy(layer.numerical_expression) layer_expression_new = [] for expr in layer_expression: new_expr = expr.subs({"x{0}".format(i): "t{0}".format(i) for i in range(len(expression))}) # Use a temporary variable to prevent overriding new_expr = new_expr.subs({"t{0}".format(i): expression[i] for i in range(len(expression))}) layer_expression_new.append(expand(new_expr)) expression = layer_expression_new # Show full expression before performing truncation: if show_before_truncate: for i, expr in enumerate(expression): print("Full expression {0}:".format(i)) pp.pprint(Poly(expr, *list(variables.values()))) print() model_ele_candidate = MLP(input_size = model[0].layer_0.input_size, struct_param = [[layer.output_size, "Symbolic_Layer", {"symbolic_expression": "x0"}]], settings = {}, is_cuda = model_ele.is_cuda, ) # Setting maximul degree for polynomial: if max_poly_degree is not None: new_expression = [] for expr in expression: expr = Poly(expr, *list(variables.values())) degree_list = [] coeff_list = [] for degree, coeff in expr.terms(): # Only use monomials with degree not larger than max_poly_degree: if sum(degree) <= max_poly_degree: degree_list.append(degree) coeff_list.append(coeff) new_expr = 0 for degree, coeff in zip(degree_list, coeff_list): new_expr += prod([variables["x{0}".format(i)] ** degree[i] for i in range(len(degree))]) * coeff new_expression.append(new_expr) expression = new_expression # Update symbolic expression for model_ele_candidate: if not is_numerical: param_dict_all = {} expression_new_all = [] for expr in expression: expression_new, param_dict = numerical_2_parameter(expr, idx = len(param_dict_all), threshold = numerical_threshold) expression_new_all.append(expression_new) param_dict_all.update(param_dict) model_ele_candidate.layer_0.set_symbolic_expression(expression_new_all, p_init = param_dict_all) else: model_ele_candidate.layer_0.set_symbolic_expression(expression) model_ele_candidate.layer_0.set_numerical(True) criteria_new, criteria_result_new = get_criteria_value(model_ele_candidate, X, y, criteria_type = simplify_criteria[0], criterion = criterion, **kwargs) if criteria_new > criteria_prev * (1 + 0.05): print("to_symbolic DL increase more than 5%! ", end = "") if force_simplification: print("Nevertheless, force simplification.") model[model_id] = model_ele_candidate else: print("Revert.") else: model[model_id] = model_ele_candidate elif mode_ele == "activation_snap": from sympy import Function def get_sign_snap_candidate(layer, activation_source, excluded_neurons = None): coeff_dict = {} for i in range(len(layer.symbolic_expression)): current_expression = [layer.symbolic_expression[i]] func_names = layer.get_function_name_list(current_expression) if activation_source in func_names: coeff = [element for element in layer.get_param_name_list(current_expression) if element[0] == "W"] coeff_dict[i] = np.mean([np.abs(value) for key, value in layer.get_param_dict().items() if key in coeff]) best_idx = None best_value = 0 for key, value in coeff_dict.items(): if value > best_value and key not in excluded_neurons: best_value = value best_idx = key return best_idx, best_value activation_source = kwargs["activation_source"] if "activation_source" in kwargs else "sigmoid" activation_target = kwargs["activation_target"] if "activation_target" in kwargs else "heaviside" activation_fun_source = Function(activation_source) activation_fun_target = Function(activation_target) for model_id, model_ele in enumerate(model): for layer_id, layer_struct_param in enumerate(model_ele.struct_param): if layer_struct_param[1] == "Symbolic_Layer": layer = getattr(model_ele, "layer_{0}".format(layer_id)) excluded_neurons = [] if activation_source not in layer.get_function_name_list(): continue performance_monitor.reset() criteria_value, criteria_result = get_criteria_value(model, X, y, criteria_type = simplify_criteria[0], criterion = criterion, **kwargs) to_stop, pivot_dict, log, _, pivot_id = performance_monitor.monitor(criteria_value, model_dict = model[model_id].model_dict, criteria_result = criteria_result) for i in range(layer_struct_param[0]): # Obtain loss before simplification: layer = getattr(model_ele, "layer_{0}".format(layer_id)) best_idx, _ = get_sign_snap_candidate(layer, activation_source, excluded_neurons = excluded_neurons) excluded_neurons.append(best_idx) new_expression = [expression.subs(activation_fun_source, activation_fun_target) if j == best_idx else expression for j, expression in enumerate(layer.symbolic_expression)] print("Pass {0}, candidate new expression: {1}".format(i, new_expression)) layer.set_symbolic_expression(new_expression) # Train: _, loss_new, data_record = train_simple(model, X, y, validation_data = validation_data, **kwargs) criteria_value, criteria_result = get_criteria_value(model, X, y, criteria_type = simplify_criteria[0], criterion = criterion, **kwargs) to_stop, pivot_dict, log, is_accept, pivot_id = performance_monitor.monitor(criteria_value, model_dict = model[model_id].model_dict, criteria_result = criteria_result) is_accept_whole.append(is_accept) # Check if the criterion after simplification and refit is worse. If it is worse than the simplify_epsilon, revert: if to_stop: model[model_id].load_model_dict(pivot_dict["model_dict"]) if verbose >= 1: print("Pass {0}, loss: {1}\tDL: {2}. New snap {3} is do not improve by {4} = {5} for {6} steps. Revert the simplification to pivot model. Continue".format( i, view_item(log, ("criteria_result", "loss")), view_item(log, ("criteria_result", "DL")), info, simplify_criteria[0], simplify_epsilon, simplify_patience)) continue mse_record_whole += data_record["mse"] data_DL_whole += data_record["data_DL"] model_DL_whole += data_record["model_DL"] if "param" in record_keys: param_record_whole += data_record["param"] if "param_grad" in record_keys: param_grad_record_whole += data_record["param_grad"] iter_end_whole.append(len(data_record["mse"])) loss_list.append(loss_new) event_list.append({mode_ele: (model_id, layer_id)}) if verbose >= 1: print("{0} succeed at (model {1}, layer {2}). loss: {3}\tDL: {4}".format( mode_ele, model_id, layer_id, view_item(log, ("criteria_result", "loss")), view_item(log, ("criteria_result", "DL")))) print("symbolic_expression: ", layer.symbolic_expression) print("numerical_expression: ", layer.numerical_expression) print() model[model_id].load_model_dict(pivot_dict["model_dict"]) elif mode_ele == "ramping-L1": loss_list_specific = [] ramping_L1_list = kwargs["ramping_L1_list"] if "ramping_L1_list" in kwargs else np.logspace(-7, -1, 30) ramping_mse_threshold = kwargs["ramping_mse_threshold"] if "ramping_mse_threshold" in kwargs else 1e-5 ramping_final_multiplier = kwargs["ramping_final_multiplier"] if "ramping_final_multiplier" in kwargs else 1e-2 layer_dict_dict = {} for i, L1_amp in enumerate(ramping_L1_list): reg_dict = {"weight": L1_amp, "bias": L1_amp, "param": L1_amp} _, loss_end, data_record = train_simple(model, X, y, reg_dict = reg_dict, patience = None, validation_data = validation_data, **kwargs) layer_dict_dict[i] = model[0].layer_0.layer_dict weight, bias = model[0].layer_0.get_weights_bias() print("L1-amp: {0}\tloss: {1}\tweight: {2}\tbias: {3}".format(L1_amp, loss_end, weight, bias)) loss_list_specific.append(loss_end) if "param" in record_keys: param_record_whole.append((weight, bias)) if loss_end > ramping_mse_threshold: if len(loss_list_specific) == 1: print("\nThe MSE after the first L1-amp={0} is already larger than the ramping_mse_threshold. Stop and use current L1-amp. The figures will look empty.".format(ramping_mse_threshold)) else: print("\nThe MSE {0} is larger than the ramping_mse_threshold {1}, stop ramping-L1 simplification".format(loss_end, ramping_mse_threshold)) break mse_record_whole.append(data_record["mse"][-1]) data_DL_whole.append(data_record["data_DL"][-1]) model_DL_whole.append(data_record["model_DL"][-1]) iter_end_whole.append(1) final_L1_amp = L1_amp * ramping_final_multiplier final_L1_idx = np.argmin(np.abs(np.array(ramping_L1_list) - final_L1_amp)) layer_dict_final = layer_dict_dict[final_L1_idx] print("Final L1_amp used: {0}".format(ramping_L1_list[final_L1_idx])) if "param" in record_keys: print("Final param value:\nweights: {0}\nbias{1}".format(param_record_whole[final_L1_idx][0], param_record_whole[final_L1_idx][1])) model[0].layer_0.load_layer_dict(layer_dict_final) mse_record_whole = mse_record_whole[: final_L1_idx + 2] data_DL_whole = data_DL_whole[: final_L1_idx + 2] model_DL_whole = model_DL_whole[: final_L1_idx + 2] iter_end_whole = iter_end_whole[: final_L1_idx + 2] if isplot: def dict_to_list(Dict): return np.array([value for value in Dict.values()]) weights_list = [] bias_list = [] for element in param_record_whole: if isinstance(element[0], dict): element_core = dict_to_list(element[0]) weights_list.append(element_core) else: element_core = to_np_array(element[0]).squeeze(1) weights_list.append(element_core) bias_list.append(to_np_array(element[1])) weights_list = np.array(weights_list) bias_list = np.array(bias_list).squeeze(1) import matplotlib.pylab as plt plt.figure(figsize = (7,5)) plt.loglog(ramping_L1_list[: len(loss_list_specific)], loss_list_specific) plt.xlabel("L1 amp", fontsize = 16) plt.ylabel("mse", fontsize = 16) plt.show() plt.figure(figsize = (7,5)) plt.semilogx(ramping_L1_list[: len(loss_list_specific)], loss_list_specific) plt.xlabel("L1 amp", fontsize = 16) plt.ylabel("mse", fontsize = 16) plt.show() plt.figure(figsize = (7,5)) for i in range(weights_list.shape[1]): plt.semilogx(ramping_L1_list[: len(loss_list_specific)], weights_list[:,i], label = "weight_{0}".format(i)) if len(bias_list) > 0: plt.semilogx(ramping_L1_list[: len(loss_list_specific)], bias_list, label = "bias") plt.xlabel("L1 amp", fontsize = 16) plt.ylabel("parameter_values", fontsize = 16) plt.legend() plt.show() plt.clf() plt.close() else: raise Exception("mode {0} not recognized!".format(mode_ele)) loss_dict[mode_ele] = {} if X is not None: loss_dict[mode_ele]["mse_record_whole"] = mse_record_whole loss_dict[mode_ele]["data_DL_whole"] = data_DL_whole loss_dict[mode_ele]["{0}_test".format(loss_type)] = loss_list loss_dict[mode_ele]["model_DL_whole"] = model_DL_whole if "param" in record_keys: loss_dict[mode_ele]["param_record_whole"] = param_record_whole if "param_grad" in record_keys: loss_dict[mode_ele]["param_grad_record_whole"] = param_grad_record_whole loss_dict[mode_ele]["iter_end_whole"] = iter_end_whole loss_dict[mode_ele]["event_list"] = event_list loss_dict[mode_ele]["is_accept_whole"] = is_accept_whole if mode_ele == "ramping-L1": loss_dict[mode_ele]["ramping_L1_list"] = ramping_L1_list loss_dict[mode_ele]["loss_list_specific"] = loss_list_specific if not is_list: model = model[0] return model, loss_dict # ## The following are different model architectures: # ## MLP: # In[3]: class MLP(nn.Module): def __init__( self, input_size, struct_param = None, W_init_list = None, # initialization for weights b_init_list = None, # initialization for bias settings = {}, # Default settings for each layer, if the settings for the layer is not provided in struct_param is_cuda = False, ): super(MLP, self).__init__() self.input_size = input_size self.is_cuda = is_cuda self.settings = deepcopy(settings) if struct_param is not None: self.num_layers = len(struct_param) self.W_init_list = W_init_list self.b_init_list = b_init_list self.info_dict = {} self.init_layers(deepcopy(struct_param)) else: self.num_layers = 0 @property def struct_param(self): return [getattr(self, "layer_{0}".format(i)).struct_param for i in range(self.num_layers)] @property def output_size(self): return self.get_layer(-1).output_size @property def structure(self): structure = OrderedDict() structure["input_size"] = self.input_size structure["output_size"] = self.output_size structure["struct_param"] = self.struct_param if hasattr(self, "struct_param") else None return structure def init_layers(self, struct_param): res_forward = self.settings["res_forward"] if "res_forward" in self.settings else False for k, layer_struct_param in enumerate(struct_param): if res_forward: num_neurons_prev = struct_param[k - 1][0] + self.input_size if k > 0 else self.input_size else: num_neurons_prev = struct_param[k - 1][0] if k > 0 else self.input_size num_neurons = layer_struct_param[0] W_init = self.W_init_list[k] if self.W_init_list is not None else None b_init = self.b_init_list[k] if self.b_init_list is not None else None # Get settings for the current layer: layer_settings = deepcopy(self.settings) if bool(self.settings) else {} layer_settings.update(layer_struct_param[2]) # Construct layer: layer = get_Layer(layer_type = layer_struct_param[1], input_size = num_neurons_prev, output_size = num_neurons, W_init = W_init, b_init = b_init, settings = layer_settings, is_cuda = self.is_cuda, ) setattr(self, "layer_{}".format(k), layer) def forward(self, *input, p_dict=None, **kwargs): kwargs = filter_kwargs(kwargs, ["res_forward", "is_res_block", "act_noise_scale"]) # only allow certain kwargs to be passed if isinstance(input, tuple): input = torch.cat(input, -1) output = input res_forward = self.settings["res_forward"] if "res_forward" in self.settings else False is_res_block = self.settings["is_res_block"] if "is_res_block" in self.settings else False for k in range(len(self.struct_param)): p_dict_ele = p_dict[k] if p_dict is not None else None if res_forward and k > 0: output = getattr(self, "layer_{}".format(k))(torch.cat([output, input], -1), p_dict=p_dict_ele, **kwargs) else: output = getattr(self, "layer_{}".format(k))(output, p_dict=p_dict_ele, **kwargs) if is_res_block: output = output + input return output def copy(self): return deepcopy(self) def simplify(self, X=None, y=None, mode="full", isplot=False, target_name=None, validation_data = None, **kwargs): new_model, _ = simplify(self, X, y, mode=mode, isplot=isplot, target_name=target_name, validation_data=validation_data, **kwargs) self.__dict__.update(new_model.__dict__) def snap(self, snap_mode="integer", top=5, **kwargs): """Generate a set of new models whose parameters are snapped, each model with a different number of snapped parameters.""" if not hasattr(self, "num_layers") or self.num_layers != 1: return False, [self] else: model_list = [] top = top if snap_mode != "unsnap" else 1 for top_ele in range(1, top + 1): new_model = self.copy() layer = new_model.layer_0 info_list = layer.simplify(mode="snap", top=top_ele, snap_mode=snap_mode) if len(info_list) > 0: new_model.reset_layer(0, layer) model_list.append(new_model) is_succeed = len(model_list) > 0 return is_succeed, model_list def get_regularization(self, source = ["weight", "bias"], mode = "L1", **kwargs): reg = to_Variable([0], is_cuda=self.is_cuda) for k in range(len(self.struct_param)): layer = getattr(self, "layer_{}".format(k)) reg = reg + layer.get_regularization(mode = mode, source = source) return reg def get_layer(self, layer_id): if layer_id < 0: layer_id += self.num_layers return getattr(self, "layer_{}".format(layer_id)) def reset_layer(self, layer_id, layer): setattr(self, "layer_{}".format(layer_id), layer) def insert_layer(self, layer_id, layer): if layer_id < 0: layer_id += self.num_layers if layer_id < self.num_layers - 1: next_layer = getattr(self, "layer_{}".format(layer_id + 1)) if next_layer.struct_param[1] == "Simple_Layer": assert next_layer.input_size == layer.output_size, "The inserted layer's output_size {0} must be compatible with next layer_{1}'s input_size {2}!" .format(layer.output_size, layer_id + 1, next_layer.input_size) for i in range(self.num_layers - 1, layer_id - 1, -1): setattr(self, "layer_{}".format(i + 1), getattr(self, "layer_{}".format(i))) setattr(self, "layer_{}".format(layer_id), layer) self.num_layers += 1 def remove_layer(self, layer_id): if layer_id < 0: layer_id += self.num_layers if layer_id < self.num_layers - 1: num_neurons_prev = self.struct_param[layer_id - 1][0] if layer_id > 0 else self.input_size replaced_layer = getattr(self, "layer_{}".format(layer_id + 1)) if replaced_layer.struct_param[1] == "Simple_Layer": assert replaced_layer.input_size == num_neurons_prev, "After deleting layer_{0}, the replaced layer's input_size {1} must be compatible with previous layer's output neurons {2}!" .format(layer_id, replaced_layer.input_size, num_neurons_prev) for i in range(layer_id, self.num_layers - 1): setattr(self, "layer_{}".format(i), getattr(self, "layer_{}".format(i + 1))) self.num_layers -= 1 def prune_neurons(self, layer_id, neuron_ids): if layer_id == "input": layer = self.get_layer(0) layer.prune_input_neurons(neuron_ids) self.input_size = layer.input_size else: if layer_id < 0: layer_id = self.num_layers + layer_id layer = getattr(self, "layer_{}".format(layer_id)) layer.prune_output_neurons(neuron_ids) self.reset_layer(layer_id, layer) if layer_id < self.num_layers - 1: next_layer = getattr(self, "layer_{}".format(layer_id + 1)) next_layer.prune_input_neurons(neuron_ids) self.reset_layer(layer_id + 1, next_layer) def add_neurons(self, layer_id, num_neurons, mode = ("imitation", "zeros")): if not isinstance(mode, list) and not isinstance(mode, tuple): mode = (mode, mode) if layer_id < 0: layer_id = self.num_layers + layer_id layer = getattr(self, "layer_{}".format(layer_id)) layer.add_output_neurons(num_neurons, mode = mode[0]) self.reset_layer(layer_id, layer) if layer_id < self.num_layers - 1: next_layer = getattr(self, "layer_{}".format(layer_id + 1)) next_layer.add_input_neurons(num_neurons, mode = mode[1]) self.reset_layer(layer_id + 1, next_layer) if layer_id == 0: self.input_size = self.get_layer(0).input_size def inspect_operation(self, input, operation_between, p_dict = None, **kwargs): output = input res_forward = self.settings["res_forward"] if "res_forward" in self.settings else False is_res_block = self.settings["is_res_block"] if "is_res_block" in self.settings else False for k in range(*operation_between): p_dict_ele = p_dict[k] if p_dict is not None else None if res_forward and k > 0: output = getattr(self, "layer_{}".format(k))(torch.cat([output, input], -1), p_dict = p_dict_ele) else: output = getattr(self, "layer_{}".format(k))(output, p_dict = p_dict_ele) if is_res_block: output = output + input return output def get_weights_bias(self, W_source = "core", b_source = "core", layer_ids = None, is_grad = False, isplot = False, verbose = False, raise_error = True): if not hasattr(self, "struct_param"): return None, None layer_ids = range(len(self.struct_param)) if layer_ids is None else layer_ids W_list = [] b_list = [] if W_source is not None: for k in range(len(self.struct_param)): if k in layer_ids: if W_source == "core": try: W, _ = getattr(self, "layer_{}".format(k)).get_weights_bias(is_grad = is_grad) except Exception as e: if raise_error: raise else: print(e) W = np.array([np.NaN]) else: raise Exception("W_source '{}' not recognized!".format(W_source)) W_list.append(W) if b_source is not None: for k in range(len(self.struct_param)): if k in layer_ids: if b_source == "core": try: _, b = getattr(self, "layer_{}".format(k)).get_weights_bias(is_grad = is_grad) except Exception as e: if raise_error: raise else: print(e) b = np.array([np.NaN]) else: raise Exception("b_source '{}' not recognized!".format(b_source)) b_list.append(b) if verbose: import pprint as pp if W_source is not None: print("weight:") pp.pprint(W_list) if b_source is not None: print("bias:") pp.pprint(b_list) if isplot: if W_source is not None: print("weight {}:".format(W_source)) plot_matrices(W_list) if b_source is not None: print("bias {}:".format(b_source)) plot_matrices(b_list) return W_list, b_list def split_to_model_ensemble(self, mode = "standardize"): num_models = self.struct_param[-1][0] model_core = deepcopy(self) if mode == "standardize": last_layer = getattr(model_core, "layer_{}".format(model_core.num_layers - 1)) last_layer.standardize(mode = "b_mean_zero") else: raise Exception("mode {} not recognized!".format(mode)) model_list = [deepcopy(model_core) for i in range(num_models)] for i, model in enumerate(model_list): to_prune = list(range(num_models)) to_prune.pop(i) model.prune_neurons(-1, to_prune) return construct_model_ensemble_from_nets(model_list) @property def model_dict(self): model_dict = {"type": self.__class__.__name__} model_dict["input_size"] = self.input_size model_dict["struct_param"] = get_full_struct_param(self.struct_param, self.settings) model_dict["weights"], model_dict["bias"] = self.get_weights_bias(W_source = "core", b_source = "core") model_dict["settings"] = deepcopy(self.settings) model_dict["net_type"] = self.__class__.__name__ return model_dict @property def DL(self): return np.sum([getattr(self, "layer_{}".format(i)).DL for i in range(self.num_layers)]) def load_model_dict(self, model_dict): new_net = load_model_dict_net(model_dict, is_cuda = self.is_cuda) self.__dict__.update(new_net.__dict__) def load(self, filename): mode = "json" if filename.endswith(".json") else "pickle" model_dict = load_model(filename, mode=mode) self.load_model_dict(model_dict) def save(self, filename): mode = "json" if filename.endswith(".json") else "pickle" save_model(self.model_dict, filename, mode=mode) def get_loss(self, input, target, criterion, **kwargs): y_pred = self(input, **kwargs) return criterion(y_pred, target) def prepare_inspection(self, X, y, **kwargs): return {} def set_cuda(self, is_cuda): for k in range(self.num_layers): getattr(self, "layer_{}".format(k)).set_cuda(is_cuda) self.is_cuda = is_cuda def set_trainable(self, is_trainable): for k in range(self.num_layers): getattr(self, "layer_{}".format(k)).set_trainable(is_trainable) def get_snap_dict(self): snap_dict = {} for k in range(len(self.struct_param)): layer = getattr(self, "layer_{}".format(k)) if hasattr(layer, "snap_dict"): recorded_layer_snap_dict = {} for key, item in layer.snap_dict.items(): recorded_layer_snap_dict[key] = {"new_value": item["new_value"]} if len(recorded_layer_snap_dict) > 0: snap_dict[k] = recorded_layer_snap_dict return snap_dict def synchronize_settings(self): snap_dict = self.get_snap_dict() if len(snap_dict) > 0: self.settings["snap_dict"] = snap_dict return self.settings def get_sympy_expression(self, verbose = True): expressions = {i: {} for i in range(self.num_layers)} for i in range(self.num_layers): layer = getattr(self, "layer_{}".format(i)) if layer.struct_param[1] == "Symbolic_Layer": if verbose: print("Layer {}, symbolic_expression: {}".format(i, layer.symbolic_expression)) print(" numerical_expression: {}".format(layer.numerical_expression)) expressions[i]["symbolic_expression"] = layer.symbolic_expression expressions[i]["numerical_expression"] = layer.numerical_expression expressions[i]["param_dict"] = layer.get_param_dict() expressions[i]["DL"] = layer.DL else: if verbose: print("Layer {} is not a symbolic layer.".format(i)) expressions[i] = None return expressions # ## Labelmix_MLP: # In[ ]: class Labelmix_MLP(nn.Module): def __init__( self, input_size, struct_param, idx_label=None, is_cuda=False, ): super(Labelmix_MLP, self).__init__() self.input_size = input_size self.struct_param = struct_param self.num_layers = len(struct_param) if idx_label is not None and len(idx_label) == input_size: idx_label = None if idx_label is not None: self.idx_label = torch.LongTensor(idx_label) idx_main = list(set(range(input_size)) - set(to_np_array(idx_label).astype(int).tolist())) self.idx_main = torch.LongTensor(idx_main) else: self.idx_label = None self.idx_main = torch.LongTensor(list(range(input_size))) num_neurons_prev = len(self.idx_main) for i, layer_struct_param in enumerate(struct_param): num_neurons = layer_struct_param[0] setattr(self, "W_{}_main".format(i), nn.Parameter(torch.randn(num_neurons_prev, num_neurons))) setattr(self, "b_{}_main".format(i), nn.Parameter(torch.zeros(num_neurons))) init_weight(getattr(self, "W_{}_main".format(i)), init=None) num_neurons_prev = num_neurons if self.idx_label is not None: setattr(self, "W_{}_mul".format(i), nn.Parameter(torch.randn(len(self.idx_label), num_neurons))) setattr(self, "W_{}_add".format(i), nn.Parameter(torch.randn(len(self.idx_label), num_neurons))) init_weight(getattr(self, "W_{}_mul".format(i)), init=None) init_weight(getattr(self, "W_{}_add".format(i)), init=None) setattr(self, "b_{}_mul".format(i), nn.Parameter(torch.zeros(num_neurons))) setattr(self, "b_{}_add".format(i), nn.Parameter(torch.zeros(num_neurons))) self.set_cuda(is_cuda) def forward(self, input): output = input[:, self.idx_main] if self.idx_label is not None: labels = input[:, self.idx_label] for i, layer_struct_param in enumerate(self.struct_param): output = torch.matmul(output, getattr(self, "W_{}_main".format(i))) + getattr(self, "b_{}_main".format(i)) if "activation" in layer_struct_param[2]: output = get_activation(layer_struct_param[2]["activation"])(output) if self.idx_label is not None: A_mul = torch.matmul(labels, getattr(self, "W_{}_mul".format(i))) + getattr(self, "b_{}_mul".format(i)) A_add = torch.matmul(labels, getattr(self, "W_{}_add".format(i))) + getattr(self, "b_{}_add".format(i)) output = output * A_mul + A_add return output def get_loss(self, X, y, criterion, **kwargs): y_pred = self(X) return criterion(y_pred, y) def set_cuda(self, is_cuda): if isinstance(is_cuda, str): self.cuda(is_cuda) else: if is_cuda: self.cuda() else: self.cpu() self.is_cuda = is_cuda def get_regularization(self, source = ["weight", "bias"], mode = "L1", **kwargs): reg = to_Variable([0], is_cuda=self.is_cuda) return reg @property def model_dict(self): model_dict = {"type": "Labelmix_MLP"} model_dict["input_size"] = self.input_size model_dict["struct_param"] = self.struct_param if self.idx_label is not None: model_dict["idx_label"] = to_np_array(self.idx_label).astype(int) model_dict["state_dict"] = to_cpu_recur(self.state_dict()) return model_dict # ## Multi_MLP (MLPs in series): # In[ ]: class Multi_MLP(nn.Module): def __init__( self, input_size, struct_param, W_init_list = None, # initialization for weights b_init_list = None, # initialization for bias settings = None, # Default settings for each layer, if the settings for the layer is not provided in struct_param is_cuda = False, ): super(Multi_MLP, self).__init__() self.input_size = input_size self.num_layers = len(struct_param) self.W_init_list = W_init_list self.b_init_list = b_init_list self.settings = deepcopy(settings) self.num_blocks = len(struct_param) self.is_cuda = is_cuda for i, struct_param_ele in enumerate(struct_param): input_size_block = input_size if i == 0 else struct_param[i - 1][-1][0] setattr(self, "block_{0}".format(i), MLP(input_size = input_size_block, struct_param = struct_param_ele, W_init_list = W_init_list[i] if W_init_list is not None else None, b_init_list = b_init_list[i] if b_init_list is not None else None, settings = self.settings[i] if self.settings is not None else {}, is_cuda = self.is_cuda, )) def forward(self, input): output = input for i in range(self.num_blocks): output = getattr(self, "block_{0}".format(i))(output) return output def get_loss(self, input, target, criterion, **kwargs): y_pred = self(input, **kwargs) return criterion(y_pred, target) def get_regularization(self, source = ["weight", "bias"], mode = "L1", **kwargs): reg = Variable(torch.FloatTensor([0]), requires_grad = False) if self.is_cuda: reg = reg.cuda() for i in range(self.num_blocks): reg = reg + getattr(self, "block_{0}".format(i)).get_regularization(mode = mode, source = source) return reg @property def struct_param(self): return [getattr(self, "block_{0}".format(i)).struct_param for i in range(self.num_blocks)] @property def model_dict(self): model_dict = {"type": self.__class__.__name__} model_dict["input_size"] = self.input_size model_dict["struct_param"] = self.struct_param model_dict["weights"], model_dict["bias"] = self.get_weights_bias(W_source = "core", b_source = "core") model_dict["settings"] = deepcopy(self.settings) model_dict["net_type"] = self.__class__.__name__ return model_dict def load_model_dict(self, model_dict): new_net = load_model_dict_Multi_MLP(model_dict, is_cuda = self.is_cuda) self.__dict__.update(new_net.__dict__) def load(self, filename): mode = "json" if filename.endswith(".json") else "pickle" model_dict = load_model(filename, mode=mode) self.load_model_dict(model_dict) def save(self, filename): mode = "json" if filename.endswith(".json") else "pickle" save_model(self.model_dict, filename, mode=mode) def get_weights_bias(self, W_source = "core", b_source = "core"): W_list = [] b_list = [] for i in range(self.num_blocks): W, b = getattr(self, "block_{0}".format(i)).get_weights_bias(W_source = W_source, b_source = b_source) W_list.append(W) b_list.append(b) return deepcopy(W_list), deepcopy(b_list) def prepare_inspection(self, X, y, **kwargs): return {} def set_cuda(self, is_cuda): for i in range(self.num_blocks): getattr(self, "block_{0}".format(i)).set_cuda(is_cuda) self.is_cuda = is_cuda def set_trainable(self, is_trainable): for i in range(self.num_blocks): getattr(self, "block_{0}".format(i)).set_trainable(is_trainable) # ## Branching_Net: # In[ ]: class Branching_Net(nn.Module): """An MLP that consists of a base network, and net_1 and net_2 that branches off from the output of the base network.""" def __init__( self, net_base_model_dict, net_1_model_dict, net_2_model_dict, is_cuda = False, ): super(Branching_Net, self).__init__() self.net_base = load_model_dict(net_base_model_dict, is_cuda = is_cuda) self.net_1 = load_model_dict(net_1_model_dict, is_cuda = is_cuda) self.net_2 = load_model_dict(net_2_model_dict, is_cuda = is_cuda) self.info_dict = {} def forward(self, X, **kwargs): shared = self.net_base(X) shared = shared.max(0, keepdim = True)[0] return self.net_1(shared)[0], self.net_2(shared)[0] def get_regularization(self, source = ["weights", "bias"], mode = "L1"): reg = self.net_base.get_regularization(source = source, mode = mode) + self.net_1.get_regularization(source = source, mode = mode) + self.net_2.get_regularization(source = source, mode = mode) return reg def set_trainable(self, is_trainable): self.net_base.set_trainable(is_trainable) self.net_1.set_trainable(is_trainable) self.net_2.set_trainable(is_trainable) def prepare_inspection(self, X, y, **kwargs): return deepcopy(self.info_dict) @property def model_dict(self): model_dict = {"type": "Branching_Net"} model_dict["net_base_model_dict"] = self.net_base.model_dict model_dict["net_1_model_dict"] = self.net_1.model_dict model_dict["net_2_model_dict"] = self.net_2.model_dict return model_dict def load(self, filename): mode = "json" if filename.endswith(".json") else "pickle" model_dict = load_model(filename, mode=mode) self.load_model_dict(model_dict) def save(self, filename): mode = "json" if filename.endswith(".json") else "pickle" save_model(self.model_dict, filename, mode=mode) class Fan_in_MLP(nn.Module): def __init__( self, model_dict_branch1, model_dict_branch2, model_dict_joint, is_cuda=False, ): super(Fan_in_MLP, self).__init__() if model_dict_branch1 is not None: self.net_branch1 = load_model_dict(model_dict_branch1, is_cuda=is_cuda) else: self.net_branch1 = None if model_dict_branch2 is not None: self.net_branch2 = load_model_dict(model_dict_branch2, is_cuda=is_cuda) else: self.net_branch2 = None self.net_joint = load_model_dict(model_dict_joint, is_cuda=is_cuda) self.is_cuda = is_cuda self.info_dict = {} def forward(self, X1, X2, is_outer=False): if is_outer: X2 = X2[...,None,:] if self.net_branch1 is not None: X1 = self.net_branch1(X1) if self.net_branch2 is not None: X2 = self.net_branch2(X2) X1, X2 = broadcast_all(X1, X2) out = torch.cat([X1, X2], -1) # if is_outer=True, then output dimension: [..., X2dim, X1dim, out_dim]: return self.net_joint(out).squeeze(-1) def get_loss(self, input, target, criterion, **kwargs): X1, X2 = input y_pred = self(X1, X2) return criterion(y_pred, target) def get_regularization(self, source = ["weight", "bias"], mode = "L1", **kwargs): reg = Variable(torch.FloatTensor([0]), requires_grad = False) if self.is_cuda: reg = reg.cuda() if self.net_branch1 is not None: reg = reg + self.net_branch1.get_regularization(source=source, mode=mode) if self.net_branch2 is not None: reg = reg + self.net_branch2.get_regularization(source=source, mode=mode) return reg def prepare_inspection(self, X, y, **kwargs): return deepcopy(self.info_dict) @property def model_dict(self): model_dict = {'type': self.__class__.__name__} model_dict["model_dict_branch1"] = self.net_branch1.model_dict if self.net_branch1 is not None else None model_dict["model_dict_branch2"] = self.net_branch2.model_dict if self.net_branch2 is not None else None model_dict["model_dict_joint"] = self.net_joint.model_dict return model_dict def load(self, filename): mode = "json" if filename.endswith(".json") else "pickle" model_dict = load_model(filename, mode=mode) self.load_model_dict(model_dict) def save(self, filename): mode = "json" if filename.endswith(".json") else "pickle" save_model(self.model_dict, filename, mode=mode) # ## Mixture_Model: # In[ ]: class Mixture_Model(nn.Module): def __init__( self, model_dict_list, weight_logits_model_dict, num_components, is_cuda=False, ): super(Mixture_Model, self).__init__() self.num_components = num_components for i in range(self.num_components): if isinstance(model_dict_list, list): setattr(self, "model_{}".format(i), load_model_dict(model_dict_list[i], is_cuda=is_cuda)) else: assert isinstance(model_dict_list, dict) setattr(self, "model_{}".format(i), load_model_dict(model_dict_list, is_cuda=is_cuda)) self.weight_logits_model = load_model_dict(weight_logits_model_dict, is_cuda=is_cuda) self.is_cuda = is_cuda def forward(self, input): output_list = [] for i in range(self.num_components): output = getattr(self, "model_{}".format(i))(input) output_list.append(output) output_list = torch.stack(output_list, -1) weight_logits = self.weight_logits_model(input) return output_list, weight_logits @property def model_dict(self): model_dict = {"type": "Mixture_Model", "model_dict_list": [getattr(self, "model_{}".format(i)).model_dict for i in range(self.num_components)], "weight_logits_model_dict": self.weight_logits_model.model_dict, "num_components": self.num_components, } return model_dict # ## Model_Ensemble: # In[ ]: class Model_Ensemble(nn.Module): """Model_Ensemble is a collection of models with the same architecture but independent parameters""" def __init__( self, num_models, input_size, struct_param, W_init_list = None, b_init_list = None, settings = None, net_type = "MLP", is_cuda = False, ): super(Model_Ensemble, self).__init__() self.num_models = num_models self.input_size = input_size self.net_type = net_type self.is_cuda = is_cuda for i in range(self.num_models): if settings is None: settings_model = {} elif isinstance(settings, list) or isinstance(settings, tuple): settings_model = settings[i] else: settings_model = settings if isinstance(struct_param, tuple): struct_param_model = struct_param[i] else: struct_param_model = struct_param if net_type == "MLP": net = MLP(input_size = self.input_size, struct_param = deepcopy(struct_param_model), W_init_list = deepcopy(W_init_list[i]) if W_init_list is not None else None, b_init_list = deepcopy(b_init_list[i]) if b_init_list is not None else None, settings = deepcopy(settings_model), is_cuda = is_cuda, ) elif net_type == "ConvNet": net = ConvNet(input_channels = self.input_size, struct_param = deepcopy(struct_param_model), settings = deepcopy(settings_model), is_cuda = is_cuda, ) else: raise Exception("Net_type {0} not recognized!".format(net_type)) setattr(self, "model_{0}".format(i), net) @property def struct_param(self): return tuple(getattr(self, "model_{0}".format(i)).struct_param for i in range(self.num_models)) @property def settings(self): return [getattr(self, "model_{0}".format(i)).settings for i in range(self.num_models)] def get_all_models(self): return [getattr(self, "model_{0}".format(i)) for i in range(self.num_models)] def init_bias_with_input(self, input, mode = "std_sqrt", neglect_last_layer = True): for i in range(self.num_models): model = getattr(self, "model_{0}".format(i)) model.init_bias_with_input(input, mode = mode, neglect_last_layer = neglect_last_layer) def initialize_param_freeze(self, update_values = True): for i in range(self.num_models): model = getattr(self, "model_{0}".format(i)) model.initialize_param_freeze(update_values = update_values) def apply_model(self, input, model_id): return fetch_model(self, model_id)(input) def fetch_model(self, model_id): return getattr(self, "model_{0}".format(model_id)) def set_trainable(self, is_trainable): for i in range(self.num_models): getattr(self, "model_{0}".format(i)).set_trainable(is_trainable) def forward(self, input): output_list = [] for i in range(self.num_models): if self.net_type == "MLP": output = getattr(self, "model_{0}".format(i))(input) elif self.net_type == "ConvNet": output = getattr(self, "model_{0}".format(i))(input)[0] else: raise Exception("Net_type {0} not recognized!".format(self.net_type)) output_list.append(output) return torch.stack(output_list, 1) def get_regularization(self, source = ["weight", "bias"], mode = "L1", **kwargs): if not isinstance(source, list): source = [source] reg = Variable(torch.FloatTensor([0]), requires_grad = False) if self.is_cuda: reg = reg.cuda() model0 = self.model_0 # Elastic_weight_reg: if "elastic_weight" in source or "elastic_bias" in source: # Setting up excluded layer: excluded_layer = kwargs["excluded_layer"] if "excluded_layer" in kwargs else [-1] if not isinstance(excluded_layer, list): excluded_layer = [excluded_layer] excluded_layer = [element + model0.num_layers if element < 0 else element for element in excluded_layer] elastic_mode = kwargs["elastic_mode"] if "elastic_mode" in kwargs else "var" # Compute the elastic_weight_reg: for k in range(model0.num_layers): if k in excluded_layer: continue W_accum_k = [] b_accum_k = [] num_neurons_prev = model0.struct_param[k - 1][0] if k > 0 else self.input_size num_neurons = model0.struct_param[k][0] for i in range(self.num_models): model = getattr(self, "model_{0}".format(i)) assert model0.num_layers == model.num_layers assert num_neurons_prev == model.struct_param[k - 1][0] if k > 0 else model.input_size, "all models' input/output size at each layer must be identical!" assert num_neurons == model.struct_param[k][0], "all models' input/output size at each layer must be identical!" layer_k = getattr(model, "layer_{0}".format(k)) if "elastic_weight" in source: W_accum_k.append(layer_k.W_core) if "elastic_bias" in source: b_accum_k.append(layer_k.b_core) if "elastic_weight" in source: if elastic_mode == "var": reg = reg + torch.stack(W_accum_k, -1).var(-1).sum() elif elastic_mode == "std": reg = reg + torch.stack(W_accum_k, -1).std(-1).sum() else: raise if "elastic_bias" in source: if elastic_mode == "var": reg = reg + torch.stack(b_accum_k, -1).var(-1).sum() elif elastic_mode == "std": reg = reg + torch.stack(b_accum_k, -1).std(-1).sum() else: raise source_core = deepcopy(source) if "elastic_weight" in source_core: source_core.remove("elastic_weight") if "elastic_bias" in source_core: source_core.remove("elastic_bias") else: source_core = source # Other regularizations: for k in range(self.num_models): reg = reg + getattr(self, "model_{0}".format(k)).get_regularization(source = source_core, mode = mode, **kwargs) return reg def get_weights_bias(self, W_source = None, b_source = None, verbose = False, isplot = False): W_list_dict = {} b_list_dict = {} for i in range(self.num_models): if verbose: print("\nmodel {0}:".format(i)) W_list_dict[i], b_list_dict[i] = getattr(self, "model_{0}".format(i)).get_weights_bias( W_source = W_source, b_source = b_source, verbose = verbose, isplot = isplot) return W_list_dict, b_list_dict def combine_to_net(self, mode = "mean", last_layer_mode = "concatenate"): model0 = self.model_0 if mode == "mean": struct_param = deepcopy(model0.struct_param) settings = deepcopy(model0.settings) W_init_list = [] b_init_list = [] for k in range(model0.num_layers): num_neurons_prev = model0.struct_param[k - 1][0] if k > 0 else self.input_size num_neurons = model0.struct_param[k][0] W_accum_k = [] b_accum_k = [] for i in range(self.num_models): model = getattr(self, "model_{0}".format(i)) assert model0.num_layers == model.num_layers assert num_neurons_prev == model.struct_param[k - 1][0] if k > 0 else model.input_size, "If mode == 'mean', all models' input/output size at each layer must be identical!" assert num_neurons == model.struct_param[k][0], "If mode == 'mean', all models' input/output size at each layer must be identical!" layer_k = getattr(model, "layer_{0}".format(k)) W_accum_k.append(layer_k.W_core) b_accum_k.append(layer_k.b_core) if k == model0.num_layers - 1: current_mode = last_layer_mode else: current_mode = mode if current_mode == "mean": W_accum_k = torch.stack(W_accum_k, -1).mean(-1) b_accum_k = torch.stack(b_accum_k, -1).mean(-1) elif current_mode == "concatenate": W_accum_k = torch.cat(W_accum_k, -1) b_accum_k = torch.cat(b_accum_k, -1) struct_param[-1][0] = sum([self.struct_param[i][-1][0] for i in range(self.num_models)]) else: raise Exception("mode {0} not recognized!".format(last_layer_mode)) W_init_list.append(W_accum_k.data.numpy()) b_init_list.append(b_accum_k.data.numpy()) # Build the net: net = MLP(input_size = self.input_size, struct_param = struct_param, W_init_list = W_init_list, b_init_list = b_init_list, settings = settings, ) else: raise Exception("mode {0} not recognized!".format(mode)) return net def remove_models(self, model_ids): if not isinstance(model_ids, list): model_ids = [model_ids] model_list = [] k = 0 for i in range(self.num_models): if i not in model_ids: if k != i: setattr(self, "model_{0}".format(k), getattr(self, "model_{0}".format(i))) k += 1 num_models_new = k for i in range(num_models_new, self.num_models): delattr(self, "model_{0}".format(i)) self.num_models = num_models_new def add_models(self, models): if not isinstance(models, list): models = [models] for i, model in enumerate(models): setattr(self, "model_{0}".format(i + self.num_models), model) self.num_models += len(models) def simplify(self, X, y, idx, mode = "full", validation_data = None, isplot = False, **kwargs): def process_idx(idx): idx = idx.byte() if len(idx.size()) == 1: idx = idx.unqueeze(1) if idx.size(1) == 1: idx = idx.repeat(1, self.num_models) return idx idx = process_idx(idx) if validation_data is not None: X_valid, y_valid, idx_valid = validation_data idx_valid = process_idx(idx_valid) loss_dict = {} for i in range(self.num_models): model = getattr(self, "model_{0}".format(i)) X_chosen = torch.masked_select(X, idx[:, i:i+1]).view(-1, X.size(1)) y_chosen = torch.masked_select(y, idx[:, i:i+1]).view(-1, y.size(1)) if validation_data is not None: X_valid_chosen = torch.masked_select(X_valid, idx_valid[:, i:i+1]).view(-1, X_valid.size(1)) y_valid_chosen = torch.masked_select(y_valid, idx_valid[:, i:i+1]).view(-1, y_valid.size(1)) if len(X_valid_chosen) == 0: validation_data_chosen = None else: validation_data_chosen = (X_valid_chosen, y_valid_chosen) else: validation_data_chosen = None if len(X_chosen) == 0: print("The {0}'th model has no corresponding data to simplify with, skip.".format(i)) else: new_model, loss_dict["model_{0}".format(i)] = simplify(model, X_chosen, y_chosen, mode = mode, validation_data = validation_data_chosen, isplot = isplot, target_name = "model_{0}".format(i), **kwargs) setattr(self, "model_{0}".format(i), new_model) return loss_dict def get_sympy_expression(self): expressions = {} for k in range(self.num_models): print("\nmodel {0}:".format(k)) expressions["model_{0}".format(k)] = getattr(self, "model_{0}".format(k)).get_sympy_expression() return expressions @property def DL(self): return np.sum([getattr(self, "model_{0}".format(i)).DL for i in range(self.num_models)]) def get_weights_bias(self, W_source = None, b_source = None, verbose = False, isplot = False): W_list_dict = {} b_list_dict = {} for i in range(self.num_models): if verbose: print("\nmodel {0}:".format(i)) W_list_dict[i], b_list_dict[i] = getattr(self, "model_{0}".format(i)).get_weights_bias(W_source = W_source, b_source = b_source, verbose = verbose, isplot = isplot) return W_list_dict, b_list_dict @property def model_dict(self): model_dict = {"type": "Model_Ensemble"} for i in range(self.num_models): model_dict["model_{0}".format(i)] = getattr(self, "model_{0}".format(i)).model_dict model_dict["input_size"] = self.input_size model_dict["struct_param"] = self.struct_param model_dict["num_models"] = self.num_models model_dict["net_type"] = self.net_type return model_dict def load_model_dict(self, model_dict): new_model_ensemble = load_model_dict(model_dict, is_cuda = self.is_cuda) self.__dict__.update(new_model_ensemble.__dict__) def load(self, filename): mode = "json" if filename.endswith(".json") else "pickle" model_dict = load_model(filename, mode=mode) self.load_model_dict(model_dict) def save(self, filename): mode = "json" if filename.endswith(".json") else "pickle" save_model(self.model_dict, filename, mode=mode) def load_model_dict_model_ensemble(model_dict, is_cuda = False): num_models = len([model_name for model_name in model_dict if model_name[:6] == "model_"]) return Model_Ensemble(num_models = num_models, input_size = model_dict["input_size"], struct_param = tuple([deepcopy(model_dict["model_{0}".format(i)]["struct_param"]) for i in range(num_models)]), W_init_list = [deepcopy(model_dict["model_{0}".format(i)]["weights"]) for i in range(num_models)], b_init_list = [deepcopy(model_dict["model_{0}".format(i)]["bias"]) for i in range(num_models)], settings = [deepcopy(model_dict["model_{0}".format(i)]["settings"]) for i in range(num_models)], net_type = model_dict["net_type"] if "net_type" in model_dict else "MLP", is_cuda = is_cuda, ) def combine_model_ensembles(model_ensembles, input_size): model_ensembles = deepcopy(model_ensembles) model_ensemble_combined = None model_id = 0 for k, model_ensemble in enumerate(model_ensembles): if model_ensemble.input_size == input_size: if model_ensemble_combined is None: model_ensemble_combined = model_ensemble else: continue for i in range(model_ensemble.num_models): model = getattr(model_ensemble, "model_{0}".format(i)) setattr(model_ensemble_combined, "model_{0}".format(model_id), model) model_id += 1 model_ensemble_combined.num_models = model_id return model_ensemble_combined def construct_model_ensemble_from_nets(nets): num_models = len(nets) if num_models is None: return None input_size = nets[0].input_size struct_param = tuple(net.struct_param for net in nets) is_cuda = False for net in nets: if net.input_size != input_size: raise Exception("The input_size for all nets must be the same!") if net.is_cuda: is_cuda = True model_ensemble = Model_Ensemble(num_models = num_models, input_size = input_size, struct_param = struct_param, is_cuda = is_cuda) for i, net in enumerate(nets): setattr(model_ensemble, "model_{0}".format(i), net) return model_ensemble # In[ ]: class Model_with_uncertainty(nn.Module): def __init__( self, model_pred, model_logstd, ): super(Model_with_uncertainty, self).__init__() self.model_pred = model_pred self.model_logstd = model_logstd def forward(self, input, noise_amp = None, **kwargs): return self.model_pred(input, noise_amp = noise_amp, **kwargs), self.model_logstd(input, **kwargs) def get_loss(self, input, target, criterion, noise_amp = None, **kwargs): pred, log_std = self(input, noise_amp = noise_amp, **kwargs) return criterion(pred = pred, target = target, log_std = log_std) def get_regularization(self, source = ["weight", "bias"], mode = "L1", **kwargs): return self.model_pred.get_regularization(source = source, mode = mode, **kwargs) + self.model_logstd.get_regularization(source = source, mode = mode, **kwargs) @property def model_dict(self): model_dict = {} model_dict["type"] = "Model_with_Uncertainty" model_dict["model_pred"] = self.model_pred.model_dict model_dict["model_logstd"] = self.model_logstd.model_dict return model_dict def load(self, filename): mode = "json" if filename.endswith(".json") else "pickle" model_dict = load_model(filename, mode=mode) self.load_model_dict(model_dict) def save(self, filename): mode = "json" if filename.endswith(".json") else "pickle" save_model(self.model_dict, filename, mode=mode) def set_cuda(self, is_cuda): self.model_pred.set_cuda(is_cuda) self.model_logstd.set_cuda(is_cuda) def set_trainable(self, is_trainable): self.model_pred.set_trainable(is_trainable) self.model_logstd.set_trainable(is_trainable) # ## RNN: # In[ ]: class RNNCellBase(nn.Module): def extra_repr(self): s = '{input_size}, {hidden_size}' if 'bias' in self.__dict__ and self.bias is not True: s += ', bias={bias}' if 'nonlinearity' in self.__dict__ and self.nonlinearity != "tanh": s += ', nonlinearity={nonlinearity}' return s.format(**self.__dict__) def check_forward_input(self, input): if input.size(1) != self.input_size: raise RuntimeError( "input has inconsistent input_size: got {}, expected {}".format( input.size(1), self.input_size)) def check_forward_hidden(self, input, hx, hidden_label=''): if input.size(0) != hx.size(0): raise RuntimeError( "Input batch size {} doesn't match hidden{} batch size {}".format( input.size(0), hidden_label, hx.size(0))) if hx.size(1) != self.hidden_size: raise RuntimeError( "hidden{} has inconsistent hidden_size: got {}, expected {}".format( hidden_label, hx.size(1), self.hidden_size)) # ### LSTM: # In[ ]: class LSTM(RNNCellBase): """a LSTM class""" def __init__( self, input_size, hidden_size, output_struct_param, output_settings = {}, bias = True, is_cuda = False, ): super(LSTM, self).__init__() self.input_size = input_size self.hidden_size = hidden_size self.bias = bias self.W_ih = nn.Parameter(torch.Tensor(4 * hidden_size, input_size)) self.W_hh = nn.Parameter(torch.Tensor(4 * hidden_size, hidden_size)) self.output_net = MLP(input_size = self.hidden_size, struct_param = output_struct_param, settings = output_settings, is_cuda = is_cuda) if bias: self.b_ih = nn.Parameter(torch.Tensor(4 * hidden_size)) self.b_hh = nn.Parameter(torch.Tensor(4 * hidden_size)) else: self.register_parameter('b_ih', None) self.register_parameter('b_hh', None) self.reset_parameters() self.is_cuda = is_cuda self.device = torch.device(self.is_cuda if isinstance(self.is_cuda, str) else "cuda" if self.is_cuda else "cpu") self.to(self.device) def reset_parameters(self): stdv = 1.0 / np.sqrt(self.hidden_size) for weight in self.parameters(): weight.data.uniform_(-stdv, stdv) def forward_one_step(self, input, hx): self.check_forward_input(input) self.check_forward_hidden(input, hx[0], '[0]') self.check_forward_hidden(input, hx[1], '[1]') return self._backend.LSTMCell( input, hx, self.W_ih, self.W_hh, self.b_ih, self.b_hh, ) def forward(self, input, hx = None): if hx is None: hx = [torch.randn(input.size(0), self.hidden_size).to(self.device), torch.randn(input.size(0), self.hidden_size).to(self.device), ] hhx, ccx = hx for i in range(input.size(1)): hhx, ccx = self.forward_one_step(input[:, i], (hhx, ccx)) output = self.output_net(hhx) return output def get_regularization(self, source, mode = "L1", **kwargs): if not isinstance(source, list): source = [source] reg = self.output_net.get_regularization(source = source, mode = mode) for source_ele in source: if source_ele == "weight": if mode == "L1": reg = reg + self.W_ih.abs().sum() + self.W_hh.abs().sum() elif mode == "L2": reg = reg + (self.W_ih ** 2).sum() + (self.W_hh ** 2).sum() else: raise Exception("mode {0} not recognized!".format(mode)) elif source_ele == "bias": if self.bias: if mode == "L1": reg = reg + self.b_ih.abs().sum() + self.b_hh.abs().sum() elif mode == "L2": reg = reg + (self.b_ih ** 2).sum() + (self.b_hh ** 2).sum() else: raise Exception("mode {0} not recognized!".format(mode)) else: raise Exception("source {0} not recognized!".format(source_ele)) return reg def get_weights_bias(self, W_source = None, b_source = None, verbose = False, isplot = False): W_dict = OrderedDict() b_dict = OrderedDict() W_o, b_o = self.output_net.get_weights_bias(W_source = W_source, b_source = b_source) if W_source == "core": W_dict["W_ih"] = self.W_ih.cpu().detach().numpy() W_dict["W_hh"] = self.W_hh.cpu().detach().numpy() W_dict["W_o"] = W_o if isplot: print("W_ih, W_hh:") plot_matrices([W_dict["W_ih"], W_dict["W_hh"]]) print("W_o:") plot_matrices(W_o) if self.bias and b_source == "core": b_dict["b_ih"] = self.b_ih.cpu().detach().numpy() b_dict["b_hh"] = self.b_hh.cpu().detach().numpy() b_dict["b_o"] = b_o if isplot: print("b_ih, b_hh:") plot_matrices([b_dict["b_ih"], b_dict["b_hh"]]) print("b_o:") plot_matrices(b_o) return W_dict, b_dict def get_loss(self, input, target, criterion, hx = None, **kwargs): y_pred = self(input, hx = hx) return criterion(y_pred, target) def prepare_inspection(self, X, y, **kwargs): return {} def load(self, filename): mode = "json" if filename.endswith(".json") else "pickle" model_dict = load_model(filename, mode=mode) self.load_model_dict(model_dict) def save(self, filename): mode = "json" if filename.endswith(".json") else "pickle" save_model(self.model_dict, filename, mode=mode) # ## Wide ResNet: # In[ ]: def conv3x3(in_planes, out_planes, stride=1): return nn.Conv2d(in_planes, out_planes, kernel_size=3, stride=stride, padding=1, bias=True) def conv_init(m): classname = m.__class__.__name__ if classname.find('Conv') != -1: init.xavier_uniform_(m.weight, gain=np.sqrt(2)) init.constant_(m.bias, 0) elif classname.find('BatchNorm') != -1: init.constant_(m.weight, 1) init.constant_(m.bias, 0) class wide_basic(nn.Module): def __init__(self, in_planes, planes, dropout_rate=None, stride=1): super(wide_basic, self).__init__() self.bn1 = nn.BatchNorm2d(in_planes) self.conv1 = nn.Conv2d(in_planes, planes, kernel_size=3, padding=1, bias=True) if dropout_rate is not None: self.dropout = nn.Dropout(p=dropout_rate) else: self.dropout = None self.bn2 = nn.BatchNorm2d(planes) self.conv2 = nn.Conv2d(planes, planes, kernel_size=3, stride=stride, padding=1, bias=True) self.shortcut = nn.Sequential() if stride != 1 or in_planes != planes: self.shortcut = nn.Sequential( nn.Conv2d(in_planes, planes, kernel_size=1, stride=stride, bias=True), ) def forward(self, x): out = self.conv1(F.relu(self.bn1(x))) if self.dropout is not None: out = self.dropout(out) out = self.conv2(F.relu(self.bn2(out))) out += self.shortcut(x) return out class Wide_ResNet(nn.Module): """Adapted from https://github.com/meliketoy/wide-resnet.pytorch/blob/master/networks/wide_resnet.py""" def __init__( self, depth, widen_factor, input_channels, output_size, dropout_rate=None, is_cuda=False, ): super(Wide_ResNet, self).__init__() self.depth = depth self.widen_factor = widen_factor self.input_channels = input_channels self.dropout_rate = dropout_rate self.output_size = output_size assert ((depth-4)%6 ==0), 'Wide-resnet depth should be 6n+4' n = (depth-4)//6 k = widen_factor nStages = [16*k, 16*k, 32*k, 64*k] self.in_planes = nStages[0] self.conv1 = conv3x3(self.input_channels,nStages[0]) self.layer1 = self._wide_layer(wide_basic, nStages[1], n, dropout_rate, stride=1) self.layer2 = self._wide_layer(wide_basic, nStages[2], n, dropout_rate, stride=2) self.layer3 = self._wide_layer(wide_basic, nStages[3], n, dropout_rate, stride=2) self.bn1 = nn.BatchNorm2d(nStages[3], momentum=0.9) self.linear = nn.Linear(nStages[3], output_size) self.set_cuda(is_cuda) def _wide_layer(self, block, planes, num_blocks, dropout_rate, stride): strides = [stride] + [1]*(int(num_blocks)-1) layers = [] for stride in strides: layers.append(block(self.in_planes, planes, dropout_rate, stride)) self.in_planes = planes return nn.Sequential(*layers) def forward(self, x): out = self.conv1(x) out = self.layer1(out) out = self.layer2(out) out = self.layer3(out) out = F.relu(self.bn1(out)) out = out.mean((-1,-2)) # replacing the out= F.avg_pool2d(out, 8) which is sensitive to the input shape. out = out.view(out.size(0), -1) out = self.linear(out) return out def set_cuda(self, is_cuda): if isinstance(is_cuda, str): self.cuda(is_cuda) else: if is_cuda: self.cuda() else: self.cpu() self.is_cuda = is_cuda @property def model_dict(self): model_dict = {"type": "Wide_ResNet"} model_dict["state_dict"] = to_cpu_recur(self.state_dict()) model_dict["depth"] = self.depth model_dict["widen_factor"] = self.widen_factor model_dict["input_channels"] = self.input_channels model_dict["output_size"] = self.output_size model_dict["dropout_rate"] = self.dropout_rate return model_dict def load(self, filename): mode = "json" if filename.endswith(".json") else "pickle" model_dict = load_model(filename, mode=mode) self.load_model_dict(model_dict) def save(self, filename): mode = "json" if filename.endswith(".json") else "pickle" save_model(self.model_dict, filename, mode=mode) def get_regularization(self, *args, **kwargs): return to_Variable([0], is_cuda = self.is_cuda) def prepare_inspection(self, *args, **kwargs): return {} # ## CNN: # In[ ]: class ConvNet(nn.Module): def __init__( self, input_channels, struct_param=None, W_init_list=None, b_init_list=None, settings={}, return_indices=False, is_cuda=False, ): super(ConvNet, self).__init__() self.input_channels = input_channels if struct_param is not None: self.struct_param = struct_param self.W_init_list = W_init_list self.b_init_list = b_init_list self.settings = settings self.num_layers = len(struct_param) self.info_dict = {} self.param_available = ["Conv2d", "ConvTranspose2d", "BatchNorm2d", "Simple_Layer"] self.return_indices = return_indices for i in range(len(self.struct_param)): if i > 0: k = 1 while self.struct_param[i - k][0] is None: k += 1 num_channels_prev = self.struct_param[i - k][0] else: num_channels_prev = input_channels k = 0 if self.struct_param[i - k][1] == "Simple_Layer" and isinstance(num_channels_prev, tuple) and len(num_channels_prev) == 3: num_channels_prev = num_channels_prev[0] num_channels = self.struct_param[i][0] layer_type = self.struct_param[i][1] layer_settings = self.struct_param[i][2] if "layer_input_size" in layer_settings and isinstance(layer_settings["layer_input_size"], tuple): num_channels_prev = layer_settings["layer_input_size"][0] if layer_type == "Conv2d": layer = nn.Conv2d(num_channels_prev, num_channels, kernel_size = layer_settings["kernel_size"], stride = layer_settings["stride"] if "stride" in layer_settings else 1, padding = layer_settings["padding"] if "padding" in layer_settings else 0, dilation = layer_settings["dilation"] if "dilation" in layer_settings else 1, ) elif layer_type == "ConvTranspose2d": layer = nn.ConvTranspose2d(num_channels_prev, num_channels, kernel_size = layer_settings["kernel_size"], stride = layer_settings["stride"] if "stride" in layer_settings else 1, padding = layer_settings["padding"] if "padding" in layer_settings else 0, output_padding = layer_settings["output_padding"] if "output_padding" in layer_settings else 0, dilation = layer_settings["dilation"] if "dilation" in layer_settings else 1, ) elif layer_type == "Simple_Layer": layer = get_Layer(layer_type = layer_type, input_size = layer_settings["layer_input_size"], output_size = num_channels, W_init = W_init_list[i] if self.W_init_list is not None and self.W_init_list[i] is not None else None, b_init = b_init_list[i] if self.b_init_list is not None and self.b_init_list[i] is not None else None, settings = layer_settings, is_cuda = is_cuda, ) elif layer_type == "MaxPool2d": layer = nn.MaxPool2d(kernel_size = layer_settings["kernel_size"], stride = layer_settings["stride"] if "stride" in layer_settings else None, padding = layer_settings["padding"] if "padding" in layer_settings else 0, return_indices = layer_settings["return_indices"] if "return_indices" in layer_settings else False, ) elif layer_type == "MaxUnpool2d": layer = nn.MaxUnpool2d(kernel_size = layer_settings["kernel_size"], stride = layer_settings["stride"] if "stride" in layer_settings else None, padding = layer_settings["padding"] if "padding" in layer_settings else 0, ) elif layer_type == "Upsample": layer = nn.Upsample(scale_factor = layer_settings["scale_factor"], mode = layer_settings["mode"] if "mode" in layer_settings else "nearest", ) elif layer_type == "BatchNorm2d": layer = nn.BatchNorm2d(num_features = num_channels) elif layer_type == "Dropout2d": layer = nn.Dropout2d(p = 0.5) elif layer_type == "Flatten": layer = Flatten() else: raise Exception("layer_type {0} not recognized!".format(layer_type)) # Initialize using provided initial values: if self.W_init_list is not None and self.W_init_list[i] is not None and layer_type not in ["Simple_Layer"]: layer.weight.data = torch.FloatTensor(self.W_init_list[i]) layer.bias.data = torch.FloatTensor(self.b_init_list[i]) setattr(self, "layer_{0}".format(i), layer) self.set_cuda(is_cuda) def forward(self, input, indices_list = None, **kwargs): return self.inspect_operation(input, operation_between = (0, self.num_layers), indices_list = indices_list) def inspect_operation(self, input, operation_between, indices_list = None): output = input if indices_list is None: indices_list = [] start_layer, end_layer = operation_between if end_layer < 0: end_layer += self.num_layers for i in range(start_layer, end_layer): if "layer_input_size" in self.struct_param[i][2]: output_size_last = output.shape[0] layer_input_size = self.struct_param[i][2]["layer_input_size"] if not isinstance(layer_input_size, tuple): layer_input_size = (layer_input_size,) output = output.view(-1, *layer_input_size) assert output.shape[0] == output_size_last, "output_size reshaped to different length. Check shape!" if "Unpool" in self.struct_param[i][1]: output_tentative = getattr(self, "layer_{0}".format(i))(output, indices_list.pop(-1)) else: output_tentative = getattr(self, "layer_{0}".format(i))(output) if isinstance(output_tentative, tuple): output, indices = output_tentative indices_list.append(indices) else: output = output_tentative if "activation" in self.struct_param[i][2]: activation = self.struct_param[i][2]["activation"] else: if "activation" in self.settings: activation = self.settings["activation"] else: activation = "linear" if "Pool" in self.struct_param[i][1] or "Unpool" in self.struct_param[i][1] or "Upsample" in self.struct_param[i][1]: activation = "linear" output = get_activation(activation)(output) if self.return_indices: return output, indices_list else: return output def get_loss(self, input, target, criterion, **kwargs): y_pred = self(input, **kwargs) if self.return_indices: y_pred = y_pred[0] return criterion(y_pred, target) def get_regularization(self, source = ["weight", "bias"], mode = "L1", **kwargs): if not isinstance(source, list): source = [source] reg = Variable(torch.FloatTensor([0]), requires_grad = False) if self.is_cuda: reg = reg.cuda() for k in range(self.num_layers): if self.struct_param[k][1] not in self.param_available: continue layer = getattr(self, "layer_{0}".format(k)) for source_ele in source: if source_ele == "weight": if self.struct_param[k][1] not in ["Simple_Layer"]: item = layer.weight else: item = layer.W_core elif source_ele == "bias": if self.struct_param[k][1] not in ["Simple_Layer"]: item = layer.bias else: item = layer.b_core if mode == "L1": reg = reg + item.abs().sum() elif mode == "L2": reg = reg + (item ** 2).sum() else: raise Exception("mode {0} not recognized!".format(mode)) return reg def get_weights_bias(self, W_source = "core", b_source = "core"): W_list = [] b_list = [] for k in range(self.num_layers): if self.struct_param[k][1] == "Simple_Layer": layer = getattr(self, "layer_{0}".format(k)) if W_source == "core": W_list.append(to_np_array(layer.W_core)) if b_source == "core": b_list.append(to_np_array(layer.b_core)) elif self.struct_param[k][1] in self.param_available: layer = getattr(self, "layer_{0}".format(k)) if W_source == "core": W_list.append(to_np_array(layer.weight)) if b_source == "core": b_list.append(to_np_array(layer.bias, full_reduce = False)) else: if W_source == "core": W_list.append(None) if b_source == "core": b_list.append(None) return W_list, b_list @property def model_dict(self): model_dict = {"type": self.__class__.__name__} model_dict["net_type"] = self.__class__.__name__ model_dict["input_channels"] = self.input_channels model_dict["struct_param"] = self.struct_param model_dict["settings"] = self.settings model_dict["weights"], model_dict["bias"] = self.get_weights_bias(W_source = "core", b_source = "core") model_dict["return_indices"] = self.return_indices return model_dict @property def output_size(self): return self.struct_param[-1][0] @property def structure(self): structure = OrderedDict() structure["input_channels"] = self.input_channels structure["output_size"] = self.output_size structure["struct_param"] = self.struct_param if hasattr(self, "struct_param") else None return structure def get_sympy_expression(self, verbose=True): expressions = {i: None for i in range(self.num_layers)} return expressions def load(self, filename): mode = "json" if filename.endswith(".json") else "pickle" model_dict = load_model(filename, mode=mode) self.load_model_dict(model_dict) def save(self, filename): mode = "json" if filename.endswith(".json") else "pickle" save_model(self.model_dict, filename, mode=mode) def DL(self): DL = 0 for k in range(self.num_layers): layer_type = self.struct_param[k][1] if layer_type in self.param_available: layer = getattr(self, "layer_{0}".format(k)) if layer_type == "Simple_Layer": DL += layer.DL else: DL += get_list_DL(to_np_array(layer.weight), "non-snapped") DL += get_list_DL(to_np_array(layer.bias), "non-snapped") return DL def load_model_dict(self, model_dict): new_net = load_model_dict_net(model_dict, is_cuda = self.is_cuda) self.__dict__.update(new_net.__dict__) def prepare_inspection(self, X, y, **kwargs): pred_prob = self(X) if self.return_indices: pred_prob = pred_prob[0] pred = pred_prob.max(1)[1] # self.info_dict["accuracy"] = get_accuracy(pred, y) return deepcopy(self.info_dict) def set_cuda(self, is_cuda): if isinstance(is_cuda, str): self.cuda(is_cuda) else: if is_cuda: self.cuda() else: self.cpu() self.is_cuda = is_cuda def set_trainable(self, is_trainable): for k in range(self.num_layers): layer = getattr(self, "layer_{0}".format(k)) if self.struct_param[k][1] == "Simple_Layer": layer.set_trainable(is_trainable) elif self.struct_param[k][1] in self.param_available: for param in layer.parameters(): param.requires_grad = is_trainable class Conv_Model(nn.Module): def __init__( self, encoder_model_dict, core_model_dict, decoder_model_dict, latent_size = 2, is_generative = True, is_res_block = True, is_cuda = False, ): """Conv_Model consists of an encoder, a core and a decoder""" super(Conv_Model, self).__init__() self.latent_size = latent_size self.is_generative = is_generative if not is_generative: self.encoder = load_model_dict(encoder_model_dict, is_cuda = is_cuda) self.core = load_model_dict(core_model_dict, is_cuda = is_cuda) self.decoder = load_model_dict(decoder_model_dict, is_cuda = is_cuda) self.is_res_block = is_res_block self.is_cuda = is_cuda self.info_dict = {} @property def num_layers(self): if self.is_generative: return 1 else: return len(self.core.model_dict["struct_param"]) def forward( self, X, latent = None, **kwargs ): if self.is_generative: if len(latent.shape) == 1: latent = latent.repeat(len(X), 1) latent = self.core(latent) else: p_dict = {k: latent if k == 0 else None for k in range(self.num_layers)} latent = self.encoder(X) latent = self.core(latent, p_dict = p_dict) output = self.decoder(latent) if self.is_res_block: output = (X + nn.Sigmoid()(output)).clamp(0, 1) return output def forward_multistep(self, X, latents, isplot = False, num_images = 1): assert len(latents.shape) == 1 length = int(len(latents) / 2) output = X for i in range(length - 1): latent = latents[i * self.latent_size: (i + 2) * self.latent_size] output = self(output, latent = latent) if isplot: plot_matrices(output[:num_images,0]) return output def get_loss(self, X, y, criterion, **kwargs): return criterion(self(X = X[0], latent = X[1]), y) def plot(self, X, y, num_images = 1): y_pred = self(X[0], latent = X[1]) idx_list = np.random.choice(len(X[0]), num_images) for idx in idx_list: matrix = torch.cat([X[0][idx], y[idx], y_pred[idx]]) plot_matrices(matrix, images_per_row = 8) def get_regularization(self, source = ["weights", "bias"], mode = "L1"): if self.is_generative: return self.core.get_regularization(source = source, mode = mode) + self.decoder.get_regularization(source = source, mode = mode) else: return self.encoder.get_regularization(source = source, mode = mode) + self.core.get_regularization(source = source, mode = mode) + self.decoder.get_regularization(source = source, mode = mode) def prepare_inspection(self, X, y, **kwargs): return deepcopy(self.info_dict) def set_trainable(self, is_trainable): if not self.is_generative: self.encoder.set_trainable(is_trainable) self.core.set_trainable(is_trainable) self.decoder.set_trainable(is_trainable) @property def model_dict(self): model_dict = {"type": "Conv_Model"} if not self.is_generative: model_dict["encoder_model_dict"] = self.encoder.model_dict model_dict["latent_size"] = self.latent_size model_dict["core_model_dict"] = self.core.model_dict model_dict["decoder_model_dict"] = self.decoder.model_dict model_dict["is_generative"] = self.is_generative model_dict["is_res_block"] = self.is_res_block return model_dict def load(self, filename): mode = "json" if filename.endswith(".json") else "pickle" model_dict = load_model(filename, mode=mode) self.load_model_dict(model_dict) def save(self, filename): mode = "json" if filename.endswith(".json") else "pickle" save_model(self.model_dict, filename, mode=mode) class Conv_Autoencoder(nn.Module): def __init__( self, input_channels_encoder, input_channels_decoder, struct_param_encoder, struct_param_decoder, latent_size = (1,2), share_model_among_steps = False, settings = {}, is_cuda = False, ): """Conv_Autoencoder consists of an encoder and a decoder""" super(Conv_Autoencoder, self).__init__() self.input_channels_encoder = input_channels_encoder self.input_channels_decoder = input_channels_decoder self.struct_param_encoder = struct_param_encoder self.struct_param_decoder = struct_param_decoder self.share_model_among_steps = share_model_among_steps self.settings = settings self.encoder = ConvNet(input_channels = input_channels_encoder, struct_param = struct_param_encoder, settings = settings, is_cuda = is_cuda) self.decoder = ConvNet(input_channels = input_channels_decoder, struct_param = struct_param_decoder, settings = settings, is_cuda = is_cuda) self.is_cuda = is_cuda def encode(self, input): if self.share_model_among_steps: latent = [] for i in range(input.shape[1]): latent_step = self.encoder(input[:, i:i+1]) latent.append(latent_step) return torch.cat(latent, 1) else: return self.encoder(input) def decode(self, latent): if self.share_model_among_steps: latent_size = self.struct_param_encoder[-1][0] latent = latent.view(latent.size(0), -1, latent_size) output = [] for i in range(latent.shape[1]): output_step = self.decoder(latent[:, i].contiguous()) output.append(output_step) return torch.cat(output, 1) else: return self.decoder(latent) def set_trainable(self, is_trainable): self.encoder.set_trainable(is_trainable) self.decoder.set_trainable(is_trainable) def forward(self, input): return self.decode(self.encode(input)) def get_loss(self, input, target, criterion, **kwargs): return criterion(self(input), target) def get_regularization(self, source = ["weight", "bias"], mode = "L1"): return self.encoder.get_regularization(source = source, mode = mode) + self.decoder.get_regularization(source = source, mode = mode) @property def model_dict(self): model_dict = {"type": "Conv_Autoencoder"} model_dict["net_type"] = "Conv_Autoencoder" model_dict["input_channels_encoder"] = self.input_channels_encoder model_dict["input_channels_decoder"] = self.input_channels_decoder model_dict["struct_param_encoder"] = self.struct_param_encoder model_dict["struct_param_decoder"] = self.struct_param_decoder model_dict["share_model_among_steps"] = self.share_model_among_steps model_dict["settings"] = self.settings model_dict["encoder"] = self.encoder.model_dict model_dict["decoder"] = self.decoder.model_dict return model_dict def load_model_dict(self, model_dict): model = load_model_dict(model_dict, is_cuda = self.is_cuda) self.__dict__.update(model.__dict__) def load(self, filename): mode = "json" if filename.endswith(".json") else "pickle" model_dict = load_model(filename, mode=mode) self.load_model_dict(model_dict) def save(self, filename): mode = "json" if filename.endswith(".json") else "pickle" save_model(self.model_dict, filename, mode=mode) def DL(self): return self.encoder.DL + self.decoder.DL class Flatten(nn.Module): def __init__(self): super(Flatten, self).__init__() def forward(self, x): return x.view(x.size(0), -1) # ## VAE: # In[ ]: class VAE(nn.Module): def __init__( self, encoder_model_dict, decoder_model_dict, is_cuda = False, ): super(VAE, self).__init__() self.encoder = load_model_dict(encoder_model_dict, is_cuda = is_cuda) self.decoder = load_model_dict(decoder_model_dict, is_cuda = is_cuda) self.is_cuda = is_cuda self.info_dict = {} def encode(self, X): Z = self.encoder(X) latent_size = int(Z.shape[-1] / 2) mu = Z[..., :latent_size] logvar = Z[..., latent_size:] return mu, logvar def reparameterize(self, mu, logvar): std = torch.exp(0.5*logvar) eps = torch.randn_like(std) return eps.mul(std).add_(mu) def decode(self, Z): return self.decoder(Z) def forward(self, X): mu, logvar = self.encode(X) Z = self.reparameterize(mu, logvar) return self.decode(Z), mu, logvar def get_loss(self, X, y = None, **kwargs): recon_X, mu, logvar = self(X) BCE = F.binary_cross_entropy(recon_X.view(recon_X.shape[0], -1), X.view(X.shape[0], -1), reduction='sum') # see Appendix B from VAE paper: # Kingma and Welling. Auto-Encoding Variational Bayes. ICLR, 2014 # https://arxiv.org/abs/1312.6114 # 0.5 * sum(1 + log(sigma^2) - mu^2 - sigma^2) KLD = -0.5 * torch.sum(1 + logvar - mu.pow(2) - logvar.exp()) loss = (BCE + KLD) / len(X) self.info_dict["KLD"] = KLD.item() / len(X) self.info_dict["BCE"] = BCE.item() / len(X) return loss def model_dict(self): model_dict = {"type": "VAE"} model_dict["encoder_model_dict"] = self.encoder.model_dict model_dict["decoder_model_dict"] = self.decoder.model_dict return model_dict def load(self, filename): mode = "json" if filename.endswith(".json") else "pickle" model_dict = load_model(filename, mode=mode) self.load_model_dict(model_dict) def save(self, filename): mode = "json" if filename.endswith(".json") else "pickle" save_model(self.model_dict, filename, mode=mode) def get_regularization(self, source = ["weight", "bias"], mode = "L1"): return self.encoder.get_regularization(source = source, mode = mode) + self.decoder.get_regularization(source = source, mode = mode) def prepare_inspection(self, X, y, **kwargs): return deepcopy(self.info_dict) # ## Reparameterization toolkit: # In[ ]: class Net_reparam(nn.Module): """Module that uses reparameterization to take into two inputs and gets a scaler""" def __init__( self, model_dict, reparam_mode, is_cuda=False, ): super(Net_reparam, self).__init__() self.model = load_model_dict(model_dict, is_cuda=is_cuda) self.reparam_mode = reparam_mode def forward(self, X, Z, is_outer=False): """ Obtaining single value using reparameterization. Args: X shape: [Bx, ...] Z shape: [S, Bz, Z] is_outer: whether to use outer product to get a tensor with shape [S, Bz, Bx]. Returns: If is_outer==True, return log_prob of shape [S, Bz, Bx] If is_outer==False, return log_prob of shape [S, Bz] (where Bz=Bx) """ dist, _ = reparameterize(self.model, X, mode=self.reparam_mode) if is_outer: log_prob = dist.log_prob(Z[...,None,:]) else: log_prob = dist.log_prob(Z) if self.reparam_mode == 'diag': log_prob = log_prob.sum(-1) return log_prob def get_regularization(self, source = ["weight", "bias"], mode = "L1", **kwargs): return self.model.get_regularization(source=source, model=mode, **kwargs) def prepare_inspection(self, X, y, **kwargs): return {} @property def model_dict(self): model_dict = {"type": "Net_reparam"} model_dict["model"] = self.model.model_dict model_dict["reparam_mode"] = self.reparam_mode return model_dict def load(self, filename): mode = "json" if filename.endswith(".json") else "pickle" model_dict = load_model(filename, mode=mode) self.load_model_dict(model_dict) def save(self, filename): mode = "json" if filename.endswith(".json") else "pickle" save_model(self.model_dict, filename, mode=mode) def reparameterize(model, input, mode="full", size=None): if mode.startswith("diag"): if model is not None and model.__class__.__name__ == "Mixture_Model": return reparameterize_mixture_diagonal(model, input, mode=mode) else: return reparameterize_diagonal(model, input, mode=mode) elif mode == "full": return reparameterize_full(model, input, size=size) else: raise Exception("Mode {} is not valid!".format(mode)) def reparameterize_diagonal(model, input, mode): if model is not None: mean_logit = model(input) else: mean_logit = input if mode.startswith("diagg"): if isinstance(mean_logit, tuple): mean = mean_logit[0] else: mean = mean_logit std = torch.ones(mean.shape).to(mean.device) dist = Normal(mean, std) return dist, (mean, std) elif mode.startswith("diag"): if isinstance(mean_logit, tuple): mean_logit = mean_logit[0] size = int(mean_logit.size(-1) / 2) mean = mean_logit[:, :size] std = F.softplus(mean_logit[:, size:], beta=1) + 1e-10 dist = Normal(mean, std) return dist, (mean, std) else: raise Exception("mode {} is not valid!".format(mode)) def reparameterize_mixture_diagonal(model, input, mode): mean_logit, weight_logits = model(input) if mode.startswith("diagg"): mean_list = mean_logit scale_list = torch.ones(mean_list.shape).to(mean_list.device) else: size = int(mean_logit.size(-2) / 2) mean_list = mean_logit[:, :size] scale_list = F.softplus(mean_logit[:, size:], beta=1) + 0.01 # Avoid the std to go to 0 dist = Mixture_Gaussian_reparam(mean_list=mean_list, scale_list=scale_list, weight_logits=weight_logits, ) return dist, (mean_list, scale_list) def reparameterize_full(model, input, size=None): if model is not None: mean_logit = model(input) else: mean_logit = input if isinstance(mean_logit, tuple): mean_logit = mean_logit[0] if size is None: dim = mean_logit.size(-1) size = int((np.sqrt(9 + 8 * dim) - 3) / 2) mean = mean_logit[:, :size] scale_tril = fill_triangular(mean_logit[:, size:], size) scale_tril = matrix_diag_transform(scale_tril, F.softplus) dist = MultivariateNormal(mean, scale_tril = scale_tril) return dist, (mean, scale_tril) def sample(dist, n=None): """Sample n instances from distribution dist""" if n is None: return dist.rsample() else: return dist.rsample((n,)) # ## Probability models: # ### Mixture of Gaussian: # In[ ]: class Mixture_Gaussian(nn.Module): def __init__( self, num_components, dim, param_mode = "full", is_cuda = False, ): super(Mixture_Gaussian, self).__init__() self.num_components = num_components self.dim = dim self.param_mode = param_mode self.is_cuda = is_cuda self.device = torch.device(self.is_cuda if isinstance(self.is_cuda, str) else "cuda" if self.is_cuda else "cpu") self.info_dict = {} def initialize(self, model_dict = None, input = None, num_samples = 100, verbose = False): if input is not None: neg_log_prob_min = np.inf loc_init_min = None scale_init_min = None for i in range(num_samples): neg_log_prob, loc_init_list, scale_init_list = self.initialize_ele(input) if verbose: print("{0}: neg_log_prob: {1:.4f}".format(i, neg_log_prob)) if neg_log_prob < neg_log_prob_min: neg_log_prob_min = neg_log_prob loc_init_min = self.loc_list.detach() scale_init_min = self.scale_list.detach() self.loc_list = nn.Parameter(loc_init_min.to(self.device)) self.scale_list = nn.Parameter(scale_init_min.to(self.device)) print("min neg_log_prob: {0:.6f}".format(to_np_array(neg_log_prob_min))) else: if model_dict is None: self.weight_logits = nn.Parameter((torch.randn(self.num_components) * np.sqrt(2 / (1 + self.dim))).to(self.device)) else: self.weight_logits = nn.Parameter((torch.FloatTensor(model_dict["weight_logits"])).to(self.device)) if self.param_mode == "full": size = self.dim * (self.dim + 1) // 2 elif self.param_mode == "diag": size = self.dim else: raise if model_dict is None: self.loc_list = nn.Parameter(torch.randn(self.num_components, self.dim).to(self.device)) self.scale_list = nn.Parameter((torch.randn(self.num_components, size) / self.dim).to(self.device)) else: self.loc_list = nn.Parameter(torch.FloatTensor(model_dict["loc_list"]).to(self.device)) self.scale_list = nn.Parameter(torch.FloatTensor(model_dict["scale_list"]).to(self.device)) def initialize_ele(self, input): if self.param_mode == "full": size = self.dim * (self.dim + 1) // 2 elif self.param_mode == "diag": size = self.dim else: raise length = len(input) self.weight_logits = nn.Parameter(torch.zeros(self.num_components).to(self.device)) self.loc_list = nn.Parameter(input[torch.multinomial(torch.ones(length) / length, self.num_components)].detach()) self.scale_list = nn.Parameter((torch.randn(self.num_components, size).to(self.device) * input.std() / 5).to(self.device)) neg_log_prob = self.get_loss(input) return neg_log_prob def prob(self, input): if len(input.shape) == 1: input = input.unsqueeze(1) assert len(input.shape) in [0, 2, 3] input = input.unsqueeze(-2) if self.param_mode == "diag": scale_list = F.softplus(self.scale_list) logits = (- (input - self.loc_list) ** 2 / 2 / scale_list ** 2 - torch.log(scale_list * np.sqrt(2 * np.pi))).sum(-1) else: raise prob = torch.matmul(torch.exp(logits), nn.Softmax(dim = 0)(self.weight_logits)) # prob_list = [] # for i in range(self.num_components): # if self.param_mode == "full": # scale_tril = fill_triangular(getattr(self, "scale_{0}".format(i)), self.dim) # scale_tril = matrix_diag_transform(scale_tril, F.softplus) # dist = MultivariateNormal(getattr(self, "loc_{0}".format(i)), scale_tril = scale_tril) # log_prob = dist.log_prob(input) # elif self.param_mode == "diag": # dist = Normal(getattr(self, "loc_{0}".format(i)).unsqueeze(0), F.softplus(getattr(self, "scale_{0}".format(i)))) # mu = getattr(self, "loc_{0}".format(i)).unsqueeze(0) # sigma = F.softplus(getattr(self, "scale_{0}".format(i))) # log_prob = (- (input - mu) ** 2 / 2 / sigma ** 2 - torch.log(sigma * np.sqrt(2 * np.pi))).sum(-1) # else: # raise # setattr(self, "component_{0}".format(i), dist) # prob = torch.exp(log_prob) # prob_list.append(prob) # prob_list = torch.stack(prob_list, -1) # prob = torch.matmul(prob_list, nn.Softmax(dim = 0)(self.weight_logits)) return prob def log_prob(self, input): return torch.log(self.prob(input) + 1e-45) def get_loss(self, X, y = None, **kwargs): """Optimize negative log-likelihood""" neg_log_prob = - self.log_prob(X).mean() / np.log(2) self.info_dict["loss"] = to_np_array(neg_log_prob) return neg_log_prob def prepare_inspection(X, y, criterion, **kwargs): return deepcopy(self.info_dict) @property def model_dict(self): model_dict = {"type": "Mixture_Gaussian"} model_dict["num_components"] = self.num_components model_dict["dim"] = self.dim model_dict["param_mode"] = self.param_mode model_dict["weight_logits"] = to_np_array(self.weight_logits) model_dict["loc_list"] = to_np_array(self.loc_list) model_dict["scale_list"] = to_np_array(self.scale_list) return model_dict def load(self, filename): mode = "json" if filename.endswith(".json") else "pickle" model_dict = load_model(filename, mode=mode) self.load_model_dict(model_dict) def save(self, filename): mode = "json" if filename.endswith(".json") else "pickle" save_model(self.model_dict, filename, mode=mode) def get_param(self): weights = to_np_array(nn.Softmax(dim = 0)(self.weight_logits)) loc_list = to_np_array(self.loc_list) scale_list = to_np_array(self.scale_list) print("weights: {0}".format(weights)) print("loc:") pp.pprint(loc_list) print("scale:") pp.pprint(scale_list) return weights, loc_list, scale_list def visualize(self, input): import scipy import matplotlib.pylab as plt std = to_np_array(input.std()) X = np.arange(to_np_array(input.min()) - 0.2 * std, to_np_array(input.max()) + 0.2 * std, 0.1) Y_dict = {} weights = nn.Softmax(dim = 0)(self.weight_logits) plt.figure(figsize=(10, 4), dpi=100).set_facecolor('white') for i in range(self.num_components): Y_dict[i] = weights[0].item() * scipy.stats.norm.pdf((X - self.loc_list[i].item()) / self.scale_list[i].item()) plt.plot(X, Y_dict[i]) Y = np.sum([item for item in Y_dict.values()], 0) plt.plot(X, Y, 'k--') plt.plot(input.data.numpy(), np.zeros(len(input)), 'k*') plt.title('Density of {0}-component mixture model'.format(self.num_components)) plt.ylabel('probability density'); def get_regularization(self, source = ["weights", "bias"], mode = "L1", **kwargs): reg = to_Variable([0], requires_grad = False).to(self.device) return reg # ### Mixture_Gaussian for reparameterization: # In[ ]: class Mixture_Gaussian_reparam(nn.Module): def __init__( self, # Use as reparamerization: mean_list=None, scale_list=None, weight_logits=None, # Use as prior: Z_size=None, n_components=None, mean_scale=0.1, scale_scale=0.1, # Mode: is_reparam=True, reparam_mode="diag", device= torch.device("cpu"), ): super(Mixture_Gaussian_reparam, self).__init__() self.is_reparam = is_reparam self.reparam_mode = reparam_mode # self.is_cuda = is_cuda self.device = device if self.is_reparam: self.mean_list = mean_list # size: [B, Z, k] self.scale_list = scale_list # size: [B, Z, k] self.weight_logits = weight_logits # size: [B, k] self.n_components = self.weight_logits.shape[-1] self.Z_size = self.mean_list.shape[-2] else: self.n_components = n_components self.Z_size = Z_size self.mean_list = nn.Parameter((torch.rand(1, Z_size, n_components) - 0.5) * mean_scale) self.scale_list = nn.Parameter(torch.log(torch.exp((torch.rand(1, Z_size, n_components) * 0.2 + 0.9) * scale_scale) - 1)) self.weight_logits = nn.Parameter(torch.zeros(1, n_components)) if mean_list is not None: self.mean_list.data = to_Variable(mean_list) self.scale_list.data = to_Variable(scale_list) self.weight_logits.data = to_Variable(weight_logits) self.to(self.device) def log_prob(self, input): """Obtain the log_prob of the input.""" input = input.unsqueeze(-1) # [S, B, Z, 1] if self.reparam_mode == "diag": if self.is_reparam: # logits: [S, B, Z, k] logits = - (input - self.mean_list) ** 2 / 2 / self.scale_list ** 2 - torch.log(self.scale_list * np.sqrt(2 * np.pi)) else: scale_list = F.softplus(self.scale_list, beta=1) logits = - (input - self.mean_list) ** 2 / 2 / scale_list ** 2 - torch.log(scale_list * np.sqrt(2 * np.pi)) else: raise # log_softmax(weight_logits): [B, k] # logits: [S, B, Z, k] # log_prob: [S, B, Z] log_prob = torch.logsumexp(logits + F.log_softmax(self.weight_logits, -1).unsqueeze(-2), axis=-1) # F(...).unsqueeze(-2): [B, 1, k] return log_prob def prob(self, Z): return torch.exp(self.log_prob(Z)) def sample(self, n=None): if n is None: n_core = 1 else: assert isinstance(n, tuple) n_core = n[0] weight_probs = F.softmax(self.weight_logits, -1) # size: [B, m] idx = torch.multinomial(weight_probs, n_core, replacement=True).unsqueeze(-2).expand(-1, self.mean_list.shape[-2], -1) # multinomial result: [B, S]; result: [B, Z, S] mean_list = torch.gather(self.mean_list, dim=-1, index=idx) # [B, Z, S] if self.is_reparam: scale_list = torch.gather(self.scale_list, dim=-1, index=idx) # [B, Z, S] else: scale_list = F.softplus(torch.gather(self.scale_list, dim=-1, index=idx), beta=1) # [B, Z, S] Z = torch.normal(mean_list, scale_list).permute(2, 0, 1) if n is None: Z = Z.squeeze(0) return Z def rsample(self, n=None): return self.sample(n=n) def __repr__(self): return "Mixture_Gaussian_reparam({}, Z_size={})".format(self.n_components, self.Z_size) @property def model_dict(self): model_dict = {"type": "Mixture_Gaussian_reparam"} model_dict["is_reparam"] = self.is_reparam model_dict["reparam_mode"] = self.reparam_mode model_dict["Z_size"] = self.Z_size model_dict["n_components"] = self.n_components model_dict["mean_list"] = to_np_array(self.mean_list) model_dict["scale_list"] = to_np_array(self.scale_list) model_dict["weight_logits"] = to_np_array(self.weight_logits) return model_dict # ### Triangular distribution: # In[ ]: class Triangular_dist(Distribution): """Probability distribution with a Triangular shape.""" def __init__(self, loc, a, b, validate_args=None): self.loc, self.a, self.b = broadcast_all(loc, a, b) batch_shape = torch.Size() if isinstance(loc, Number) else self.loc.size() super(Triangular_dist, self).__init__(batch_shape, validate_args=validate_args) @property def mean(self): return self.loc + (self.b - self.a) / 3 @property def variance(self): return (self.a ** 2 + self.b ** 2 + self.a * self.b) / 18 @property def stddev(self): return torch.sqrt(self.variance) def expand(self, batch_shape, _instance=None): new = self._get_checked_instance(PieceWise, _instance) batch_shape = torch.Size(batch_shape) new.loc = self.loc.expand(batch_shape) new.a = self.a.expand(batch_shape) new.b = self.b.expand(batch_shape) super(Triangular_dist, new).__init__(batch_shape, validate_args=False) new._validate_args = self._validate_args return new @constraints.dependent_property def support(self): return constraints.interval(self.loc - self.a, self.loc + self.b) def sample(self, sample_shape=torch.Size()): shape = self._extended_shape(sample_shape) rand = torch.rand(shape, dtype=self.loc.dtype, device=self.loc.device) with torch.no_grad(): return self.icdf(rand) def rsample(self, sample_shape=torch.Size()): """Sample with reparameterization.""" shape = self._extended_shape(sample_shape) rand = torch.rand(shape, dtype=self.loc.dtype, device=self.loc.device) return self.icdf(rand) def icdf(self, value): """Inverse cdf.""" if self._validate_args: self._validate_sample(value) assert value.min() >= 0 and value.max() <= 1 value, loc, a, b = broadcast_all(value, self.loc, self.a, self.b) a_plus_b = a + b idx = value < a / a_plus_b iidx = ~idx out = torch.ones_like(value) out[idx] = loc[idx] - a[idx] + torch.sqrt(a[idx] * a_plus_b[idx] * value[idx]) out[iidx] = loc[iidx] + b[iidx] - torch.sqrt(b[iidx] * a_plus_b[iidx] * (1 - value[iidx]) ) return out def prob(self, value): """Get probability.""" if self._validate_args: self._validate_sample(value) # compute the variance value, loc, a, b = broadcast_all(value, self.loc, self.a, self.b) idx1 = (loc - a <= value) & (value <= loc) idx2 = (loc < value) & (value <= loc + b) a_plus_b = a + b out = torch.zeros_like(value) out[idx1] = 2 * (value[idx1] - loc[idx1] + a[idx1]) / a[idx1] / a_plus_b[idx1] out[idx2] = -2 * (value[idx2] - loc[idx2] - b[idx2]) / b[idx2] / a_plus_b[idx2] return out def log_prob(self, value): """Get log probability.""" return torch.log(self.prob(value)) @property def model_dict(self): model_dict = {"type": "Triangular_dist"} model_dict["loc"] = to_np_array(self.loc) model_dict["a"] = to_np_array(self.a) model_dict["b"] = to_np_array(self.b) return model_dict def load(self, filename): mode = "json" if filename.endswith(".json") else "pickle" model_dict = load_model(filename, mode=mode) self.load_model_dict(model_dict) def save(self, filename): mode = "json" if filename.endswith(".json") else "pickle" save_model(self.model_dict, filename, mode=mode) # In[ ]: def load_model_dict_distribution(model_dict, is_cuda = False): if model_dict["type"] == "Mixture_Gaussian": model = Mixture_Gaussian( num_components=model_dict["num_components"], dim=model_dict["dim"], param_mode=model_dict["param_mode"], is_cuda=is_cuda, ) model.initialize(model_dict = model_dict) elif model_dict["type"] == "Mixture_Gaussian_reparam": model = Mixture_Gaussian_reparam( is_reparam=model_dict["is_reparam"], reparam_mode=model_dict["reparam_mode"], mean_list=model_dict["mean_list"], scale_list=model_dict["scale_list"], weight_logits=model_dict["weight_logits"], Z_size=model_dict["Z_size"], n_components=model_dict["n_components"], is_cuda=is_cuda, ) elif model_dict["type"] == "Triangular_dist": model = Triangular_dist( loc=model_dict["loc"], a=model_dict["a"], b=model_dict["b"], ) else: raise Exception("Type {} is not valid!".format(model_dict["type"])) return model
208,307
46.428962
300
py
theedhum-nandrum
theedhum-nandrum-master/src/tn/multiclassrnnclassifier.py
""" @author sanjeethr, oligoglot Thanks to Susan Li for this step by step guide: https://towardsdatascience.com/multi-class-text-classification-with-lstm-1590bee1bd17 """ import pandas as pd import matplotlib.pyplot as plt import numpy as np import sys, os from keras.preprocessing.text import Tokenizer from keras.preprocessing.sequence import pad_sequences from keras import Sequential from keras.layers import Embedding, SpatialDropout1D, LSTM, Dense from keras.callbacks import EarlyStopping from keras.optimizers import Adam from sklearn.model_selection import train_test_split from sklearn.preprocessing import LabelBinarizer from sklearn.metrics import classification_report from libindic.soundex import Soundex from lib.feature_utils import load_docs, get_emojis_from_text, get_doc_len_range sys.path.append(os.path.join(os.path.dirname(sys.path[0]),'extern', 'indic_nlp_library')) from indicnlp.normalize.indic_normalize import BaseNormalizer try: from indictrans import Transliterator except ImportError: print('Please install indic-trans from git: https://github.com/libindic/indic-trans') ta_trans = Transliterator(source='eng', target='tam', build_lookup=True) ml_trans = Transliterator(source='eng', target='mal', build_lookup=True) # The maximum number of words to be used. (most frequent) MAX_NB_WORDS = 50000 # Max number of words in each review. MAX_SEQUENCE_LENGTH = 150 # This is fixed. EMBEDDING_DIM = 100 tokenizer = Tokenizer(num_words=MAX_NB_WORDS, filters='!"#$%&()*+,-./:;<=>?@[\]^_`{|}~', lower=True) soundexer = Soundex() def load_language_maps(mapfile): lmap = {} with open(mapfile, 'r') as mapf: for line in mapf: text, lang, conf = line.rstrip().split('\t') lmap[text] = (lang, float(conf)) return lmap def get_language_tag(text): return lmap.get(text, ('unknown', 0.0)) def append_language_tag(text): p_lang, conf = get_language_tag(text) if p_lang == lang or p_lang == (lang + 'en'): # google agrees with some confidence agreement = 1 elif conf < 0.5: # google says not-tamil, but weakly agreement = 0.5 else: # google clearly says not-tamil agreement = 0 return ' '.join((' ', text, p_lang, lang, str(agreement), ' ')) def append_emoji_sentiment(text): emojis, sentiment = get_emojis_from_text(text) return ' '.join((' ', text, str(emojis), sentiment, ' ')) def append_soundex(text): if lang == 'ta': text = ta_trans.transform(text) if lang == 'ml': text = ml_trans.transform(text) soundexes = [soundexer.soundex(word) for word in text.split()] return ' ' + text + ' ' + ' '.join(soundexes) + ' ' def append_doc_len_range(text): return ' ' + get_doc_len_range(text) + ' ' def load_data(df, mode, lb = None): df.info() df = df.reset_index(drop=True) df['text'] = df['text'].apply(append_emoji_sentiment) df['text'] = df['text'].apply(append_language_tag) df['text'] = df['text'].apply(append_soundex) df['text'] = df['text'].apply(append_doc_len_range) tokenizer.fit_on_texts([normalizer.normalize (text) for text in df.text.values]) word_index = tokenizer.word_index print('Found %s unique tokens.' % len(word_index)) X = tokenizer.texts_to_sequences(df.text.values) X = pad_sequences(X, maxlen=MAX_SEQUENCE_LENGTH) print('Shape of data tensor:', X.shape) if mode == 'pred': Y = df.id.values else: print(df.category.value_counts()) if lb is None: lb = LabelBinarizer() Y = lb.fit_transform(df.category.values.reshape(-1, 1)) else: Y = lb.transform(df.category.values.reshape(-1, 1)) print('Shape of label tensor:', Y.shape) return (X, Y, lb) lang, train_file, test_file, predict_file, outfile = sys.argv[1:6] normalizer = BaseNormalizer(lang) lmap = load_language_maps('../../resources/data/alltextslang.txt') #train_file = '../../resources/data/tamil_train.tsv' train_df = pd.read_csv(train_file, sep='\t') X_train, Y_train, lb = load_data(train_df, 'train') #test_file = '../../resources/data/tamil_dev.tsv' test_df = pd.read_csv(test_file, sep='\t') X_test, Y_test, lb = load_data(test_df, 'test', lb) # X_train, X_test, Y_train, Y_test = train_test_split(X,Y, test_size = 0.10, random_state = 42) print(X_train.shape,Y_train.shape) print(X_test.shape,Y_test.shape) if lang == 'ta': model = Sequential() model.add(Embedding(MAX_NB_WORDS, EMBEDDING_DIM, input_length=X_train.shape[1])) model.add(SpatialDropout1D(0.8)) model.add(LSTM(100, dropout=0.7, recurrent_dropout=0.5)) model.add(Dense(5, activation='softmax')) model.compile(loss='categorical_crossentropy', optimizer=Adam(learning_rate=0.0001), metrics=['accuracy']) epochs = 15 batch_size = 64 if lang == 'ml': model = Sequential() model.add(Embedding(MAX_NB_WORDS, EMBEDDING_DIM, input_length=X_train.shape[1])) model.add(SpatialDropout1D(0.5)) #model.add(LSTM(100, dropout=0.3, recurrent_dropout=0.3, return_sequences=True)) model.add(LSTM(100, dropout=0.3, recurrent_dropout=0.3)) model.add(Dense(5, activation='softmax')) model.compile(loss='categorical_crossentropy', optimizer=Adam(learning_rate=0.0001), metrics=['accuracy']) epochs = 10 batch_size = 64 history = model.fit(X_train, Y_train, epochs=epochs, batch_size=batch_size,validation_split=0.1,callbacks=[EarlyStopping(monitor='val_loss', patience=3, min_delta=0.0001)]) # accr = model.evaluate(X_test,Y_test) # print('Test set\n Loss: {:0.3f}\n Accuracy: {:0.3f}'.format(accr[0],accr[1])) Y_test_idx = np.argmax(Y_test, axis=1) # Convert one-hot to index Y_pred = model.predict_classes(X_test) print(classification_report(Y_test_idx, Y_pred)) new_review = ['Thalaiva superstar Rajinikanth number one mass Hero'] seq = tokenizer.texts_to_sequences(new_review) padded = pad_sequences(seq, maxlen=MAX_SEQUENCE_LENGTH) pred = model.predict(padded) print(pred, lb.inverse_transform(pred)) with open(outfile, 'w') as outf: test_df = pd.read_csv(predict_file, sep='\t') X_pred, ID_pred, lb = load_data(test_df, 'pred', lb) Y_pred = lb.inverse_transform(model.predict(X_pred)).flatten() outf.write('id\ttext\tlabel\n') for idx, text, pred_category in zip(ID_pred, test_df.text.values, Y_pred): #print(idx, text, pred_category) outf.write('\t'.join((idx, text, pred_category)) + '\n')
6,528
38.331325
172
py
deephyper
deephyper-master/setup.py
#!/usr/bin/env python # -*- coding: utf-8 -*- # Note: To use the 'upload' functionality of this file, you must: # $ pip install twine import os import platform import sys from shutil import rmtree from setuptools import Command, setup # path of the directory where this file is located here = os.path.abspath(os.path.dirname(__file__)) # query platform informations, e.g. 'macOS-12.0.1-arm64-arm-64bit' platform_infos = platform.platform() # What packages are required for this module to be executed? REQUIRED = [ "ConfigSpace>=0.4.20", "dm-tree", "Jinja2<3.1", "numpy", # ==1.19.4", # working with 1.20.1 "pandas>=0.24.2", "packaging", "parse", "scikit-learn>=0.23.1", "scipy>=1.7", "tqdm>=4.64.0", "pyyaml", ] # !Requirements for Neural Architecture Search (NAS) REQUIRED_NAS = ["networkx", "pydot"] REQUIRED_NAS_PLATFORM = { "default": ["tensorflow>=2.0.0", "tensorflow_probability"], "macOS-arm64": ["tensorflow_probability~=0.14"], } # if "macOS" in platform_infos and "arm64" in platform_infos: # REQUIRED_NAS = REQUIRED_NAS + REQUIRED_NAS_PLATFORM["macOS-arm64"] # else: # x86_64 REQUIRED_NAS = REQUIRED_NAS + REQUIRED_NAS_PLATFORM["default"] # !Requirements for Pipeline Optimization for ML (popt) REQUIRED_POPT = ["xgboost"] # !Requirements for Automated Deep Ensemble with Uncertainty Quantification (AutoDEUQ) REQUIRED_AUTODEUQ = REQUIRED_NAS + ["ray[default]>=1.3.0"] # !Transfer Learning for Bayesian Optimization with SVD REQUIRED_TL_SDV = ["sdv>=0.17.1"] # What packages are optional? EXTRAS = { "autodeuq": REQUIRED_AUTODEUQ, # automated deep ensemble with uncertainty quantification "automl": ["xgboost"], # for automl with scikit-learn "jax-cpu": ["jax[cpu]>=0.3.25", "numpyro[cpu]"], "jax-cuda": ["jax[cuda]>=0.3.25", "numpyro[cuda]"], "hps": [], # hyperparameter search (already the base requirements) "nas": REQUIRED_NAS, # neural architecture search "hps-tl": REQUIRED_TL_SDV, # transfer learning for bayesian optimization, "mpi": ["mpi4py>=3.1.3"], "ray": ["ray[default]>=1.3.0"], "redis": ["redis[hiredis]"], "dev": [ # Test "codecov", "pytest", "pytest-cov", # Packaging "twine", # Formatter and Linter "black==22.6.0", "flake8==5.0.4", "pre-commit", "rstcheck", # Documentation "GitPython", "ipython", "nbsphinx", "Sphinx~=3.5.4", "sphinx-book-theme==0.3.2", "sphinx-copybutton", "sphinx-gallery", "sphinx_lfs_content", "sphinx-togglebutton", ], "analytics": [ "altair", "jupyter", "jupyter_contrib_nbextensions>=0.5.1", "nbconvert<6", "streamlit", "streamlit-aggrid", "tinydb", ], "hvd": ["horovod>=0.21.3", "mpi4py>=3.0.0"], } # Default dependencies for DeepHyper DEFAULT_DEPENDENCIES = REQUIRED[:] DEFAULT_DEPENDENCIES += EXTRAS["nas"] DEFAULT_DEPENDENCIES += EXTRAS["autodeuq"] DEFAULT_DEPENDENCIES += EXTRAS["hps-tl"] DEFAULT_DEPENDENCIES += EXTRAS["jax-cpu"] EXTRAS["default"] = DEFAULT_DEPENDENCIES # Useful commands to build/upload the wheel to PyPI class UploadCommand(Command): """Support setup.py upload.""" description = "Build and publish the package." user_options = [] @staticmethod def status(s): """Prints things in bold.""" print("\033[1m{0}\033[0m".format(s)) def initialize_options(self): pass def finalize_options(self): pass def run(self): try: self.status("Removing previous builds…") rmtree(os.path.join(here, "dist")) except OSError: pass self.status("Building Source and Wheel (universal) distribution…") os.system("{0} setup.py sdist bdist_wheel --universal".format(sys.executable)) self.status("Uploading the package to PyPI via Twine…") os.system("twine upload dist/*") sys.exit() class TestUploadCommand(Command): """Support setup.py testupload.""" description = "Build and publish the package to test.pypi." user_options = [] @staticmethod def status(s): """Prints things in bold.""" print("\033[1m{0}\033[0m".format(s)) def initialize_options(self): pass def finalize_options(self): pass def run(self): try: self.status("Removing previous builds…") rmtree(os.path.join(here, "dist")) except OSError: pass self.status("Building Source and Wheel (universal) distribution…") os.system("{0} setup.py sdist bdist_wheel --universal".format(sys.executable)) self.status("Uploading the package to PyPI via Twine…") os.system("twine upload --repository-url https://test.pypi.org/legacy/ dist/*") sys.exit() class TestInstallCommand(Command): """Support setup.py testinstall""" description = "Install deephyper from TestPyPI." user_options = [] @staticmethod def status(s): """Prints things in bold.""" print("\033[1m{0}\033[0m".format(s)) def initialize_options(self): pass def finalize_options(self): pass def run(self): self.status("Downloading the package from Test PyPI and installing it") os.system("pip install --index-url https://test.pypi.org/simple/ deephyper") sys.exit() # Where the magic happens: setup( install_requires=REQUIRED, extras_require=EXTRAS, cmdclass={ "upload": UploadCommand, "testupload": TestUploadCommand, "testinstall": TestInstallCommand, }, )
5,761
25.552995
93
py
deephyper
deephyper-master/tests/deephyper/nas/test_node.py
import unittest import pytest @pytest.mark.fast @pytest.mark.nas class NodeTest(unittest.TestCase): def test_mirror_node(self): import tensorflow as tf from deephyper.nas.node import MirrorNode, VariableNode from deephyper.nas.operation import operation Dense = operation(tf.keras.layers.Dense) vnode = VariableNode() vop = Dense(10) vnode.add_op(vop) vnode.add_op(Dense(20)) mnode = MirrorNode(vnode) vnode.set_op(0) assert vnode.op == vop assert mnode.op == vop def test_mime_node(self): import tensorflow as tf from deephyper.nas.node import MimeNode, VariableNode from deephyper.nas.operation import operation Dense = operation(tf.keras.layers.Dense) vnode = VariableNode() vop = Dense(10) vnode.add_op(vop) vnode.add_op(Dense(20)) mnode = MimeNode(vnode) mop = Dense(30) mnode.add_op(mop) mnode.add_op(Dense(40)) vnode.set_op(0) assert vnode.op == vop assert mnode.op == mop
1,118
21.836735
63
py
deephyper
deephyper-master/tests/deephyper/nas/test_new_api.py
import pytest @pytest.mark.fast @pytest.mark.nas def test_basic_space(verbose=0): import tensorflow as tf from deephyper.nas import KSearchSpace from deephyper.nas.node import VariableNode, ConstantNode from deephyper.nas.operation import operation, Identity Dense = operation(tf.keras.layers.Dense) class BasicSpace(KSearchSpace): def __init__(self, input_shape, output_shape, batch_size=None, *args, **kwargs): super().__init__( input_shape, output_shape, batch_size=batch_size, *args, **kwargs ) def build(self): input_node = self.input[0] dense = VariableNode() dense.add_op(Identity()) for i in range(1, 1000): dense.add_op(Dense(i)) self.connect(input_node, dense) output_node = ConstantNode(Dense(self.output_shape[0])) self.connect(dense, output_node) space = BasicSpace(input_shape=(1,), output_shape=(1,)) space.build() model_1 = space.sample([1]) if verbose: model_1.summary() model_2 = space.sample() if verbose: model_2.summary()
1,177
24.608696
88
py
deephyper
deephyper-master/tests/deephyper/nas/test_trainer_keras_regressor.py
import unittest import pytest @pytest.mark.slow @pytest.mark.nas class TrainerKerasRegressorTest(unittest.TestCase): def test_trainer_regressor_train_valid_with_one_input(self): import sys from random import random import deephyper.core.utils import numpy as np from deephyper.nas.trainer import BaseTrainer from deephyper.test.nas.linearReg.problem import Problem config = Problem.space config["hyperparameters"]["num_epochs"] = 2 # load functions load_data = deephyper.core.utils.load_attr(config["load_data"]["func"]) config["load_data"]["func"] = load_data config["search_space"]["class"] = deephyper.core.utils.load_attr( config["search_space"]["class"] ) # Loading data kwargs = config["load_data"].get("kwargs") (tX, ty), (vX, vy) = load_data() if kwargs is None else load_data(**kwargs) # Set data shape input_shape = np.shape(tX)[1:] # interested in shape of data not in length output_shape = np.shape(ty)[1:] config["data"] = {"train_X": tX, "train_Y": ty, "valid_X": vX, "valid_Y": vy} search_space = config["search_space"]["class"]( input_shape, output_shape, **config["search_space"]["kwargs"] ).build() arch_seq = [random() for i in range(search_space.num_nodes)] search_space.set_ops(arch_seq) search_space.plot("trainer_keras_regressor_test.dot") if config.get("preprocessing") is not None: preprocessing = deephyper.core.utils.load_attr( config["preprocessing"]["func"] ) config["preprocessing"]["func"] = preprocessing else: config["preprocessing"] = None model = search_space.create_model() trainer = BaseTrainer(config=config, model=model) res = trainer.train() assert res != sys.float_info.max and type(res) is dict def test_trainer_regressor_train_valid_with_multiple_ndarray_inputs(self): import sys from random import random import deephyper.core.utils import numpy as np from deephyper.nas.trainer import BaseTrainer from deephyper.test.nas.linearRegMultiInputs.problem import Problem config = Problem.space config["hyperparameters"]["num_epochs"] = 2 # load functions load_data = deephyper.core.utils.load_attr(config["load_data"]["func"]) config["load_data"]["func"] = load_data config["search_space"]["class"] = deephyper.core.utils.load_attr( config["search_space"]["class"] ) # Loading data kwargs = config["load_data"].get("kwargs") (tX, ty), (vX, vy) = load_data() if kwargs is None else load_data(**kwargs) # Set data shape # interested in shape of data not in length input_shape = [np.shape(itX)[1:] for itX in tX] output_shape = np.shape(ty)[1:] config["data"] = {"train_X": tX, "train_Y": ty, "valid_X": vX, "valid_Y": vy} search_space = config["search_space"]["class"]( input_shape, output_shape, **config["search_space"]["kwargs"] ).build() arch_seq = [random() for i in range(search_space.num_nodes)] search_space.set_ops(arch_seq) search_space.plot("trainer_keras_regressor_test.dot") if config.get("preprocessing") is not None: preprocessing = deephyper.core.utils.load_attr( config["preprocessing"]["func"] ) config["preprocessing"]["func"] = preprocessing else: config["preprocessing"] = None model = search_space.create_model() trainer = BaseTrainer(config=config, model=model) res = trainer.train() assert res != sys.float_info.max and type(res) is dict def test_trainer_regressor_train_valid_with_multiple_generator_inputs(self): import sys from deephyper.nas.run._util import get_search_space, load_config, setup_data from deephyper.nas.trainer import BaseTrainer from deephyper.test.nas.linearReg.problem import Problem from deephyper.test.nas.linearRegMultiInputsGen import Problem config = Problem.space load_config(config) input_shape, output_shape = setup_data(config) search_space = get_search_space(config, input_shape, output_shape, 42) config["hyperparameters"]["num_epochs"] = 2 model = search_space.sample() trainer = BaseTrainer(config=config, model=model) res = trainer.train() assert res != sys.float_info.max and type(res) is dict if __name__ == "__main__": test = TrainerKerasRegressorTest() test.test_trainer_regressor_train_valid_with_multiple_ndarray_inputs()
4,879
33.125874
85
py
deephyper
deephyper-master/tests/deephyper/nas/test_keras_search_space.py
import unittest import pytest @pytest.mark.nas class TestKSearchSpace(unittest.TestCase): def test_create(self): import tensorflow as tf from deephyper.nas import KSearchSpace from deephyper.nas.node import VariableNode from deephyper.nas.operation import operation Dense = operation(tf.keras.layers.Dense) class TestSpace(KSearchSpace): def __init__(self, input_shape, output_shape): super().__init__(input_shape, output_shape) def build(self): vnode = VariableNode() self.connect(self.input_nodes[0], vnode) vnode.add_op(Dense(1)) return self space = TestSpace((5,), (1,)).build() model = space.sample() def test_create_more_nodes(self): import tensorflow as tf from deephyper.nas import KSearchSpace from deephyper.nas.node import VariableNode from deephyper.nas.operation import operation Dense = operation(tf.keras.layers.Dense) class TestSpace(KSearchSpace): def __init__(self, input_shape, output_shape): super().__init__(input_shape, output_shape) def build(self): vnode1 = VariableNode() self.connect(self.input_nodes[0], vnode1) vnode1.add_op(Dense(10)) vnode2 = VariableNode() vnode2.add_op(Dense(1)) self.connect(vnode1, vnode2) return self space = TestSpace((5,), (1,)).build() model = space.sample() def test_create_multiple_inputs_with_one_vnode(self): import tensorflow as tf from deephyper.nas import KSearchSpace from deephyper.nas.node import ConstantNode, VariableNode from deephyper.nas.operation import operation, Concatenate Dense = operation(tf.keras.layers.Dense) class TestSpace(KSearchSpace): def __init__(self, input_shape, output_shape): super().__init__(input_shape, output_shape) def build(self): merge = ConstantNode() merge.set_op(Concatenate(self, self.input_nodes)) vnode1 = VariableNode() self.connect(merge, vnode1) vnode1.add_op(Dense(1)) return self space = TestSpace([(5,), (5,)], (1,)).build() model = space.sample()
2,481
28.2
66
py
deephyper
deephyper-master/tests/deephyper/keras/layers/padding_test.py
import pytest @pytest.mark.fast @pytest.mark.nas def test_padding_layer(): import tensorflow as tf import numpy as np from deephyper.keras.layers import Padding model = tf.keras.Sequential() model.add(Padding([[1, 1]])) data = np.random.random((3, 1)) shape_data = np.shape(data) assert shape_data == (3, 1) res = model.predict(data, batch_size=1) res_shape = np.shape(res) assert res_shape == (3, 3)
450
20.47619
46
py
deephyper
deephyper-master/docs/conf.py
# -*- coding: utf-8 -*- # # Configuration file for the Sphinx documentation builder. # # This file does only contain a selection of the most common options. For a # full list see the documentation: # http://www.sphinx-doc.org/en/master/config # -- Path setup -------------------------------------------------------------- # If extensions (or modules to document with autodoc) are in another directory, # add these directories to sys.path here. If the directory is relative to the # documentation root, use os.path.abspath to make it absolute, like shown here. # import os import sys import git import sphinx_book_theme sys.path.insert(0, os.path.abspath("..")) # -- Project information ----------------------------------------------------- project = "DeepHyper" copyright = "2018-2022, Argonne" author = "Argonne" # The short X.Y version about = {} with open(f"../deephyper/__version__.py") as f: exec(f.read(), about) version = about["__version__"] # The full version, including alpha/beta/rc tags if about["__version__"] == "": release = f'v{about["__version__"]}' else: release = f'v{about["__version__"]}-{about["__version_suffix__"]}' # PULL Tutorials branch_name_map = {"master": "main", "latest": "main", "develop": "develop"} if os.environ.get("READTHEDOCS"): doc_version = os.environ["READTHEDOCS_VERSION"] else: github_repo = git.Repo(search_parent_directories=True) doc_version = github_repo.active_branch.name tutorial_branch = branch_name_map.get(doc_version, "develop") tutorials_github_link = "https://github.com/deephyper/tutorials.git" tutorials_dest_dir = "tutorials" def pull_tutorials(github_link, dest_dir, tutorial_branch): os.system(f"rm -rf {dest_dir}/") os.system( f"git clone --depth=1 --branch={tutorial_branch} {github_link} {dest_dir}" ) os.system(f"rm -rf {dest_dir}/.git") pull_tutorials(tutorials_github_link, tutorials_dest_dir, tutorial_branch) # -- General configuration --------------------------------------------------- # If your documentation needs a minimal Sphinx version, state it here. # # needs_sphinx = '1.0' # Add any Sphinx extension module names here, as strings. They can be # extensions coming with Sphinx (named 'sphinx.ext.*') or your custom # ones. extensions = [ "nbsphinx", "sphinx_book_theme", "sphinx_copybutton", "sphinx_gallery.gen_gallery", "sphinx_lfs_content", "sphinx.ext.autodoc", "sphinx.ext.autosummary", "sphinx.ext.coverage", "sphinx.ext.doctest", "sphinx.ext.githubpages", "sphinx.ext.ifconfig", "sphinx.ext.intersphinx", "sphinx.ext.mathjax", "sphinx.ext.napoleon", "sphinx.ext.todo", "sphinx.ext.viewcode", ] autosummary_generate = True autosummary_imported_members = True # Add any paths that contain templates here, relative to this directory. templates_path = [ "_templates", os.path.join(sphinx_book_theme.get_html_theme_path(), "components"), ] # The suffix(es) of source filenames. # You can specify multiple suffix as a list of string: source_suffix = {".rst": "restructuredtext"} # The master toctree document. master_doc = "index" # The language for content autogenerated by Sphinx. Refer to documentation # for a list of supported languages. # # This is also used if you do content translation via gettext catalogs. # Usually you set "language" from the command line for these cases. language = None # List of patterns, relative to source directory, that match files and # directories to ignore when looking for source files. # This pattern also affects html_static_path and html_extra_path . exclude_patterns = [ "_build", "_templates", "Thumbs.db", ".DS_Store", "examples/*.ipynb", ] # The name of the Pygments (syntax highlighting) style to use. pygments_style = "sphinx" # -- Options for HTML output ------------------------------------------------- # The theme to use for HTML and HTML Help pages. See the documentation for # a list of builtin themes. # html_theme = "sphinx_book_theme" html_theme_path = [sphinx_book_theme.get_html_theme_path()] # Theme options are theme-specific and customize the look and feel of a theme # further. For a list of options available for each theme, see the # documentation. # html_logo = "_static/logo/medium.png" html_theme_options = { # header settings "repository_url": "https://github.com/deephyper/deephyper", "use_repository_button": True, "use_issues_button": True, "use_edit_page_button": True, "repository_branch": "develop", "path_to_docs": "docs", "use_download_button": True, # sidebar settings "show_navbar_depth": 1, "logo_only": True, } # Add any paths that contain custom static files (such as style sheets) here, # relative to this directory. They are copied after the builtin static files, # so a file named "default.css" will overwrite the builtin "default.css". html_static_path = ["_static"] # Custom sidebar templates, must be a dictionary that maps document names # to template names. # # The default sidebars (for documents that don't match any pattern) are # defined by theme itself. Builtin themes are using these templates by # default: ``['localtoc.html', 'relations.html', 'sourcelink.html', # 'searchbox.html']``. # # html_sidebars = {} # -- Options for HTMLHelp output --------------------------------------------- # Output file base name for HTML help builder. htmlhelp_basename = "deephyperdoc" # CopyButton Settings copybutton_prompt_text = r">>> |\.\.\. |\$ |In \[\d*\]: | {2,5}\.\.\.: | {5,8}: " copybutton_prompt_is_regexp = True # -- Options for LaTeX output ------------------------------------------------ latex_elements = { # The paper size ('letterpaper' or 'a4paper'). # # 'papersize': 'letterpaper', # The font size ('10pt', '11pt' or '12pt'). # # 'pointsize': '10pt', # Additional stuff for the LaTeX preamble. # # 'preamble': '', # Latex figure (float) alignment # # 'figure_align': 'htbp', } # Grouping the document tree into LaTeX files. List of tuples # (source start file, target name, title, # author, documentclass [howto, manual, or own class]). latex_documents = [ (master_doc, "deephyper.tex", "deephyper Documentation", "ArgonneMCS", "manual") ] # -- Options for manual page output ------------------------------------------ # One entry per manual page. List of tuples # (source start file, name, description, authors, manual section). man_pages = [(master_doc, "deephyper", "deephyper Documentation", [author], 1)] # -- Options for Texinfo output ---------------------------------------------- # Grouping the document tree into Texinfo files. List of tuples # (source start file, target name, title, author, # dir menu entry, description, category) texinfo_documents = [ ( master_doc, "deephyper", "deephyper Documentation", author, "Automated Machine Learning Software for HPC", "Miscellaneous", ) ] # -- Options for Epub output ------------------------------------------------- # Bibliographic Dublin Core info. epub_title = project epub_author = author epub_publisher = author epub_copyright = copyright # The unique identifier of the text. This can be a ISBN number # or the project homepage. # # epub_identifier = '' # A unique identification for the text. # # epub_uid = '' # A list of files that should not be packed into the epub file. epub_exclude_files = ["search.html"] # -- Extension configuration ------------------------------------------------- # -- Options for intersphinx extension --------------------------------------- # Example configuration for intersphinx: refer to the Python standard library. intersphinx_mapping = {"python": ("https://docs.python.org/3", None)} # -- Options for todo extension ---------------------------------------------- # If true, `todo` and `todoList` produce output, else they produce nothing. todo_include_todos = True # makes sphinx do a mock import of mpi4py so it’s not broken when you try to do auto-docs and import mpi4py autodoc_mock_imports = [ "horovod", "joblib", "matplotlib", "mpi4py", "nbformat", "networkx", "ray", "redis", "sklearn", "skopt", "tensorflow_probability", "tensorflow", "tqdm", "xgboost", ] autosummary_mock_imports = autodoc_mock_imports # Remove <BLANKLINE> trim_doctest_flags = True # Sphinx Gallery sphinx_gallery_conf = { "examples_dirs": "../examples", # path to your example scripts "gallery_dirs": "examples", # path to where to save gallery generated output "filename_pattern": "/plot_", "ignore_pattern": r"_util\.py", } def setup(app): app.add_css_file("custom.css") app.add_js_file("custom.js")
8,809
27.79085
107
py
deephyper
deephyper-master/deephyper/stopper/_lcmodel_stopper.py
import sys from functools import partial import jax import jax.numpy as jnp import numpy as np import numpyro import numpyro.distributions as dist from numpyro.infer import MCMC, NUTS from scipy.optimize import least_squares from sklearn.base import BaseEstimator, RegressorMixin from sklearn.utils import check_random_state from sklearn.utils.validation import check_array, check_is_fitted, check_X_y from deephyper.stopper._stopper import Stopper # Budget allocation models def b_lin2(z, nu=[1, 1]): return nu[1] * (z - 1) + nu[0] def b_exp2(z, nu=[1, 2]): return nu[0] * jnp.power(nu[1], z - 1) # Learning curves models def f_lin2(z, b, rho): return rho[1] * b(z) + rho[0] def f_loglin2(z, b, rho): Z = jnp.log(b(z)) Y = rho[1] * Z + rho[0] y = jnp.exp(Y) return -y def f_loglin3(z, b, rho): Z = jnp.log(b(z)) Y = rho[2] * jnp.power(Z, 2) + rho[1] * Z + rho[0] y = jnp.exp(Y) return -y def f_loglin4(z, b, rho): Z = jnp.log(b(z)) Y = rho[3] * jnp.power(Z, 3) + rho[2] * jnp.power(Z, 2) + rho[1] * Z + rho[0] y = jnp.exp(Y) return -y def f_pow3(z, b, rho): return rho[0] - rho[1] * b(z) ** -rho[2] def f_mmf4(z, b, rho): return (rho[0] * rho[1] + rho[2] * jnp.power(b(z), rho[3])) / ( rho[1] + jnp.power(b(z), rho[3]) ) def f_vapor3(z, b, rho): return rho[0] + rho[1] / b(z) + rho[2] * np.log(b(z)) def f_logloglin2(z, b, rho): return jnp.log(rho[0] * jnp.log(b(z)) + rho[1]) def f_hill3(z, b, rho): ymax, eta, kappa = rho return ymax * (b(z) ** eta) / (kappa * eta + b(z) ** eta) def f_logpow3(z, b, rho): return rho[0] / (1 + (b(z) / jnp.exp(rho[1])) ** rho[2]) def f_pow4(z, b, rho): return rho[2] - (rho[0] * b(z) + rho[1]) ** (-rho[3]) def f_exp4(z, b, rho): return rho[2] - jnp.exp(-rho[0] * (b(z) ** rho[3]) + rho[1]) def f_janoschek4(z, b, rho): return rho[0] - (rho[0] - rho[1]) * jnp.exp(-rho[2] * (b(z) ** rho[3])) def f_weibull4(z, b, rho): return rho[0] - (rho[0] - rho[1]) * jnp.exp(-((rho[2] * b(z)) ** rho[3])) def f_ilog2(z, b, rho): return rho[1] - (rho[0] / jnp.log(b(z) + 1)) # Utility to estimate parameters of learning curve model # The combination of "partial" and "static_argnums" is necessary # with the "f" lambda function passed as argument @partial(jax.jit, static_argnums=(1,)) def residual_least_square(rho, f, z, y): """Residual for least squares.""" return f(z, rho) - y def prob_model(z=None, y=None, f=None, rho_mu_prior=None, num_obs=None): rho_mu_prior = jnp.array(rho_mu_prior) rho_sigma_prior = 1.0 rho = numpyro.sample("rho", dist.Normal(rho_mu_prior, rho_sigma_prior)) sigma = numpyro.sample("sigma", dist.Exponential(1.0)) # introducing noise # sigma = 0.1 mu = f(z[:num_obs], rho) numpyro.sample("obs", dist.Normal(mu, sigma), obs=y[:num_obs]) @partial(jax.jit, static_argnums=(0,)) def predict_moments_from_posterior(f, X, posterior_samples): vf_model = jax.vmap(f, in_axes=(None, 0)) posterior_mu = vf_model(X, posterior_samples) mean_mu = jnp.mean(posterior_mu, axis=0) std_mu = jnp.std(posterior_mu, axis=0) return mean_mu, std_mu class BayesianLearningCurveRegressor(BaseEstimator, RegressorMixin): def __init__( self, f_model=f_loglin3, f_model_num_params=3, b_model=b_lin2, max_trials_ls_fit=10, mcmc_num_warmup=200, mcmc_num_samples=1_000, n_jobs=-1, random_state=None, verbose=0, batch_size=100, ): self.b_model = b_model self.f_model = lambda z, rho: f_model(z, self.b_model, rho) self.f_nparams = f_model_num_params self.mcmc_num_warmup = mcmc_num_warmup self.mcmc_num_samples = mcmc_num_samples self.max_trials_ls_fit = max_trials_ls_fit self.n_jobs = n_jobs self.random_state = check_random_state(random_state) self.verbose = verbose self.rho_mu_prior_ = np.zeros((self.f_nparams,)) self.batch_size = batch_size self.X_ = np.zeros((self.batch_size,)) self.y_ = np.zeros((self.batch_size,)) def fit(self, X, y, update_prior=True): check_X_y(X, y, ensure_2d=False) # !Trick for performance to avoid performign JIT again and again # !This will fix the shape of inputs of the model for numpyro # !see https://github.com/pyro-ppl/numpyro/issues/441 num_samples = len(X) assert num_samples <= self.batch_size self.X_[:num_samples] = X[:] self.y_[:num_samples] = y[:] if update_prior: self.rho_mu_prior_[:] = self._fit_learning_curve_model_least_square(X, y)[:] if not (hasattr(self, "kernel_")): self.kernel_ = NUTS( model=lambda z, y, rho_mu_prior: prob_model( z, y, self.f_model, rho_mu_prior, num_obs=num_samples ), ) self.mcmc_ = MCMC( self.kernel_, num_warmup=self.mcmc_num_warmup, num_samples=self.mcmc_num_samples, progress_bar=self.verbose, jit_model_args=True, ) seed = self.random_state.randint(low=0, high=2**32) rng_key = jax.random.PRNGKey(seed) self.mcmc_.run(rng_key, z=self.X_, y=self.y_, rho_mu_prior=self.rho_mu_prior_) if self.verbose: self.mcmc_.print_summary() return self def predict(self, X, return_std=True): posterior_samples = self.predict_posterior_samples(X) mean_mu = jnp.mean(posterior_samples, axis=0) if return_std: std_mu = jnp.std(posterior_samples, axis=0) return mean_mu, std_mu return mean_mu def predict_posterior_samples(self, X): # Check if fit has been called check_is_fitted(self) # Input validation X = check_array(X, ensure_2d=False) posterior_samples = self.mcmc_.get_samples() vf_model = jax.vmap(self.f_model, in_axes=(None, 0)) posterior_mu = vf_model(X, posterior_samples["rho"]) return posterior_mu def prob(self, X, condition): """Compute the approximate probability of P(cond(m(X_i), y_i)) where m is the current fitted model and cond a condition. Args: X (np.array): An array of inputs. condition (callable): A function defining the condition to test. Returns: array: an array of shape X. """ # Check if fit has been called check_is_fitted(self) # Input validation X = check_array(X, ensure_2d=False) posterior_samples = self.mcmc_.get_samples() vf_model = jax.vmap(self.f_model, in_axes=(None, 0)) posterior_mu = vf_model(X, posterior_samples["rho"]) prob = jnp.mean(condition(posterior_mu), axis=0) return prob def _fit_learning_curve_model_least_square( self, z_train, y_train, ): """The learning curve model is assumed to be modeled by 'f' with interface f(z, rho). """ seed = self.random_state.randint(low=0, high=2**32) random_state = check_random_state(seed) z_train = np.asarray(z_train) y_train = np.asarray(y_train) # compute the jacobian # using the true jacobian is important to avoid problems # with numerical errors and approximations! indeed the scale matters # a lot when approximating with finite differences def fun_wrapper(rho, f, z, y): return np.array(residual_least_square(rho, f, z, y)) if not (hasattr(self, "jac_residual_ls_")): self.jac_residual_ls_ = partial(jax.jit, static_argnums=(1,))( jax.jacfwd(residual_least_square, argnums=0) ) def jac_wrapper(rho, f, z, y): return np.array(self.jac_residual_ls_(rho, f, z, y)) results = [] mse_hist = [] for _ in range(self.max_trials_ls_fit): rho_init = random_state.randn(self.f_nparams) try: res_lsq = least_squares( fun_wrapper, rho_init, args=(self.f_model, z_train, y_train), method="lm", jac=jac_wrapper, ) except ValueError: continue mse_res_lsq = np.mean(res_lsq.fun**2) mse_hist.append(mse_res_lsq) results.append(res_lsq.x) i_best = np.nanargmin(mse_hist) res = results[i_best] return res def area_learning_curve(z, f, z_max) -> float: assert len(z) == len(f) assert z[-1] <= z_max area = 0 for i in range(1, len(z)): # z: is always monotinic increasing but not f! area += (z[i] - z[i - 1]) * f[i - 1] if z[-1] < z_max: area += (z_max - z[-1]) * f[-1] return area class LCModelStopper(Stopper): """Stopper based on learning curve extrapolation (LCE) to evaluate if the iterations of the learning algorithm should be stopped. The LCE is based on a parametric learning curve model (LCM) which is modeling the score as a function of the number of training steps. Training steps can correspond to the number of training epochs, the number of training batches, the number of observed samples or any other quantity that is iterated through during the training process. The LCE is based on the following steps: 1. An early stopping condition is always checked first. If the early stopping condition is met, the LCE is not applied. 2. Then, some safeguard conditions are checked to ensure that the LCE can be applied (number of observed steps must be greater or equal to the number of parameters of the LCM). 3. If the LCM cannot be fitted (number of observed steps is less than number of parameters of the model), then the last observed step is compared to hitorical performance of others at the same step to check if it is a low-performing outlier (outlier in the direction of performing worse!) using the IQR criterion. 4. If the LCM can be fitted, a least square fit is performed to estimate the parameters of the LCM. 5. The probability of the current LC to perform worse than the best observed score at the maximum iteration is computed using Monte-Carlo Markov Chain (MCMC). To use this stopper, you need to install the following dependencies: .. code-block:: bash $ jax>=0.3.25 $ numpyro Args: max_steps (int): The maximum number of training steps which can be performed. min_steps (int, optional): The minimum number of training steps which can be performed. Defaults to ``1``. lc_model (str, optional): The parameteric learning model to use. It should be a string in the following list: ``["lin2", "loglin2", "loglin3", "loglin4", "pow3","mmf4", "vapor3", "logloglin2", "hill3", "logpow3", "pow4", "exp4", "janoschek4", "weibull4", "ilog2"]``. The number in the name corresponds to the number of parameters of the parametric model. Defaults to ``"mmf4"``. min_done_for_outlier_detection (int, optional): The minimum number of observed scores at the same step to check for if it is a lower-bound outlier. Defaults to ``10``. iqr_factor_for_outlier_detection (float, optional): The IQR factor for outlier detection. The higher it is the more inclusive the condition will be (i.e. if set very large it is likely not going to detect any outliers). Defaults to ``1.5``. prob_promotion (float, optional): The threshold probabily to stop the iterations. If the current learning curve has a probability greater than ``prob_promotion`` to be worse that the best observed score accross all evaluations then the current iterations are stopped. Defaults to ``0.9`` (i.e. probability of 0.9 of being worse). early_stopping_patience (float, optional): The patience of the early stopping condition. If it is an ``int`` it is directly corresponding to a number of iterations. If it is a ``float`` then it is corresponding to a proportion between [0,1] w.r.t. ``max_steps``. Defaults to ``0.25`` (i.e. 25% of ``max_steps``). objective_returned (str, optional): The returned objective. It can be a value in ``["last", "max", "alc"]`` where ``"last"`` corresponds to the last observed score, ``"max"`` corresponds to the maximum observed score and ``"alc"`` corresponds to the area under the learning curve. Defaults to "last". random_state (int or np.RandomState, optional): The random state of estimation process. Defaults to ``None``. Raises: ValueError: parameters are not valid. """ def __init__( self, max_steps: int, min_steps: int = 1, lc_model="mmf4", min_done_for_outlier_detection=10, iqr_factor_for_outlier_detection=1.5, prob_promotion=0.9, early_stopping_patience=0.25, objective_returned="last", random_state=None, ) -> None: super().__init__(max_steps=max_steps) self.min_steps = min_steps lc_model = "f_" + lc_model lc_model_num_params = int(lc_model[-1]) lc_model = getattr(sys.modules[__name__], lc_model) self.min_obs_to_fit = lc_model_num_params self.min_done_for_outlier_detection = min_done_for_outlier_detection self.iqr_factor_for_outlier_detection = iqr_factor_for_outlier_detection self.prob_promotion = prob_promotion if type(early_stopping_patience) is int: self.early_stopping_patience = early_stopping_patience elif type(early_stopping_patience) is float: self.early_stopping_patience = int(early_stopping_patience * self.max_steps) else: raise ValueError("early_stopping_patience must be int or float") self.objective_returned = objective_returned self._rung = 0 # compute the step at which to stop based on steps allocation policy max_rung = np.floor( np.log(self.max_steps / self.min_steps) / np.log(self.min_obs_to_fit) ) self.max_steps_ = int(self.min_steps * self.min_obs_to_fit**max_rung) self.lc_model = BayesianLearningCurveRegressor( f_model=lc_model, f_model_num_params=lc_model_num_params, random_state=random_state, batch_size=self.max_steps_, ) self._lc_objectives = [] def _compute_halting_step(self): return self.min_steps * self.min_obs_to_fit**self._rung def _retrieve_best_objective(self) -> float: search_id, _ = self.job.id.split(".") objectives = [] for obj in self.job.storage.load_out_from_all_jobs(search_id): try: objectives.append(float(obj)) except ValueError: pass if len(objectives) > 0: return np.max(objectives) else: return np.max(self.observations[1]) def _get_competiting_objectives(self, rung) -> list: search_id, _ = self.job.id.split(".") values = self.job.storage.load_metadata_from_all_jobs( search_id, f"_completed_rung_{rung}" ) values = [float(v) for v in values] return values def observe(self, budget: float, objective: float): super().observe(budget, objective) self._budget = self.observed_budgets[-1] self._lc_objectives.append(self.objective) self._objective = self._lc_objectives[-1] # For Early-Stopping based on Patience if ( not (hasattr(self, "_local_best_objective")) or self._objective > self._local_best_objective ): self._local_best_objective = self._objective self._local_best_step = self.step halting_step = self._compute_halting_step() if self._budget >= halting_step: self.job.storage.store_job_metadata( self.job.id, f"_completed_rung_{self._rung}", str(self._objective) ) def stop(self) -> bool: # Enforce Pre-conditions Before Learning-Curve based Early Discarding if super().stop(): print("Stopped after reaching the maximum number of steps.") self.infos_stopped = "max steps reached" return True if self.step - self._local_best_step >= self.early_stopping_patience: print( f"Stopped after reaching {self.early_stopping_patience} steps without improvement." ) self.infos_stopped = "early stopping" return True # This condition will enforce the stopper to stop the evaluation at the first step # for the first evaluation (The FABOLAS method does the same, bias the first samples with # small budgets) self.best_objective = self._retrieve_best_objective() halting_step = self._compute_halting_step() if self.step < max(self.min_steps, self.min_obs_to_fit): if self.step >= halting_step: competing_objectives = self._get_competiting_objectives(self._rung) if len(competing_objectives) > self.min_done_for_outlier_detection: q1 = np.quantile( competing_objectives, q=0.25, ) q3 = np.quantile( competing_objectives, q=0.75, ) iqr = q3 - q1 # lower than the minimum of a box plot if ( self._objective < q1 - self.iqr_factor_for_outlier_detection * iqr ): print( f"Stopped early because of abnormally low objective: {self._objective}" ) self.infos_stopped = "outlier" return True self._rung += 1 return False # Check if the halting budget condition is met if self.step < halting_step: return False # Check if the evaluation should be stopped based on LC-Model # Fit and predict the performance of the learning curve model z_train = self.observed_budgets y_train = self._lc_objectives z_train, y_train = np.asarray(z_train), np.asarray(y_train) self.lc_model.fit(z_train, y_train, update_prior=True) # Check if the configuration is promotable based on its predicted objective value p = self.lc_model.prob( X=[self.max_steps], condition=lambda y_hat: y_hat <= self.best_objective )[0] # Return whether the configuration should be stopped if p <= self.prob_promotion: self._rung += 1 else: print( f"Stopped because the probability of performing worse is {p} > {self.prob_promotion}" ) self.infos_stopped = f"prob={p:.3f}" return True @property def objective(self): if self.objective_returned == "last": return self.observations[-1][-1] elif self.objective_returned == "max": return max(self.observations[-1]) elif self.objective_returned == "alc": z, y = self.observations return area_learning_curve(z, y, z_max=self.max_steps) else: raise ValueError("objective_returned must be one of 'last', 'best', 'alc'")
19,786
36.263653
401
py
deephyper
deephyper-master/deephyper/test/nas/linearRegHybrid/problem.py
from deephyper.nas.spacelib.tabular import OneLayerSpace from deephyper.problem import NaProblem from deephyper.test.nas.linearReg.load_data import load_data Problem = NaProblem() Problem.load_data(load_data) Problem.search_space(OneLayerSpace) Problem.hyperparameters( batch_size=Problem.add_hyperparameter((1, 100), "batch_size"), learning_rate=Problem.add_hyperparameter( (1e-4, 1e-1, "log-uniform"), "learning_rate" ), optimizer=Problem.add_hyperparameter(["adam", "nadam", "rmsprop"], "optimizer"), num_epochs=1, ) Problem.loss("mse") Problem.metrics(["r2"]) Problem.objective("val_r2") # Just to print your problem, to test its definition and imports in the current python environment. if __name__ == "__main__": print(Problem) model = Problem.get_keras_model([1])
817
24.5625
99
py
deephyper
deephyper-master/deephyper/test/nas/linearReg/problem.py
from deephyper.nas.spacelib.tabular import OneLayerSpace from deephyper.problem import NaProblem from deephyper.test.nas.linearReg.load_data import load_data Problem = NaProblem() Problem.load_data(load_data) Problem.search_space(OneLayerSpace) Problem.hyperparameters( batch_size=100, learning_rate=0.1, optimizer="adam", num_epochs=1 ) Problem.loss("mse") Problem.metrics(["r2"]) Problem.objective("val_r2") # Just to print your problem, to test its definition and imports in the current python environment. if __name__ == "__main__": print(Problem) model = Problem.get_keras_model([1])
611
21.666667
99
py
deephyper
deephyper-master/deephyper/nas/_nx_search_space.py
import abc import traceback from collections.abc import Iterable import networkx as nx from deephyper.core.exceptions.nas.space import ( NodeAlreadyAdded, StructureHasACycle, WrongSequenceToSetOperations, ) from deephyper.nas.node import MimeNode, Node, VariableNode class NxSearchSpace(abc.ABC): """A NxSearchSpace is an search_space based on a networkx graph.""" def __init__(self, seed=None, **kwargs): self.graph = nx.DiGraph() self.seed = seed self.output_node = None def plot(self, path): with open(path, "w") as f: try: nx.nx_agraph.write_dot(self.graph, f) except Exception: print("Error: can't create graphviz file...") traceback.print_exc() def __len__(self): """Number of VariableNodes in the current search_space. Returns: int: number of variable nodes in the current search_space. """ return len(self.nodes) @property def nodes(self): """Nodes of the current KSearchSpace. Returns: iterator: nodes of the current KSearchSpace. """ return list(self.graph.nodes) def add_node(self, node): """Add a new node to the search_space. Args: node (Node): node to add to the search_space. Raises: TypeError: if 'node' is not an instance of Node. NodeAlreadyAdded: if 'node' has already been added to the search_space. """ if not isinstance(node, Node): raise TypeError("'node' argument should be an instance of Node!") if node in self.nodes: raise NodeAlreadyAdded(node) self.graph.add_node(node) def connect(self, node1, node2): """Create a new connection in the KSearchSpace graph. The edge created corresponds to : node1 -> node2. Args: node1 (Node) node2 (Node) Raise: StructureHasACycle: if the new edge is creating a cycle. """ assert isinstance(node1, Node) assert isinstance(node2, Node) self.graph.add_edge(node1, node2) if not (nx.is_directed_acyclic_graph(self.graph)): raise StructureHasACycle( f"the connection between {node1} -> {node2} is creating a cycle in the search_space's graph." ) @property def size(self): """Size of the search space define by the search_space""" s = 0 for n in filter(lambda n: isinstance(n, VariableNode), self.nodes): if n.num_ops != 0: if s == 0: s = n.num_ops else: s *= n.num_ops return s @property def max_num_ops(self): """Returns the maximum number of operations accross all VariableNodes of the struct. Returns: int: maximum number of Operations for a VariableNode in the current Structure. """ return max(map(lambda n: n.num_ops, self.variable_nodes)) @property def num_nodes(self): """Returns the number of VariableNodes in the current Structure. Returns: int: number of VariableNodes in the current Structure. """ return len(list(self.variable_nodes)) @property def variable_nodes(self): """Iterator of VariableNodes of the search_space. Returns: (Iterator(VariableNode)): generator of VariablesNodes of the search_space. """ return filter(lambda n: isinstance(n, VariableNode), self.nodes) @property def mime_nodes(self): """Iterator of MimeNodes of the search_space. Returns: (Iterator(MimeNode)): iterator of MimeNodes of the search_space. """ return filter(lambda n: isinstance(n, MimeNode), self.nodes) def denormalize(self, indexes): """Denormalize a sequence of normalized indexes to get a sequence of absolute indexes. Useful when you want to compare the number of different search_spaces. Args: indexes (Iterable): a sequence of normalized indexes. Returns: list: A list of absolute indexes corresponding to operations choosen with relative indexes of `indexes`. """ assert isinstance( indexes, Iterable ), 'Wrong argument, "indexes" should be of Iterable.' if len(indexes) != self.num_nodes: raise WrongSequenceToSetOperations(indexes, list(self.variable_nodes)) return [ vnode.denormalize(op_i) for op_i, vnode in zip(indexes, self.variable_nodes) ] def set_output_node(self): """Set the output node of the search_space. :meta private: """ if self.output_node is None: nodes = list(self.graph.nodes()) self.output_node = [] for n in nodes: if len(list(self.graph.successors(n))) == 0: self.output_node.append(n) if len(self.output_node) == 1: self.output_node = self.output_node[0] def create_tensor_aux(self, g, n, train=None): """Recursive function to create the tensors from the graph. :meta private: Args: g (nx.DiGraph): a graph n (nx.Node): a node train (bool): True if the network is built for training, False if the network is built for validation/testing (for example False will deactivate Dropout). Return: the tensor represented by n. """ try: if n._tensor is not None: output_tensor = n._tensor else: pred = list(g.predecessors(n)) if len(pred) == 0: output_tensor = n.create_tensor(train=train, seed=self.seed) else: tensor_list = list() for s_i in pred: tmp = self.create_tensor_aux(g, s_i, train=train) if type(tmp) is list: tensor_list.extend(tmp) else: tensor_list.append(tmp) output_tensor = n.create_tensor( tensor_list, train=train, seed=self.seed ) return output_tensor except TypeError: raise RuntimeError(f"Failed to build tensors from :{n}") @abc.abstractmethod def choices(self): """Gives the possible choices for each decision variable of the search space. Returns: list: A list of tuple where each element corresponds to a discrete variable represented by ``(low, high)``. """ @abc.abstractmethod def sample(self, choice=None): """Sample a ``tf.keras.Model`` from the search space. Args: choice (list, optional): A list of decision for the operations of this search space. Defaults to None, will generate a random sample. Returns: tf.keras.Model: A Tensorflow Keras model. """ @abc.abstractmethod def build(self): """Build the current graph search space."""
7,301
30.747826
166
py
deephyper
deephyper-master/deephyper/nas/losses.py
"""This module provides different loss functions. A loss can be defined by a keyword (str) or a callable following the ``tensorflow.keras`` interface. If it is a keyword it has to be available in ``tensorflow.keras`` or in ``deephyper.losses``. The loss functions availble in ``deephyper.losses`` are: * Negative Log Likelihood (compatible with Tensorflow Probability): ``tfp_negloglik`` or ``tfp_nll`` """ from collections import OrderedDict import tensorflow as tf from deephyper.core.utils import load_attr def tfp_negloglik(y, rv_y): """Negative log likelihood for Tensorflow probability.""" return -rv_y.log_prob(y) losses_func = OrderedDict() losses_func["tfp_negloglik"] = losses_func["tfp_nll"] = tfp_negloglik losses_obj = OrderedDict() def selectLoss(name: str): """Return the loss defined by name. Args: name (str): a string referenced in DeepHyper, one referenced in keras or an attribute name to import. Returns: str or callable: a string suppossing it is referenced in the keras framework or a callable taking (y_true, y_pred) as inputs and returning a tensor. """ if callable(name): return name if losses_func.get(name) is None and losses_obj.get(name) is None: try: loaded_obj = load_attr(name) return loaded_obj except Exception: return tf.keras.losses.get( name ) # supposing it is referenced in keras losses else: if name in losses_func: return losses_func[name] else: return losses_obj[name]()
1,605
34.688889
301
py
deephyper
deephyper-master/deephyper/nas/node.py
"""This module provides the available node types to build a ``KSearchSpace``. """ import tensorflow as tf import deephyper.core.exceptions from deephyper.nas.operation import Operation class Node: """Represents a node of a ``KSearchSpace``. Args: name (str): node name. """ # Number of 'Node' instances created num = 0 def __init__(self, name="", *args, **kwargs): Node.num += 1 self._num = Node.num self._tensor = None self.name = name def __str__(self): return f"{self.name}[id={self._num}]" @property def id(self): return self._num @property def op(self): raise NotImplementedError def create_tensor(self, *args, **kwargs): raise NotImplementedError @staticmethod def verify_operation(op): if isinstance(op, Operation): return op elif isinstance(op, tf.keras.layers.Layer): return Operation(op) else: raise RuntimeError( f"Can't add this operation '{op.__name__}'. An operation should be either of type Operation or tf.keras.layers.Layer when is of type: {type(op)}" ) class OperationNode(Node): def __init__(self, name="", *args, **kwargs): super().__init__(name=name, *args, **kwargs) def create_tensor(self, inputs=None, train=True, seed=None, **kwargs): if self._tensor is None: if inputs is None: try: self._tensor = self.op(train=train, seed=None) except TypeError: raise RuntimeError( f'Verify if node: "{self}" has incoming connexions!' ) else: self._tensor = self.op(inputs, train=train) return self._tensor class VariableNode(OperationNode): """This class represents a node of a graph where you have a set of possible operations. It means the agent will have to act to choose one of these operations. >>> import tensorflow as tf >>> from deephyper.nas.space.node import VariableNode >>> vnode = VariableNode("VNode1") >>> from deephyper.nas.space.op.op1d import Dense >>> vnode.add_op(Dense( ... units=10, ... activation=tf.nn.relu)) >>> vnode.num_ops 1 >>> vnode.add_op(Dense( ... units=1000, ... activation=tf.nn.tanh)) >>> vnode.num_ops 2 >>> vnode.set_op(0) >>> vnode.op.units 10 Args: name (str): node name. """ def __init__(self, name=""): super().__init__(name=name) self._ops = list() self._index = None def __str__(self): if self._index is not None: return f"{super().__str__()}(Variable[{str(self.op)}])" else: return f"{super().__str__()}(Variable[?])" def add_op(self, op): self._ops.append(self.verify_operation(op)) @property def num_ops(self): return len(self._ops) def set_op(self, index): self.get_op(index).init(self) def get_op(self, index): assert "float" in str(type(index)) or "int" in str( type(index) ), f"found type is : {type(index)}" if "float" in str(type(index)): self._index = self.denormalize(index) else: assert 0 <= index and index < len( self._ops ), f"Number of possible operations is: {len(self._ops)}, but index given is: {index} (index starts from 0)!" self._index = index return self.op def denormalize(self, index): """Denormalize a normalized index to get an absolute indexes. Useful when you want to compare the number of different search_spaces. Args: indexes (float|int): a normalized index. Returns: int: An absolute indexes corresponding to the operation choosen with the relative index of `index`. """ if type(index) is int: return index else: assert 0.0 <= index and index <= 1.0 res = int(index * len(self._ops)) if index == 1.0: res -= 1 return res @property def op(self): if len(self._ops) == 0: raise RuntimeError("This VariableNode doesn't have any operation yet.") elif self._index is None: raise RuntimeError( 'This VariableNode doesn\'t have any set operation, please use "set_op(index)" if you want to set one' ) else: return self._ops[self._index] @property def ops(self): return self._ops class ConstantNode(OperationNode): """A ConstantNode represents a node with a fixed operation. It means the agent will not make any new decision for this node. The common use case for this node is to add a tensor in the graph. >>> import tensorflow as tf >>> from deephyper.nas.space.node import ConstantNode >>> from deephyper.nas.space.op.op1d import Dense >>> cnode = ConstantNode(op=Dense(units=100, activation=tf.nn.relu), name='CNode1') >>> cnode.op Dense_100_relu Args: op (Operation, optional): operation to fix for this node. Defaults to None. name (str, optional): node name. Defaults to ``''``. """ def __init__(self, op=None, name="", *args, **kwargs): super().__init__(name=name) if op is not None: op = self.verify_operation(op) op.init(self) # set operation self._op = op def set_op(self, op): op = self.verify_operation(op) op.init(self) self._op = op def __str__(self): return f"{super().__str__()}(Constant[{str(self.op)}])" @property def op(self): return self._op class MirrorNode(OperationNode): """A MirrorNode is a node which reuse an other, it enable the reuse of tf.keras layers. This node will not add operations to choose. Args: node (Node): The targeted node to mirror. >>> from deephyper.nas.space.node import VariableNode, MirrorNode >>> from deephyper.nas.space.op.op1d import Dense >>> vnode = VariableNode() >>> vnode.add_op(Dense(10)) >>> vnode.add_op(Dense(20)) >>> mnode = MirrorNode(vnode) >>> vnode.set_op(0) >>> vnode.op Dense_10 >>> mnode.op Dense_10 """ def __init__(self, node): super().__init__(name=f"Mirror[{str(node)}]") self._node = node @property def op(self): return self._node.op class MimeNode(OperationNode): """A MimeNode is a node which reuse an the choice made for an VariableNode, it enable the definition of a Cell based search_space. This node reuse the operation from the mimed VariableNode but only the choice made. Args: node (VariableNode): the VariableNode to mime. >>> from deephyper.nas.space.node import VariableNode, MimeNode >>> from deephyper.nas.space.op.op1d import Dense >>> vnode = VariableNode() >>> vnode.add_op(Dense(10)) >>> vnode.add_op(Dense(20)) >>> mnode = MimeNode(vnode) >>> mnode.add_op(Dense(30)) >>> mnode.add_op(Dense(40)) >>> vnode.set_op(0) >>> vnode.op Dense_10 >>> mnode.op Dense_30 """ def __init__(self, node, name=""): super().__init__(name=f"Mime[{name}][src={str(node)}]") self.node = node self._ops = list() def add_op(self, op): self._ops.append(self.verify_operation(op)) @property def num_ops(self): return len(self._ops) def set_op(self): if self.node._index is None: raise deephyper.core.exceptions.DeephyperRuntimeError( f"{str(self)} cannot be initialized because its source {str(self.node)} is not initialized!" ) self._ops[self.node._index].init(self) @property def op(self): if self.num_ops != self.node.num_ops: raise deephyper.core.exceptions.DeephyperRuntimeError( f"{str(self)} and {str(self.node)} should have the same number of opertions, when {str(self)} has {self.num_ops} and {str(self.node)} has {self.node.num_ops}!" ) else: return self._ops[self.node._index] @property def ops(self): return self._ops
8,363
29.086331
218
py
deephyper
deephyper-master/deephyper/nas/_keras_search_space.py
import copy import logging import warnings import networkx as nx import numpy as np import tensorflow as tf from deephyper.core.exceptions.nas.space import ( InputShapeOfWrongType, WrongSequenceToSetOperations, ) from deephyper.nas._nx_search_space import NxSearchSpace from deephyper.nas.node import ConstantNode from deephyper.nas.operation import Tensor from tensorflow import keras from tensorflow.python.keras.utils.vis_utils import model_to_dot logger = logging.getLogger(__name__) class KSearchSpace(NxSearchSpace): """A KSearchSpace represents a search space of neural networks. >>> import tensorflow as tf >>> from deephyper.nas import KSearchSpace >>> from deephyper.nas.node import ConstantNode, VariableNode >>> from deephyper.nas.operation import operation, Identity >>> Dense = operation(tf.keras.layers.Dense) >>> Dropout = operation(tf.keras.layers.Dropout) >>> class ExampleSpace(KSearchSpace): ... def build(self): ... # input nodes are automatically built based on `input_shape` ... input_node = self.input_nodes[0] ... # we want 4 layers maximum (Identity corresponds to not adding a layer) ... for i in range(4): ... node = VariableNode() ... self.connect(input_node, node) ... # we add 3 possible operations for each node ... node.add_op(Identity()) ... node.add_op(Dense(100, "relu")) ... node.add_op(Dropout(0.2)) ... input_node = node ... output = ConstantNode(op=Dense(self.output_shape[0])) ... self.connect(input_node, output) ... return self ... >>> >>> space = ExampleSpace(input_shape=(1,), output_shape=(1,)).build() >>> space.sample().summary() Args: input_shape (list(tuple(int))): list of shapes of all inputs. output_shape (tuple(int)): shape of output. batch_size (list(tuple(int))): batch size of the input layer. If ``input_shape`` is defining a list of inputs, ``batch_size`` should also define a list of inputs. Raises: InputShapeOfWrongType: [description] """ def __init__( self, input_shape, output_shape, batch_size=None, seed=None, *args, **kwargs ): super().__init__() self._random = np.random.RandomState(seed) self.input_shape = input_shape if type(input_shape) is tuple: # we have only one input tensor here op = Tensor( keras.layers.Input(input_shape, name="input_0", batch_size=batch_size) ) self.input_nodes = [ConstantNode(op=op, name="Input_0")] elif type(input_shape) is list and all( map(lambda x: type(x) is tuple, input_shape) ): # we have a list of input tensors here self.input_nodes = list() for i in range(len(input_shape)): batch_size = batch_size[i] if type(batch_size) is list else None op = Tensor( keras.layers.Input( input_shape[i], name=f"input_{i}", batch_size=batch_size ) ) inode = ConstantNode(op=op, name=f"Input_{i}") self.input_nodes.append(inode) else: raise InputShapeOfWrongType(input_shape) for node in self.input_nodes: self.graph.add_node(node) self.output_shape = output_shape self.output_node = None self._model = None @property def input(self): return self.input_nodes @property def output(self): return self.output_node @property def depth(self): if self._model is None: raise RuntimeError("Can't compute depth of model without creating a model.") return len(self.longest_path) @property def longest_path(self): if self._model is None: raise RuntimeError( "Can't compute longest path of model without creating a model." ) nx_graph = nx.drawing.nx_pydot.from_pydot(model_to_dot(self._model)) return nx.algorithms.dag.dag_longest_path(nx_graph) def set_ops(self, indexes): """Set the operations for each node of each cell of the search_space. :meta private: Args: indexes (list): element of list can be float in [0, 1] or int. Raises: WrongSequenceToSetOperations: raised when 'indexes' is of a wrong length. """ if len(indexes) != len(list(self.variable_nodes)): raise WrongSequenceToSetOperations(indexes, list(self.variable_nodes)) for op_i, node in zip(indexes, self.variable_nodes): node.set_op(op_i) for node in self.mime_nodes: node.set_op() self.set_output_node() def create_model(self): """Create the tensors corresponding to the search_space. :meta private: Returns: A keras.Model for the current search_space with the corresponding set of operations. """ # !the output layer does not have to be of the same shape as the data # !this depends on the loss if type(self.output_node) is list: output_tensors = [ self.create_tensor_aux(self.graph, out) for out in self.output_node ] for out_T in output_tensors: output_n = int(out_T.name.split("/")[0].split("_")[-1]) out_S = self.output_shape[output_n] if tf.keras.backend.is_keras_tensor(out_T): out_T_shape = out_T.type_spec.shape if out_T_shape[1:] != out_S: warnings.warn( f"The output tensor of shape {out_T_shape} doesn't match the expected shape {out_S}!", RuntimeWarning, ) input_tensors = [inode._tensor for inode in self.input_nodes] self._model = keras.Model(inputs=input_tensors, outputs=output_tensors) else: output_tensors = self.create_tensor_aux(self.graph, self.output_node) if tf.keras.backend.is_keras_tensor(output_tensors): output_tensors_shape = output_tensors.type_spec.shape if output_tensors_shape[1:] != self.output_shape: warnings.warn( f"The output tensor of shape {output_tensors_shape} doesn't match the expected shape {self.output_shape}!", RuntimeWarning, ) input_tensors = [inode._tensor for inode in self.input_nodes] self._model = keras.Model(inputs=input_tensors, outputs=[output_tensors]) return self._model def choices(self): """Gives the possible choices for each decision variable of the search space. Returns: list: A list of tuple where each element corresponds to a discrete variable represented by ``(low, high)``. """ return [(0, vnode.num_ops - 1) for vnode in self.variable_nodes] def sample(self, choice=None): """Sample a ``tf.keras.Model`` from the search space. Args: choice (list, optional): A list of decision for the operations of this search space. Defaults to None, will generate a random sample. Returns: tf.keras.Model: A Tensorflow Keras model. """ if choice is None: choice = [self._random.randint(c[0], c[1] + 1) for c in self.choices()] self_copy = copy.deepcopy(self) self_copy.set_ops(choice) model = self_copy.create_model() return model
7,837
34.789954
170
py
deephyper
deephyper-master/deephyper/nas/metrics.py
"""This module provides different metric functions. A metric can be defined by a keyword (str) or a callable. If it is a keyword it has to be available in ``tensorflow.keras`` or in ``deephyper.netrics``. The loss functions availble in ``deephyper.metrics`` are: * Sparse Perplexity: ``sparse_perplexity`` * R2: ``r2`` * AUC ROC: ``auroc`` * AUC Precision-Recall: ``aucpr`` """ import functools from collections import OrderedDict import tensorflow as tf from deephyper.core.utils import load_attr def r2(y_true, y_pred): SS_res = tf.math.reduce_sum(tf.math.square(y_true - y_pred), axis=0) SS_tot = tf.math.reduce_sum( tf.math.square(y_true - tf.math.reduce_mean(y_true, axis=0)), axis=0 ) output_scores = 1 - SS_res / (SS_tot + tf.keras.backend.epsilon()) r2 = tf.math.reduce_mean(output_scores) return r2 def mae(y_true, y_pred): return tf.keras.metrics.mean_absolute_error(y_true, y_pred) def mse(y_true, y_pred): return tf.keras.metrics.mean_squared_error(y_true, y_pred) def rmse(y_true, y_pred): return tf.math.sqrt(tf.math.reduce_mean(tf.math.square(y_pred - y_true))) def acc(y_true, y_pred): return tf.keras.metrics.categorical_accuracy(y_true, y_pred) def sparse_perplexity(y_true, y_pred): cross_entropy = tf.keras.losses.sparse_categorical_crossentropy(y_true, y_pred) perplexity = tf.pow(2.0, cross_entropy) return perplexity def to_tfp(metric_func): """Convert a regular tensorflow-keras metric for tensorflow probability where the output is a distribution. Args: metric_func (func): A regular tensorflow-keras metric function. """ @functools.wraps(metric_func) def wrapper(y_true, y_pred): return metric_func(y_true, y_pred.mean()) wrapper.__name__ = f"tfp_{metric_func.__name__}" return wrapper # convert some metrics for Tensorflow Probability where the output of the model is # a distribution tfp_r2 = to_tfp(r2) tfp_mae = to_tfp(mae) tfp_mse = to_tfp(mse) tfp_rmse = to_tfp(rmse) metrics_func = OrderedDict() metrics_func["mean_absolute_error"] = metrics_func["mae"] = mae metrics_func["r2"] = r2 metrics_func["mean_squared_error"] = metrics_func["mse"] = mse metrics_func["root_mean_squared_error"] = metrics_func["rmse"] = rmse metrics_func["accuracy"] = metrics_func["acc"] = acc metrics_func["sparse_perplexity"] = sparse_perplexity metrics_func["tfp_r2"] = tfp_r2 metrics_func["tfp_mse"] = tfp_mse metrics_func["tfp_mae"] = tfp_mae metrics_func["tfp_rmse"] = tfp_rmse metrics_obj = OrderedDict() metrics_obj["auroc"] = lambda: tf.keras.metrics.AUC(name="auroc", curve="ROC") metrics_obj["aucpr"] = lambda: tf.keras.metrics.AUC(name="aucpr", curve="PR") def selectMetric(name: str): """Return the metric defined by name. Args: name (str): a string referenced in DeepHyper, one referenced in keras or an attribute name to import. Returns: str or callable: a string suppossing it is referenced in the keras framework or a callable taking (y_true, y_pred) as inputs and returning a tensor. """ if callable(name): return name if metrics_func.get(name) is None and metrics_obj.get(name) is None: try: return load_attr(name) except Exception: return name # supposing it is referenced in keras metrics else: if name in metrics_func: return metrics_func[name] else: return metrics_obj[name]()
3,474
31.175926
262
py
deephyper
deephyper-master/deephyper/nas/__init__.py
from ._nx_search_space import NxSearchSpace from ._keras_search_space import KSearchSpace __all__ = ["NxSearchSpace", "KSearchSpace"]
135
26.2
45
py
deephyper
deephyper-master/deephyper/nas/trainer/_utils.py
from collections import OrderedDict import tensorflow as tf optimizers_keras = OrderedDict() optimizers_keras["sgd"] = tf.keras.optimizers.SGD optimizers_keras["rmsprop"] = tf.keras.optimizers.RMSprop optimizers_keras["adagrad"] = tf.keras.optimizers.Adagrad optimizers_keras["adam"] = tf.keras.optimizers.Adam optimizers_keras["adadelta"] = tf.keras.optimizers.Adadelta optimizers_keras["adamax"] = tf.keras.optimizers.Adamax optimizers_keras["nadam"] = tf.keras.optimizers.Nadam def selectOptimizer_keras(name): """Return the optimizer defined by name.""" if optimizers_keras.get(name) is None: raise RuntimeError('"{0}" is not a defined optimizer for keras.'.format(name)) else: return optimizers_keras[name] def check_data_config(data_dict): gen_keys = ["train_gen", "train_size", "valid_gen", "valid_size", "types", "shapes"] ndarray_keys = ["train_X", "train_Y", "valid_X", "valid_Y"] if all([k in data_dict.keys() for k in gen_keys]): return "gen" elif all([k in data_dict.keys() for k in ndarray_keys]): return "ndarray" else: raise RuntimeError("Wrong data config...")
1,156
35.15625
88
py
deephyper
deephyper-master/deephyper/nas/trainer/_horovod.py
import logging import time from inspect import signature import deephyper.nas.trainer._arch as a import deephyper.nas.trainer._utils as U import horovod.tensorflow.keras as hvd import numpy as np import tensorflow as tf from deephyper.core.exceptions import DeephyperRuntimeError from deephyper.nas.losses import selectLoss from deephyper.nas.metrics import selectMetric logger = logging.getLogger(__name__) AUTOTUNE = tf.data.experimental.AUTOTUNE class HorovodTrainer: def __init__(self, config, model): self.cname = self.__class__.__name__ self.config = config self.model = model self.callbacks = [] self.data = self.config[a.data] # hyperparameters self.config_hp = self.config[a.hyperparameters] self.optimizer_name = self.config_hp.get(a.optimizer, "adam") self.optimizer_eps = self.config_hp.get("epsilon", None) self.batch_size = self.config_hp.get(a.batch_size, 32) self.clipvalue = self.config_hp.get("clipvalue", None) self.learning_rate = self.config_hp.get(a.learning_rate, 1e-3) # augmentation strategy if not self.config.get("augment", None) is None: if not self.config["augment"].get("kwargs", None) is None: self.augment_func = lambda inputs, outputs: self.config["augment"][ "func" ](inputs, outputs, **self.config["augment"]["kwargs"]) else: self.augment_func = self.config["augment"]["func"] self.num_epochs = self.config_hp[a.num_epochs] self.verbose = ( self.config_hp.get("verbose", 1) if self.config_hp.get("verbose", 1) and hvd.rank() == 0 else 0 ) self.setup_losses_and_metrics() # DATA loading self.data_config_type = None self.train_size = None self.valid_size = None self.train_steps_per_epoch = None self.valid_steps_per_epoch = None self.load_data() # DATA preprocessing self.preprocessing_func = None if self.config.get("preprocessing"): self.preprocessing_func = self.config["preprocessing"]["func"] self.preprocessor = None self.preprocess_data() # Dataset self.dataset_train = None self.set_dataset_train() self.dataset_valid = None self.set_dataset_valid() self.model_compile() self.train_history = None self.init_history() # Test on validation after each epoch if self.verbose == 1: logger.info("KerasTrainer instantiated") model.summary(print_fn=logger.info) def init_history(self): self.train_history = dict() self.train_history["n_parameters"] = self.model.count_params() def setup_losses_and_metrics(self): def selectL(loss): if type(loss) is dict: loss = {k: selectLoss(v) for k, v in loss.items()} else: loss = selectLoss(loss) return loss self.loss_metrics = selectL(self.config[a.loss_metric]) self.loss_weights = self.config.get("loss_weights") self.class_weights = self.config.get("class_weights") if self.loss_weights is None and type(self.loss_metrics) is dict: self.loss_weights = [1.0 for _ in range(len(self.loss_metrics))] if type(self.config[a.metrics]) is list: self.metrics_name = [selectMetric(m) for m in self.config[a.metrics]] else: def selectM(metric): if type(metric) is list: return [selectMetric(m_i) for m_i in metric] else: return selectMetric(metric) self.metrics_name = { n: selectM(m) for n, m in self.config[a.metrics].items() } def load_data(self): logger.debug("load_data") self.data_config_type = U.check_data_config(self.data) logger.debug(f"data config type: {self.data_config_type}") if self.data_config_type == "gen": self.load_data_generator() elif self.data_config_type == "ndarray": self.load_data_ndarray() else: raise DeephyperRuntimeError( f"Data config is not supported by this Trainer: '{self.data_config_type}'!" ) # prepare number of steps for training and validation self.train_steps_per_epoch = self.train_size // self.batch_size if self.train_steps_per_epoch * self.batch_size < self.train_size: self.train_steps_per_epoch += 1 self.valid_steps_per_epoch = self.valid_size // self.batch_size if self.valid_steps_per_epoch * self.batch_size < self.valid_size: self.valid_steps_per_epoch += 1 self.train_steps_per_epoch //= hvd.size() self.valid_steps_per_epoch //= hvd.size() def load_data_generator(self): self.train_gen = self.data["train_gen"] self.valid_gen = self.data["valid_gen"] self.data_types = self.data["types"] self.data_shapes = self.data["shapes"] self.train_size = self.data["train_size"] self.valid_size = self.data["valid_size"] def load_data_ndarray(self): def f(x): return type(x) is np.ndarray # check data type # Output data if ( type(self.config[a.data][a.train_Y]) is np.ndarray and type(self.config[a.data][a.valid_Y]) is np.ndarray ): self.train_Y = self.config[a.data][a.train_Y] self.valid_Y = self.config[a.data][a.valid_Y] elif ( type(self.config[a.data][a.train_Y]) is list and type(self.config[a.data][a.valid_Y]) is list ): if not all(map(f, self.config[a.data][a.train_Y])) or not all( map(f, self.config[a.data][a.valid_Y]) ): raise DeephyperRuntimeError( "all outputs data should be of type np.ndarray !" ) if ( len(self.config[a.data][a.train_Y]) > 1 and len(self.config[a.data][a.valid_Y]) > 1 ): self.train_Y = self.config[a.data][a.train_Y] self.valid_Y = self.config[a.data][a.valid_Y] else: self.train_Y = self.config[a.data][a.train_Y][0] self.valid_Y = self.config[a.data][a.valid_Y][0] else: raise DeephyperRuntimeError( f"Data are of an unsupported type and should be of same type: type(self.config['data']['train_Y'])=={type(self.config[a.data][a.train_Y])} and type(self.config['data']['valid_Y'])=={type(self.config[a.valid_Y][a.valid_X])} !" ) # Input data if ( type(self.config[a.data][a.train_X]) is np.ndarray and type(self.config[a.data][a.valid_X]) is np.ndarray ): self.train_X = [self.config[a.data][a.train_X]] self.valid_X = [self.config[a.data][a.valid_X]] elif ( type(self.config[a.data][a.train_X]) is list and type(self.config[a.data][a.valid_X]) is list ): if not all(map(f, self.config[a.data][a.train_X])) or not all( map(f, self.config[a.data][a.valid_X]) ): raise DeephyperRuntimeError( "all inputs data should be of type np.ndarray !" ) if ( len(self.config[a.data][a.train_X]) > 1 and len(self.config[a.data][a.valid_X]) > 1 ): self.train_X = self.config[a.data][a.train_X] self.valid_X = self.config[a.data][a.valid_X] else: self.train_X = self.config[a.data][a.train_X][0] self.valid_X = self.config[a.data][a.valid_X][0] else: raise DeephyperRuntimeError( f"Data are of an unsupported type and should be of same type: type(self.config['data']['train_X'])=={type(self.config[a.data][a.train_X])} and type(self.config['data']['valid_X'])=={type(self.config[a.data][a.valid_X])} !" ) # check data length self.train_size = np.shape(self.train_X[0])[0] if not all(map(lambda x: np.shape(x)[0] == self.train_size, self.train_X)): raise DeephyperRuntimeError( "All training inputs data should have same length!" ) self.valid_size = np.shape(self.valid_X[0])[0] if not all(map(lambda x: np.shape(x)[0] == self.valid_size, self.valid_X)): raise DeephyperRuntimeError( "All validation inputs data should have same length!" ) def preprocess_data(self): if self.data_config_type == "gen": return if self.preprocessor is not None: raise DeephyperRuntimeError("You can only preprocess data one time.") if self.preprocessing_func: logger.debug(f"preprocess_data with: {str(self.preprocessing_func)}") if len(np.shape(self.train_Y)) == 2: data_train = np.concatenate((*self.train_X, self.train_Y), axis=1) data_valid = np.concatenate((*self.valid_X, self.valid_Y), axis=1) data = np.concatenate((data_train, data_valid), axis=0) self.preprocessor = self.preprocessing_func() dt_shp = np.shape(data_train) tX_shp = [np.shape(x) for x in self.train_X] preproc_data = self.preprocessor.fit_transform(data) acc, self.train_X = 0, list() for shp in tX_shp: self.train_X.append(preproc_data[: dt_shp[0], acc : acc + shp[1]]) acc += shp[1] self.train_Y = preproc_data[: dt_shp[0], acc:] acc, self.valid_X = 0, list() for shp in tX_shp: self.valid_X.append(preproc_data[dt_shp[0] :, acc : acc + shp[1]]) acc += shp[1] self.valid_Y = preproc_data[dt_shp[0] :, acc:] else: logger.info("no preprocessing function") def set_dataset_train(self): if self.data_config_type == "ndarray": if type(self.train_Y) is list: output_mapping = { f"output_{i}": tY for i, tY in enumerate(self.train_Y) } else: output_mapping = self.train_Y self.dataset_train = tf.data.Dataset.from_tensor_slices( ( {f"input_{i}": tX for i, tX in enumerate(self.train_X)}, output_mapping, ) ) else: # self.data_config_type == "gen" self.dataset_train = tf.data.Dataset.from_generator( self.train_gen, output_types=self.data_types, output_shapes=( { f"input_{i}": tf.TensorShape( [*self.data_shapes[0][f"input_{i}"]] ) for i in range(len(self.data_shapes[0])) }, tf.TensorShape([*self.data_shapes[1]]), ), ) self.dataset_train = self.dataset_train.shard( num_shards=hvd.size(), index=hvd.rank() ) self.dataset_train = self.dataset_train.shuffle( self.train_size // hvd.size(), reshuffle_each_iteration=True ) self.dataset_train = self.dataset_train.repeat(self.num_epochs) if hasattr(self, "augment_func"): logger.info("Data augmentation set.") self.dataset_train = self.dataset_train.map( self.augment_func, num_parallel_calls=AUTOTUNE ) self.dataset_train = self.dataset_train.batch(self.batch_size) self.dataset_train = self.dataset_train.prefetch(AUTOTUNE) # self.dataset_train = self.dataset_train.repeat() def set_dataset_valid(self): if self.data_config_type == "ndarray": if type(self.valid_Y) is list: output_mapping = { f"output_{i}": vY for i, vY in enumerate(self.valid_Y) } else: output_mapping = self.valid_Y self.dataset_valid = tf.data.Dataset.from_tensor_slices( ( {f"input_{i}": vX for i, vX in enumerate(self.valid_X)}, output_mapping, ) ) else: self.dataset_valid = tf.data.Dataset.from_generator( self.valid_gen, output_types=self.data_types, output_shapes=( { f"input_{i}": tf.TensorShape( [*self.data_shapes[0][f"input_{i}"]] ) for i in range(len(self.data_shapes[0])) }, tf.TensorShape([*self.data_shapes[1]]), ), ) self.dataset_valid = self.dataset_valid.batch(self.batch_size).repeat() def model_compile(self): optimizer_fn = U.selectOptimizer_keras(self.optimizer_name) opti_parameters = signature(optimizer_fn).parameters params = {} # "lr" and "learning_rate" is checked depending if Keras or Tensorflow optimizer is used if "lr" in opti_parameters: params["lr"] = self.learning_rate elif "learning_rate" in opti_parameters: params["learning_rate"] = self.learning_rate else: raise DeephyperRuntimeError( f"The learning_rate parameter is not found amoung optimiser arguments: {opti_parameters}" ) if "epsilon" in opti_parameters: params["epsilon"] = self.optimizer_eps if self.clipvalue is not None: params["clipvalue"] = self.clipvalue self.optimizer = hvd.DistributedOptimizer(optimizer_fn(**params)) if type(self.loss_metrics) is dict: self.model.compile( optimizer=self.optimizer, loss=self.loss_metrics, loss_weights=self.loss_weights, metrics=self.metrics_name, ) else: self.model.compile( optimizer=self.optimizer, loss=self.loss_metrics, metrics=self.metrics_name, ) def predict(self, dataset: str = "valid", keep_normalize: bool = False) -> tuple: """[summary] Args: dataset (str, optional): 'valid' or 'train'. Defaults to 'valid'. keep_normalize (bool, optional): if False then the preprocessing will be reversed after prediction. if True nothing will be reversed. Defaults to False. Raises: DeephyperRuntimeError: [description] Returns: tuple: (y_true, y_pred) """ if not (dataset == "valid" or dataset == "train"): raise DeephyperRuntimeError( "dataset parameter should be equal to: 'valid' or 'train'" ) if dataset == "valid": valid_steps = self.valid_size // self.batch_size if valid_steps * self.batch_size < self.valid_size: valid_steps += 1 y_pred = self.model.predict(self.dataset_valid, steps=valid_steps) else: # dataset == 'train' y_pred = self.model.predict( self.dataset_train, steps=self.train_steps_per_epoch ) if ( self.preprocessing_func and not keep_normalize and not self.data_config_type == "gen" ): if dataset == "valid": data_X, data_Y = self.valid_X, self.valid_Y else: # dataset == 'train' data_X, data_Y = self.train_X, self.train_Y val_pred = np.concatenate((*data_X, y_pred), axis=1) val_orig = np.concatenate((*data_X, data_Y), axis=1) val_pred_trans = self.preprocessor.inverse_transform(val_pred) val_orig_trans = self.preprocessor.inverse_transform(val_orig) y_orig = val_orig_trans[:, -np.shape(data_Y)[1] :] y_pred = val_pred_trans[:, -np.shape(data_Y)[1] :] else: if self.data_config_type == "ndarray": y_orig = self.valid_Y if dataset == "valid" else self.train_Y else: gen = self.valid_gen() if dataset == "valid" else self.train_gen() y_orig = np.array([e[-1] for e in gen]) return y_orig, y_pred def evaluate(self, dataset="train"): """Evaluate the performance of your model for the same configuration. Args: dataset (str, optional): must be "train" or "valid". If "train" then metrics will be evaluated on the training dataset. If "valid" then metrics will be evaluated on the "validation" dataset. Defaults to 'train'. Returns: list: a list of scalar values corresponding do config loss & metrics. """ if dataset == "train": return self.model.evaluate( self.dataset_train, steps=self.train_steps_per_epoch ) else: return self.model.evaluate( self.dataset_valid, steps=self.valid_steps_per_epoch ) def train( self, num_epochs: int = None, with_pred: bool = False, last_only: bool = False ): """Train the model. Args: num_epochs (int, optional): override the num_epochs passed to init the Trainer. Defaults to None, will use the num_epochs passed to init the Trainer. with_pred (bool, optional): will compute a prediction after the training and will add ('y_true', 'y_pred') to the output history. Defaults to False, will skip it (use it to save compute time). last_only (bool, optional): will compute metrics after the last epoch only. Defaults to False, will compute metrics after each training epoch (use it to save compute time). Raises: DeephyperRuntimeError: raised when the ``num_epochs < 0``. Returns: dict: a dictionnary corresponding to the training. """ num_epochs = self.num_epochs if num_epochs is None else num_epochs self.init_history() if num_epochs > 0: time_start_training = time.time() # TIMING if not last_only: logger.info( "Trainer is computing metrics on validation after each training epoch." ) history = self.model.fit( self.dataset_train, verbose=self.verbose, epochs=num_epochs, steps_per_epoch=self.train_steps_per_epoch, callbacks=self.callbacks, validation_data=self.dataset_valid, validation_steps=self.valid_steps_per_epoch, ) else: logger.info( "Trainer is computing metrics on validation after the last training epoch." ) if num_epochs > 1: self.model.fit( self.dataset_train, verbose=self.verbose, epochs=num_epochs - 1, steps_per_epoch=self.train_steps_per_epoch, callbacks=self.callbacks, ) history = self.model.fit( self.dataset_train, epochs=1, verbose=self.verbose, steps_per_epoch=self.train_steps_per_epoch, callbacks=self.callbacks, validation_data=self.dataset_valid, validation_steps=self.valid_steps_per_epoch, ) time_end_training = time.time() # TIMING self.train_history["training_time"] = ( time_end_training - time_start_training ) self.train_history.update(history.history) elif num_epochs < 0: raise DeephyperRuntimeError( f"Trainer: number of epochs should be >= 0: {num_epochs}" ) if with_pred: time_start_predict = time.time() y_true, y_pred = self.predict(dataset="valid") time_end_predict = time.time() self.train_history["val_predict_time"] = ( time_end_predict - time_start_predict ) self.train_history["y_true"] = y_true self.train_history["y_pred"] = y_pred return self.train_history
21,070
37.733456
241
py
deephyper
deephyper-master/deephyper/nas/trainer/_base.py
import inspect import logging import time from inspect import signature import deephyper.nas.trainer._arch as a import deephyper.nas.trainer._utils as U import numpy as np import tensorflow as tf from deephyper.core.exceptions import DeephyperRuntimeError from deephyper.nas.losses import selectLoss from deephyper.nas.metrics import selectMetric logger = logging.getLogger(__name__) class BaseTrainer: def __init__(self, config, model): self.cname = self.__class__.__name__ self.config = config self.model = model self.callbacks = [] self.data = self.config[a.data] self.config_hp = self.config[a.hyperparameters] self.optimizer_name = self.config_hp.get(a.optimizer, "adam") self.optimizer_eps = self.config_hp.get("epsilon", None) self.batch_size = self.config_hp.get(a.batch_size, 32) self.learning_rate = self.config_hp.get(a.learning_rate, 1e-3) self.num_epochs = self.config_hp.get(a.num_epochs, 1) self.shuffle_data = self.config_hp.get(a.shuffle_data, True) self.cache_data = self.config_hp.get(a.cache_data, True) self.batch = self.config_hp.get("batch", True) self.momentum = self.config_hp.get("momentum", 0.0) self.nesterov = self.config_hp.get("nesterov", False) self.label_smoothing = self.config_hp.get("label_smoothing", 0.0) self.verbose = self.config_hp.get("verbose", 1) # self.balanced = self.config_hp.get("balanced", False) self.setup_losses_and_metrics() # DATA loading self.data_config_type = None self.train_size = None self.valid_size = None self.train_steps_per_epoch = None self.valid_steps_per_epoch = None self.load_data() # DATA preprocessing self.preprocessing_func = None if self.config.get("preprocessing"): self.preprocessing_func = self.config["preprocessing"]["func"] self.preprocessor = None self.preprocess_data() # Dataset self.dataset_train = None self.set_dataset_train() self.dataset_valid = None self.set_dataset_valid() self.model_compile() self.train_history = None self.init_history() # Test on validation after each epoch if self.verbose == 1: logger.info("KerasTrainer instantiated") model.summary(print_fn=logger.info) def init_history(self): self.train_history = dict() self.train_history["n_parameters"] = self.model.count_params() def _select_loss(self, loss): if type(loss) is dict: loss = {k: selectLoss(v) for k, v in loss.items()} else: loss = selectLoss(loss) if inspect.isclass(loss): loss_parameters = signature(loss).parameters params = {} if "label_smoothing" in loss_parameters: params["label_smoothing"] = self.label_smoothing loss = loss(**params) return loss def setup_losses_and_metrics(self): self.loss_metrics = self._select_loss(self.config[a.loss_metric]) self.loss_weights = self.config.get("loss_weights") self.class_weights = self.config.get("class_weights") if self.loss_weights is None and type(self.loss_metrics) is dict: self.loss_weights = [1.0 for _ in range(len(self.loss_metrics))] if type(self.config[a.metrics]) is list: self.metrics_name = [selectMetric(m) for m in self.config[a.metrics]] else: def selectM(metric): if type(metric) is list: return [selectMetric(m_i) for m_i in metric] else: return selectMetric(metric) self.metrics_name = { n: selectM(m) for n, m in self.config[a.metrics].items() } def load_data(self): logger.debug("load_data") self.data_config_type = U.check_data_config(self.data) logger.debug(f"data config type: {self.data_config_type}") if self.data_config_type == "gen": self.load_data_generator() elif self.data_config_type == "ndarray": self.load_data_ndarray() else: raise DeephyperRuntimeError( f"Data config is not supported by this Trainer: '{self.data_config_type}'!" ) # prepare number of steps for training and validation self.train_steps_per_epoch = self.train_size // self.batch_size if self.train_steps_per_epoch * self.batch_size < self.train_size: self.train_steps_per_epoch += 1 self.valid_steps_per_epoch = self.valid_size // self.batch_size if self.valid_steps_per_epoch * self.batch_size < self.valid_size: self.valid_steps_per_epoch += 1 def load_data_generator(self): self.train_gen = self.data["train_gen"] self.valid_gen = self.data["valid_gen"] self.data_types = self.data["types"] self.data_shapes = self.data["shapes"] self.train_size = self.data["train_size"] self.valid_size = self.data["valid_size"] def load_data_ndarray(self): def f(x): return type(x) is np.ndarray # check data type # Output data if ( type(self.config[a.data][a.train_Y]) is np.ndarray and type(self.config[a.data][a.valid_Y]) is np.ndarray ): self.train_Y = self.config[a.data][a.train_Y] self.valid_Y = self.config[a.data][a.valid_Y] elif ( type(self.config[a.data][a.train_Y]) is list and type(self.config[a.data][a.valid_Y]) is list ): if not all(map(f, self.config[a.data][a.train_Y])) or not all( map(f, self.config[a.data][a.valid_Y]) ): raise DeephyperRuntimeError( "all outputs data should be of type np.ndarray !" ) if ( len(self.config[a.data][a.train_Y]) > 1 and len(self.config[a.data][a.valid_Y]) > 1 ): self.train_Y = self.config[a.data][a.train_Y] self.valid_Y = self.config[a.data][a.valid_Y] else: self.train_Y = self.config[a.data][a.train_Y][0] self.valid_Y = self.config[a.data][a.valid_Y][0] else: raise DeephyperRuntimeError( f"Data are of an unsupported type and should be of same type: type(self.config['data']['train_Y'])=={type(self.config[a.data][a.train_Y])} and type(self.config['data']['valid_Y'])=={type(self.config[a.valid_Y][a.valid_X])} !" ) # Input data if ( type(self.config[a.data][a.train_X]) is np.ndarray and type(self.config[a.data][a.valid_X]) is np.ndarray ): self.train_X = [self.config[a.data][a.train_X]] self.valid_X = [self.config[a.data][a.valid_X]] elif ( type(self.config[a.data][a.train_X]) is list and type(self.config[a.data][a.valid_X]) is list ): if not all(map(f, self.config[a.data][a.train_X])) or not all( map(f, self.config[a.data][a.valid_X]) ): raise DeephyperRuntimeError( "all inputs data should be of type np.ndarray !" ) if ( len(self.config[a.data][a.train_X]) > 1 and len(self.config[a.data][a.valid_X]) > 1 ): self.train_X = self.config[a.data][a.train_X] self.valid_X = self.config[a.data][a.valid_X] else: self.train_X = self.config[a.data][a.train_X][0] self.valid_X = self.config[a.data][a.valid_X][0] else: raise DeephyperRuntimeError( f"Data are of an unsupported type and should be of same type: type(self.config['data']['train_X'])=={type(self.config[a.data][a.train_X])} and type(self.config['data']['valid_X'])=={type(self.config[a.data][a.valid_X])} !" ) logger.debug(f"{self.cname}: {len(self.train_X)} inputs") # check data length self.train_size = np.shape(self.train_X[0])[0] if not all(map(lambda x: np.shape(x)[0] == self.train_size, self.train_X)): raise DeephyperRuntimeError( "All training inputs data should have same length!" ) self.valid_size = np.shape(self.valid_X[0])[0] if not all(map(lambda x: np.shape(x)[0] == self.valid_size, self.valid_X)): raise DeephyperRuntimeError( "All validation inputs data should have same length!" ) def preprocess_data(self): logger.debug("Starting preprocess of data") if self.data_config_type == "gen": logger.warn("Cannot preprocess data with generator!") return if self.preprocessor is not None: raise DeephyperRuntimeError("You can only preprocess data one time.") if self.preprocessing_func: logger.debug(f"preprocess_data with: {str(self.preprocessing_func)}") if all( [ len(np.shape(tX)) == len(np.shape(self.train_Y)) for tX in self.train_X ] ): data_train = np.concatenate((*self.train_X, self.train_Y), axis=-1) data_valid = np.concatenate((*self.valid_X, self.valid_Y), axis=-1) self.preprocessor = self.preprocessing_func() tX_shp = [np.shape(x) for x in self.train_X] preproc_data_train = self.preprocessor.fit_transform(data_train) preproc_data_valid = self.preprocessor.transform(data_valid) acc, self.train_X = 0, list() for shp in tX_shp: self.train_X.append(preproc_data_train[..., acc : acc + shp[1]]) acc += shp[1] self.train_Y = preproc_data_train[..., acc:] acc, self.valid_X = 0, list() for shp in tX_shp: self.valid_X.append(preproc_data_valid[..., acc : acc + shp[1]]) acc += shp[1] self.valid_Y = preproc_data_valid[..., acc:] else: logger.warn( f"Skipped preprocess because shape {np.shape(self.train_Y)} is not handled!" ) else: logger.info("Skipped preprocess of data because no function is defined!") def set_dataset_train(self): if self.data_config_type == "ndarray": if type(self.train_Y) is list: output_mapping = { f"output_{i}": tY for i, tY in enumerate(self.train_Y) } else: output_mapping = self.train_Y self.dataset_train = tf.data.Dataset.from_tensor_slices( ( {f"input_{i}": tX for i, tX in enumerate(self.train_X)}, output_mapping, ) ) else: # self.data_config_type == "gen" self.dataset_train = tf.data.Dataset.from_generator( self.train_gen, output_signature=self._get_output_signatures(), ) if self.cache_data: self.dataset_train = self.dataset_train.cache() if self.shuffle_data: self.dataset_train = self.dataset_train.shuffle( self.train_size, reshuffle_each_iteration=True ) if self.batch: self.dataset_train = self.dataset_train.batch(self.batch_size) self.dataset_train = self.dataset_train.prefetch(tf.data.AUTOTUNE).repeat( self.num_epochs ) def set_dataset_valid(self): if self.data_config_type == "ndarray": if type(self.valid_Y) is list: output_mapping = { f"output_{i}": vY for i, vY in enumerate(self.valid_Y) } else: output_mapping = self.valid_Y self.dataset_valid = tf.data.Dataset.from_tensor_slices( ( {f"input_{i}": vX for i, vX in enumerate(self.valid_X)}, output_mapping, ) ) else: self.dataset_valid = tf.data.Dataset.from_generator( self.valid_gen, output_signature=self._get_output_signatures(valid=True), ) self.dataset_valid = self.dataset_valid.cache() self.dataset_valid = self.dataset_valid.batch(self.batch_size) self.dataset_valid = self.dataset_valid.prefetch(tf.data.AUTOTUNE).repeat( self.num_epochs ) def _get_output_signatures(self, valid=False): if self.batch or valid: return ( { f"input_{i}": tf.TensorSpec( shape=(*self.data_shapes[0][f"input_{i}"],), dtype=self.data_types[0][f"input_{i}"], ) for i in range(len(self.data_shapes[0])) }, tf.TensorSpec( shape=(*self.data_shapes[1],), dtype=self.data_types[1], ), ) else: return ( { f"input_{i}": tf.TensorSpec( shape=( None, *self.data_shapes[0][f"input_{i}"], ), dtype=self.data_types[0][f"input_{i}"], ) for i in range(len(self.data_shapes[0])) }, tf.TensorSpec( shape=(None, *self.data_shapes[1]), dtype=self.data_types[1], ), ) def _setup_optimizer(self): optimizer_fn = U.selectOptimizer_keras(self.optimizer_name) opti_parameters = signature(optimizer_fn).parameters params = {} if "lr" in opti_parameters: params["lr"] = self.learning_rate elif "learning_rate" in opti_parameters: params["learning_rate"] = self.learning_rate else: raise DeephyperRuntimeError( f"The learning_rate parameter is not found amoung optimiser arguments: {opti_parameters}" ) if "epsilon" in opti_parameters: params["epsilon"] = self.optimizer_eps if "momentum" in opti_parameters: params["momentum"] = self.momentum if "nesterov" in opti_parameters: params["nesterov"] = self.nesterov self.optimizer = optimizer_fn(**params) def model_compile(self): self._setup_optimizer() if type(self.loss_metrics) is dict: self.model.compile( optimizer=self.optimizer, loss=self.loss_metrics, loss_weights=self.loss_weights, metrics=self.metrics_name, ) else: self.model.compile( optimizer=self.optimizer, loss=self.loss_metrics, metrics=self.metrics_name, ) def predict(self, dataset: str = "valid", keep_normalize: bool = False) -> tuple: """[summary] Args: dataset (str, optional): 'valid' or 'train'. Defaults to 'valid'. keep_normalize (bool, optional): if False then the preprocessing will be reversed after prediction. if True nothing will be reversed. Defaults to False. Raises: DeephyperRuntimeError: [description] Returns: tuple: (y_true, y_pred) """ if not (dataset == "valid" or dataset == "train"): raise DeephyperRuntimeError( "dataset parameter should be equal to: 'valid' or 'train'" ) if dataset == "valid": y_pred = self.model.predict( self.dataset_valid, steps=self.valid_steps_per_epoch ) else: # dataset == 'train' y_pred = self.model.predict( self.dataset_train, steps=self.train_steps_per_epoch ) if ( self.preprocessing_func and not keep_normalize and not self.data_config_type == "gen" ): if dataset == "valid": data_X, data_Y = self.valid_X, self.valid_Y else: # dataset == 'train' data_X, data_Y = self.train_X, self.train_Y val_pred = np.concatenate((*data_X, y_pred), axis=1) val_orig = np.concatenate((*data_X, data_Y), axis=1) val_pred_trans = self.preprocessor.inverse_transform(val_pred) val_orig_trans = self.preprocessor.inverse_transform(val_orig) y_orig = val_orig_trans[:, -np.shape(data_Y)[1] :] y_pred = val_pred_trans[:, -np.shape(data_Y)[1] :] else: if self.data_config_type == "ndarray": y_orig = self.valid_Y if dataset == "valid" else self.train_Y else: gen = self.valid_gen() if dataset == "valid" else self.train_gen() y_orig = np.array([e[-1] for e in gen]) return y_orig, y_pred def evaluate(self, dataset="train"): """Evaluate the performance of your model for the same configuration. Args: dataset (str, optional): must be "train" or "valid". If "train" then metrics will be evaluated on the training dataset. If "valid" then metrics will be evaluated on the "validation" dataset. Defaults to 'train'. Returns: list: a list of scalar values corresponding do config loss & metrics. """ if dataset == "train": return self.model.evaluate( self.dataset_train, steps=self.train_steps_per_epoch ) else: return self.model.evaluate( self.dataset_valid, steps=self.valid_steps_per_epoch ) def train( self, num_epochs: int = None, with_pred: bool = False, last_only: bool = False ): """Train the model. Args: num_epochs (int, optional): override the num_epochs passed to init the Trainer. Defaults to None, will use the num_epochs passed to init the Trainer. with_pred (bool, optional): will compute a prediction after the training and will add ('y_true', 'y_pred') to the output history. Defaults to False, will skip it (use it to save compute time). last_only (bool, optional): will compute metrics after the last epoch only. Defaults to False, will compute metrics after each training epoch (use it to save compute time). Raises: DeephyperRuntimeError: raised when the ``num_epochs < 0``. Returns: dict: a dictionnary corresponding to the training. """ num_epochs = self.num_epochs if num_epochs is None else num_epochs self.init_history() if num_epochs > 0: time_start_training = time.time() # TIMING if not last_only: logger.info( "Trainer is computing metrics on validation after each training epoch." ) history = self.model.fit( self.dataset_train, verbose=self.verbose, epochs=num_epochs, steps_per_epoch=self.train_steps_per_epoch, callbacks=self.callbacks, validation_data=self.dataset_valid, validation_steps=self.valid_steps_per_epoch, class_weight=self.class_weights, ) else: logger.info( "Trainer is computing metrics on validation after the last training epoch." ) if num_epochs > 1: self.model.fit( self.dataset_train, verbose=self.verbose, epochs=num_epochs - 1, steps_per_epoch=self.train_steps_per_epoch, callbacks=self.callbacks, class_weight=self.class_weights, ) history = self.model.fit( self.dataset_train, epochs=1, verbose=self.verbose, steps_per_epoch=self.train_steps_per_epoch, callbacks=self.callbacks, validation_data=self.dataset_valid, validation_steps=self.valid_steps_per_epoch, class_weight=self.class_weights, ) time_end_training = time.time() # TIMING self.train_history["training_time"] = ( time_end_training - time_start_training ) self.train_history.update(history.history) elif num_epochs < 0: raise DeephyperRuntimeError( f"Trainer: number of epochs should be >= 0: {num_epochs}" ) if with_pred: time_start_predict = time.time() y_true, y_pred = self.predict(dataset="valid") time_end_predict = time.time() self.train_history["val_predict_time"] = ( time_end_predict - time_start_predict ) self.train_history["y_true"] = y_true self.train_history["y_pred"] = y_pred return self.train_history
21,923
37.0625
241
py
deephyper
deephyper-master/deephyper/nas/spacelib/tabular/one_layer.py
import tensorflow as tf from deephyper.nas import KSearchSpace from deephyper.nas.node import ConstantNode, VariableNode from deephyper.nas.operation import operation, Concatenate Dense = operation(tf.keras.layers.Dense) Dropout = operation(tf.keras.layers.Dropout) class OneLayerSpace(KSearchSpace): def __init__( self, input_shape, output_shape, batch_size=None, seed=None, regression=True ): super().__init__(input_shape, output_shape, batch_size=batch_size, seed=seed) self.regression = regression def build(self): if type(self.input_shape) is list: vnodes = [] for i in range(len(self.input_shape)): vn = self.gen_vnode() vnodes.append(vn) self.connect(self.input_nodes[i], vn) print(i) prev_node = ConstantNode(Concatenate(self, vnodes)) else: prev_node = self.gen_vnode() self.connect(self.input_nodes[0], prev_node) output_node = ConstantNode( Dense( self.output_shape[0], activation=None if self.regression else "softmax" ) ) self.connect(prev_node, output_node) return self def gen_vnode(self) -> VariableNode: vnode = VariableNode() for i in range(1, 1000): vnode.add_op(Dense(i, tf.nn.relu)) return vnode if __name__ == "__main__": from tensorflow.keras.utils import plot_model shapes = dict(input_shape=[(10,), (10,)], output_shape=(1,)) space = OneLayerSpace(**shapes).build() model = space.sample() plot_model(model)
1,655
27.551724
87
py
deephyper
deephyper-master/deephyper/nas/spacelib/tabular/supervised_reg_auto_encoder.py
import tensorflow as tf from deephyper.nas import KSearchSpace from deephyper.nas.node import ConstantNode, VariableNode from deephyper.nas.operation import Identity, operation Dense = operation(tf.keras.layers.Dense) class SupervisedRegAutoEncoderSpace(KSearchSpace): def __init__( self, input_shape, output_shape, batch_size=None, seed=None, units=[128, 64, 32, 16, 8, 16, 32, 64, 128], num_layers=5, ): super().__init__(input_shape, output_shape, batch_size=batch_size, seed=seed) self.units = units self.num_layers = num_layers def build(self): inp = self.input_nodes[0] # auto-encoder units = [128, 64, 32, 16, 8, 16, 32, 64, 128] prev_node = inp d = 1 for i in range(len(units)): vnode = VariableNode() vnode.add_op(Identity()) if d == 1 and units[i] < units[i + 1]: d = -1 for u in range(min(2, units[i]), max(2, units[i]) + 1, 2): vnode.add_op(Dense(u, tf.nn.relu)) latente_space = vnode else: for u in range( min(units[i], units[i + d]), max(units[i], units[i + d]) + 1, 2 ): vnode.add_op(Dense(u, tf.nn.relu)) self.connect(prev_node, vnode) prev_node = vnode out2 = ConstantNode(op=Dense(self.output_shape[0][0], name="output_0")) self.connect(prev_node, out2) # regressor prev_node = latente_space # prev_node = inp for _ in range(self.num_layers): vnode = VariableNode() for i in range(16, 129, 16): vnode.add_op(Dense(i, tf.nn.relu)) self.connect(prev_node, vnode) prev_node = vnode out1 = ConstantNode(op=Dense(self.output_shape[1][0], name="output_1")) self.connect(prev_node, out1) return self if __name__ == "__main__": from tensorflow.keras.utils import plot_model shapes = dict(input_shape=(100,), output_shape=[(100,), (10,)]) space = SupervisedRegAutoEncoderSpace(**shapes).build() model = space.sample() plot_model(model)
2,268
28.855263
85
py
deephyper
deephyper-master/deephyper/nas/spacelib/tabular/feed_forward.py
import tensorflow as tf from deephyper.nas import KSearchSpace from deephyper.nas.node import ConstantNode, VariableNode from deephyper.nas.operation import Identity, operation Dense = operation(tf.keras.layers.Dense) class FeedForwardSpace(KSearchSpace): """Simple search space for a feed-forward neural network. No skip-connection. Looking over the number of units per layer and the number of layers. Args: input_shape (tuple, optional): True shape of inputs (no batch size dimension). Defaults to (2,). output_shape (tuple, optional): True shape of outputs (no batch size dimension).. Defaults to (1,). num_layers (int, optional): Maximum number of layers to have. Defaults to 10. num_units (tuple, optional): Range of number of units such as range(start, end, step_size). Defaults to (1, 11). regression (bool, optional): A boolean defining if the model is a regressor or a classifier. Defaults to True. """ def __init__( self, input_shape, output_shape, batch_size=None, seed=None, regression=True, num_units=(1, 11), num_layers=10, ): super().__init__(input_shape, output_shape, batch_size=batch_size, seed=seed) self.regression = regression self.num_units = num_units self.num_layers = num_layers def build(self): prev_node = self.input_nodes[0] for _ in range(self.num_layers): vnode = VariableNode() vnode.add_op(Identity()) for i in range(*self.num_units): vnode.add_op(Dense(i, tf.nn.relu)) self.connect(prev_node, vnode) prev_node = vnode output_node = ConstantNode( Dense( self.output_shape[0], activation=None if self.regression else "softmax" ) ) self.connect(prev_node, output_node) return self if __name__ == "__main__": from tensorflow.keras.utils import plot_model shapes = dict(input_shape=(10,), output_shape=(1,)) space = FeedForwardSpace(**shapes).build() model = space.sample() plot_model(model)
2,179
32.030303
150
py
deephyper
deephyper-master/deephyper/nas/spacelib/tabular/dense_skipco.py
import collections import tensorflow as tf from deephyper.nas import KSearchSpace from deephyper.nas.node import ConstantNode, VariableNode from deephyper.nas.operation import operation, Zero, Connect, AddByProjecting, Identity Dense = operation(tf.keras.layers.Dense) Dropout = operation(tf.keras.layers.Dropout) class DenseSkipCoSpace(KSearchSpace): def __init__( self, input_shape, output_shape, batch_size=None, seed=None, regression=True, num_layers=10, dropout=0.0, ): super().__init__(input_shape, output_shape, batch_size=batch_size, seed=seed) self.regression = regression self.num_layers = num_layers self.dropout = dropout def build(self): source = prev_input = self.input_nodes[0] # look over skip connections within a range of the 3 previous nodes anchor_points = collections.deque([source], maxlen=3) for _ in range(self.num_layers): vnode = VariableNode() self.add_dense_to_(vnode) self.connect(prev_input, vnode) # * Cell output cell_output = vnode cmerge = ConstantNode() cmerge.set_op(AddByProjecting(self, [cell_output], activation="relu")) for anchor in anchor_points: skipco = VariableNode() skipco.add_op(Zero()) skipco.add_op(Connect(self, anchor)) self.connect(skipco, cmerge) prev_input = cmerge # ! for next iter anchor_points.append(prev_input) if self.dropout >= 0.0: dropout_node = ConstantNode(op=Dropout(rate=self.dropout)) self.connect(prev_input, dropout_node) prev_input = dropout_node output_node = ConstantNode( Dense( self.output_shape[0], activation=None if self.regression else "softmax" ) ) self.connect(prev_input, output_node) return self def add_dense_to_(self, node): node.add_op(Identity()) # we do not want to create a layer in this case activations = [None, tf.nn.swish, tf.nn.relu, tf.nn.tanh, tf.nn.sigmoid] for units in range(16, 97, 16): for activation in activations: node.add_op(Dense(units=units, activation=activation)) if __name__ == "__main__": from tensorflow.keras.utils import plot_model shapes = dict(input_shape=(10,), output_shape=(1,)) space = DenseSkipCoSpace(**shapes).build() model = space.sample() plot_model(model)
2,637
28.311111
87
py
deephyper
deephyper-master/deephyper/nas/run/_run_horovod.py
"""The :func:`deephyper.nas.run.horovod.run` function is used to evaluate a deep neural network by enabling data-parallelism with Horovod to the :func:`deephyper.nas.run.alpha.run` function. This function will automatically apply the linear scaling rule to the learning rate and batch size given the current number of ranks (i.e., the initial learning rate and batch size are scaled by the number of ranks). """ import os import traceback import logging import numpy as np import tensorflow as tf from deephyper.keras.callbacks import import_callback import horovod.tensorflow.keras as hvd import deephyper.nas.trainer._arch as a from deephyper.nas.trainer import HorovodTrainer from deephyper.nas.run._util import ( compute_objective, load_config, preproc_trainer, save_history, setup_data, get_search_space, ) logger = logging.getLogger(__name__) # Default callbacks parameters default_callbacks_config = { "EarlyStopping": dict( monitor="val_loss", min_delta=0, mode="min", verbose=0, patience=0 ), "ModelCheckpoint": dict( monitor="val_loss", mode="min", save_best_only=True, verbose=1, filepath="model.h5", save_weights_only=False, ), "TensorBoard": dict( log_dir="", histogram_freq=0, batch_size=32, write_graph=False, write_grads=False, write_images=False, update_freq="epoch", ), "CSVLogger": dict(filename="training.csv", append=True), "CSVExtendedLogger": dict(filename="training.csv", append=True), "TimeStopping": dict(), "ReduceLROnPlateau": dict(patience=5, verbose=0), } # Name of Callbacks reserved for root node hvd_root_cb = ["ModelCheckpoint", "Tensorboard", "CSVLogger", "CSVExtendedLogger"] def run_horovod(config: dict) -> float: hvd.init() # Threading configuration if os.environ.get("OMP_NUM_THREADS", None) is not None: logger.debug(f"OMP_NUM_THREADS is {os.environ.get('OMP_NUM_THREADS')}") num_intra = int(os.environ.get("OMP_NUM_THREADS")) tf.config.threading.set_intra_op_parallelism_threads(num_intra) tf.config.threading.set_inter_op_parallelism_threads(2) if os.environ.get("CUDA_VISIBLE_DEVICES") is not None: devices = os.environ.get("CUDA_VISIBLE_DEVICES").split(",") os.environ["CUDA_VISIBLE_DEVICES"] = devices[hvd.rank()] config["seed"] seed = config["seed"] if seed is not None: np.random.seed(seed) tf.random.set_seed(seed) load_config(config) # Scale batch size and learning rate according to the number of ranks initial_lr = config[a.hyperparameters][a.learning_rate] batch_size = config[a.hyperparameters][a.batch_size] * hvd.size() learning_rate = config[a.hyperparameters][a.learning_rate] * hvd.size() logger.info( f"Scaled: 'batch_size' from {config[a.hyperparameters][a.batch_size]} to {batch_size} " ) logger.info( f"Scaled: 'learning_rate' from {config[a.hyperparameters][a.learning_rate]} to {learning_rate} " ) config[a.hyperparameters][a.batch_size] = batch_size config[a.hyperparameters][a.learning_rate] = learning_rate input_shape, output_shape = setup_data(config) search_space = get_search_space(config, input_shape, output_shape, seed=seed) # Initialize Horovod model_created = False try: model = search_space.sample(config["arch_seq"]) model_created = True except Exception: logger.info("Error: Model creation failed...") logger.info(traceback.format_exc()) if model_created: # Setup callbacks only callbacks = [ # Horovod: broadcast initial variable states from rank 0 to all other processes. # This is necessary to ensure consistent initialization of all workers when # training is started with random weights or restored from a checkpoint. hvd.callbacks.BroadcastGlobalVariablesCallback(0), # Horovod: average metrics among workers at the end of every epoch. # # Note: This callback must be in the list before the ReduceLROnPlateau, # TensorBoard or other metrics-based callbacks. hvd.callbacks.MetricAverageCallback(), # Horovod: using `lr = 1.0 * hvd.size()` from the very beginning leads to worse final # accuracy. Scale the learning rate `lr = 1.0` ---> `lr = 1.0 * hvd.size()` during # the first five epochs. See https://arxiv.org/abs/1706.02677 for details. # !initial_lr argument is not available in horovod==0.19.0 hvd.callbacks.LearningRateWarmupCallback( warmup_epochs=5, verbose=0, initial_lr=initial_lr ), ] cb_requires_valid = False # Callbacks requires validation data callbacks_config = config[a.hyperparameters].get(a.callbacks, {}) if callbacks_config is not None: for cb_name, cb_conf in callbacks_config.items(): if cb_name in default_callbacks_config: # cb_bame in hvd_root_cb implies hvd.rank() == 0 if not (cb_name in hvd_root_cb) or hvd.rank() == 0: default_callbacks_config[cb_name].update(cb_conf) # Import and create corresponding callback Callback = import_callback(cb_name) callbacks.append(Callback(**default_callbacks_config[cb_name])) if cb_name in ["EarlyStopping"]: cb_requires_valid = "val" in cb_conf["monitor"].split("_") else: logger.error(f"'{cb_name}' is not an accepted callback!") trainer = HorovodTrainer(config=config, model=model) trainer.callbacks.extend(callbacks) last_only, with_pred = preproc_trainer(config) last_only = last_only and not cb_requires_valid history = trainer.train(with_pred=with_pred, last_only=last_only) # save history if hvd.rank() == 0: save_history(config.get("log_dir", None), history, config) result = compute_objective(config["objective"], history) else: # penalising actions if model cannot be created result = -1 if result < -10: result = -10 return result
6,420
37.680723
407
py
deephyper
deephyper-master/deephyper/nas/run/_run_distributed_base_trainer.py
"""The :func:`deephyper.nas.run.tf_distributed.run` function is used to deploy a data-distributed training (on a single node) with ``tensorflow.distribute.MirroredStrategy``. It follows the same training pipeline as :func:`deephyper.nas.run.alpha.run`. Two hyperparameters arguments can be used to activate or deactivate the linear scaling rule (aka ``lsr``) for the learning rate and batch size, respectively: .. code-block:: python Problem.hyperparameters( ... lsr_batch_size=True, lsr_learning_rate=True, warmup_lr=True, warmup_epochs=5 ... ) The hyperparameters of the form ``patience_{Callback}`` such as ``patience_EarlyStopping`` and ``patience_ReduceLROnPlateau`` are valid when the corresponding callback is declared: .. code-block:: python Problem.hyperparameters( ... patience_ReduceLROnPlateau=5, patience_EarlyStopping=10, callbacks=dict( ReduceLROnPlateau=dict(monitor="val_r2", mode="max", verbose=0), EarlyStopping=dict(monitor="val_r2", min_delta=0, mode="max", verbose=0), ), ... ) """ import traceback import logging import numpy as np import tensorflow as tf from deephyper.keras.callbacks import import_callback from deephyper.keras.callbacks import LearningRateWarmupCallback from deephyper.nas.run._util import ( compute_objective, load_config, preproc_trainer, save_history, setup_data, get_search_space, default_callbacks_config, ) from deephyper.nas.trainer import BaseTrainer import deephyper.nas.trainer._arch as a logger = logging.getLogger(__name__) def run_distributed_base_trainer(config): physical_devices = tf.config.list_physical_devices("GPU") try: for i in range(len(physical_devices)): tf.config.experimental.set_memory_growth(physical_devices[i], True) except Exception: # Invalid device or cannot modify virtual devices once initialized. pass distributed_strategy = tf.distribute.MirroredStrategy() n_replicas = distributed_strategy.num_replicas_in_sync seed = config["seed"] if seed is not None: np.random.seed(seed) tf.random.set_seed(seed) load_config(config) # Scale batch size and learning rate according to the number of ranks initial_lr = config[a.hyperparameters][a.learning_rate] if config[a.hyperparameters].get("lsr_batch_size"): batch_size = config[a.hyperparameters][a.batch_size] * n_replicas else: batch_size = config[a.hyperparameters][a.batch_size] if config[a.hyperparameters].get("lsr_learning_rate"): learning_rate = config[a.hyperparameters][a.learning_rate] * n_replicas else: learning_rate = config[a.hyperparameters][a.learning_rate] logger.info( f"Scaled: 'batch_size' from {config[a.hyperparameters][a.batch_size]} to {batch_size} " ) logger.info( f"Scaled: 'learning_rate' from {config[a.hyperparameters][a.learning_rate]} to {learning_rate} " ) config[a.hyperparameters][a.batch_size] = batch_size config[a.hyperparameters][a.learning_rate] = learning_rate input_shape, output_shape = setup_data(config) search_space = get_search_space(config, input_shape, output_shape, seed=seed) model_created = False with distributed_strategy.scope(): try: model = search_space.sample(config["arch_seq"]) model_created = True except Exception: logger.info("Error: Model creation failed...") logger.info(traceback.format_exc()) else: # Setup callbacks callbacks = [] cb_requires_valid = False # Callbacks requires validation data callbacks_config = config["hyperparameters"].get("callbacks") if callbacks_config is not None: for cb_name, cb_conf in callbacks_config.items(): if cb_name in default_callbacks_config: default_callbacks_config[cb_name].update(cb_conf) # Special dynamic parameters for callbacks if cb_name == "ModelCheckpoint": default_callbacks_config[cb_name][ "filepath" ] = f'best_model_{config["id"]}.h5' # replace patience hyperparameter if "patience" in default_callbacks_config[cb_name]: patience = config["hyperparameters"].get( f"patience_{cb_name}" ) if patience is not None: default_callbacks_config[cb_name]["patience"] = patience # Import and create corresponding callback Callback = import_callback(cb_name) callbacks.append(Callback(**default_callbacks_config[cb_name])) if cb_name in ["EarlyStopping"]: cb_requires_valid = "val" in cb_conf["monitor"].split("_") else: logger.error(f"'{cb_name}' is not an accepted callback!") # WarmupLR if config[a.hyperparameters].get("warmup_lr"): warmup_epochs = config[a.hyperparameters].get("warmup_epochs", 5) callbacks.append( LearningRateWarmupCallback( n_replicas=n_replicas, warmup_epochs=warmup_epochs, verbose=0, initial_lr=initial_lr, ) ) trainer = BaseTrainer(config=config, model=model) trainer.callbacks.extend(callbacks) last_only, with_pred = preproc_trainer(config) last_only = last_only and not cb_requires_valid if model_created: history = trainer.train(with_pred=with_pred, last_only=last_only) # save history save_history(config.get("log_dir", None), history, config) result = compute_objective(config["objective"], history) else: # penalising actions if model cannot be created result = -1 if result < -10 or np.isnan(result): result = -10 return result
6,405
37.359281
410
py
deephyper
deephyper-master/deephyper/nas/run/_run_base_trainer.py
"""The :func:`deephyper.nas.run.alpha.run` function is used to evaluate a deep neural network by loading the data, building the model, training the model and returning a scalar value corresponding to the objective defined in the used :class:`deephyper.problem.NaProblem`. """ import os import traceback import logging import numpy as np import tensorflow as tf from deephyper.keras.callbacks import import_callback from deephyper.nas.run._util import ( compute_objective, load_config, preproc_trainer, setup_data, get_search_space, default_callbacks_config, HistorySaver, ) from deephyper.nas.trainer import BaseTrainer logger = logging.getLogger(__name__) def run_base_trainer(job): config = job.parameters config["job_id"] = job.id tf.keras.backend.clear_session() # tf.config.optimizer.set_jit(True) # setup history saver if config.get("log_dir") is None: config["log_dir"] = "." save_dir = os.path.join(config["log_dir"], "save") saver = HistorySaver(config, save_dir) saver.write_config() saver.write_model(None) # GPU Configuration if available physical_devices = tf.config.list_physical_devices("GPU") try: for i in range(len(physical_devices)): tf.config.experimental.set_memory_growth(physical_devices[i], True) except Exception: # Invalid device or cannot modify virtual devices once initialized. logger.info("error memory growth for GPU device") # Threading configuration if ( len(physical_devices) == 0 and os.environ.get("OMP_NUM_THREADS", None) is not None ): logger.info(f"OMP_NUM_THREADS is {os.environ.get('OMP_NUM_THREADS')}") num_intra = int(os.environ.get("OMP_NUM_THREADS")) try: tf.config.threading.set_intra_op_parallelism_threads(num_intra) tf.config.threading.set_inter_op_parallelism_threads(2) except RuntimeError: # Session already initialized pass tf.config.set_soft_device_placement(True) seed = config.get("seed") if seed is not None: np.random.seed(seed) tf.random.set_seed(seed) load_config(config) input_shape, output_shape = setup_data(config) search_space = get_search_space(config, input_shape, output_shape, seed=seed) model_created = False try: model = search_space.sample(config["arch_seq"]) model_created = True except Exception: logger.info("Error: Model creation failed...") logger.info(traceback.format_exc()) if model_created: # Setup callbacks callbacks = [] cb_requires_valid = False # Callbacks requires validation data callbacks_config = config["hyperparameters"].get("callbacks") if callbacks_config is not None: for cb_name, cb_conf in callbacks_config.items(): if cb_name in default_callbacks_config: default_callbacks_config[cb_name].update(cb_conf) # Special dynamic parameters for callbacks if cb_name == "ModelCheckpoint": default_callbacks_config[cb_name]["filepath"] = saver.model_path # replace patience hyperparameter if "patience" in default_callbacks_config[cb_name]: patience = config["hyperparameters"].get(f"patience_{cb_name}") if patience is not None: default_callbacks_config[cb_name]["patience"] = patience # Import and create corresponding callback Callback = import_callback(cb_name) callbacks.append(Callback(**default_callbacks_config[cb_name])) if cb_name in ["EarlyStopping"]: cb_requires_valid = "val" in cb_conf["monitor"].split("_") else: logger.error(f"'{cb_name}' is not an accepted callback!") trainer = BaseTrainer(config=config, model=model) trainer.callbacks.extend(callbacks) last_only, with_pred = preproc_trainer(config) last_only = last_only and not cb_requires_valid history = trainer.train(with_pred=with_pred, last_only=last_only) # save history saver.write_history(history) result = compute_objective(config["objective"], history) else: # penalising actions if model cannot be created logger.info("Model could not be created returning -Inf!") result = -float("inf") if np.isnan(result): logger.info("Computed objective is NaN returning -Inf instead!") result = -float("inf") return result
4,751
34.2
271
py
deephyper
deephyper-master/deephyper/nas/operation/_merge.py
import deephyper as dh import tensorflow as tf from ._base import Operation class Concatenate(Operation): """Concatenate operation. Args: graph: node (Node): stacked_nodes (list(Node)): nodes to concatenate axis (int): axis to concatenate """ def __init__(self, search_space, stacked_nodes=None, axis=-1): self.search_space = search_space self.node = None # current_node of the operation self.stacked_nodes = stacked_nodes self.axis = axis def __str__(self): return "Concatenate" def init(self, current_node): self.node = current_node if self.stacked_nodes is not None: for n in self.stacked_nodes: self.search_space.connect(n, self.node) def __call__(self, values, **kwargs): # case where there is no inputs if len(values) == 0: return [] len_shp = max([len(x.get_shape()) for x in values]) if len_shp > 4: raise RuntimeError( f"This concatenation is for 2D or 3D tensors only but a {len_shp-1}D is passed!" ) # zeros padding if len(values) > 1: if all( map( lambda x: len(x.get_shape()) == len_shp or len(x.get_shape()) == (len_shp - 1), values, ) ): # all tensors should have same number of dimensions 2d or 3d, but we can also accept a mix of 2d en 3d tensors # we have a mix of 2d and 3d tensors so we are expanding 2d tensors to be 3d with last_dim==1 for i, v in enumerate(values): if len(v.get_shape()) < len_shp: values[i] = tf.keras.layers.Reshape( (*tuple(v.get_shape()[1:]), 1) )(v) # for 3d tensors concatenation is applied along last dim (axis=-1), so we are applying a zero padding to make 2nd dimensions (ie. shape()[1]) equals if len_shp == 3: max_len = max(map(lambda x: int(x.get_shape()[1]), values)) paddings = map(lambda x: max_len - int(x.get_shape()[1]), values) for i, (p, v) in enumerate(zip(paddings, values)): lp = p // 2 rp = p - lp values[i] = tf.keras.layers.ZeroPadding1D(padding=(lp, rp))(v) # elif len_shp == 2 nothing to do else: raise RuntimeError( f"All inputs of concatenation operation should have same shape length:\n" f"number_of_inputs=={len(values)}\n" f"shape_of_inputs=={[str(x.get_shape()) for x in values]}" ) # concatenation if len(values) > 1: out = tf.keras.layers.Concatenate(axis=-1)(values) else: out = values[0] return out class AddByPadding(Operation): """Add operation. If tensor are of different shapes a padding will be applied before adding them. Args: search_space (KSearchSpace): [description]. Defaults to None. activation ([type], optional): Activation function to apply after adding ('relu', tanh', 'sigmoid'...). Defaults to None. stacked_nodes (list(Node)): nodes to add. axis (int): axis to concatenate. """ def __init__(self, search_space, stacked_nodes=None, activation=None, axis=-1): self.search_space = search_space self.node = None # current_node of the operation self.stacked_nodes = stacked_nodes self.activation = activation self.axis = axis def init(self, current_node): self.node = current_node if self.stacked_nodes is not None: for n in self.stacked_nodes: self.search_space.connect(n, self.node) def __call__(self, values, **kwargs): # case where there is no inputs if len(values) == 0: return [] values = values[:] max_len_shp = max([len(x.get_shape()) for x in values]) # zeros padding if len(values) > 1: for i, v in enumerate(values): if len(v.get_shape()) < max_len_shp: values[i] = tf.keras.layers.Reshape( ( *tuple(v.get_shape()[1:]), *tuple(1 for i in range(max_len_shp - len(v.get_shape()))), ) )(v) def max_dim_i(i): return max(map(lambda x: int(x.get_shape()[i]), values)) max_dims = [None] + list(map(max_dim_i, range(1, max_len_shp))) def paddings_dim_i(i): return list(map(lambda x: max_dims[i] - int(x.get_shape()[i]), values)) paddings_dim = list(map(paddings_dim_i, range(1, max_len_shp))) for i in range(len(values)): paddings = list() for j in range(len(paddings_dim)): p = paddings_dim[j][i] lp = p // 2 rp = p - lp paddings.append([lp, rp]) if sum(map(sum, paddings)) != 0: values[i] = dh.layers.Padding(paddings)(values[i]) # concatenation if len(values) > 1: out = tf.keras.layers.Add()(values) if self.activation is not None: out = tf.keras.layers.Activation(self.activation)(out) else: out = values[0] return out class AddByProjecting(Operation): """Add operation. If tensors are of different shapes a projection will be applied before adding them. Args: search_space (KSearchSpace): [description]. Defaults to None. activation ([type], optional): Activation function to apply after adding ('relu', tanh', 'sigmoid'...). Defaults to None. stacked_nodes (list(Node)): nodes to add. axis (int): axis to concatenate. """ def __init__(self, search_space, stacked_nodes=None, activation=None, axis=-1): self.search_space = search_space self.node = None # current_node of the operation self.stacked_nodes = stacked_nodes self.activation = activation self.axis = axis def init(self, current_node): self.node = current_node if self.stacked_nodes is not None: for n in self.stacked_nodes: self.search_space.connect(n, self.node) def __call__(self, values, seed=None, **kwargs): # case where there is no inputs if len(values) == 0: return [] values = values[:] max_len_shp = max([len(x.get_shape()) for x in values]) # projection if len(values) > 1: for i, v in enumerate(values): if len(v.get_shape()) < max_len_shp: values[i] = tf.keras.layers.Reshape( ( *tuple(v.get_shape()[1:]), *tuple(1 for i in range(max_len_shp - len(v.get_shape()))), ) )(v) proj_size = values[0].get_shape()[self.axis] for i in range(len(values)): if values[i].get_shape()[self.axis] != proj_size: values[i] = tf.keras.layers.Dense( units=proj_size, kernel_initializer=tf.keras.initializers.glorot_uniform( seed=seed ), )(values[i]) # concatenation if len(values) > 1: out = tf.keras.layers.Add()(values) if self.activation is not None: out = tf.keras.layers.Activation(self.activation)(out) else: out = values[0] return out
8,004
35.221719
164
py
deephyper
deephyper-master/deephyper/nas/operation/_base.py
import tensorflow as tf class Operation: """Interface of an operation. >>> import tensorflow as tf >>> from deephyper.nas.space.op import Operation >>> Operation(layer=tf.keras.layers.Dense(10)) Dense Args: layer (Layer): a ``tensorflow.keras.layers.Layer``. """ def __init__(self, layer: tf.keras.layers.Layer): assert isinstance(layer, tf.keras.layers.Layer) self.from_keras_layer = True self._layer = layer def __str__(self): return self.__repr__() def __repr__(self): if hasattr(self, "from_keras_layer"): return type(self._layer).__name__ else: try: return str(self) except Exception: return type(self).__name__ def __call__(self, tensors: list, seed: int = None, **kwargs): """ Args: tensors (list): a list of incoming tensors. Returns: tensor: an output tensor. """ if len(tensors) == 1: out = self._layer(tensors[0]) else: out = self._layer(tensors) return out def init(self, current_node): """Preprocess the current operation.""" def operation(cls): """Dynamically creates a sub-class of Operation from a Keras layer. Args: cls (tf.keras.layers.Layer): takes a Keras layer class as input and return an operation class corresponding to this layer. """ def __init__(self, *args, **kwargs): self._args = args self._kwargs = kwargs self._layer = None def __repr__(self): return cls.__name__ def __call__(self, inputs, **kwargs): if self._layer is None: self._layer = cls(*self._args, **self._kwargs) if len(inputs) == 1: out = self._layer(inputs[0]) else: out = self._layer(inputs) return out cls_attrs = dict(__init__=__init__, __repr__=__repr__, __call__=__call__) op_class = type(cls.__name__, (Operation,), cls_attrs) return op_class class Identity(Operation): def __init__(self): pass def __call__(self, inputs, **kwargs): assert ( len(inputs) == 1 ), f"{type(self).__name__} as {len(inputs)} inputs when 1 is required." return inputs[0] class Tensor(Operation): def __init__(self, tensor, *args, **kwargs): self.tensor = tensor def __str__(self): return str(self.tensor) def __call__(self, *args, **kwargs): return self.tensor class Zero(Operation): def __init__(self): self.tensor = [] def __str__(self): return "Zero" def __call__(self, *args, **kwargs): return self.tensor class Connect(Operation): """Connection node. Represents a possibility to create a connection between n1 -> n2. Args: graph (nx.DiGraph): a graph source_node (Node): source """ def __init__(self, search_space, source_node, *args, **kwargs): self.search_space = search_space self.source_node = source_node self.destin_node = None def __str__(self): if type(self.source_node) is list: if len(self.source_node) > 0: ids = str(self.source_node[0].id) for n in self.source_node[1:]: ids += "," + str(n.id) else: ids = "None" else: ids = self.source_node.id if self.destin_node is None: return f"{type(self).__name__}_{ids}->?" else: return f"{type(self).__name__}_{ids}->{self.destin_node.id}" def init(self, current_node): """Set the connection in the search_space graph from n1 -> n2.""" self.destin_node = current_node if type(self.source_node) is list: for n in self.source_node: self.search_space.connect(n, self.destin_node) else: self.search_space.connect(self.source_node, self.destin_node) def __call__(self, value, *args, **kwargs): return value
4,141
25.382166
130
py
deephyper
deephyper-master/deephyper/problem/_neuralarchitecture.py
from collections import OrderedDict from copy import deepcopy from inspect import signature import ConfigSpace.hyperparameters as csh import tensorflow as tf from deephyper.core.exceptions.problem import ( NaProblemError, ProblemLoadDataIsNotCallable, ProblemPreprocessingIsNotCallable, SearchSpaceBuilderMissingParameter, WrongProblemObjective, ) from deephyper.nas.run._util import get_search_space, setup_data from deephyper.problem import HpProblem class NaProblem: """A Neural Architecture Problem specification for Neural Architecture Search. >>> from deephyper.problem import NaProblem >>> from deephyper.nas.preprocessing import minmaxstdscaler >>> from deepspace.tabular import OneLayerSpace >>> Problem = NaProblem() >>> Problem.load_data(load_data) >>> Problem.preprocessing(minmaxstdscaler) >>> Problem.search_space(OneLayerSpace) >>> Problem.hyperparameters( ... batch_size=100, ... learning_rate=0.1, ... optimizer='adam', ... num_epochs=10, ... callbacks=dict( ... EarlyStopping=dict( ... monitor='val_r2', ... mode='max', ... verbose=0, ... patience=5 ... ) ... ) ... ) >>> Problem.loss('mse') >>> Problem.metrics(['r2']) >>> Problem.objective('val_r2__last') """ def __init__(self): self._space = OrderedDict() self._hp_space = HpProblem() self._space["metrics"] = [] self._space["hyperparameters"] = dict(verbose=0) def __repr__(self): preprocessing = ( None if self._space.get("preprocessing") is None else module_location(self._space["preprocessing"]["func"]) ) hps = "".join( [ f"\n * {h}: {self._space['hyperparameters'][h]}" for h in self._space["hyperparameters"] ] ) if type(self._space["metrics"]) is list: metrics = "".join([f"\n * {m}" for m in self._space["metrics"]]) else: metrics = "".join( [f"\n * {m[0]}: {m[1]}" for m in self._space["metrics"].items()] ) objective = self._space["objective"] if not type(objective) is str: objective = module_location(objective) out = ( f"Problem is:\n" f" - search space : {module_location(self._space['search_space']['class'])}\n" f" - data loading : {module_location(self._space['load_data']['func'])}\n" f" - preprocessing : {preprocessing}\n" f" - hyperparameters: {hps}\n" f" - loss : {self._space['loss']}\n" f" - metrics : {metrics}\n" f" - objective : {objective}\n" ) return out def load_data(self, func: callable, **kwargs): """Define the function loading the data. .. code-block:: python Problem.load_data(load_data, load_data_kwargs) This ``load_data`` callable can follow two different interfaces: Numpy arrays or generators. 1. **Numpy arrays**: In the case of Numpy arrays, the callable passed to ``Problem.load_data(...)`` has to return the following tuple: ``(X_train, y_train), (X_valid, y_valid)``. In the most simple case where the model takes a single input, each of these elements is a Numpy array. Generally, ``X_train`` and ``y_train`` have to be of the same length (i.e., same ``array.shape[0]``) which is also the case for ``X_valid`` and ``y_valid``. Similarly, the shape of the elements of ``X_train`` and ``X_valid`` which is also the case for ``y_train`` and ``y_valid``. An example ``load_data`` function can be .. code-block:: python import numpy as np def load_data(N=100): X = np.zeros((N, 1)) y = np.zeros((N,1)) return (X, y), (X, y) It is also possible for the model to take several inputs. In fact, experimentaly it can be notices that separating some inputs with different inputs can significantly help the learning of the model. Also, sometimes different inputs may be of the "types" for example two molecular fingerprints. In this case, it can be very interesting to share the weights of the model to process these two inputs. In the case of multi-inputs models the ``load_data`` function will also return ``(X_train, y_train), (X_valid, y_valid)`` bu where ``X_train`` and ``X_valid`` are two lists of Numpy arrays. For example, the following is correct: .. code-block:: python import numpy as np def load_data(N=100): X = np.zeros((N, 1)) y = np.zeros((N,1)) return ([X, X], y), ([X, X], y) 2. **Generators**: Returning generators with a single input: .. code-block:: python def load_data(N=100): tX, ty = np.zeros((N,1)), np.zeros((N,1)) vX, vy = np.zeros((N,1)), np.zeros((N,1)) def train_gen(): for x, y in zip(tX, ty): yield ({"input_0": x}, y) def valid_gen(): for x, y in zip(vX, vy): yield ({"input_0": x}, y) res = { "train_gen": train_gen, "train_size": N, "valid_gen": valid_gen, "valid_size": N, "types": ({"input_0": tf.float64}, tf.float64), "shapes": ({"input_0": (1, )}, (1, )) } return res Returning generators with multiple inputs: .. code-block:: python def load_data(N=100): tX0, tX1, ty = np.zeros((N,1)), np.zeros((N,1)), np.zeros((N,1)), vX0, vX1, vy = np.zeros((N,1)), np.zeros((N,1)), np.zeros((N,1)), def train_gen(): for x0, x1, y in zip(tX0, tX1, ty): yield ({ "input_0": x0, "input_1": x1 }, y) def valid_gen(): for x0, x1, y in zip(vX0, vX1, vy): yield ({ "input_0": x0, "input_1": x1 }, y) res = { "train_gen": train_gen, "train_size": N, "valid_gen": valid_gen, "valid_size": N, "types": ({"input_0": tf.float64, "input_1": tf.float64}, tf.float64), "shapes": ({"input_0": (5, ), "input_1": (5, )}, (1, )) } return res Args: func (callable): the load data function. """ if not callable(func): raise ProblemLoadDataIsNotCallable(func) self._space["load_data"] = {"func": func, "kwargs": kwargs} def augment(self, func: callable, **kwargs): """ :meta private: """ if not callable(func): raise ProblemLoadDataIsNotCallable(func) self._space["augment"] = {"func": func, "kwargs": kwargs} def search_space(self, space_class, **kwargs): """Set a search space for neural architecture search. Args: space_class (KSearchSpace): an object of type ``KSearchSpace`` which has to implement the ``build()`` method. Raises: SearchSpaceBuilderMissingParameter: raised when either of ``(input_shape, output_shape)`` are missing parameters of ``func``. """ sign = signature(space_class) if "input_shape" not in sign.parameters: raise SearchSpaceBuilderMissingParameter("input_shape") if "output_shape" not in sign.parameters: raise SearchSpaceBuilderMissingParameter("output_shape") self._space["search_space"] = {"class": space_class, "kwargs": kwargs} def add_hyperparameter( self, value, name: str = None, default_value=None ) -> csh.Hyperparameter: """Add hyperparameters to search the neural architecture search problem. >>> Problem.hyperparameters( ... batch_size=problem.add_hyperparameter((32, 256), "batch_size") ... ) Args: value: a hyperparameter description. name: a name of the defined hyperparameter, the same as the current key. default_value (Optional): a default value of the hyperparameter. Returns: Hyperparameter: the defined hyperparameter. """ return self._hp_space.add_hyperparameter(value, name, default_value) def preprocessing(self, func: callable): """Define how to preprocess your data. Args: func (callable): a function which returns a preprocessing scikit-learn ``Pipeline``. """ if not callable(func): raise ProblemPreprocessingIsNotCallable(func) self._space["preprocessing"] = {"func": func} def hyperparameters(self, **kwargs): """Define hyperparameters used to evaluate generated architectures. Hyperparameters can be defined such as: .. code-block:: python Problem.hyperparameters( batch_size=256, learning_rate=0.01, optimizer="adam", num_epochs=20, verbose=0, callbacks=dict(...), ) """ if self._space.get("hyperparameters") is None: self._space["hyperparameters"] = dict() self._space["hyperparameters"].update(kwargs) def loss(self, loss, loss_weights=None, class_weights=None): """Define the loss used to train generated architectures. It can be a ``str`` corresponding to a Keras loss function: .. code-block:: python problem.loss("categorical_crossentropy") A custom loss function can also be defined: .. code-block:: python def NLL(y, rv_y): return -rv_y.log_prob(y) problem.loss(NLL) The loss can be automatically searched: .. code-block:: python problem.loss( problem.add_hyperparameter( ["mae", "mse", "huber_loss", "log_cosh", "mape", "msle"], "loss" ) ) It is possible to define a different loss for each output: .. code-block:: python problem.loss( loss={"output_0": "mse", "output_1": "mse"}, loss_weights={"output_0": 0.0, "output_1": 1.0}, ) Args: loss (str or callable orlist): a string indicating a specific loss function. loss_weights (list): Optional. class_weights (dict): Optional. """ if not (type(loss) is csh.CategoricalHyperparameter): if not type(loss) is str and not callable(loss) and not type(loss) is dict: raise RuntimeError( f"The loss should be either a str, dict or a callable when it's of type {type(loss)}" ) if ( type(loss) is dict and loss_weights is not None and len(loss) != len(loss_weights) ): raise RuntimeError( f"The losses list (len={len(loss)}) and the weights list (len={len(loss_weights)}) should be of same length!" ) self._space["loss"] = loss if loss_weights is not None: self._space["loss_weights"] = loss_weights if class_weights is not None: self._space["class_weights"] = class_weights def metrics(self, metrics=None): """Define a list of metrics for the training of generated architectures. A list of metrics can be defined to be monitored or used as an objective. It can be a keyword or a callable. For example, if it is a keyword: .. code-block:: python problem.metrics(["acc"]) In case you need multiple metrics: .. code-block:: python problem.metrics["mae", "mse"] In case you want to use a custom metric: .. code-block:: python def sparse_perplexity(y_true, y_pred): cross_entropy = tf.keras.losses.sparse_categorical_crossentropy(y_true, y_pred) perplexity = tf.pow(2.0, cross_entropy) return perplexity problem.metrics([sparse_perplexity]) Args: metrics (list(str or callable) or dict): If ``str`` the metric should be defined in Keras or in DeepHyper. If ``callable`` it should take 2 arguments ``(y_pred, y_true)`` which are a prediction and a true value respectively. """ if metrics is None: metrics = [] self._space["metrics"] = metrics def check_objective(self, objective): """ :meta private: """ if not type(objective) is str and not callable(objective): raise WrongProblemObjective(objective) # elif type(objective) is str: # list_suffix = ["__min", "__max", "__last"] # for suffix in list_suffix: # if suffix in objective: # objective = objective.replace(suffix, "") # break # only one suffix autorized # objective = objective.replace("val_", "") # possible_names = list() # if type(self._space["metrics"]) is dict: # metrics = list(self._space["metrics"].values()) # for k in self._space["metrics"].keys(): # objective = objective.replace(f"{k}_", "") # else: # assuming it s a list # metrics = self._space["metrics"] # for val in ["loss"] + metrics: # if callable(val): # possible_names.append(val.__name__) # else: # possible_names.append(val) # if not (objective in possible_names): # raise WrongProblemObjective(objective, possible_names) def objective(self, objective): """Define the objective you want to maximize for the search. If you want to use the validation accuracy at the last epoch: .. code-block:: python problem.objective("val_acc") .. note:: Be sure to define ``acc`` in the ``problem.metrics(["acc"])``. It can accept some prefix and suffix such as ``__min, __max, __last``: .. code-block:: python problem.objective("-val_acc__max") It can be a ``callable``: .. code-block:: python def myobjective(history: dict) -> float: return history["val_acc"][-1] problem.objective(myobjective) Args: objective (str or callable): The objective will be maximized. If ``objective`` is ``str`` then it should be either 'loss' or a defined metric. You can use the ``'val_'`` prefix when you want to select the objective on the validation set. You can use one of ``['__min', '__max', '__last']`` which respectively means you want to select the min, max or last value among all epochs. Using '__last' will save a consequent compute time because the evaluation will not compute metrics on validation set for all epochs but the last. If ``objective`` is callable it should return a scalar value (i.e. float) and it will take a ``dict`` parameter. The ``dict`` will contain keys corresponding to loss and metrics such as ``['loss', 'val_loss', 'r2', 'val_r2]``, it will also contains ``'n_parameters'`` which corresponds to the number of trainable parameters of the current model, ``'training_time'`` which corresponds to the time required to train the model, ``'predict_time'`` which corresponds to the time required to make a prediction over the whole validation set. If this callable has a ``'__last'`` suffix then the evaluation will only compute validation loss/metrics for the last epoch. If this callable has contains 'with_pred' in its name then the ``dict`` will have two other keys ``['y_pred', 'y_true']`` where ``'y_pred`` corresponds to prediction of the model on validation set and ``'y_true'`` corresponds to real prediction. Raise: WrongProblemObjective: raised when the objective is of a wrong definition. """ if ( not self._space.get("loss") is None and not self._space.get("metrics") is None ): self.check_objective(objective) else: raise NaProblemError( ".loss and .metrics should be defined before .objective!" ) self._space["objective"] = objective @property def space(self): keys = list(self._space.keys()) keys.sort() space = OrderedDict(**{d: self._space[d] for d in keys}) return space def build_search_space(self, seed=None): """Build and return a search space object using the infered data shapes after loading data. Returns: KSearchSpace: A search space instance. """ config = self.space input_shape, output_shape, _ = setup_data(config, add_to_config=False) search_space = get_search_space(config, input_shape, output_shape, seed=seed) return search_space def get_keras_model(self, arch_seq: list) -> tf.keras.Model: """Get a keras model object from a set of decisions in the current search space. Args: arch_seq (list): a list of int of floats describing a choice of operations for the search space defined in the current problem. """ search_space = self.build_search_space() return search_space.sample(arch_seq) def gen_config(self, arch_seq: list, hp_values: list) -> dict: """Generate a ``dict`` configuration from the ``arch_seq`` and ``hp_values`` passed. Args: arch_seq (list): a valid embedding of a neural network described by the search space of the current ``NaProblem``. hp_values (list): a valid list of hyperparameters corresponding to the defined hyperparameters of the current ``NaProblem``. """ config = deepcopy(self.space) # architecture DNA config["arch_seq"] = arch_seq # replace hp values in the config hp_names = self._hp_space._space.get_hyperparameter_names() for hp_name, hp_value in zip(hp_names, hp_values): if hp_name == "loss": config["loss"] = hp_value else: config["hyperparameters"][hp_name] = hp_value return config def extract_hp_values(self, config): """Extract the value of hyperparameters present in ``config`` based on the defined hyperparameters in the current ``NaProblem``""" hp_names = self.hyperparameter_names hp_values = [] for hp_name in hp_names: if hp_name == "loss": hp_values.append(config["loss"]) else: hp_values.append(config["hyperparameters"][hp_name]) return hp_values @property def hyperparameter_names(self): """The list of hyperparameters names.""" return self._hp_space.hyperparameter_names @property def default_hp_configuration(self): """The default configuration as a dictionnary.""" return self._hp_space.default_configuration def module_location(attr): """ :meta private: """ return f"{attr.__module__}.{attr.__name__}"
19,985
36.287313
1,442
py
deephyper
deephyper-master/deephyper/keras/callbacks/learning_rate_warmup.py
""" Adapted from Horovod implementation: https://github.com/horovod/horovod/blob/master/horovod/keras/callbacks.py """ import tensorflow as tf class LearningRateScheduleCallback(tf.keras.callbacks.Callback): def __init__( self, initial_lr, multiplier, start_epoch=0, end_epoch=None, staircase=True, momentum_correction=True, steps_per_epoch=None, *args ): super(LearningRateScheduleCallback, self).__init__(*args) self.start_epoch = start_epoch self.end_epoch = end_epoch self.staircase = staircase self.momentum_correction = momentum_correction self.initial_lr = initial_lr self.restore_momentum = None self.steps_per_epoch = steps_per_epoch self.current_epoch = None if not callable(multiplier): self.staircase = True self.multiplier = lambda epoch: multiplier else: self.multiplier = multiplier if self.initial_lr is None: raise ValueError("Parameter `initial_lr` is required") def _autodetect_steps_per_epoch(self): if self.params.get("steps"): # The number of steps is provided in the parameters. return self.params["steps"] elif self.params.get("samples") and self.params.get("batch_size"): # Compute the number of steps per epoch using # of samples and a batch size. return self.params["samples"] // self.params["batch_size"] else: raise ValueError( "Could not autodetect the number of steps per epoch. " "Please specify the steps_per_epoch parameter to the " "%s() or upgrade to the latest version of Keras." % self.__class__.__name__ ) def _adjust_learning_rate(self, epoch): old_lr = tf.keras.backend.get_value(self.model.optimizer.lr) new_lr = self.initial_lr * self.multiplier(epoch) tf.keras.backend.set_value(self.model.optimizer.lr, new_lr) if hasattr(self.model.optimizer, "momentum") and self.momentum_correction: # See the paper cited above for more information about momentum correction. self.restore_momentum = tf.keras.backend.get_value( self.model.optimizer.momentum ) tf.keras.backend.set_value( self.model.optimizer.momentum, self.restore_momentum * new_lr / old_lr ) def _restore_momentum_if_needed(self): if self.restore_momentum: tf.keras.backend.set_value( self.model.optimizer.momentum, self.restore_momentum ) self.restore_momentum = None def on_train_begin(self, logs=None): if self.initial_lr is None: self.initial_lr = tf.keras.backend.get_value(self.model.optimizer.lr) if not self.staircase and not self.steps_per_epoch: self.steps_per_epoch = self._autodetect_steps_per_epoch() def on_epoch_begin(self, epoch, logs=None): self.current_epoch = epoch def on_batch_begin(self, batch, logs=None): if self.current_epoch < self.start_epoch or ( self.end_epoch is not None and self.current_epoch >= self.end_epoch ): # Outside of the adjustment scope. return if self.staircase and batch == 0: # Do on first batch of every epoch. self._adjust_learning_rate(self.current_epoch) elif not self.staircase: epoch = self.current_epoch + float(batch) / self.steps_per_epoch self._adjust_learning_rate(epoch) def on_batch_end(self, batch, logs=None): self._restore_momentum_if_needed() def on_epoch_end(self, epoch, logs=None): if logs is not None: # Log current learning rate. logs["lr"] = tf.keras.backend.get_value(self.model.optimizer.lr) class LearningRateWarmupCallback(LearningRateScheduleCallback): def __init__( self, n_replicas, initial_lr, warmup_epochs=5, momentum_correction=True, steps_per_epoch=None, verbose=0, *args ): def multiplier(epoch): # Adjust epoch to produce round numbers at the end of each epoch, so that TensorBoard # learning rate graphs look better. epoch += 1.0 / self.steps_per_epoch return 1.0 / n_replicas * (epoch * (n_replicas - 1) / warmup_epochs + 1) super(LearningRateWarmupCallback, self).__init__( initial_lr, multiplier, start_epoch=0, end_epoch=warmup_epochs, staircase=False, momentum_correction=momentum_correction, steps_per_epoch=steps_per_epoch, *args ) self.verbose = verbose def on_epoch_end(self, epoch, logs=None): super(LearningRateWarmupCallback, self).on_epoch_end(epoch, logs) if epoch == self.end_epoch - 1 and self.verbose > 0: new_lr = tf.keras.backend.get_value(self.model.optimizer.lr) print( "\nEpoch %d: finished gradual learning rate warmup to %g." % (epoch + 1, new_lr) )
5,317
35.930556
110
py
deephyper
deephyper-master/deephyper/keras/callbacks/utils.py
from typing import Type import deephyper import deephyper.core.exceptions import tensorflow as tf def import_callback(cb_name: str) -> Type[tf.keras.callbacks.Callback]: """Import a callback class from its name. Args: cb_name (str): class name of the callback to import fron ``tensorflow.keras.callbacks`` or ``deephyper.keras.callbacks``. Raises: DeephyperRuntimeError: raised if the class name of the callback is not registered in corresponding packages. Returns: tensorflow.keras.callbacks.Callback: the class corresponding to the given class name. """ if cb_name in dir(tf.keras.callbacks): return getattr(tf.keras.callbacks, cb_name) elif cb_name in dir(deephyper.keras.callbacks): return getattr(deephyper.keras.callbacks, cb_name) else: raise deephyper.core.exceptions.DeephyperRuntimeError( f"Callback '{cb_name}' is not registered in tensorflow.keras and deephyper.keras.callbacks." )
1,000
34.75
129
py
deephyper
deephyper-master/deephyper/keras/callbacks/stop_if_unfeasible.py
import time import tensorflow as tf class StopIfUnfeasible(tf.keras.callbacks.Callback): def __init__(self, time_limit=600, patience=20): super().__init__() self.time_limit = time_limit self.timing = list() self.stopped = False # boolean set to True if the model training has been stopped due to time_limit condition self.patience = patience def set_params(self, params): self.params = params if self.params["steps"] is None: self.steps = self.params["samples"] // self.params["batch_size"] self.steps = self.params["samples"] // self.params["batch_size"] if self.steps * self.params["batch_size"] < self.params["samples"]: self.steps += 1 else: self.steps = self.params["steps"] def on_batch_begin(self, batch, logs=None): """Called at the beginning of a training batch in `fit` methods. Subclasses should override for any actions to run. Args: batch (int): index of batch within the current epoch. logs (dict): has keys `batch` and `size` representing the current batch number and the size of the batch. """ self.timing.append(time.time()) def on_batch_end(self, batch, logs=None): """Called at the end of a training batch in `fit` methods. Subclasses should override for any actions to run. Args: batch (int): index of batch within the current epoch. logs (dict): metric results for this batch. """ self.timing[-1] = time.time() - self.timing[-1] self.avr_batch_time = sum(self.timing) / len(self.timing) self.estimate_training_time = sum(self.timing) + self.avr_batch_time * ( self.steps - len(self.timing) ) if ( len(self.timing) >= self.patience and self.estimate_training_time > self.time_limit ): self.stopped = True self.model.stop_training = True
2,047
36.236364
118
py
deephyper
deephyper-master/deephyper/keras/callbacks/stop_on_timeout.py
from datetime import datetime from tensorflow.keras.callbacks import Callback class TerminateOnTimeOut(Callback): def __init__(self, timeout_in_min=10): super(TerminateOnTimeOut, self).__init__() self.run_timestamp = None self.timeout_in_sec = timeout_in_min * 60 # self.validation_data = validation_data def on_train_begin(self, logs={}): self.run_timestamp = datetime.now() def on_batch_end(self, epoch, logs={}): run_end = datetime.now() run_duration = run_end - self.run_timestamp run_in_sec = run_duration.total_seconds() # / (60 * 60) # print(' - current training time = %2.3fs/%2.3fs' % (run_in_sec, self.timeout_in_sec)) if self.timeout_in_sec != -1: if run_in_sec >= self.timeout_in_sec: print( " - timeout: training time = %2.3fs/%2.3fs" % (run_in_sec, self.timeout_in_sec) ) # print('TimeoutRuntime: %2.3fs, Maxtime: %2.3fs' % (run_in_sec, self.timeout_in_sec)) self.model.stop_training = True # if self.validation_data is not None: # x, y = self.validation_data[0], self.validation_data[1] # loss, acc = self.model.evaluate(x,y) # #print(self.model.history.keys())
1,364
40.363636
102
py
deephyper
deephyper-master/deephyper/keras/callbacks/csv_extended_logger.py
import collections import io import time import csv import numpy as np import six import tensorflow as tf from tensorflow.python.lib.io import file_io from tensorflow.python.util.compat import collections_abc class CSVExtendedLogger(tf.keras.callbacks.Callback): """Callback that streams epoch results to a csv file. Supports all values that can be represented as a string, including 1D iterables such as np.ndarray. Example: .. code-block:: python csv_logger = CSVLogger('training.log') model.fit(X_train, Y_train, callbacks=[csv_logger]) Args: filename: filename of the csv file, e.g. 'run/log.csv'. separator: string used to separate elements in the csv file. append: True: append if file exists (useful for continuing training). False: overwrite existing file, """ def __init__(self, filename, separator=",", append=False): self.sep = separator self.filename = filename self.append = append self.writer = None self.keys = None self.append_header = True if six.PY2: self.file_flags = "b" self._open_args = {} else: self.file_flags = "" self._open_args = {"newline": "\n"} self.timestamp = None super(CSVExtendedLogger, self).__init__() def on_train_begin(self, logs=None): if self.append: if file_io.file_exists(self.filename): with open(self.filename, "r" + self.file_flags) as f: self.append_header = not bool(len(f.readline())) mode = "a" else: mode = "w" self.csv_file = io.open( self.filename, mode + self.file_flags, **self._open_args ) def on_epoch_begin(self, epoch, logs=None): self.timestamp = time.time() def on_epoch_end(self, epoch, logs=None): timestamp = time.time() duration = timestamp - self.timestamp # duration of curent epoch logs = logs or {} def handle_value(k): is_zero_dim_ndarray = isinstance(k, np.ndarray) and k.ndim == 0 if isinstance(k, six.string_types): return k elif isinstance(k, collections_abc.Iterable) and not is_zero_dim_ndarray: return '"[%s]"' % (", ".join(map(str, k))) else: return k if self.keys is None: self.keys = sorted(logs.keys()) if self.model.stop_training: # We set NA so that csv parsers do not fail for this last epoch. logs = {k: logs[k] if k in logs else (k, "NA") for k in self.keys} if not self.writer: class CustomDialect(csv.excel): delimiter = self.sep fieldnames = ["epoch", "timestamp", "duration"] + self.keys if six.PY2: fieldnames = [f"{x}" for x in fieldnames] self.writer = csv.DictWriter( self.csv_file, fieldnames=fieldnames, dialect=CustomDialect ) if self.append_header: self.writer.writeheader() row_dict = collections.OrderedDict( {"epoch": epoch, "timestamp": timestamp, "duration": duration} ) row_dict.update((key, handle_value(logs[key])) for key in self.keys) self.writer.writerow(row_dict) self.csv_file.flush() def on_train_end(self, logs=None): self.csv_file.close() self.writer = None
3,539
30.891892
85
py
deephyper
deephyper-master/deephyper/keras/callbacks/__init__.py
from deephyper.keras.callbacks.utils import import_callback from deephyper.keras.callbacks.stop_if_unfeasible import StopIfUnfeasible from deephyper.keras.callbacks.csv_extended_logger import CSVExtendedLogger from deephyper.keras.callbacks.time_stopping import TimeStopping from deephyper.keras.callbacks.learning_rate_warmup import ( LearningRateScheduleCallback, LearningRateWarmupCallback, ) __all__ = [ "import_callback", "StopIfUnfeasible", "CSVExtendedLogger", "TimeStopping", "LearningRateScheduleCallback", "LearningRateWarmupCallback", ]
581
31.333333
75
py
deephyper
deephyper-master/deephyper/keras/callbacks/time_stopping.py
"""Callback that stops training when a specified amount of time has passed. source: https://github.com/tensorflow/addons/blob/master/tensorflow_addons/callbacks/time_stopping.py """ import datetime import time import tensorflow as tf class TimeStopping(tf.keras.callbacks.Callback): """Stop training when a specified amount of time has passed. Args: seconds: maximum amount of time before stopping. Defaults to 86400 (1 day). verbose: verbosity mode. Defaults to 0. """ def __init__(self, seconds: int = 86400, verbose: int = 0): super().__init__() self.seconds = seconds self.verbose = verbose self.stopped_epoch = None def on_train_begin(self, logs=None): self.stopping_time = time.time() + self.seconds def on_epoch_end(self, epoch, logs={}): if time.time() >= self.stopping_time: self.model.stop_training = True self.stopped_epoch = epoch def on_train_end(self, logs=None): if self.stopped_epoch is not None and self.verbose > 0: formatted_time = datetime.timedelta(seconds=self.seconds) msg = "Timed stopping at epoch {} after training for {}".format( self.stopped_epoch + 1, formatted_time ) print(msg) def get_config(self): config = {"seconds": self.seconds, "verbose": self.verbose} base_config = super().get_config() return {**base_config, **config}
1,499
30.25
101
py
deephyper
deephyper-master/deephyper/keras/layers/_mpnn.py
import tensorflow as tf import tensorflow.keras.backend as K from tensorflow.keras import activations from tensorflow.keras.layers import Dense class SparseMPNN(tf.keras.layers.Layer): """Message passing cell. Args: state_dim (int): number of output channels. T (int): number of message passing repetition. attn_heads (int): number of attention heads. attn_method (str): type of attention methods. aggr_method (str): type of aggregation methods. activation (str): type of activation functions. update_method (str): type of update functions. """ def __init__( self, state_dim, T, aggr_method, attn_method, update_method, attn_head, activation, ): super(SparseMPNN, self).__init__(self) self.state_dim = state_dim self.T = T self.activation = activations.get(activation) self.aggr_method = aggr_method self.attn_method = attn_method self.attn_head = attn_head self.update_method = update_method def build(self, input_shape): self.embed = tf.keras.layers.Dense(self.state_dim, activation=self.activation) self.MP = MessagePassing( self.state_dim, self.aggr_method, self.activation, self.attn_method, self.attn_head, self.update_method, ) self.built = True def call(self, inputs, **kwargs): """Apply the layer on input tensors. Args: inputs (list): X (tensor): node feature tensor (batch size * # nodes * # node features) A (tensor): edge pair tensor (batch size * # edges * 2), one is source ID, one is target ID E (tensor): edge feature tensor (batch size * # edges * # edge features) mask (tensor): node mask tensor to mask out non-existent nodes (batch size * # nodes) degree (tensor): node degree tensor for GCN attention (batch size * # edges) Returns: X (tensor): results after several repetitions of edge network, attention, aggregation and update function (batch size * # nodes * # node features) """ # the input contains a list of five tensors X, A, E, mask, degree = inputs # edge pair needs to be in the int format A = tf.cast(A, tf.int32) # this is a limitation of MPNN in general, the node feature is mapped to (batch size * # nodes * # node # features) X = self.embed(X) # run T times message passing for _ in range(self.T): X = self.MP([X, A, E, mask, degree]) return X class MessagePassing(tf.keras.layers.Layer): """Message passing layer. Args: state_dim (int): number of output channels. attn_heads (int): number of attention heads. attn_method (str): type of attention methods. aggr_method (str): type of aggregation methods. activation (str): type of activation functions. update_method (str): type of update functions. """ def __init__( self, state_dim, aggr_method, activation, attn_method, attn_head, update_method ): super(MessagePassing, self).__init__(self) self.state_dim = state_dim self.aggr_method = aggr_method self.activation = activation self.attn_method = attn_method self.attn_head = attn_head self.update_method = update_method def build(self, input_shape): self.message_passer = MessagePasserNNM( self.state_dim, self.attn_head, self.attn_method, self.aggr_method, self.activation, ) if self.update_method == "gru": self.update_functions = UpdateFuncGRU(self.state_dim) elif self.update_method == "mlp": self.update_functions = UpdateFuncMLP(self.state_dim, self.activation) self.built = True def call(self, inputs, **kwargs): """Apply the layer on input tensors. Args: inputs (list): X (tensor): node feature tensor (batch size * # nodes * state dimension) A (tensor): edge pair tensor (batch size * # edges * 2), one is source ID, one is target ID E (tensor): edge feature tensor (batch size * # edges * # edge features) mask (tensor): node mask tensor to mask out non-existent nodes (batch size * # nodes) degree (tensor): node degree tensor for GCN attention (batch size * # edges) Returns: updated_nodes (tensor): results after edge network, attention, aggregation and update function (batch size * # nodes * state dimension) """ # the input contains a list of five tensors X, A, E, mask, degree = inputs # use the message passing to generate aggregated results # agg_m (batch size * # nodes * state dimension) agg_m = self.message_passer([X, A, E, degree]) # expand the mask to (batch size * # nodes * state dimension) mask = tf.tile(mask[..., None], [1, 1, self.state_dim]) # use the mask to screen out non-existent nodes # agg_m (batch size * # nodes * state dimension) agg_m = tf.multiply(agg_m, mask) # update function using the old node feature X and new aggregated node feature agg_m # updated_nodes (batch size * # nodes * state dimension) updated_nodes = self.update_functions([X, agg_m]) # use the mask to screen out non-existent nodes # updated_nodes (batch size * # nodes * state dimension) updated_nodes = tf.multiply(updated_nodes, mask) return updated_nodes class MessagePasserNNM(tf.keras.layers.Layer): """Message passing kernel. Args: state_dim (int): number of output channels. attn_heads (int): number of attention heads. attn_method (str): type of attention methods. aggr_method (str): type of aggregation methods. activation (str): type of activation functions. """ def __init__(self, state_dim, attn_heads, attn_method, aggr_method, activation): super(MessagePasserNNM, self).__init__() self.state_dim = state_dim self.attn_heads = attn_heads self.attn_method = attn_method self.aggr_method = aggr_method self.activation = activation def build(self, input_shape): self.nn1 = tf.keras.layers.Dense(units=32, activation=tf.nn.relu) self.nn2 = tf.keras.layers.Dense(units=32, activation=tf.nn.relu) self.nn3 = tf.keras.layers.Dense( units=self.attn_heads * self.state_dim * self.state_dim, activation=tf.nn.relu, ) if self.attn_method == "gat": self.attn_func = AttentionGAT(self.state_dim, self.attn_heads) elif self.attn_method == "sym-gat": self.attn_func = AttentionSymGAT(self.state_dim, self.attn_heads) elif self.attn_method == "cos": self.attn_func = AttentionCOS(self.state_dim, self.attn_heads) elif self.attn_method == "linear": self.attn_func = AttentionLinear(self.state_dim, self.attn_heads) elif self.attn_method == "gen-linear": self.attn_func = AttentionGenLinear(self.state_dim, self.attn_heads) elif self.attn_method == "const": self.attn_func = AttentionConst(self.state_dim, self.attn_heads) elif self.attn_method == "gcn": self.attn_func = AttentionGCN(self.state_dim, self.attn_heads) self.bias = self.add_weight( name="attn_bias", shape=[self.state_dim], initializer="zeros" ) self.built = True def call(self, inputs, **kwargs): """Apply the layer on input tensors. Args: inputs (list): X (tensor): node feature tensor (batch size * # nodes * state dimension) A (tensor): edge pair tensor (batch size * # edges * 2), one is source ID, one is target ID E (tensor): edge feature tensor (batch size * # edges * # edge features) degree (tensor): node degree tensor for GCN attention (batch size * # edges) Returns: output (tensor): results after edge network, attention and aggregation (batch size * # nodes * state dimension) """ # Edge network to transform edge information to message weight # the input contains a list of four tensors X, A, E, degree = inputs # N is the number of nodes (scalar) N = K.int_shape(X)[1] # extract target and source IDs from the edge pair # targets (batch size * # edges) # sources (batch size * # edges) targets, sources = A[..., -2], A[..., -1] # the first edge network layer that maps edge features to a weight tensor W # W (batch size * # edges * 128) W = self.nn1(E) # W (batch size * # edges * 128) W = self.nn2(W) # W (batch size * # edges * state dimension ** 2) W = self.nn3(W) # reshape W to (batch size * # edges * #attention heads * state dimension * state dimension) W = tf.reshape( W, [-1, tf.shape(E)[1], self.attn_heads, self.state_dim, self.state_dim] ) # expand the dimension of node features to # (batch size * # nodes * state dimension * #attention heads) X = tf.tile(X[..., None], [1, 1, 1, self.attn_heads]) # transpose the node features to # (batch size * # nodes * #attention heads * node features) X = tf.transpose(X, [0, 1, 3, 2]) # attention added to the message weight # attn_coef (batch size * # edges * #attention heads * state dimension) attn_coef = self.attn_func([X, N, targets, sources, degree]) # gather source node features # The batch_dims argument lets you gather different items from each element of a batch. # Using batch_dims=1 is equivalent to having an outer loop over the first axis of params and indices: # Here is an example from https://www.tensorflow.org/api_docs/python/tf/gather # params = tf.constant([ # [0, 0, 1, 0, 2], # [3, 0, 0, 0, 4], # [0, 5, 0, 6, 0]]) # indices = tf.constant([ # [2, 4], # [0, 4], # [1, 3]]) # tf.gather(params, indices, axis=1, batch_dims=1).numpy() # array([[1, 2], # [3, 4], # [5, 6]], dtype=int32) # messages (batch size * # edges * #attention heads * state dimension) messages = tf.gather(X, sources, batch_dims=1, axis=1) # messages (batch size * # edges * #attention heads * state dimension * 1) messages = messages[..., None] # W (batch size * # edges * #attention heads * state dimension * state dimension) # messages (batch size * # edges * #attention heads * state dimension * 1) # --> messages (batch size * # edges * #attention heads * state dimension * 1) messages = tf.matmul(W, messages) # messages (batch size * # edges * #attention heads * state dimension) messages = messages[..., 0] # attn_coef (batch size * # edges * # attention heads * state dimension) # messages (batch size * # edges * # attention heads * state dimension) # --> output (batch size * # edges * # attention heads * state dimension) output = attn_coef * messages # batch size num_rows = tf.shape(targets)[0] # [0, ..., batch size] (batch size) rows_idx = tf.range(num_rows) # N is # nodes, add this to distinguish each batch segment_ids_per_row = targets + N * tf.expand_dims(rows_idx, axis=1) # Aggregation to summarize neighboring node messages # output (batch size * # nodes * # attention heads * state dimension) if self.aggr_method == "max": output = tf.math.unsorted_segment_max( output, segment_ids_per_row, N * num_rows ) elif self.aggr_method == "mean": output = tf.math.unsorted_segment_mean( output, segment_ids_per_row, N * num_rows ) elif self.aggr_method == "sum": output = tf.math.unsorted_segment_sum( output, segment_ids_per_row, N * num_rows ) # output the mean of all attention heads # output (batch size * # nodes * # attention heads * state dimension) output = tf.reshape(output, [-1, N, self.attn_heads, self.state_dim]) # output (batch size * # nodes * state dimension) output = tf.reduce_mean(output, axis=-2) # add bias, output (batch size * # nodes * state dimension) output = K.bias_add(output, self.bias) return output class UpdateFuncGRU(tf.keras.layers.Layer): """Gated recurrent unit update function. Check details here https://arxiv.org/abs/1412.3555 Args: state_dim (int): number of output channels. """ def __init__(self, state_dim): super(UpdateFuncGRU, self).__init__() self.state_dim = state_dim def build(self, input_shape): self.concat_layer = tf.keras.layers.Concatenate(axis=1) self.GRU = tf.keras.layers.GRU(self.state_dim) self.built = True def call(self, inputs, **kwargs): """Apply the layer on input tensors. Args: inputs (list): old_state (tensor): node hidden feature tensor (batch size * # nodes * state dimension) agg_messages (tensor): node hidden feature tensor (batch size * # nodes * state dimension) Returns: activation (tensor): activated tensor from update function (batch size * # nodes * state dimension) """ # Remember node dim # old_state (batch size * # nodes * state dimension) # agg_messages (batch size * # nodes * state dimension) old_state, agg_messages = inputs # B is batch size # N is # nodes # F is # node features = state dimension B, N, F = K.int_shape(old_state) # similar to B, N, F B1, N1, F1 = K.int_shape(agg_messages) # reshape so GRU can be applied, concat so old_state and messages are in sequence # old_state (batch size * # nodes * 1 * state dimension) old_state = tf.reshape(old_state, [-1, 1, F]) # agg_messages (batch size * # nodes * 1 * state dimension) agg_messages = tf.reshape(agg_messages, [-1, 1, F1]) # agg_messages (batch size * # nodes * 2 * state dimension) concat = self.concat_layer([old_state, agg_messages]) # Apply GRU and then reshape so it can be returned # activation (batch size * # nodes * state dimension) activation = self.GRU(concat) activation = tf.reshape(activation, [-1, N, F]) return activation class UpdateFuncMLP(tf.keras.layers.Layer): """Multi-layer perceptron update function. Args: state_dim (int): number of output channels. activation (str): the type of activation functions. """ def __init__(self, state_dim, activation): super(UpdateFuncMLP, self).__init__() self.state_dim = state_dim self.activation = activation def build(self, input_shape): self.concat_layer = tf.keras.layers.Concatenate(axis=-1) self.dense = tf.keras.layers.Dense( self.state_dim, activation=self.activation, kernel_initializer="zeros" ) def call(self, inputs, **kwargs): """Apply the layer on input tensors. Args: inputs (list): old_state (tensor): node hidden feature tensor agg_messages (tensor): node hidden feature tensor Returns: activation (tensor): activated tensor from update function ( """ old_state, agg_messages = inputs concat = self.concat_layer([old_state, agg_messages]) activation = self.dense(concat) return activation class AttentionGAT(tf.keras.layers.Layer): """GAT Attention. Check details here https://arxiv.org/abs/1710.10903 The attention coefficient between node :math:`i` and :math:`j` is calculated as: .. math:: \\text{LeakyReLU}(\\textbf{a}(\\textbf{Wh}_i||\\textbf{Wh}_j)) where :math:`\\textbf{a}` is a trainable vector, and :math:`||` represents concatenation. Args: state_dim (int): number of output channels. attn_heads (int): number of attention heads. """ def __init__(self, state_dim, attn_heads): super(AttentionGAT, self).__init__() self.state_dim = state_dim self.attn_heads = attn_heads def build(self, input_shape): self.attn_kernel_self = self.add_weight( name="attn_kernel_self", shape=[self.state_dim, self.attn_heads, 1], initializer="glorot_uniform", ) self.attn_kernel_adjc = self.add_weight( name="attn_kernel_adjc", shape=[self.state_dim, self.attn_heads, 1], initializer="glorot_uniform", ) self.built = True def call(self, inputs, **kwargs): """Apply the layer on input tensors. Args: inputs (list): X (tensor): node feature tensor N (int): number of nodes targets (tensor): target node index tensor sources (tensor): source node index tensor degree (tensor): node degree sqrt tensor (for GCN attention) Returns: attn_coef (tensor): attention coefficient tensor """ X, N, targets, sources, _ = inputs attn_kernel_self = tf.transpose(self.attn_kernel_self, (2, 1, 0)) attn_kernel_adjc = tf.transpose(self.attn_kernel_adjc, (2, 1, 0)) attn_for_self = tf.reduce_sum(X * attn_kernel_self[None, ...], -1) attn_for_self = tf.gather(attn_for_self, targets, batch_dims=1) attn_for_adjc = tf.reduce_sum(X * attn_kernel_adjc[None, ...], -1) attn_for_adjc = tf.gather(attn_for_adjc, sources, batch_dims=1) attn_coef = attn_for_self + attn_for_adjc attn_coef = tf.nn.leaky_relu(attn_coef, alpha=0.2) attn_coef = tf.exp( attn_coef - tf.gather(tf.math.unsorted_segment_max(attn_coef, targets, N), targets) ) attn_coef /= tf.gather( tf.math.unsorted_segment_max(attn_coef, targets, N) + 1e-9, targets ) attn_coef = tf.nn.dropout(attn_coef, 0.5) attn_coef = attn_coef[..., None] return attn_coef class AttentionSymGAT(tf.keras.layers.Layer): """GAT Symmetry Attention. The attention coefficient between node :math:`i` and :math:`j` is calculated as: .. math:: \\alpha_{ij} + \\alpha_{ij} based on GAT. Args: state_dim (int): number of output channels. attn_heads (int): number of attention heads. """ def __init__(self, state_dim, attn_heads): super(AttentionSymGAT, self).__init__() self.state_dim = state_dim self.attn_heads = attn_heads def build(self, input_shape): self.attn_kernel_self = self.add_weight( name="attn_kernel_self", shape=[self.state_dim, self.attn_heads, 1], initializer="glorot_uniform", ) self.attn_kernel_adjc = self.add_weight( name="attn_kernel_adjc", shape=[self.state_dim, self.attn_heads, 1], initializer="glorot_uniform", ) self.built = True def call(self, inputs, **kwargs): """Apply the layer on input tensors. Args: inputs (list): X (tensor): node feature tensor N (int): number of nodes targets (tensor): target node index tensor sources (tensor): source node index tensor degree (tensor): node degree sqrt tensor (for GCN attention) Returns: attn_coef (tensor): attention coefficient tensor """ X, N, targets, sources, _ = inputs attn_kernel_self = tf.transpose(self.attn_kernel_self, (2, 1, 0)) attn_kernel_adjc = tf.transpose(self.attn_kernel_adjc, (2, 1, 0)) attn_for_self = tf.reduce_sum(X * attn_kernel_self[None, ...], -1) attn_for_self = tf.gather(attn_for_self, targets, batch_dims=1) attn_for_adjc = tf.reduce_sum(X * attn_kernel_adjc[None, ...], -1) attn_for_adjc = tf.gather(attn_for_adjc, sources, batch_dims=1) attn_for_self_reverse = tf.gather(attn_for_self, sources, batch_dims=1) attn_for_adjc_reverse = tf.gather(attn_for_self, targets, batch_dims=1) attn_coef = ( attn_for_self + attn_for_adjc + attn_for_self_reverse + attn_for_adjc_reverse ) attn_coef = tf.nn.leaky_relu(attn_coef, alpha=0.2) attn_coef = tf.exp( attn_coef - tf.gather(tf.math.unsorted_segment_max(attn_coef, targets, N), targets) ) attn_coef /= tf.gather( tf.math.unsorted_segment_max(attn_coef, targets, N) + 1e-9, targets ) attn_coef = tf.nn.dropout(attn_coef, 0.5) attn_coef = attn_coef[..., None] return attn_coef class AttentionCOS(tf.keras.layers.Layer): """COS Attention. Check details here https://arxiv.org/abs/1803.07294 The attention coefficient between node $i$ and $j$ is calculated as: .. math:: \\textbf{a}(\\textbf{Wh}_i || \\textbf{Wh}_j) where :math:`\\textbf{a}` is a trainable vector. Args: state_dim (int): number of output channels. attn_heads (int): number of attention heads. """ def __init__(self, state_dim, attn_heads): super(AttentionCOS, self).__init__() self.state_dim = state_dim self.attn_heads = attn_heads def build(self, input_shape): self.attn_kernel_self = self.add_weight( name="attn_kernel_self", shape=[self.state_dim, self.attn_heads, 1], initializer="glorot_uniform", ) self.attn_kernel_adjc = self.add_weight( name="attn_kernel_adjc", shape=[self.state_dim, self.attn_heads, 1], initializer="glorot_uniform", ) self.built = True def call(self, inputs, **kwargs): """Apply the layer on input tensors. Args: inputs (list): X (tensor): node feature tensor N (int): number of nodes targets (tensor): target node index tensor sources (tensor): source node index tensor degree (tensor): node degree sqrt tensor (for GCN attention) Returns: attn_coef (tensor): attention coefficient tensor (batch, E, H, 1) """ X, N, targets, sources, _ = inputs attn_kernel_self = tf.transpose(self.attn_kernel_self, (2, 1, 0)) attn_kernel_adjc = tf.transpose(self.attn_kernel_adjc, (2, 1, 0)) attn_for_self = tf.reduce_sum(X * attn_kernel_self[None, ...], -1) attn_for_self = tf.gather(attn_for_self, targets, batch_dims=1) attn_for_adjc = tf.reduce_sum(X * attn_kernel_adjc[None, ...], -1) attn_for_adjc = tf.gather(attn_for_adjc, sources, batch_dims=1) attn_coef = tf.multiply(attn_for_self, attn_for_adjc) attn_coef = tf.nn.leaky_relu(attn_coef, alpha=0.2) attn_coef = tf.exp( attn_coef - tf.gather(tf.math.unsorted_segment_max(attn_coef, targets, N), targets) ) attn_coef /= tf.gather( tf.math.unsorted_segment_max(attn_coef, targets, N) + 1e-9, targets ) attn_coef = tf.nn.dropout(attn_coef, 0.5) attn_coef = attn_coef[..., None] return attn_coef class AttentionLinear(tf.keras.layers.Layer): """Linear Attention. The attention coefficient between node :math:`i` and :math:`j` is calculated as: .. math:: \\text{tanh} (\\textbf{a}_l\\textbf{Wh}_i + \\textbf{a}_r\\textbf{Wh}_j) where :math:`\\textbf{a}_l` and :math:`\\textbf{a}_r` are trainable vectors. Args: state_dim (int): number of output channels. attn_heads (int): number of attention heads. """ def __init__(self, state_dim, attn_heads): super(AttentionLinear, self).__init__() self.state_dim = state_dim self.attn_heads = attn_heads def build(self, input_shape): self.attn_kernel_adjc = self.add_weight( name="attn_kernel_adjc", shape=[self.state_dim, self.attn_heads, 1], initializer="glorot_uniform", ) self.built = True def call(self, inputs, **kwargs): """Apply the layer on input tensors. Args: inputs (list): X (tensor): node feature tensor N (int): number of nodes targets (tensor): target node index tensor sources (tensor): source node index tensor degree (tensor): node degree sqrt tensor (for GCN attention) Returns: attn_coef (tensor): attention coefficient tensor """ X, N, targets, sources, _ = inputs attn_kernel_adjc = tf.transpose(self.attn_kernel_adjc, (2, 1, 0)) attn_for_adjc = tf.reduce_sum(X * attn_kernel_adjc[None, ...], -1) attn_for_adjc = tf.gather(attn_for_adjc, sources, batch_dims=1) attn_coef = attn_for_adjc attn_coef = tf.nn.tanh(attn_coef) attn_coef = tf.exp( attn_coef - tf.gather(tf.math.unsorted_segment_max(attn_coef, targets, N), targets) ) attn_coef /= tf.gather( tf.math.unsorted_segment_max(attn_coef, targets, N) + 1e-9, targets ) attn_coef = tf.nn.dropout(attn_coef, 0.5) attn_coef = attn_coef[..., None] return attn_coef class AttentionGenLinear(tf.keras.layers.Layer): """Generalized Linear Attention. Check details here https://arxiv.org/abs/1802.00910 The attention coefficient between node :math:`i` and :math:`j` is calculated as: .. math:: \\textbf{W}_G \\text{tanh} (\\textbf{Wh}_i + \\textbf{Wh}_j) where :math:`\\textbf{W}_G` is a trainable matrix. Args: state_dim (int): number of output channels. attn_heads (int): number of attention heads. """ def __init__(self, state_dim, attn_heads): super(AttentionGenLinear, self).__init__() self.state_dim = state_dim self.attn_heads = attn_heads def build(self, input_shape): self.attn_kernel_self = self.add_weight( name="attn_kernel_self", shape=[self.state_dim, self.attn_heads, 1], initializer="glorot_uniform", ) self.attn_kernel_adjc = self.add_weight( name="attn_kernel_adjc", shape=[self.state_dim, self.attn_heads, 1], initializer="glorot_uniform", ) self.gen_nn = tf.keras.layers.Dense( units=self.attn_heads, kernel_initializer="glorot_uniform", use_bias=False ) self.built = True def call(self, inputs, **kwargs): """Apply the layer on input tensors. Args: inputs (list): X (tensor): node feature tensor N (int): number of nodes targets (tensor): target node index tensor sources (tensor): source node index tensor degree (tensor): node degree sqrt tensor (for GCN attention) Returns: attn_coef (tensor): attention coefficient tensor """ X, N, targets, sources, _ = inputs attn_kernel_self = tf.transpose(self.attn_kernel_self, (2, 1, 0)) attn_kernel_adjc = tf.transpose(self.attn_kernel_adjc, (2, 1, 0)) attn_for_self = tf.reduce_sum(X * attn_kernel_self[None, ...], -1) attn_for_self = tf.gather(attn_for_self, targets, batch_dims=1) attn_for_adjc = tf.reduce_sum(X * attn_kernel_adjc[None, ...], -1) attn_for_adjc = tf.gather(attn_for_adjc, sources, batch_dims=1) attn_coef = attn_for_self + attn_for_adjc attn_coef = tf.nn.tanh(attn_coef) attn_coef = self.gen_nn(attn_coef) attn_coef = tf.exp( attn_coef - tf.gather(tf.math.unsorted_segment_max(attn_coef, targets, N), targets) ) attn_coef /= tf.gather( tf.math.unsorted_segment_max(attn_coef, targets, N) + 1e-9, targets ) attn_coef = tf.nn.dropout(attn_coef, 0.5) attn_coef = attn_coef[..., None] return attn_coef class AttentionGCN(tf.keras.layers.Layer): """GCN Attention. The attention coefficient between node :math:`i` and :math:`j` is calculated as: .. math:: \\frac{1}{\sqrt{|\mathcal{N}(i)||\mathcal{N}(j)|}} where :math:`\mathcal{N}(i)` is the number of neighboring nodes of node :math:`i`. Args: state_dim (int): number of output channels. attn_heads (int): number of attention heads. """ def __init__(self, state_dim, attn_heads): super(AttentionGCN, self).__init__() self.state_dim = state_dim self.attn_heads = attn_heads def call(self, inputs, **kwargs): """Apply the layer on input tensors. Args: inputs (list): X (tensor): node feature tensor N (int): number of nodes targets (tensor): target node index tensor sources (tensor): source node index tensor degree (tensor): node degree sqrt tensor (for GCN attention) Returns: attn_coef (tensor): attention coefficient tensor """ _, _, _, _, degree = inputs attn_coef = degree[..., None, None] attn_coef = tf.tile(attn_coef, [1, 1, self.attn_heads, 1]) return attn_coef class AttentionConst(tf.keras.layers.Layer): """Constant Attention. The attention coefficient between node :math:`i` and :math:`j` is calculated as: .. math:: \\alpha_{ij} = 1 Args: state_dim (int): number of output channels. attn_heads (int): number of attention heads. """ def __init__(self, state_dim, attn_heads): super(AttentionConst, self).__init__() self.state_dim = state_dim self.attn_heads = attn_heads def call(self, inputs, **kwargs): """Apply the layer on input tensors. Args: inputs (list): X (tensor): node feature tensor N (int): number of nodes targets (tensor): target node index tensor sources (tensor): source node index tensor degree (tensor): node degree sqrt tensor (for GCN attention) Returns: attn_coef (tensor): attention coefficient tensor """ _, _, targets, _, degree = inputs attn_coef = tf.ones( (tf.shape(targets)[0], tf.shape(targets)[1], self.attn_heads, 1) ) return attn_coef class GlobalAttentionPool(tf.keras.layers.Layer): """Global Attention Pool. A gated attention global pooling layer as presented by [Li et al. (2017)](https://arxiv.org/abs/1511.05493). Details can be seen from https://github.com/danielegrattarola/spektral Args: state_dim (int): number of output channels. """ def __init__(self, state_dim, **kwargs): super(GlobalAttentionPool, self).__init__() self.state_dim = state_dim self.kwargs = kwargs def __str__(self): return "GlobalAttentionPool" def build(self, input_shape): self.features_layer = Dense(self.state_dim, name="features_layer") self.attention_layer = Dense( self.state_dim, name="attention_layer", activation="sigmoid" ) self.built = True def call(self, inputs, **kwargs): """Apply the layer on input tensors. Args: inputs (tensor): the node feature tensor Returns: GlobalAttentionPool tensor (tensor) """ inputs_linear = self.features_layer(inputs) attn = self.attention_layer(inputs) masked_inputs = inputs_linear * attn output = K.sum(masked_inputs, axis=-2, keepdims=False) return output class GlobalAttentionSumPool(tf.keras.layers.Layer): """Global Attention Summation Pool. Pools a graph by learning attention coefficients to sum node features. Details can be seen from https://github.com/danielegrattarola/spektral """ def __init__(self, **kwargs): super(GlobalAttentionSumPool, self).__init__() self.kwargs = kwargs def __str__(self): return "GlobalAttentionSumPool" def build(self, input_shape): F = int(input_shape[-1]) # Attention kernels self.attn_kernel = self.add_weight( shape=(F, 1), initializer="glorot_uniform", name="attn_kernel" ) self.built = True def call(self, inputs, **kwargs): """Apply the layer on input tensors. Args: inputs (tensor): the node feature tensor Returns: GlobalAttentionSumPool tensor (tensor) """ X = inputs attn_coeff = K.dot(X, self.attn_kernel) attn_coeff = K.squeeze(attn_coeff, -1) attn_coeff = K.softmax(attn_coeff) output = K.batch_dot(attn_coeff, X) return output class GlobalAvgPool(tf.keras.layers.Layer): """Global Average Pool. Takes the average over all the nodes or features. Details can be seen from https://github.com/danielegrattarola/spektral Args: axis (int): the axis to take average. """ def __init__(self, axis=-2, **kwargs): super(GlobalAvgPool, self).__init__() self.axis = axis self.kwargs = kwargs def __str__(self): return "GlobalAvgPool" def call(self, inputs, **kwargs): """Apply the layer on input tensors. Args: inputs (tensor): the node feature tensor Returns: GlobalAvgPool tensor (tensor) """ return tf.reduce_mean(inputs, axis=self.axis) class GlobalMaxPool(tf.keras.layers.Layer): """Global Max Pool. Takes the max value over all the nodes or features. Details can be seen from https://github.com/danielegrattarola/spektral Args: axis (int): the axis to take the max value. """ def __init__(self, axis=-2, **kwargs): super(GlobalMaxPool, self).__init__() self.axis = axis self.kwargs = kwargs def __str__(self): return "GlobalMaxPool" def call(self, inputs, **kwargs): """Apply the layer on input tensors. Args: inputs (tensor): the node feature tensor Returns: GlobalMaxPool tensor (tensor) """ return tf.reduce_max(inputs, axis=self.axis) class GlobalSumPool(tf.keras.layers.Layer): """Global Summation Pool. Takes the summation over all the nodes or features. Details can be seen from https://github.com/danielegrattarola/spektral Args: axis (int): the axis to take summation. """ def __init__(self, axis=-2, **kwargs): super(GlobalSumPool, self).__init__() self.axis = axis self.kwargs = kwargs def __str__(self): return "GlobalSumPool" def call(self, inputs, **kwargs): """Apply the layer on input tensors. Args: inputs (tensor): the node feature tensor Returns: GlobalSumPool tensor (tensor) """ return tf.reduce_sum(inputs, axis=self.axis)
36,142
35.471241
183
py
deephyper
deephyper-master/deephyper/keras/layers/__init__.py
from deephyper.keras.layers._mpnn import ( AttentionConst, AttentionCOS, AttentionGAT, AttentionGCN, AttentionGenLinear, AttentionLinear, AttentionSymGAT, GlobalAttentionPool, GlobalAttentionSumPool, GlobalAvgPool, GlobalMaxPool, GlobalSumPool, MessagePasserNNM, MessagePassing, SparseMPNN, UpdateFuncGRU, UpdateFuncMLP, ) from deephyper.keras.layers._padding import Padding __all__ = [ "AttentionConst", "AttentionCOS", "AttentionGAT", "AttentionGCN", "AttentionGenLinear", "AttentionLinear", "AttentionSymGAT", "GlobalAttentionPool", "GlobalAttentionSumPool", "GlobalAvgPool", "GlobalMaxPool", "GlobalSumPool", "MessagePasserNNM", "MessagePassing", "SparseMPNN", "UpdateFuncGRU", "UpdateFuncMLP", ] # When loading models with: "model.load('file.h5', custom_objects=custom_objects)" custom_objects = {"Padding": Padding}
960
20.840909
82
py
deephyper
deephyper-master/deephyper/keras/layers/_padding.py
import tensorflow as tf class Padding(tf.keras.layers.Layer): """Multi-dimensions padding layer. This operation pads a tensor according to the paddings you specify. paddings is an integer tensor with shape [n-1, 2], where n is the rank of tensor. For each dimension D of input, paddings[D, 0] indicates how many values to add before the contents of tensor in that dimension, and paddings[D, 1] indicates how many values to add after the contents of tensor in that dimension. The first dimension corresponding to the batch size cannot be padded. Args: padding (list(list(int))): e.g. [[1, 1]] mode (str): 'CONSTANT', 'REFLECT' or 'SYMMETRIC' """ def __init__(self, padding, mode="CONSTANT", constant_values=0, **kwargs): super(Padding, self).__init__(**kwargs) self.padding = [[0, 0]] + padding self.mode = mode self.constant_values = constant_values def call(self, x, mask=None): padding = tf.constant(self.padding) return tf.pad( tensor=x, paddings=padding, mode=self.mode, constant_values=self.constant_values, ) def compute_output_shape(self, input_shape): return tf.TensorShape( [ input_shape[i] + sum(self.padding[i]) if not input_shape[i] is None else None for i in range(len(input_shape)) ] ) def get_config(self): config = { "padding": self.padding[1:], "mode": self.mode, "constant_values": self.constant_values, } base_config = super(Padding, self).get_config() return dict(list(base_config.items()) + list(config.items()))
1,788
32.12963
89
py
deephyper
deephyper-master/deephyper/sklearn/classifier/_autosklearn1.py
""" This module provides ``problem_autosklearn1`` and ``run_autosklearn`` for classification tasks. """ import warnings from inspect import signature import ConfigSpace as cs from deephyper.problem import HpProblem from sklearn.ensemble import AdaBoostClassifier, RandomForestClassifier from sklearn.linear_model import LogisticRegression from sklearn.metrics import accuracy_score from sklearn.model_selection import train_test_split from sklearn.neighbors import KNeighborsClassifier from sklearn.neural_network import MLPClassifier from sklearn.pipeline import Pipeline from sklearn.preprocessing import MinMaxScaler, StandardScaler from sklearn.svm import SVC from xgboost import XGBClassifier def minmaxstdscaler() -> Pipeline: """MinMax preprocesssing followed by Standard normalization. Returns: Pipeline: a pipeline with two steps ``[MinMaxScaler, StandardScaler]``. """ preprocessor = Pipeline( [ ("minmaxscaler", MinMaxScaler()), ("stdscaler", StandardScaler()), ] ) return preprocessor problem_autosklearn1 = HpProblem() classifier = problem_autosklearn1.add_hyperparameter( name="classifier", value=[ "RandomForest", "Logistic", "AdaBoost", "KNeighbors", "MLP", "SVC", "XGBoost", ], ) # n_estimators n_estimators = problem_autosklearn1.add_hyperparameter( name="n_estimators", value=(1, 2000, "log-uniform") ) cond_n_estimators = cs.OrConjunction( cs.EqualsCondition(n_estimators, classifier, "RandomForest"), cs.EqualsCondition(n_estimators, classifier, "AdaBoost"), ) problem_autosklearn1.add_condition(cond_n_estimators) # max_depth max_depth = problem_autosklearn1.add_hyperparameter( name="max_depth", value=(2, 100, "log-uniform") ) cond_max_depth = cs.EqualsCondition(max_depth, classifier, "RandomForest") problem_autosklearn1.add_condition(cond_max_depth) # n_neighbors n_neighbors = problem_autosklearn1.add_hyperparameter( name="n_neighbors", value=(1, 100) ) cond_n_neighbors = cs.EqualsCondition(n_neighbors, classifier, "KNeighbors") problem_autosklearn1.add_condition(cond_n_neighbors) # alpha alpha = problem_autosklearn1.add_hyperparameter( name="alpha", value=(1e-5, 10.0, "log-uniform") ) cond_alpha = cs.EqualsCondition(alpha, classifier, "MLP") problem_autosklearn1.add_condition(cond_alpha) # C C = problem_autosklearn1.add_hyperparameter(name="C", value=(1e-5, 10.0, "log-uniform")) cond_C = cs.OrConjunction( cs.EqualsCondition(C, classifier, "Logistic"), cs.EqualsCondition(C, classifier, "SVC"), ) problem_autosklearn1.add_condition(cond_C) # kernel kernel = problem_autosklearn1.add_hyperparameter( name="kernel", value=["linear", "poly", "rbf", "sigmoid"] ) cond_kernel = cs.EqualsCondition(kernel, classifier, "SVC") problem_autosklearn1.add_condition(cond_kernel) # gamma gamma = problem_autosklearn1.add_hyperparameter( name="gamma", value=(1e-5, 10.0, "log-uniform") ) cond_gamma = cs.OrConjunction( cs.EqualsCondition(gamma, kernel, "rbf"), cs.EqualsCondition(gamma, kernel, "poly"), cs.EqualsCondition(gamma, kernel, "sigmoid"), ) problem_autosklearn1.add_condition(cond_gamma) # Mapping available classifiers CLASSIFIERS = { "RandomForest": RandomForestClassifier, "Logistic": LogisticRegression, "AdaBoost": AdaBoostClassifier, "KNeighbors": KNeighborsClassifier, "MLP": MLPClassifier, "SVC": SVC, "XGBoost": XGBClassifier, } def run_autosklearn1(config: dict, load_data: callable) -> float: """Run function which can be used for AutoML classification. It has to be used with the ``deephyper.sklearn.classifier.problem_autosklearn1`` problem definition which corresponds to: .. code-block:: Configuration space object: Hyperparameters: C, Type: UniformFloat, Range: [1e-05, 10.0], Default: 0.01, on log-scale alpha, Type: UniformFloat, Range: [1e-05, 10.0], Default: 0.01, on log-scale classifier, Type: Categorical, Choices: {RandomForest, Logistic, AdaBoost, KNeighbors, MLP, SVC, XGBoost}, Default: RandomForest gamma, Type: UniformFloat, Range: [1e-05, 10.0], Default: 0.01, on log-scale kernel, Type: Categorical, Choices: {linear, poly, rbf, sigmoid}, Default: linear max_depth, Type: UniformInteger, Range: [2, 100], Default: 14, on log-scale n_estimators, Type: UniformInteger, Range: [1, 2000], Default: 45, on log-scale n_neighbors, Type: UniformInteger, Range: [1, 100], Default: 50 Conditions: (C | classifier == 'Logistic' || C | classifier == 'SVC') (gamma | kernel == 'rbf' || gamma | kernel == 'poly' || gamma | kernel == 'sigmoid') (n_estimators | classifier == 'RandomForest' || n_estimators | classifier == 'AdaBoost') alpha | classifier == 'MLP' kernel | classifier == 'SVC' max_depth | classifier == 'RandomForest' n_neighbors | classifier == 'KNeighbors' Args: config (dict): an hyperparameter configuration ``dict`` corresponding to the ``deephyper.sklearn.classifier.problem_autosklearn1``. load_data (callable): a function returning data as Numpy arrays ``(X, y)``. Returns: float: returns the accuracy on the validation set. """ config["random_state"] = config.get("random_state", 42) config["n_jobs"] = config.get("n_jobs", 1) X, y = load_data() X_train, X_test, y_train, y_test = train_test_split( X, y, test_size=0.33, random_state=config["random_state"] ) preproc = minmaxstdscaler() X_train = preproc.fit_transform(X_train) X_test = preproc.transform(X_test) mapping = CLASSIFIERS clf_class = mapping[config["classifier"]] # keep parameters possible for the current classifier sig = signature(clf_class) clf_allowed_params = list(sig.parameters.keys()) clf_params = { k: v for k, v in config.items() if k in clf_allowed_params and not (v in ["nan", "NA"]) } try: # good practice to manage the fail value yourself... clf = clf_class(**clf_params) with warnings.catch_warnings(): warnings.simplefilter("ignore") clf.fit(X_train, y_train) fit_is_complete = True except: # noqa: E722 fit_is_complete = False if fit_is_complete: y_pred = clf.predict(X_test) acc = accuracy_score(y_test, y_pred) else: acc = -1.0 return acc if __name__ == "__main__": print(problem_autosklearn1)
6,739
30.495327
144
py
deephyper
deephyper-master/deephyper/sklearn/regressor/_autosklearn1.py
""" This module provides ``problem_autosklearn1`` and ``run_autosklearn`` for regression tasks. """ import warnings from inspect import signature import ConfigSpace as cs from deephyper.problem import HpProblem from sklearn.ensemble import AdaBoostRegressor, RandomForestRegressor from sklearn.linear_model import LinearRegression from sklearn.metrics import r2_score from sklearn.model_selection import train_test_split from sklearn.neighbors import KNeighborsRegressor from sklearn.neural_network import MLPRegressor from sklearn.pipeline import Pipeline from sklearn.preprocessing import MinMaxScaler, StandardScaler from sklearn.svm import SVR from xgboost import XGBRegressor def minmaxstdscaler() -> Pipeline: """MinMax preprocesssing followed by Standard normalization. Returns: Pipeline: a pipeline with two steps ``[MinMaxScaler, StandardScaler]``. """ preprocessor = Pipeline( [ ("minmaxscaler", MinMaxScaler()), ("stdscaler", StandardScaler()), ] ) return preprocessor REGRESSORS = { "RandomForest": RandomForestRegressor, "Linear": LinearRegression, "AdaBoost": AdaBoostRegressor, "KNeighbors": KNeighborsRegressor, "MLP": MLPRegressor, "SVR": SVR, "XGBoost": XGBRegressor, } problem_autosklearn1 = HpProblem() regressor = problem_autosklearn1.add_hyperparameter( name="regressor", value=["RandomForest", "Linear", "AdaBoost", "KNeighbors", "MLP", "SVR", "XGBoost"], ) # n_estimators n_estimators = problem_autosklearn1.add_hyperparameter( name="n_estimators", value=(1, 2000, "log-uniform") ) cond_n_estimators = cs.OrConjunction( cs.EqualsCondition(n_estimators, regressor, "RandomForest"), cs.EqualsCondition(n_estimators, regressor, "AdaBoost"), ) problem_autosklearn1.add_condition(cond_n_estimators) # max_depth max_depth = problem_autosklearn1.add_hyperparameter( name="max_depth", value=(2, 100, "log-uniform") ) cond_max_depth = cs.EqualsCondition(max_depth, regressor, "RandomForest") problem_autosklearn1.add_condition(cond_max_depth) # n_neighbors n_neighbors = problem_autosklearn1.add_hyperparameter( name="n_neighbors", value=(1, 100) ) cond_n_neighbors = cs.EqualsCondition(n_neighbors, regressor, "KNeighbors") problem_autosklearn1.add_condition(cond_n_neighbors) # alpha alpha = problem_autosklearn1.add_hyperparameter( name="alpha", value=(1e-5, 10.0, "log-uniform") ) cond_alpha = cs.EqualsCondition(alpha, regressor, "MLP") problem_autosklearn1.add_condition(cond_alpha) # C C = problem_autosklearn1.add_hyperparameter(name="C", value=(1e-5, 10.0, "log-uniform")) cond_C = cs.EqualsCondition(C, regressor, "SVR") problem_autosklearn1.add_condition(cond_C) # kernel kernel = problem_autosklearn1.add_hyperparameter( name="kernel", value=["linear", "poly", "rbf", "sigmoid"] ) cond_kernel = cs.EqualsCondition(kernel, regressor, "SVR") problem_autosklearn1.add_condition(cond_kernel) # gamma gamma = problem_autosklearn1.add_hyperparameter( name="gamma", value=(1e-5, 10.0, "log-uniform") ) cond_gamma = cs.OrConjunction( cs.EqualsCondition(gamma, kernel, "rbf"), cs.EqualsCondition(gamma, kernel, "poly"), cs.EqualsCondition(gamma, kernel, "sigmoid"), ) problem_autosklearn1.add_condition(cond_gamma) def run_autosklearn1(config: dict, load_data: callable) -> float: """Run function which can be used for AutoML regression. It has to be used with the ``deephyper.sklearn.regressor.problem_autosklearn1`` problem definition which corresponds to: .. code-block:: Configuration space object: Hyperparameters: C, Type: UniformFloat, Range: [1e-05, 10.0], Default: 0.01, on log-scale alpha, Type: UniformFloat, Range: [1e-05, 10.0], Default: 0.01, on log-scale gamma, Type: UniformFloat, Range: [1e-05, 10.0], Default: 0.01, on log-scale kernel, Type: Categorical, Choices: {linear, poly, rbf, sigmoid}, Default: linear max_depth, Type: UniformInteger, Range: [2, 100], Default: 14, on log-scale n_estimators, Type: UniformInteger, Range: [1, 2000], Default: 45, on log-scale n_neighbors, Type: UniformInteger, Range: [1, 100], Default: 50 regressor, Type: Categorical, Choices: {RandomForest, Linear, AdaBoost, KNeighbors, MLP, SVR, XGBoost}, Default: RandomForest Conditions: (gamma | kernel == 'rbf' || gamma | kernel == 'poly' || gamma | kernel == 'sigmoid') (n_estimators | regressor == 'RandomForest' || n_estimators | regressor == 'AdaBoost') C | regressor == 'SVR' alpha | regressor == 'MLP' kernel | regressor == 'SVR' max_depth | regressor == 'RandomForest' n_neighbors | regressor == 'KNeighbors' Args: config (dict): an hyperparameter configuration ``dict`` corresponding to the ``deephyper.sklearn.regressor.problem_autosklearn1``. load_data (callable): a function returning data as Numpy arrays ``(X, y)``. Returns: float: returns the :math:`R^2` on the validation set. """ config["random_state"] = config.get("random_state", 42) config["n_jobs"] = config.get("n_jobs", 1) X, y = load_data() X_train, X_test, y_train, y_test = train_test_split( X, y, test_size=0.33, random_state=config["random_state"] ) preproc = minmaxstdscaler() X_train = preproc.fit_transform(X_train) X_test = preproc.transform(X_test) mapping = REGRESSORS clf_class = mapping[config["regressor"]] # keep parameters possible for the current regressor sig = signature(clf_class) clf_allowed_params = list(sig.parameters.keys()) clf_params = { k: v for k, v in config.items() if k in clf_allowed_params and not (v in ["nan", "NA"]) } try: # good practice to manage the fail value yourself... clf = clf_class(**clf_params) with warnings.catch_warnings(): warnings.simplefilter("ignore") clf.fit(X_train, y_train) fit_is_complete = True except: # noqa: E722 fit_is_complete = False if fit_is_complete: y_pred = clf.predict(X_test) r2 = r2_score(y_test, y_pred) else: r2 = -1.0 return r2 if __name__ == "__main__": print(problem_autosklearn1)
6,472
30.8867
141
py
deephyper
deephyper-master/deephyper/ensemble/_bagging_ensemble.py
import os import traceback import tensorflow as tf import numpy as np import ray from deephyper.nas.metrics import selectMetric from deephyper.ensemble import BaseEnsemble from deephyper.nas.run._util import set_memory_growth_for_visible_gpus def mse(y_true, y_pred): return tf.square(y_true - y_pred) @ray.remote(num_cpus=1) def model_predict(model_path, X, batch_size=32, verbose=0): """Perform an inference of the model located at ``model_path``. :meta private: Args: model_path (str): Path to the ``h5`` file to load to perform the inferencec. X (array): array of input data for which we perform the inference. batch_size (int, optional): Batch size used to perform the inferencec. Defaults to 32. verbose (int, optional): Verbose option. Defaults to 0. Returns: array: The prediction based on the provided input data. """ # GPU Configuration if available set_memory_growth_for_visible_gpus(True) tf.keras.backend.clear_session() model_file = model_path.split("/")[-1] try: if verbose: print(f"Loading model {model_file}", flush=True) model = tf.keras.models.load_model(model_path, compile=False) except Exception: if verbose: print(f"Could not load model {model_file}", flush=True) traceback.print_exc() model = None if model: y = model.predict(X, batch_size=batch_size) else: y = None return y class BaggingEnsemble(BaseEnsemble): """Ensemble based on uniform averaging of the predictions of each members. :meta private: Args: model_dir (str): Path to directory containing saved Keras models in .h5 format. loss (callable): a callable taking (y_true, y_pred) as input. size (int, optional): Number of unique models used in the ensemble. Defaults to 5. verbose (bool, optional): Verbose mode. Defaults to True. ray_address (str, optional): Address of the Ray cluster. If "auto" it will try to connect to an existing cluster. If "" it will start a local Ray cluster. Defaults to "". num_cpus (int, optional): Number of CPUs allocated to load one model and predict. Defaults to 1. num_gpus (int, optional): Number of GPUs allocated to load one model and predict. Defaults to None. batch_size (int, optional): Batch size used batchify the inference of loaded models. Defaults to 32. selection (str, optional): Selection strategy to build the ensemble. Value in ``["topk"]``. Default to ``topk``. mode (str, optional): Value in ``["regression", "classification"]``. Default to ``"regression"``. """ def __init__( self, model_dir, loss=mse, size=5, verbose=True, ray_address="", num_cpus=1, num_gpus=None, batch_size=32, selection="topk", mode="regression", ): super().__init__( model_dir, loss, size, verbose, ray_address, num_cpus, num_gpus, batch_size, ) assert selection in ["topk"] self.selection = selection assert mode in ["regression", "classification"] self.mode = mode def __repr__(self) -> str: out = super().__repr__() out += f"Mode: {self.mode}\n" out += f"Selection: {self.selection}\n" return out def fit(self, X, y): """Fit the current algorithm to the provided data. Args: X (array): The input data. y (array): The output data. Returns: BaseEnsemble: The current fitted instance. """ X_id = ray.put(X) model_files = self._list_files_in_model_dir() def model_path(f): return os.path.join(self.model_dir, f) y_pred = ray.get( [ model_predict.options( num_cpus=self.num_cpus, num_gpus=self.num_gpus ).remote(model_path(f), X_id, self.batch_size, self.verbose) for f in model_files ] ) y_pred = np.array([arr for arr in y_pred if arr is not None]) members_indexes = topk(self.loss, y_true=y, y_pred=y_pred, k=self.size) self.members_files = [model_files[i] for i in members_indexes] return self def predict(self, X) -> np.ndarray: """Execute an inference of the ensemble for the provided data. Args: X (array): An array of input data. Returns: array: The prediction. """ # make predictions X_id = ray.put(X) def model_path(f): os.path.join(self.model_dir, f) y_pred = ray.get( [ model_predict.options( num_cpus=self.num_cpus, num_gpus=self.num_gpus ).remote(model_path(f), X_id, self.batch_size, self.verbose) for f in self.members_files ] ) y_pred = np.array([arr for arr in y_pred if arr is not None]) y = aggregate_predictions(y_pred, regression=(self.mode == "regression")) return y def evaluate(self, X, y, metrics=None): """Compute metrics based on the provided data. Args: X (array): An array of input data. y (array): An array of true output data. metrics (callable, optional): A metric. Defaults to None. """ scores = {} y_pred = self.predict(X) scores["loss"] = tf.reduce_mean(self.loss(y, y_pred)).numpy() if metrics: for metric_name in metrics: scores[metric_name] = apply_metric(metric_name, y, y_pred) return scores class BaggingEnsembleRegressor(BaggingEnsemble): """Ensemble for regression based on uniform averaging of the predictions of each members. Args: model_dir (str): Path to directory containing saved Keras models in .h5 format. loss (callable): a callable taking (y_true, y_pred) as input. size (int, optional): Number of unique models used in the ensemble. Defaults to 5. verbose (bool, optional): Verbose mode. Defaults to True. ray_address (str, optional): Address of the Ray cluster. If "auto" it will try to connect to an existing cluster. If "" it will start a local Ray cluster. Defaults to "". num_cpus (int, optional): Number of CPUs allocated to load one model and predict. Defaults to 1. num_gpus (int, optional): Number of GPUs allocated to load one model and predict. Defaults to None. batch_size (int, optional): Batch size used batchify the inference of loaded models. Defaults to 32. selection (str, optional): Selection strategy to build the ensemble. Value in ``["topk"]``. Default to ``topk``. """ def __init__( self, model_dir, loss=mse, size=5, verbose=True, ray_address="", num_cpus=1, num_gpus=None, selection="topk", ): super().__init__( model_dir, loss, size, verbose, ray_address, num_cpus, num_gpus, selection, mode="regression", ) class BaggingEnsembleClassifier(BaggingEnsemble): """Ensemble for classification based on uniform averaging of the predictions of each members. Args: model_dir (str): Path to directory containing saved Keras models in .h5 format. loss (callable): a callable taking (y_true, y_pred) as input. size (int, optional): Number of unique models used in the ensemble. Defaults to 5. verbose (bool, optional): Verbose mode. Defaults to True. ray_address (str, optional): Address of the Ray cluster. If "auto" it will try to connect to an existing cluster. If "" it will start a local Ray cluster. Defaults to "". num_cpus (int, optional): Number of CPUs allocated to load one model and predict. Defaults to 1. num_gpus (int, optional): Number of GPUs allocated to load one model and predict. Defaults to None. batch_size (int, optional): Batch size used batchify the inference of loaded models. Defaults to 32. selection (str, optional): Selection strategy to build the ensemble. Value in ``["topk"]``. Default to ``topk``. """ def __init__( self, model_dir, loss=mse, size=5, verbose=True, ray_address="", num_cpus=1, num_gpus=None, selection="topk", ): super().__init__( model_dir, loss, size, verbose, ray_address, num_cpus, num_gpus, selection, mode="classification", ) def apply_metric(metric_name, y_true, y_pred) -> float: """Perform the computation of provided metric. :meta private: Args: metric_name (str|callable): If ``str`` then it needs to be a metric available in ``deephyper.nas.metrics``. y_true (array): Array of true predictions. y_pred (array): Array of predicted predictions Returns: float: a scalar value of the computed metric. """ metric_func = selectMetric(metric_name) metric = tf.reduce_mean( metric_func( tf.convert_to_tensor(y_true, dtype=np.float32), tf.convert_to_tensor(y_pred, dtype=np.float32), ) ).numpy() return metric def aggregate_predictions(y_pred, regression=True): """Build an ensemble from predictions. :meta private: Args: ensemble_members (np.array): Indexes of selected members in the axis-0 of y_pred. y_pred (np.array): Predictions array of shape (n_models, n_samples, n_outputs). regression (bool): Boolean (True) if it is a regression (False) if it is a classification. Return: A TFP Normal Distribution in the case of regression and a np.array with average probabilities in the case of classification. """ n = np.shape(y_pred)[0] y_pred = np.sum(y_pred, axis=0) if regression: agg_y_pred = y_pred / n else: # classification agg_y_pred = np.argmax(y_pred, axis=1) return agg_y_pred def topk(loss_func, y_true, y_pred, k=2) -> list: """Select the Top-k models to be part of the ensemble. A model can appear only once in the ensemble for this strategy. :meta private: Args: loss_func (callable): loss function. y_true (array): Array of true predictions. y_pred (array): Array of predicted predictions k (int, optional): Number of models composing the ensemble. Defaults to 2. Returns: list: a list of model indexes composing the ensembles. """ # losses is of shape: (n_models, n_outputs) losses = tf.reduce_mean(loss_func(y_true, y_pred), axis=1).numpy() ensemble_members = np.argsort(losses, axis=0)[:k].reshape(-1).tolist() return ensemble_members
11,171
32.752266
178
py
deephyper
deephyper-master/deephyper/ensemble/__init__.py
"""The ``ensemble`` module provides a way to build ensembles of checkpointed deep neural networks from ``tensorflow.keras``, with ``.h5`` format, to regularize and boost predictive performance as well as estimate better uncertainties. """ from deephyper.ensemble._base_ensemble import BaseEnsemble from deephyper.ensemble._bagging_ensemble import ( BaggingEnsembleRegressor, BaggingEnsembleClassifier, ) from deephyper.ensemble._uq_bagging_ensemble import ( UQBaggingEnsembleRegressor, UQBaggingEnsembleClassifier, ) __all__ = [ "BaseEnsemble", "BaggingEnsembleRegressor", "BaggingEnsembleClassifier", "UQBaggingEnsembleRegressor", "UQBaggingEnsembleClassifier", ]
702
34.15
234
py
deephyper
deephyper-master/deephyper/ensemble/_uq_bagging_ensemble.py
import os import traceback import numpy as np import ray import tensorflow as tf import tensorflow_probability as tfp from deephyper.ensemble import BaseEnsemble from deephyper.nas.metrics import selectMetric from deephyper.nas.run._util import set_memory_growth_for_visible_gpus from deephyper.core.exceptions import DeephyperRuntimeError from pandas import DataFrame def nll(y, rv_y): """Negative log likelihood loss for Tensorflow probability.""" return -rv_y.log_prob(y) cce_obj = tf.keras.losses.CategoricalCrossentropy( reduction=tf.keras.losses.Reduction.NONE ) def cce(y_true, y_pred): """Categorical cross-entropy loss.""" return cce_obj(tf.broadcast_to(y_true, y_pred.shape), y_pred) @ray.remote(num_cpus=1) def model_predict(model_path, X, batch_size=32, verbose=0): """Perform an inference of the model located at ``model_path``. :meta private: Args: model_path (str): Path to the ``h5`` file to load to perform the inferencec. X (array): array of input data for which we perform the inference. batch_size (int, optional): Batch size used to perform the inferencec. Defaults to 32. verbose (int, optional): Verbose option. Defaults to 0. Returns: array: The prediction based on the provided input data. """ import tensorflow as tf import tensorflow_probability as tfp # GPU Configuration if available set_memory_growth_for_visible_gpus(True) tf.keras.backend.clear_session() model_file = model_path.split("/")[-1] try: if verbose: print(f"Loading model {model_file}", end="\n", flush=True) model = tf.keras.models.load_model(model_path, compile=False) except Exception: if verbose: print(f"Could not load model {model_file}", flush=True) traceback.print_exc() model = None if model is None: return None # dataset if type(X) is list: dataset = tf.data.Dataset.from_tensor_slices( {f"input_{i}": Xi for i, Xi in enumerate(X)} ) else: dataset = tf.data.Dataset.from_tensor_slices(X) dataset = dataset.batch(batch_size) def batch_predict(dataset, convert_func=lambda x: x): y_list = [] for batch in dataset: y = model(batch, training=False) y_list.append(convert_func(y)) y = np.concatenate(y_list, axis=0) return y y_dist = model( next(iter(dataset)), training=False ) # just to test the type of the output if isinstance(y_dist, tfp.distributions.Distribution): if hasattr(y_dist, "loc") and hasattr(y_dist, "scale"): def convert_func(y_dist): return np.concatenate([y_dist.loc, y_dist.scale], axis=-1) y = batch_predict(dataset, convert_func) else: raise DeephyperRuntimeError( "Distribution doesn't have 'loc' or 'scale' attributes!" ) else: y = model.predict(X, batch_size=batch_size) return y class UQBaggingEnsemble(BaseEnsemble): """Ensemble with uncertainty quantification based on uniform averaging of the predictions of each members. :meta private: Args: model_dir (str): Path to directory containing saved Keras models in .h5 format. loss (callable): a callable taking (y_true, y_pred) as input. size (int, optional): Number of unique models used in the ensemble. Defaults to 5. verbose (bool, optional): Verbose mode. Defaults to True. ray_address (str, optional): Address of the Ray cluster. If "auto" it will try to connect to an existing cluster. If "" it will start a local Ray cluster. Defaults to "". num_cpus (int, optional): Number of CPUs allocated to load one model and predict. Defaults to 1. num_gpus (int, optional): Number of GPUs allocated to load one model and predict. Defaults to None. batch_size (int, optional): Batch size used batchify the inference of loaded models. Defaults to 32. selection (str, optional): Selection strategy to build the ensemble. Value in ``["topk", "caruana"]``. Default to ``topk``. mode (str, optional): Value in ``["regression", "classification"]``. Default to ``"regression"``. """ def __init__( self, model_dir, loss=nll, size=5, verbose=True, ray_address="", num_cpus=1, num_gpus=None, batch_size=32, selection="topk", mode="regression", ): super().__init__( model_dir, loss, size, verbose, ray_address, num_cpus, num_gpus, batch_size, ) assert selection in ["topk", "caruana"] self.selection = selection assert mode in ["regression", "classification"] self.mode = mode def __repr__(self) -> str: out = super().__repr__() out += f"Mode: {self.mode}\n" out += f"Selection: {self.selection}\n" return out def _select_members(self, loss_func, y_true, y_pred, k=2, verbose=0): if self.selection == "topk": func = topk elif self.selection == "caruana": func = greedy_caruana else: raise NotImplementedError return func(loss_func, y_true, y_pred, k, verbose) def fit(self, X, y): X_id = ray.put(X) model_files = self._list_files_in_model_dir() def model_path(f): return os.path.join(self.model_dir, f) y_pred = ray.get( [ model_predict.options( num_cpus=self.num_cpus, num_gpus=self.num_gpus ).remote(model_path(f), X_id, self.batch_size, self.verbose) for f in model_files ] ) y_pred = np.array([arr for arr in y_pred if arr is not None]) self._members_indexes = self._select_members( self.loss, y_true=y, y_pred=y_pred, k=self.size ) self.members_files = [model_files[i] for i in self._members_indexes] def predict(self, X) -> np.ndarray: # make predictions X_id = ray.put(X) def model_path(f): return os.path.join(self.model_dir, f) y_pred = ray.get( [ model_predict.options( num_cpus=self.num_cpus, num_gpus=self.num_gpus ).remote(model_path(f), X_id, self.batch_size, self.verbose) for f in self.members_files ] ) y_pred = np.array([arr for arr in y_pred if arr is not None]) y = aggregate_predictions(y_pred, regression=(self.mode == "regression")) return y def evaluate(self, X, y, metrics=None, scaler_y=None): scores = {} y_pred = self.predict(X) if scaler_y: y_pred = scaler_y(y_pred) y = scaler_y(y) scores["loss"] = tf.reduce_mean(self.loss(y, y_pred)).numpy() if metrics: if type(metrics) is list: for metric in metrics: if callable(metric): metric_name = metric.__name__ else: metric_name = metric scores[metric_name] = apply_metric(metric, y, y_pred) elif type(metrics) is dict: for metric_name, metric in metrics.items(): scores[metric_name] = apply_metric(metric, y, y_pred) else: raise ValueError("Metrics should be of type list or dict.") return scores class UQBaggingEnsembleRegressor(UQBaggingEnsemble): """Ensemble with uncertainty quantification for regression based on uniform averaging of the predictions of each members. Args: model_dir (str): Path to directory containing saved Keras models in .h5 format. loss (callable): a callable taking (y_true, y_pred) as input. size (int, optional): Number of unique models used in the ensemble. Defaults to 5. verbose (bool, optional): Verbose mode. Defaults to True. ray_address (str, optional): Address of the Ray cluster. If "auto" it will try to connect to an existing cluster. If "" it will start a local Ray cluster. Defaults to "". num_cpus (int, optional): Number of CPUs allocated to load one model and predict. Defaults to 1. num_gpus (int, optional): Number of GPUs allocated to load one model and predict. Defaults to None. batch_size (int, optional): Batch size used batchify the inference of loaded models. Defaults to 32. selection (str, optional): Selection strategy to build the ensemble. Value in ``[["topk", "caruana"]``. Default to ``topk``. """ def __init__( self, model_dir, loss=nll, size=5, verbose=True, ray_address="", num_cpus=1, num_gpus=None, batch_size=32, selection="topk", ): super().__init__( model_dir, loss, size, verbose, ray_address, num_cpus, num_gpus, batch_size, selection, mode="regression", ) def predict_var_decomposition(self, X): """Execute an inference of the ensemble for the provided data with uncertainty quantification estimates. The **aleatoric uncertainty** corresponds to the expected value of learned variance of each model composing the ensemble :math:`\mathbf{E}[\sigma_\\theta^2(\mathbf{x})]`. The **epistemic uncertainty** corresponds to the variance of learned mean estimates of each model composing the ensemble :math:`\mathbf{V}[\mu_\\theta(\mathbf{x})]`. Args: X (array): An array of input data. Returns: y, u1, u2: where ``y`` is the mixture distribution, ``u1`` is the aleatoric component of the variance of ``y`` and ``u2`` is the epistemic component of the variance of ``y``. """ # make predictions X_id = ray.put(X) def model_path(f): return os.path.join(self.model_dir, f) y_pred = ray.get( [ model_predict.options( num_cpus=self.num_cpus, num_gpus=self.num_gpus ).remote(model_path(f), X_id, self.batch_size, self.verbose) for f in self.members_files ] ) y_pred = np.array([arr for arr in y_pred if arr is not None]) y = aggregate_predictions(y_pred, regression=(self.mode == "regression")) # variance decomposition mid = np.shape(y_pred)[-1] // 2 selection = [slice(0, s) for s in np.shape(y_pred)] selection_loc = selection[:] selection_std = selection[:] selection_loc[-1] = slice(0, mid) selection_std[-1] = slice(mid, np.shape(y_pred)[-1]) loc = y_pred[tuple(selection_loc)] scale = y_pred[tuple(selection_std)] aleatoric_unc = np.mean(np.square(scale), axis=0) epistemic_unc = np.square(np.std(loc, axis=0)) # dist, aleatoric uq, epistemic uq return y, aleatoric_unc, epistemic_unc class UQBaggingEnsembleClassifier(UQBaggingEnsemble): """Ensemble with uncertainty quantification for classification based on uniform averaging of the predictions of each members. Args: model_dir (str): Path to directory containing saved Keras models in .h5 format. loss (callable): a callable taking (y_true, y_pred) as input. size (int, optional): Number of unique models used in the ensemble. Defaults to 5. verbose (bool, optional): Verbose mode. Defaults to True. ray_address (str, optional): Address of the Ray cluster. If "auto" it will try to connect to an existing cluster. If "" it will start a local Ray cluster. Defaults to "". num_cpus (int, optional): Number of CPUs allocated to load one model and predict. Defaults to 1. num_gpus (int, optional): Number of GPUs allocated to load one model and predict. Defaults to None. batch_size (int, optional): Batch size used batchify the inference of loaded models. Defaults to 32. selection (str, optional): Selection strategy to build the ensemble. Value in ``[["topk", "caruana"]``. Default to ``topk``. """ def __init__( self, model_dir, loss=cce, size=5, verbose=True, ray_address="", num_cpus=1, num_gpus=None, batch_size=32, selection="topk", ): super().__init__( model_dir, loss, size, verbose, ray_address, num_cpus, num_gpus, batch_size, selection, mode="classification", ) def apply_metric(metric_name, y_true, y_pred) -> float: """Perform the computation of provided metric. :meta private: Args: metric_name (str|callable): If ``str`` then it needs to be a metric available in ``deephyper.nas.metrics``. y_true (array): Array of true predictions. y_pred (array): Array of predicted predictions Returns: float: a scalar value of the computed metric. """ metric_func = selectMetric(metric_name) if type(y_true) is np.ndarray: y_true = tf.convert_to_tensor(y_true, dtype=np.float32) if type(y_pred) is np.ndarray: y_pred = tf.convert_to_tensor(y_pred, dtype=np.float32) metric = metric_func(y_true, y_pred) if tf.size(metric) >= 1: metric = tf.reduce_mean(metric) return metric.numpy() def aggregate_predictions(y_pred, regression=True): """Build an ensemble from predictions. :meta private: Args: ensemble_members (np.array): Indexes of selected members in the axis-0 of y_pred. y_pred (np.array): Predictions array of shape (n_models, n_samples, n_outputs). regression (bool): Boolean (True) if it is a regression (False) if it is a classification. Return: A TFP Normal Distribution in the case of regression and a np.array with average probabilities in the case of classification. """ if regression: # assuming first half are means, second half are std mid = np.shape(y_pred)[-1] // 2 selection = [slice(0, s) for s in np.shape(y_pred)] selection_loc = selection[:] selection_std = selection[:] selection_loc[-1] = slice(0, mid) selection_std[-1] = slice(mid, np.shape(y_pred)[-1]) loc = y_pred[tuple(selection_loc)] scale = y_pred[tuple(selection_std)] mean_loc = np.mean(loc, axis=0) sum_loc_scale = np.square(loc) + np.square(scale) mean_scale = np.sqrt(np.mean(sum_loc_scale, axis=0) - np.square(mean_loc)) return tfp.distributions.Normal(loc=mean_loc, scale=mean_scale) else: # classification agg_y_pred = np.mean(y_pred[:, :, :], axis=0) return agg_y_pred def topk(loss_func, y_true, y_pred, k=2, verbose=0): """Select the top-k models to be part of the ensemble. A model can appear only once in the ensemble for this strategy. :meta private: """ if np.shape(y_true)[-1] * 2 == np.shape(y_pred)[-1]: # regression mid = np.shape(y_true)[-1] y_pred = tfp.distributions.Normal( loc=y_pred[:, :, :mid], scale=y_pred[:, :, mid:] ) # losses is of shape: (n_models, n_outputs) losses = tf.reduce_mean(loss_func(y_true, y_pred), axis=1).numpy() if verbose: print(f"Top-{k} losses: {losses.reshape(-1)[:k]}") ensemble_members = np.argsort(losses, axis=0)[:k].reshape(-1).tolist() return ensemble_members def greedy_caruana(loss_func, y_true, y_pred, k=2, verbose=0): """Select the top-k models to be part of the ensemble. A model can appear only once in the ensemble for this strategy. :meta private: """ regression = np.shape(y_true)[-1] * 2 == np.shape(y_pred)[-1] n_models = np.shape(y_pred)[0] if regression: # regression mid = np.shape(y_true)[-1] selection = [slice(0, s) for s in np.shape(y_pred)] selection_loc = selection[:] selection_std = selection[:] selection_loc[-1] = slice(0, mid) selection_std[-1] = slice(mid, np.shape(y_pred)[-1]) y_pred_ = tfp.distributions.Normal( loc=y_pred[tuple(selection_loc)], scale=y_pred[tuple(selection_std)], ) else: y_pred_ = y_pred losses = tf.reduce_mean( tf.reshape(loss_func(y_true, y_pred_), [n_models, -1]), axis=1 ).numpy() assert n_models == np.shape(losses)[0] i_min = np.nanargmin(losses) loss_min = losses[i_min] ensemble_members = [i_min] if verbose: print(f"Loss: {loss_min:.3f} - Ensemble: {ensemble_members}") def loss(y_true, y_pred): return tf.reduce_mean(loss_func(y_true, y_pred)).numpy() while len(np.unique(ensemble_members)) < k: losses = [ loss( y_true, aggregate_predictions( y_pred[ensemble_members + [i]], regression=regression ), ) for i in range(n_models) # iterate over all models ] i_min_ = np.nanargmin(losses) loss_min_ = losses[i_min_] if loss_min_ < loss_min: if ( len(np.unique(ensemble_members)) == 1 and ensemble_members[0] == i_min_ ): # numerical errors... return ensemble_members loss_min = loss_min_ ensemble_members.append(i_min_) if verbose: print(f"Loss: {loss_min:.3f} - Ensemble: {ensemble_members}") else: return ensemble_members return ensemble_members def __convert_to_block_df(a, y_col=None, group_col=None, block_col=None, melted=False): # TODO: refactor conversion of block data to DataFrame if melted and not all([i is not None for i in [block_col, group_col, y_col]]): raise ValueError( "`block_col`, `group_col`, `y_col` should be explicitly specified if using melted data" ) if isinstance(a, DataFrame) and not melted: x = a.copy(deep=True) group_col = "groups" block_col = "blocks" y_col = "y" x.columns.name = group_col x.index.name = block_col x = x.reset_index().melt( id_vars=block_col, var_name=group_col, value_name=y_col ) elif isinstance(a, DataFrame) and melted: x = DataFrame.from_dict( {"groups": a[group_col], "blocks": a[block_col], "y": a[y_col]} ) elif not isinstance(a, DataFrame): x = np.array(a) x = DataFrame(x, index=np.arange(x.shape[0]), columns=np.arange(x.shape[1])) if not melted: group_col = "groups" block_col = "blocks" y_col = "y" x.columns.name = group_col x.index.name = block_col x = x.reset_index().melt( id_vars=block_col, var_name=group_col, value_name=y_col ) else: x.rename( columns={group_col: "groups", block_col: "blocks", y_col: "y"}, inplace=True, ) group_col = "groups" block_col = "blocks" y_col = "y" return x, y_col, group_col, block_col
19,600
34.703097
449
py
Pyramid-Attention-Networks
Pyramid-Attention-Networks-master/Demosaic/code/main.py
import torch import utility import data import model import loss from option import args from trainer import Trainer torch.manual_seed(args.seed) checkpoint = utility.checkpoint(args) def main(): global model if args.data_test == ['video']: from videotester import VideoTester model = model.Model(args,checkpoint) print('total params: %.2fM' % (sum(p.numel() for p in model.parameters())/1000000.0)) t = VideoTester(args, model, checkpoint) t.test() else: if checkpoint.ok: loader = data.Data(args) _model = model.Model(args, checkpoint) print('total params:%.2fM' % (sum(p.numel() for p in _model.parameters())/1000000.0)) _loss = loss.Loss(args, checkpoint) if not args.test_only else None t = Trainer(args, loader, _model, _loss, checkpoint) while not t.terminate(): t.train() t.test() checkpoint.done() if __name__ == '__main__': main()
1,026
27.527778
97
py
Pyramid-Attention-Networks
Pyramid-Attention-Networks-master/Demosaic/code/utility.py
import os import math import time import datetime from multiprocessing import Process from multiprocessing import Queue import matplotlib matplotlib.use('Agg') import matplotlib.pyplot as plt import numpy as np import imageio import torch import torch.optim as optim import torch.optim.lr_scheduler as lrs class timer(): def __init__(self): self.acc = 0 self.tic() def tic(self): self.t0 = time.time() def toc(self, restart=False): diff = time.time() - self.t0 if restart: self.t0 = time.time() return diff def hold(self): self.acc += self.toc() def release(self): ret = self.acc self.acc = 0 return ret def reset(self): self.acc = 0 class checkpoint(): def __init__(self, args): self.args = args self.ok = True self.log = torch.Tensor() now = datetime.datetime.now().strftime('%Y-%m-%d-%H:%M:%S') if not args.load: if not args.save: args.save = now self.dir = os.path.join('..', 'experiment', args.save) else: self.dir = os.path.join('..', 'experiment', args.load) if os.path.exists(self.dir): self.log = torch.load(self.get_path('psnr_log.pt')) print('Continue from epoch {}...'.format(len(self.log))) else: args.load = '' if args.reset: os.system('rm -rf ' + self.dir) args.load = '' os.makedirs(self.dir, exist_ok=True) os.makedirs(self.get_path('model'), exist_ok=True) for d in args.data_test: os.makedirs(self.get_path('results-{}'.format(d)), exist_ok=True) open_type = 'a' if os.path.exists(self.get_path('log.txt'))else 'w' self.log_file = open(self.get_path('log.txt'), open_type) with open(self.get_path('config.txt'), open_type) as f: f.write(now + '\n\n') for arg in vars(args): f.write('{}: {}\n'.format(arg, getattr(args, arg))) f.write('\n') self.n_processes = 8 def get_path(self, *subdir): return os.path.join(self.dir, *subdir) def save(self, trainer, epoch, is_best=False): trainer.model.save(self.get_path('model'), epoch, is_best=is_best) trainer.loss.save(self.dir) trainer.loss.plot_loss(self.dir, epoch) self.plot_psnr(epoch) trainer.optimizer.save(self.dir) torch.save(self.log, self.get_path('psnr_log.pt')) def add_log(self, log): self.log = torch.cat([self.log, log]) def write_log(self, log, refresh=False): print(log) self.log_file.write(log + '\n') if refresh: self.log_file.close() self.log_file = open(self.get_path('log.txt'), 'a') def done(self): self.log_file.close() def plot_psnr(self, epoch): axis = np.linspace(1, epoch, epoch) for idx_data, d in enumerate(self.args.data_test): label = 'SR on {}'.format(d) fig = plt.figure() plt.title(label) for idx_scale, scale in enumerate(self.args.scale): plt.plot( axis, self.log[:, idx_data, idx_scale].numpy(), label='Scale {}'.format(scale) ) plt.legend() plt.xlabel('Epochs') plt.ylabel('PSNR') plt.grid(True) plt.savefig(self.get_path('test_{}.pdf'.format(d))) plt.close(fig) def begin_background(self): self.queue = Queue() def bg_target(queue): while True: if not queue.empty(): filename, tensor = queue.get() if filename is None: break imageio.imwrite(filename, tensor.numpy()) self.process = [ Process(target=bg_target, args=(self.queue,)) \ for _ in range(self.n_processes) ] for p in self.process: p.start() def end_background(self): for _ in range(self.n_processes): self.queue.put((None, None)) while not self.queue.empty(): time.sleep(1) for p in self.process: p.join() def save_results(self, dataset, filename, save_list, scale): if self.args.save_results: filename = self.get_path( 'results-{}'.format(dataset.dataset.name), '{}_x{}_'.format(filename, scale) ) postfix = ('DM', 'LQ', 'HQ') for v, p in zip(save_list, postfix): normalized = v[0].mul(255 / self.args.rgb_range) tensor_cpu = normalized.byte().permute(1, 2, 0).cpu() self.queue.put(('{}{}.png'.format(filename, p), tensor_cpu)) def quantize(img, rgb_range): pixel_range = 255 / rgb_range return img.mul(pixel_range).clamp(0, 255).round().div(pixel_range) def calc_psnr(sr, hr, scale, rgb_range, dataset=None): if hr.nelement() == 1: return 0 diff = (sr - hr) / rgb_range if dataset and dataset.dataset.benchmark: shave = scale if diff.size(1) > 5: gray_coeffs = [65.738, 129.057, 25.064] convert = diff.new_tensor(gray_coeffs).view(1, 3, 1, 1) / 256 diff = diff.mul(convert).sum(dim=1) else: shave = scale + 6 valid = diff[..., :, :] mse = valid.pow(2).mean() return -10 * math.log10(mse) def make_optimizer(args, target): ''' make optimizer and scheduler together ''' # optimizer trainable = filter(lambda x: x.requires_grad, target.parameters()) kwargs_optimizer = {'lr': args.lr, 'weight_decay': args.weight_decay} if args.optimizer == 'SGD': optimizer_class = optim.SGD kwargs_optimizer['momentum'] = args.momentum elif args.optimizer == 'ADAM': optimizer_class = optim.Adam kwargs_optimizer['betas'] = args.betas kwargs_optimizer['eps'] = args.epsilon elif args.optimizer == 'RMSprop': optimizer_class = optim.RMSprop kwargs_optimizer['eps'] = args.epsilon # scheduler milestones = list(map(lambda x: int(x), args.decay.split('-'))) kwargs_scheduler = {'milestones': milestones, 'gamma': args.gamma} scheduler_class = lrs.MultiStepLR class CustomOptimizer(optimizer_class): def __init__(self, *args, **kwargs): super(CustomOptimizer, self).__init__(*args, **kwargs) def _register_scheduler(self, scheduler_class, **kwargs): self.scheduler = scheduler_class(self, **kwargs) def save(self, save_dir): torch.save(self.state_dict(), self.get_dir(save_dir)) def load(self, load_dir, epoch=1): self.load_state_dict(torch.load(self.get_dir(load_dir))) if epoch > 1: for _ in range(epoch): self.scheduler.step() def get_dir(self, dir_path): return os.path.join(dir_path, 'optimizer.pt') def schedule(self): self.scheduler.step() def get_lr(self): return self.scheduler.get_lr()[0] def get_last_epoch(self): return self.scheduler.last_epoch optimizer = CustomOptimizer(trainable, **kwargs_optimizer) optimizer._register_scheduler(scheduler_class, **kwargs_scheduler) return optimizer
7,458
30.340336
77
py
Pyramid-Attention-Networks
Pyramid-Attention-Networks-master/Demosaic/code/dataloader.py
import threading import random import torch import torch.multiprocessing as multiprocessing from torch.utils.data import DataLoader from torch.utils.data import SequentialSampler from torch.utils.data import RandomSampler from torch.utils.data import BatchSampler from torch.utils.data import _utils from torch.utils.data.dataloader import _DataLoaderIter from torch.utils.data._utils import collate from torch.utils.data._utils import signal_handling from torch.utils.data._utils import MP_STATUS_CHECK_INTERVAL from torch.utils.data._utils import ExceptionWrapper from torch.utils.data._utils import IS_WINDOWS from torch.utils.data._utils.worker import ManagerWatchdog from torch._six import queue def _ms_loop(dataset, index_queue, data_queue, done_event, collate_fn, scale, seed, init_fn, worker_id): try: collate._use_shared_memory = True signal_handling._set_worker_signal_handlers() torch.set_num_threads(1) random.seed(seed) torch.manual_seed(seed) data_queue.cancel_join_thread() if init_fn is not None: init_fn(worker_id) watchdog = ManagerWatchdog() while watchdog.is_alive(): try: r = index_queue.get(timeout=MP_STATUS_CHECK_INTERVAL) except queue.Empty: continue if r is None: assert done_event.is_set() return elif done_event.is_set(): continue idx, batch_indices = r try: idx_scale = 0 if len(scale) > 1 and dataset.train: idx_scale = random.randrange(0, len(scale)) dataset.set_scale(idx_scale) samples = collate_fn([dataset[i] for i in batch_indices]) samples.append(idx_scale) except Exception: data_queue.put((idx, ExceptionWrapper(sys.exc_info()))) else: data_queue.put((idx, samples)) del samples except KeyboardInterrupt: pass class _MSDataLoaderIter(_DataLoaderIter): def __init__(self, loader): self.dataset = loader.dataset self.scale = loader.scale self.collate_fn = loader.collate_fn self.batch_sampler = loader.batch_sampler self.num_workers = loader.num_workers self.pin_memory = loader.pin_memory and torch.cuda.is_available() self.timeout = loader.timeout self.sample_iter = iter(self.batch_sampler) base_seed = torch.LongTensor(1).random_().item() if self.num_workers > 0: self.worker_init_fn = loader.worker_init_fn self.worker_queue_idx = 0 self.worker_result_queue = multiprocessing.Queue() self.batches_outstanding = 0 self.worker_pids_set = False self.shutdown = False self.send_idx = 0 self.rcvd_idx = 0 self.reorder_dict = {} self.done_event = multiprocessing.Event() base_seed = torch.LongTensor(1).random_()[0] self.index_queues = [] self.workers = [] for i in range(self.num_workers): index_queue = multiprocessing.Queue() index_queue.cancel_join_thread() w = multiprocessing.Process( target=_ms_loop, args=( self.dataset, index_queue, self.worker_result_queue, self.done_event, self.collate_fn, self.scale, base_seed + i, self.worker_init_fn, i ) ) w.daemon = True w.start() self.index_queues.append(index_queue) self.workers.append(w) if self.pin_memory: self.data_queue = queue.Queue() pin_memory_thread = threading.Thread( target=_utils.pin_memory._pin_memory_loop, args=( self.worker_result_queue, self.data_queue, torch.cuda.current_device(), self.done_event ) ) pin_memory_thread.daemon = True pin_memory_thread.start() self.pin_memory_thread = pin_memory_thread else: self.data_queue = self.worker_result_queue _utils.signal_handling._set_worker_pids( id(self), tuple(w.pid for w in self.workers) ) _utils.signal_handling._set_SIGCHLD_handler() self.worker_pids_set = True for _ in range(2 * self.num_workers): self._put_indices() class MSDataLoader(DataLoader): def __init__(self, cfg, *args, **kwargs): super(MSDataLoader, self).__init__( *args, **kwargs, num_workers=cfg.n_threads ) self.scale = cfg.scale def __iter__(self): return _MSDataLoaderIter(self)
5,259
32.081761
104
py
Pyramid-Attention-Networks
Pyramid-Attention-Networks-master/Demosaic/code/videotester.py
import os import math import utility from data import common import torch import cv2 from tqdm import tqdm class VideoTester(): def __init__(self, args, my_model, ckp): self.args = args self.scale = args.scale self.ckp = ckp self.model = my_model self.filename, _ = os.path.splitext(os.path.basename(args.dir_demo)) def test(self): torch.set_grad_enabled(False) self.ckp.write_log('\nEvaluation on video:') self.model.eval() timer_test = utility.timer() for idx_scale, scale in enumerate(self.scale): vidcap = cv2.VideoCapture(self.args.dir_demo) total_frames = int(vidcap.get(cv2.CAP_PROP_FRAME_COUNT)) vidwri = cv2.VideoWriter( self.ckp.get_path('{}_x{}.avi'.format(self.filename, scale)), cv2.VideoWriter_fourcc(*'XVID'), vidcap.get(cv2.CAP_PROP_FPS), ( int(scale * vidcap.get(cv2.CAP_PROP_FRAME_WIDTH)), int(scale * vidcap.get(cv2.CAP_PROP_FRAME_HEIGHT)) ) ) tqdm_test = tqdm(range(total_frames), ncols=80) for _ in tqdm_test: success, lr = vidcap.read() if not success: break lr, = common.set_channel(lr, n_channels=self.args.n_colors) lr, = common.np2Tensor(lr, rgb_range=self.args.rgb_range) lr, = self.prepare(lr.unsqueeze(0)) sr = self.model(lr, idx_scale) sr = utility.quantize(sr, self.args.rgb_range).squeeze(0) normalized = sr * 255 / self.args.rgb_range ndarr = normalized.byte().permute(1, 2, 0).cpu().numpy() vidwri.write(ndarr) vidcap.release() vidwri.release() self.ckp.write_log( 'Total: {:.2f}s\n'.format(timer_test.toc()), refresh=True ) torch.set_grad_enabled(True) def prepare(self, *args): device = torch.device('cpu' if self.args.cpu else 'cuda') def _prepare(tensor): if self.args.precision == 'half': tensor = tensor.half() return tensor.to(device) return [_prepare(a) for a in args]
2,280
30.246575
77
py
Pyramid-Attention-Networks
Pyramid-Attention-Networks-master/Demosaic/code/trainer.py
import os import math from decimal import Decimal import utility import torch import torch.nn.utils as utils from tqdm import tqdm class Trainer(): def __init__(self, args, loader, my_model, my_loss, ckp): self.args = args self.scale = args.scale self.ckp = ckp self.loader_train = loader.loader_train self.loader_test = loader.loader_test self.model = my_model self.loss = my_loss self.optimizer = utility.make_optimizer(args, self.model) if self.args.load != '': self.optimizer.load(ckp.dir, epoch=len(ckp.log)) self.error_last = 1e8 def train(self): self.loss.step() epoch = self.optimizer.get_last_epoch() + 1 lr = self.optimizer.get_lr() self.ckp.write_log( '[Epoch {}]\tLearning rate: {:.2e}'.format(epoch, Decimal(lr)) ) self.loss.start_log() self.model.train() timer_data, timer_model = utility.timer(), utility.timer() # TEMP self.loader_train.dataset.set_scale(0) for batch, (lr, hr, _,) in enumerate(self.loader_train): lr, hr = self.prepare(lr, hr) timer_data.hold() timer_model.tic() self.optimizer.zero_grad() sr = self.model(lr, 0) loss = self.loss(sr, hr) loss.backward() if self.args.gclip > 0: utils.clip_grad_value_( self.model.parameters(), self.args.gclip ) self.optimizer.step() timer_model.hold() if (batch + 1) % self.args.print_every == 0: self.ckp.write_log('[{}/{}]\t{}\t{:.1f}+{:.1f}s'.format( (batch + 1) * self.args.batch_size, len(self.loader_train.dataset), self.loss.display_loss(batch), timer_model.release(), timer_data.release())) timer_data.tic() self.loss.end_log(len(self.loader_train)) self.error_last = self.loss.log[-1, -1] self.optimizer.schedule() def test(self): torch.set_grad_enabled(False) epoch = self.optimizer.get_last_epoch() self.ckp.write_log('\nEvaluation:') self.ckp.add_log( torch.zeros(1, len(self.loader_test), len(self.scale)) ) self.model.eval() timer_test = utility.timer() if self.args.save_results: self.ckp.begin_background() for idx_data, d in enumerate(self.loader_test): for idx_scale, scale in enumerate(self.scale): d.dataset.set_scale(idx_scale) for lr, hr, filename in tqdm(d, ncols=80): lr, hr = self.prepare(lr, hr) sr = self.model(lr, idx_scale) sr = utility.quantize(sr, self.args.rgb_range) save_list = [sr] self.ckp.log[-1, idx_data, idx_scale] += utility.calc_psnr( sr, hr, scale, self.args.rgb_range, dataset=d ) if self.args.save_gt: save_list.extend([lr, hr]) if self.args.save_results: self.ckp.save_results(d, filename[0], save_list, scale) self.ckp.log[-1, idx_data, idx_scale] /= len(d) best = self.ckp.log.max(0) self.ckp.write_log( '[{} x{}]\tPSNR: {:.3f} (Best: {:.3f} @epoch {})'.format( d.dataset.name, scale, self.ckp.log[-1, idx_data, idx_scale], best[0][idx_data, idx_scale], best[1][idx_data, idx_scale] + 1 ) ) self.ckp.write_log('Forward: {:.2f}s\n'.format(timer_test.toc())) self.ckp.write_log('Saving...') if self.args.save_results: self.ckp.end_background() if not self.args.test_only: self.ckp.save(self, epoch, is_best=(best[1][0, 0] + 1 == epoch)) self.ckp.write_log( 'Total: {:.2f}s\n'.format(timer_test.toc()), refresh=True ) torch.set_grad_enabled(True) def prepare(self, *args): device = torch.device('cpu' if self.args.cpu else 'cuda') def _prepare(tensor): if self.args.precision == 'half': tensor = tensor.half() return tensor.to(device) return [_prepare(a) for a in args] def terminate(self): if self.args.test_only: self.test() return True else: epoch = self.optimizer.get_last_epoch() + 1 return epoch >= self.args.epochs
4,820
31.795918
79
py
Pyramid-Attention-Networks
Pyramid-Attention-Networks-master/Demosaic/code/loss/adversarial.py
import utility from types import SimpleNamespace from model import common from loss import discriminator import torch import torch.nn as nn import torch.nn.functional as F import torch.optim as optim class Adversarial(nn.Module): def __init__(self, args, gan_type): super(Adversarial, self).__init__() self.gan_type = gan_type self.gan_k = args.gan_k self.dis = discriminator.Discriminator(args) if gan_type == 'WGAN_GP': # see https://arxiv.org/pdf/1704.00028.pdf pp.4 optim_dict = { 'optimizer': 'ADAM', 'betas': (0, 0.9), 'epsilon': 1e-8, 'lr': 1e-5, 'weight_decay': args.weight_decay, 'decay': args.decay, 'gamma': args.gamma } optim_args = SimpleNamespace(**optim_dict) else: optim_args = args self.optimizer = utility.make_optimizer(optim_args, self.dis) def forward(self, fake, real): # updating discriminator... self.loss = 0 fake_detach = fake.detach() # do not backpropagate through G for _ in range(self.gan_k): self.optimizer.zero_grad() # d: B x 1 tensor d_fake = self.dis(fake_detach) d_real = self.dis(real) retain_graph = False if self.gan_type == 'GAN': loss_d = self.bce(d_real, d_fake) elif self.gan_type.find('WGAN') >= 0: loss_d = (d_fake - d_real).mean() if self.gan_type.find('GP') >= 0: epsilon = torch.rand_like(fake).view(-1, 1, 1, 1) hat = fake_detach.mul(1 - epsilon) + real.mul(epsilon) hat.requires_grad = True d_hat = self.dis(hat) gradients = torch.autograd.grad( outputs=d_hat.sum(), inputs=hat, retain_graph=True, create_graph=True, only_inputs=True )[0] gradients = gradients.view(gradients.size(0), -1) gradient_norm = gradients.norm(2, dim=1) gradient_penalty = 10 * gradient_norm.sub(1).pow(2).mean() loss_d += gradient_penalty # from ESRGAN: Enhanced Super-Resolution Generative Adversarial Networks elif self.gan_type == 'RGAN': better_real = d_real - d_fake.mean(dim=0, keepdim=True) better_fake = d_fake - d_real.mean(dim=0, keepdim=True) loss_d = self.bce(better_real, better_fake) retain_graph = True # Discriminator update self.loss += loss_d.item() loss_d.backward(retain_graph=retain_graph) self.optimizer.step() if self.gan_type == 'WGAN': for p in self.dis.parameters(): p.data.clamp_(-1, 1) self.loss /= self.gan_k # updating generator... d_fake_bp = self.dis(fake) # for backpropagation, use fake as it is if self.gan_type == 'GAN': label_real = torch.ones_like(d_fake_bp) loss_g = F.binary_cross_entropy_with_logits(d_fake_bp, label_real) elif self.gan_type.find('WGAN') >= 0: loss_g = -d_fake_bp.mean() elif self.gan_type == 'RGAN': better_real = d_real - d_fake_bp.mean(dim=0, keepdim=True) better_fake = d_fake_bp - d_real.mean(dim=0, keepdim=True) loss_g = self.bce(better_fake, better_real) # Generator loss return loss_g def state_dict(self, *args, **kwargs): state_discriminator = self.dis.state_dict(*args, **kwargs) state_optimizer = self.optimizer.state_dict() return dict(**state_discriminator, **state_optimizer) def bce(self, real, fake): label_real = torch.ones_like(real) label_fake = torch.zeros_like(fake) bce_real = F.binary_cross_entropy_with_logits(real, label_real) bce_fake = F.binary_cross_entropy_with_logits(fake, label_fake) bce_loss = bce_real + bce_fake return bce_loss # Some references # https://github.com/kuc2477/pytorch-wgan-gp/blob/master/model.py # OR # https://github.com/caogang/wgan-gp/blob/master/gan_cifar10.py
4,393
37.884956
84
py
Pyramid-Attention-Networks
Pyramid-Attention-Networks-master/Demosaic/code/loss/discriminator.py
from model import common import torch.nn as nn class Discriminator(nn.Module): ''' output is not normalized ''' def __init__(self, args): super(Discriminator, self).__init__() in_channels = args.n_colors out_channels = 64 depth = 7 def _block(_in_channels, _out_channels, stride=1): return nn.Sequential( nn.Conv2d( _in_channels, _out_channels, 3, padding=1, stride=stride, bias=False ), nn.BatchNorm2d(_out_channels), nn.LeakyReLU(negative_slope=0.2, inplace=True) ) m_features = [_block(in_channels, out_channels)] for i in range(depth): in_channels = out_channels if i % 2 == 1: stride = 1 out_channels *= 2 else: stride = 2 m_features.append(_block(in_channels, out_channels, stride=stride)) patch_size = args.patch_size // (2**((depth + 1) // 2)) m_classifier = [ nn.Linear(out_channels * patch_size**2, 1024), nn.LeakyReLU(negative_slope=0.2, inplace=True), nn.Linear(1024, 1) ] self.features = nn.Sequential(*m_features) self.classifier = nn.Sequential(*m_classifier) def forward(self, x): features = self.features(x) output = self.classifier(features.view(features.size(0), -1)) return output
1,595
27.5
79
py
Pyramid-Attention-Networks
Pyramid-Attention-Networks-master/Demosaic/code/loss/vgg.py
from model import common import torch import torch.nn as nn import torch.nn.functional as F import torchvision.models as models class VGG(nn.Module): def __init__(self, conv_index, rgb_range=1): super(VGG, self).__init__() vgg_features = models.vgg19(pretrained=True).features modules = [m for m in vgg_features] if conv_index.find('22') >= 0: self.vgg = nn.Sequential(*modules[:8]) elif conv_index.find('54') >= 0: self.vgg = nn.Sequential(*modules[:35]) vgg_mean = (0.485, 0.456, 0.406) vgg_std = (0.229 * rgb_range, 0.224 * rgb_range, 0.225 * rgb_range) self.sub_mean = common.MeanShift(rgb_range, vgg_mean, vgg_std) for p in self.parameters(): p.requires_grad = False def forward(self, sr, hr): def _forward(x): x = self.sub_mean(x) x = self.vgg(x) return x vgg_sr = _forward(sr) with torch.no_grad(): vgg_hr = _forward(hr.detach()) loss = F.mse_loss(vgg_sr, vgg_hr) return loss
1,106
28.918919
75
py
Pyramid-Attention-Networks
Pyramid-Attention-Networks-master/Demosaic/code/loss/__init__.py
import os from importlib import import_module import matplotlib matplotlib.use('Agg') import matplotlib.pyplot as plt import numpy as np import torch import torch.nn as nn import torch.nn.functional as F class Loss(nn.modules.loss._Loss): def __init__(self, args, ckp): super(Loss, self).__init__() print('Preparing loss function:') self.n_GPUs = args.n_GPUs self.loss = [] self.loss_module = nn.ModuleList() for loss in args.loss.split('+'): weight, loss_type = loss.split('*') if loss_type == 'MSE': loss_function = nn.MSELoss() elif loss_type == 'L1': loss_function = nn.L1Loss() elif loss_type.find('VGG') >= 0: module = import_module('loss.vgg') loss_function = getattr(module, 'VGG')( loss_type[3:], rgb_range=args.rgb_range ) elif loss_type.find('GAN') >= 0: module = import_module('loss.adversarial') loss_function = getattr(module, 'Adversarial')( args, loss_type ) self.loss.append({ 'type': loss_type, 'weight': float(weight), 'function': loss_function} ) if loss_type.find('GAN') >= 0: self.loss.append({'type': 'DIS', 'weight': 1, 'function': None}) if len(self.loss) > 1: self.loss.append({'type': 'Total', 'weight': 0, 'function': None}) for l in self.loss: if l['function'] is not None: print('{:.3f} * {}'.format(l['weight'], l['type'])) self.loss_module.append(l['function']) self.log = torch.Tensor() device = torch.device('cpu' if args.cpu else 'cuda') self.loss_module.to(device) if args.precision == 'half': self.loss_module.half() if not args.cpu and args.n_GPUs > 1: self.loss_module = nn.DataParallel( self.loss_module, range(args.n_GPUs) ) if args.load != '': self.load(ckp.dir, cpu=args.cpu) def forward(self, sr, hr): losses = [] for i, l in enumerate(self.loss): if l['function'] is not None: loss = l['function'](sr, hr) effective_loss = l['weight'] * loss losses.append(effective_loss) self.log[-1, i] += effective_loss.item() elif l['type'] == 'DIS': self.log[-1, i] += self.loss[i - 1]['function'].loss loss_sum = sum(losses) if len(self.loss) > 1: self.log[-1, -1] += loss_sum.item() return loss_sum def step(self): for l in self.get_loss_module(): if hasattr(l, 'scheduler'): l.scheduler.step() def start_log(self): self.log = torch.cat((self.log, torch.zeros(1, len(self.loss)))) def end_log(self, n_batches): self.log[-1].div_(n_batches) def display_loss(self, batch): n_samples = batch + 1 log = [] for l, c in zip(self.loss, self.log[-1]): log.append('[{}: {:.4f}]'.format(l['type'], c / n_samples)) return ''.join(log) def plot_loss(self, apath, epoch): axis = np.linspace(1, epoch, epoch) for i, l in enumerate(self.loss): label = '{} Loss'.format(l['type']) fig = plt.figure() plt.title(label) plt.plot(axis, self.log[:, i].numpy(), label=label) plt.legend() plt.xlabel('Epochs') plt.ylabel('Loss') plt.grid(True) plt.savefig(os.path.join(apath, 'loss_{}.pdf'.format(l['type']))) plt.close(fig) def get_loss_module(self): if self.n_GPUs == 1: return self.loss_module else: return self.loss_module.module def save(self, apath): torch.save(self.state_dict(), os.path.join(apath, 'loss.pt')) torch.save(self.log, os.path.join(apath, 'loss_log.pt')) def load(self, apath, cpu=False): if cpu: kwargs = {'map_location': lambda storage, loc: storage} else: kwargs = {} self.load_state_dict(torch.load( os.path.join(apath, 'loss.pt'), **kwargs )) self.log = torch.load(os.path.join(apath, 'loss_log.pt')) for l in self.get_loss_module(): if hasattr(l, 'scheduler'): for _ in range(len(self.log)): l.scheduler.step()
4,659
31.361111
80
py
Pyramid-Attention-Networks
Pyramid-Attention-Networks-master/Demosaic/code/utils/tools.py
import os import torch import numpy as np from PIL import Image import torch.nn.functional as F def normalize(x): return x.mul_(2).add_(-1) def same_padding(images, ksizes, strides, rates): assert len(images.size()) == 4 batch_size, channel, rows, cols = images.size() out_rows = (rows + strides[0] - 1) // strides[0] out_cols = (cols + strides[1] - 1) // strides[1] effective_k_row = (ksizes[0] - 1) * rates[0] + 1 effective_k_col = (ksizes[1] - 1) * rates[1] + 1 padding_rows = max(0, (out_rows-1)*strides[0]+effective_k_row-rows) padding_cols = max(0, (out_cols-1)*strides[1]+effective_k_col-cols) # Pad the input padding_top = int(padding_rows / 2.) padding_left = int(padding_cols / 2.) padding_bottom = padding_rows - padding_top padding_right = padding_cols - padding_left paddings = (padding_left, padding_right, padding_top, padding_bottom) images = torch.nn.ZeroPad2d(paddings)(images) return images def extract_image_patches(images, ksizes, strides, rates, padding='same'): """ Extract patches from images and put them in the C output dimension. :param padding: :param images: [batch, channels, in_rows, in_cols]. A 4-D Tensor with shape :param ksizes: [ksize_rows, ksize_cols]. The size of the sliding window for each dimension of images :param strides: [stride_rows, stride_cols] :param rates: [dilation_rows, dilation_cols] :return: A Tensor """ assert len(images.size()) == 4 assert padding in ['same', 'valid'] batch_size, channel, height, width = images.size() if padding == 'same': images = same_padding(images, ksizes, strides, rates) elif padding == 'valid': pass else: raise NotImplementedError('Unsupported padding type: {}.\ Only "same" or "valid" are supported.'.format(padding)) unfold = torch.nn.Unfold(kernel_size=ksizes, dilation=rates, padding=0, stride=strides) patches = unfold(images) return patches # [N, C*k*k, L], L is the total number of such blocks def reduce_mean(x, axis=None, keepdim=False): if not axis: axis = range(len(x.shape)) for i in sorted(axis, reverse=True): x = torch.mean(x, dim=i, keepdim=keepdim) return x def reduce_std(x, axis=None, keepdim=False): if not axis: axis = range(len(x.shape)) for i in sorted(axis, reverse=True): x = torch.std(x, dim=i, keepdim=keepdim) return x def reduce_sum(x, axis=None, keepdim=False): if not axis: axis = range(len(x.shape)) for i in sorted(axis, reverse=True): x = torch.sum(x, dim=i, keepdim=keepdim) return x
2,777
32.878049
79
py
Pyramid-Attention-Networks
Pyramid-Attention-Networks-master/Demosaic/code/data/benchmark.py
import os from data import common from data import srdata import numpy as np import torch import torch.utils.data as data class Benchmark(srdata.SRData): def __init__(self, args, name='', train=True, benchmark=True): super(Benchmark, self).__init__( args, name=name, train=train, benchmark=True ) def _set_filesystem(self, dir_data): self.apath = os.path.join(dir_data, 'benchmark', self.name) self.dir_hr = os.path.join(self.apath, 'HR') if self.input_large: self.dir_lr = os.path.join(self.apath, 'LR_bicubicL') else: self.dir_lr = os.path.join(self.apath, 'LR_bicubic') self.ext = ('', '.png')
703
26.076923
67
py
Pyramid-Attention-Networks
Pyramid-Attention-Networks-master/Demosaic/code/data/video.py
import os from data import common import cv2 import numpy as np import imageio import torch import torch.utils.data as data class Video(data.Dataset): def __init__(self, args, name='Video', train=False, benchmark=False): self.args = args self.name = name self.scale = args.scale self.idx_scale = 0 self.train = False self.do_eval = False self.benchmark = benchmark self.filename, _ = os.path.splitext(os.path.basename(args.dir_demo)) self.vidcap = cv2.VideoCapture(args.dir_demo) self.n_frames = 0 self.total_frames = int(self.vidcap.get(cv2.CAP_PROP_FRAME_COUNT)) def __getitem__(self, idx): success, lr = self.vidcap.read() if success: self.n_frames += 1 lr, = common.set_channel(lr, n_channels=self.args.n_colors) lr_t, = common.np2Tensor(lr, rgb_range=self.args.rgb_range) return lr_t, -1, '{}_{:0>5}'.format(self.filename, self.n_frames) else: vidcap.release() return None def __len__(self): return self.total_frames def set_scale(self, idx_scale): self.idx_scale = idx_scale
1,207
25.844444
77
py
Pyramid-Attention-Networks
Pyramid-Attention-Networks-master/Demosaic/code/data/srdata.py
import os import glob import random import pickle from data import common import numpy as np import imageio import torch import torch.utils.data as data class SRData(data.Dataset): def __init__(self, args, name='', train=True, benchmark=False): self.args = args self.name = name self.train = train self.split = 'train' if train else 'test' self.do_eval = True self.benchmark = benchmark self.input_large = (args.model == 'VDSR') self.scale = args.scale self.idx_scale = 0 self._set_filesystem(args.dir_data) if args.ext.find('img') < 0: path_bin = os.path.join(self.apath, 'bin') os.makedirs(path_bin, exist_ok=True) list_hr, list_lr = self._scan() if args.ext.find('img') >= 0 or benchmark: self.images_hr, self.images_lr = list_hr, list_lr elif args.ext.find('sep') >= 0: os.makedirs( self.dir_hr.replace(self.apath, path_bin), exist_ok=True ) for s in self.scale: os.makedirs( os.path.join( self.dir_lr.replace(self.apath, path_bin), 'X{}'.format(s) ), exist_ok=True ) self.images_hr, self.images_lr = [], [[] for _ in self.scale] for h in list_hr: b = h.replace(self.apath, path_bin) b = b.replace(self.ext[0], '.pt') self.images_hr.append(b) self._check_and_load(args.ext, h, b, verbose=True) for i, ll in enumerate(list_lr): for l in ll: b = l.replace(self.apath, path_bin) b = b.replace(self.ext[1], '.pt') self.images_lr[i].append(b) self._check_and_load(args.ext, l, b, verbose=True) if train: n_patches = args.batch_size * args.test_every n_images = len(args.data_train) * len(self.images_hr) if n_images == 0: self.repeat = 0 else: self.repeat = max(n_patches // n_images, 1) # Below functions as used to prepare images def _scan(self): names_hr = sorted( glob.glob(os.path.join(self.dir_hr, '*' + self.ext[0])) ) names_lr = [[] for _ in self.scale] for f in names_hr: filename, _ = os.path.splitext(os.path.basename(f)) for si, s in enumerate(self.scale): names_lr[si].append(os.path.join( self.dir_lr, 'X{}/{}{}'.format( s, filename, self.ext[1] ) )) return names_hr, names_lr def _set_filesystem(self, dir_data): self.apath = os.path.join(dir_data, self.name) self.dir_hr = os.path.join(self.apath, 'HR') self.dir_lr = os.path.join(self.apath, 'LR_bicubic') if self.input_large: self.dir_lr += 'L' self.ext = ('.png', '.png') def _check_and_load(self, ext, img, f, verbose=True): if not os.path.isfile(f) or ext.find('reset') >= 0: if verbose: print('Making a binary: {}'.format(f)) with open(f, 'wb') as _f: pickle.dump(imageio.imread(img), _f) def __getitem__(self, idx): lr, hr, filename = self._load_file(idx) pair = self.get_patch(lr, hr) pair = common.set_channel(*pair, n_channels=self.args.n_colors) pair_t = common.np2Tensor(*pair, rgb_range=self.args.rgb_range) return pair_t[0], pair_t[1], filename def __len__(self): if self.train: return len(self.images_hr) * self.repeat else: return len(self.images_hr) def _get_index(self, idx): if self.train: return idx % len(self.images_hr) else: return idx def _load_file(self, idx): idx = self._get_index(idx) f_hr = self.images_hr[idx] f_lr = self.images_lr[self.idx_scale][idx] filename, _ = os.path.splitext(os.path.basename(f_hr)) if self.args.ext == 'img' or self.benchmark: hr = imageio.imread(f_hr) lr = imageio.imread(f_lr) elif self.args.ext.find('sep') >= 0: with open(f_hr, 'rb') as _f: hr = pickle.load(_f) with open(f_lr, 'rb') as _f: lr = pickle.load(_f) return lr, hr, filename def get_patch(self, lr, hr): scale = self.scale[self.idx_scale] if self.train: lr, hr = common.get_patch( lr, hr, patch_size=self.args.patch_size, scale=scale, multi=(len(self.scale) > 1), input_large=self.input_large ) if not self.args.no_augment: lr, hr = common.augment(lr, hr) else: ih, iw = lr.shape[:2] hr = hr[0:ih * scale, 0:iw * scale] return lr, hr def set_scale(self, idx_scale): if not self.input_large: self.idx_scale = idx_scale else: self.idx_scale = random.randint(0, len(self.scale) - 1)
5,337
32.78481
73
py
Pyramid-Attention-Networks
Pyramid-Attention-Networks-master/Demosaic/code/data/demo.py
import os from data import common import numpy as np import imageio import torch import torch.utils.data as data class Demo(data.Dataset): def __init__(self, args, name='Demo', train=False, benchmark=False): self.args = args self.name = name self.scale = args.scale self.idx_scale = 0 self.train = False self.benchmark = benchmark self.filelist = [] for f in os.listdir(args.dir_demo): if f.find('.png') >= 0 or f.find('.jp') >= 0: self.filelist.append(os.path.join(args.dir_demo, f)) self.filelist.sort() def __getitem__(self, idx): filename = os.path.splitext(os.path.basename(self.filelist[idx]))[0] lr = imageio.imread(self.filelist[idx]) lr, = common.set_channel(lr, n_channels=self.args.n_colors) lr_t, = common.np2Tensor(lr, rgb_range=self.args.rgb_range) return lr_t, -1, filename def __len__(self): return len(self.filelist) def set_scale(self, idx_scale): self.idx_scale = idx_scale
1,075
25.9
76
py
Pyramid-Attention-Networks
Pyramid-Attention-Networks-master/Demosaic/code/data/common.py
import random import numpy as np import skimage.color as sc import torch def get_patch(*args, patch_size=96, scale=1, multi=False, input_large=False): ih, iw = args[0].shape[:2] if not input_large: p = 1 if multi else 1 tp = p * patch_size ip = tp // 1 else: tp = patch_size ip = patch_size ix = random.randrange(0, iw - ip + 1) iy = random.randrange(0, ih - ip + 1) if not input_large: tx, ty = 1 * ix, 1 * iy else: tx, ty = ix, iy ret = [ args[0][iy:iy + ip, ix:ix + ip, :], *[a[ty:ty + tp, tx:tx + tp, :] for a in args[1:]] ] return ret def set_channel(*args, n_channels=3): def _set_channel(img): if img.ndim == 2: img = np.expand_dims(img, axis=2) c = img.shape[2] if n_channels == 1 and c == 3: img = np.expand_dims(sc.rgb2ycbcr(img)[:, :, 0], 2) elif n_channels == 3 and c == 1: img = np.concatenate([img] * n_channels, 2) return img return [_set_channel(a) for a in args] def np2Tensor(*args, rgb_range=255): def _np2Tensor(img): np_transpose = np.ascontiguousarray(img.transpose((2, 0, 1))) tensor = torch.from_numpy(np_transpose).float() tensor.mul_(rgb_range / 255) return tensor return [_np2Tensor(a) for a in args] def augment(*args, hflip=True, rot=True): hflip = hflip and random.random() < 0.5 vflip = rot and random.random() < 0.5 rot90 = rot and random.random() < 0.5 def _augment(img): if hflip: img = img[:, ::-1, :] if vflip: img = img[::-1, :, :] if rot90: img = img.transpose(1, 0, 2) return img return [_augment(a) for a in args]
1,770
23.260274
77
py
Pyramid-Attention-Networks
Pyramid-Attention-Networks-master/Demosaic/code/data/__init__.py
from importlib import import_module #from dataloader import MSDataLoader from torch.utils.data import dataloader from torch.utils.data import ConcatDataset # This is a simple wrapper function for ConcatDataset class MyConcatDataset(ConcatDataset): def __init__(self, datasets): super(MyConcatDataset, self).__init__(datasets) self.train = datasets[0].train def set_scale(self, idx_scale): for d in self.datasets: if hasattr(d, 'set_scale'): d.set_scale(idx_scale) class Data: def __init__(self, args): self.loader_train = None if not args.test_only: datasets = [] for d in args.data_train: module_name = d if d.find('DIV2K-Q') < 0 else 'DIV2KJPEG' m = import_module('data.' + module_name.lower()) datasets.append(getattr(m, module_name)(args, name=d)) self.loader_train = dataloader.DataLoader( MyConcatDataset(datasets), batch_size=args.batch_size, shuffle=True, pin_memory=not args.cpu, num_workers=args.n_threads, ) self.loader_test = [] for d in args.data_test: if d in ['CBSD68','Kodak24','McM','Set5', 'Set14', 'B100', 'Urban100']: m = import_module('data.benchmark') testset = getattr(m, 'Benchmark')(args, train=False, name=d) else: module_name = d if d.find('DIV2K-Q') < 0 else 'DIV2KJPEG' m = import_module('data.' + module_name.lower()) testset = getattr(m, module_name)(args, train=False, name=d) self.loader_test.append( dataloader.DataLoader( testset, batch_size=1, shuffle=False, pin_memory=not args.cpu, num_workers=args.n_threads, ) )
1,974
36.264151
83
py
Pyramid-Attention-Networks
Pyramid-Attention-Networks-master/Demosaic/code/model/rcan.py
## ECCV-2018-Image Super-Resolution Using Very Deep Residual Channel Attention Networks ## https://arxiv.org/abs/1807.02758 from model import common import torch.nn as nn def make_model(args, parent=False): return RCAN(args) ## Channel Attention (CA) Layer class CALayer(nn.Module): def __init__(self, channel, reduction=16): super(CALayer, self).__init__() # global average pooling: feature --> point self.avg_pool = nn.AdaptiveAvgPool2d(1) # feature channel downscale and upscale --> channel weight self.conv_du = nn.Sequential( nn.Conv2d(channel, channel // reduction, 1, padding=0, bias=True), nn.ReLU(inplace=True), nn.Conv2d(channel // reduction, channel, 1, padding=0, bias=True), nn.Sigmoid() ) def forward(self, x): y = self.avg_pool(x) y = self.conv_du(y) return x * y ## Residual Channel Attention Block (RCAB) class RCAB(nn.Module): def __init__( self, conv, n_feat, kernel_size, reduction, bias=True, bn=False, act=nn.ReLU(True), res_scale=1): super(RCAB, self).__init__() modules_body = [] for i in range(2): modules_body.append(conv(n_feat, n_feat, kernel_size, bias=bias)) if bn: modules_body.append(nn.BatchNorm2d(n_feat)) if i == 0: modules_body.append(act) modules_body.append(CALayer(n_feat, reduction)) self.body = nn.Sequential(*modules_body) self.res_scale = res_scale def forward(self, x): res = self.body(x) #res = self.body(x).mul(self.res_scale) res += x return res ## Residual Group (RG) class ResidualGroup(nn.Module): def __init__(self, conv, n_feat, kernel_size, reduction, act, res_scale, n_resblocks): super(ResidualGroup, self).__init__() modules_body = [] modules_body = [ RCAB( conv, n_feat, kernel_size, reduction, bias=True, bn=False, act=nn.ReLU(True), res_scale=1) \ for _ in range(n_resblocks)] modules_body.append(conv(n_feat, n_feat, kernel_size)) self.body = nn.Sequential(*modules_body) def forward(self, x): res = self.body(x) res += x return res ## Residual Channel Attention Network (RCAN) class RCAN(nn.Module): def __init__(self, args, conv=common.default_conv): super(RCAN, self).__init__() n_resgroups = args.n_resgroups n_resblocks = args.n_resblocks n_feats = args.n_feats kernel_size = 3 reduction = args.reduction scale = args.scale[0] act = nn.ReLU(True) # RGB mean for DIV2K rgb_mean = (0.4488, 0.4371, 0.4040) rgb_std = (1.0, 1.0, 1.0) self.sub_mean = common.MeanShift(args.rgb_range, rgb_mean, rgb_std) # define head module modules_head = [conv(args.n_colors, n_feats, kernel_size)] # define body module modules_body = [ ResidualGroup( conv, n_feats, kernel_size, reduction, act=act, res_scale=args.res_scale, n_resblocks=n_resblocks) \ for _ in range(n_resgroups)] modules_body.append(conv(n_feats, n_feats, kernel_size)) # define tail module modules_tail = [ common.Upsampler(conv, scale, n_feats, act=False), conv(n_feats, args.n_colors, kernel_size)] self.add_mean = common.MeanShift(args.rgb_range, rgb_mean, rgb_std, 1) self.head = nn.Sequential(*modules_head) self.body = nn.Sequential(*modules_body) self.tail = nn.Sequential(*modules_tail) def forward(self, x): x = self.sub_mean(x) x = self.head(x) res = self.body(x) res += x x = self.tail(res) x = self.add_mean(x) return x def load_state_dict(self, state_dict, strict=False): own_state = self.state_dict() for name, param in state_dict.items(): if name in own_state: if isinstance(param, nn.Parameter): param = param.data try: own_state[name].copy_(param) except Exception: if name.find('tail') >= 0: print('Replace pre-trained upsampler to new one...') else: raise RuntimeError('While copying the parameter named {}, ' 'whose dimensions in the model are {} and ' 'whose dimensions in the checkpoint are {}.' .format(name, own_state[name].size(), param.size())) elif strict: if name.find('tail') == -1: raise KeyError('unexpected key "{}" in state_dict' .format(name)) if strict: missing = set(own_state.keys()) - set(state_dict.keys()) if len(missing) > 0: raise KeyError('missing keys in state_dict: "{}"'.format(missing))
5,178
34.717241
116
py
Pyramid-Attention-Networks
Pyramid-Attention-Networks-master/Demosaic/code/model/ddbpn.py
# Deep Back-Projection Networks For Super-Resolution # https://arxiv.org/abs/1803.02735 from model import common import torch import torch.nn as nn def make_model(args, parent=False): return DDBPN(args) def projection_conv(in_channels, out_channels, scale, up=True): kernel_size, stride, padding = { 2: (6, 2, 2), 4: (8, 4, 2), 8: (12, 8, 2) }[scale] if up: conv_f = nn.ConvTranspose2d else: conv_f = nn.Conv2d return conv_f( in_channels, out_channels, kernel_size, stride=stride, padding=padding ) class DenseProjection(nn.Module): def __init__(self, in_channels, nr, scale, up=True, bottleneck=True): super(DenseProjection, self).__init__() if bottleneck: self.bottleneck = nn.Sequential(*[ nn.Conv2d(in_channels, nr, 1), nn.PReLU(nr) ]) inter_channels = nr else: self.bottleneck = None inter_channels = in_channels self.conv_1 = nn.Sequential(*[ projection_conv(inter_channels, nr, scale, up), nn.PReLU(nr) ]) self.conv_2 = nn.Sequential(*[ projection_conv(nr, inter_channels, scale, not up), nn.PReLU(inter_channels) ]) self.conv_3 = nn.Sequential(*[ projection_conv(inter_channels, nr, scale, up), nn.PReLU(nr) ]) def forward(self, x): if self.bottleneck is not None: x = self.bottleneck(x) a_0 = self.conv_1(x) b_0 = self.conv_2(a_0) e = b_0.sub(x) a_1 = self.conv_3(e) out = a_0.add(a_1) return out class DDBPN(nn.Module): def __init__(self, args): super(DDBPN, self).__init__() scale = args.scale[0] n0 = 128 nr = 32 self.depth = 6 rgb_mean = (0.4488, 0.4371, 0.4040) rgb_std = (1.0, 1.0, 1.0) self.sub_mean = common.MeanShift(args.rgb_range, rgb_mean, rgb_std) initial = [ nn.Conv2d(args.n_colors, n0, 3, padding=1), nn.PReLU(n0), nn.Conv2d(n0, nr, 1), nn.PReLU(nr) ] self.initial = nn.Sequential(*initial) self.upmodules = nn.ModuleList() self.downmodules = nn.ModuleList() channels = nr for i in range(self.depth): self.upmodules.append( DenseProjection(channels, nr, scale, True, i > 1) ) if i != 0: channels += nr channels = nr for i in range(self.depth - 1): self.downmodules.append( DenseProjection(channels, nr, scale, False, i != 0) ) channels += nr reconstruction = [ nn.Conv2d(self.depth * nr, args.n_colors, 3, padding=1) ] self.reconstruction = nn.Sequential(*reconstruction) self.add_mean = common.MeanShift(args.rgb_range, rgb_mean, rgb_std, 1) def forward(self, x): x = self.sub_mean(x) x = self.initial(x) h_list = [] l_list = [] for i in range(self.depth - 1): if i == 0: l = x else: l = torch.cat(l_list, dim=1) h_list.append(self.upmodules[i](l)) l_list.append(self.downmodules[i](torch.cat(h_list, dim=1))) h_list.append(self.upmodules[-1](torch.cat(l_list, dim=1))) out = self.reconstruction(torch.cat(h_list, dim=1)) out = self.add_mean(out) return out
3,629
26.5
78
py
Pyramid-Attention-Networks
Pyramid-Attention-Networks-master/Demosaic/code/model/rdn.py
# Residual Dense Network for Image Super-Resolution # https://arxiv.org/abs/1802.08797 from model import common import torch import torch.nn as nn def make_model(args, parent=False): return RDN(args) class RDB_Conv(nn.Module): def __init__(self, inChannels, growRate, kSize=3): super(RDB_Conv, self).__init__() Cin = inChannels G = growRate self.conv = nn.Sequential(*[ nn.Conv2d(Cin, G, kSize, padding=(kSize-1)//2, stride=1), nn.ReLU() ]) def forward(self, x): out = self.conv(x) return torch.cat((x, out), 1) class RDB(nn.Module): def __init__(self, growRate0, growRate, nConvLayers, kSize=3): super(RDB, self).__init__() G0 = growRate0 G = growRate C = nConvLayers convs = [] for c in range(C): convs.append(RDB_Conv(G0 + c*G, G)) self.convs = nn.Sequential(*convs) # Local Feature Fusion self.LFF = nn.Conv2d(G0 + C*G, G0, 1, padding=0, stride=1) def forward(self, x): return self.LFF(self.convs(x)) + x class RDN(nn.Module): def __init__(self, args): super(RDN, self).__init__() r = args.scale[0] G0 = args.G0 kSize = args.RDNkSize # number of RDB blocks, conv layers, out channels self.D, C, G = { 'A': (20, 6, 32), 'B': (16, 8, 64), }[args.RDNconfig] # Shallow feature extraction net self.SFENet1 = nn.Conv2d(args.n_colors, G0, kSize, padding=(kSize-1)//2, stride=1) self.SFENet2 = nn.Conv2d(G0, G0, kSize, padding=(kSize-1)//2, stride=1) # Redidual dense blocks and dense feature fusion self.RDBs = nn.ModuleList() for i in range(self.D): self.RDBs.append( RDB(growRate0 = G0, growRate = G, nConvLayers = C) ) # Global Feature Fusion self.GFF = nn.Sequential(*[ nn.Conv2d(self.D * G0, G0, 1, padding=0, stride=1), nn.Conv2d(G0, G0, kSize, padding=(kSize-1)//2, stride=1) ]) # Up-sampling net if r == 2 or r == 3: self.UPNet = nn.Sequential(*[ nn.Conv2d(G0, G * r * r, kSize, padding=(kSize-1)//2, stride=1), nn.PixelShuffle(r), nn.Conv2d(G, args.n_colors, kSize, padding=(kSize-1)//2, stride=1) ]) elif r == 4: self.UPNet = nn.Sequential(*[ nn.Conv2d(G0, G * 4, kSize, padding=(kSize-1)//2, stride=1), nn.PixelShuffle(2), nn.Conv2d(G, G * 4, kSize, padding=(kSize-1)//2, stride=1), nn.PixelShuffle(2), nn.Conv2d(G, args.n_colors, kSize, padding=(kSize-1)//2, stride=1) ]) else: raise ValueError("scale must be 2 or 3 or 4.") def forward(self, x): f__1 = self.SFENet1(x) x = self.SFENet2(f__1) RDBs_out = [] for i in range(self.D): x = self.RDBs[i](x) RDBs_out.append(x) x = self.GFF(torch.cat(RDBs_out,1)) x += f__1 return self.UPNet(x)
3,202
29.216981
90
py
Pyramid-Attention-Networks
Pyramid-Attention-Networks-master/Demosaic/code/model/mdsr.py
from model import common import torch.nn as nn def make_model(args, parent=False): return MDSR(args) class MDSR(nn.Module): def __init__(self, args, conv=common.default_conv): super(MDSR, self).__init__() n_resblocks = args.n_resblocks n_feats = args.n_feats kernel_size = 3 self.scale_idx = 0 act = nn.ReLU(True) rgb_mean = (0.4488, 0.4371, 0.4040) rgb_std = (1.0, 1.0, 1.0) self.sub_mean = common.MeanShift(args.rgb_range, rgb_mean, rgb_std) m_head = [conv(args.n_colors, n_feats, kernel_size)] self.pre_process = nn.ModuleList([ nn.Sequential( common.ResBlock(conv, n_feats, 5, act=act), common.ResBlock(conv, n_feats, 5, act=act) ) for _ in args.scale ]) m_body = [ common.ResBlock( conv, n_feats, kernel_size, act=act ) for _ in range(n_resblocks) ] m_body.append(conv(n_feats, n_feats, kernel_size)) self.upsample = nn.ModuleList([ common.Upsampler( conv, s, n_feats, act=False ) for s in args.scale ]) m_tail = [conv(n_feats, args.n_colors, kernel_size)] self.add_mean = common.MeanShift(args.rgb_range, rgb_mean, rgb_std, 1) self.head = nn.Sequential(*m_head) self.body = nn.Sequential(*m_body) self.tail = nn.Sequential(*m_tail) def forward(self, x): x = self.sub_mean(x) x = self.head(x) x = self.pre_process[self.scale_idx](x) res = self.body(x) res += x x = self.upsample[self.scale_idx](res) x = self.tail(x) x = self.add_mean(x) return x def set_scale(self, scale_idx): self.scale_idx = scale_idx
1,837
25.637681
78
py
Pyramid-Attention-Networks
Pyramid-Attention-Networks-master/Demosaic/code/model/common.py
import math import torch import torch.nn as nn import torch.nn.functional as F def default_conv(in_channels, out_channels, kernel_size,stride=1, bias=True): return nn.Conv2d( in_channels, out_channels, kernel_size, padding=(kernel_size//2),stride=stride, bias=bias) class MeanShift(nn.Conv2d): def __init__( self, rgb_range, rgb_mean=(0.4488, 0.4371, 0.4040), rgb_std=(1.0, 1.0, 1.0), sign=-1): super(MeanShift, self).__init__(3, 3, kernel_size=1) std = torch.Tensor(rgb_std) self.weight.data = torch.eye(3).view(3, 3, 1, 1) / std.view(3, 1, 1, 1) self.bias.data = sign * rgb_range * torch.Tensor(rgb_mean) / std for p in self.parameters(): p.requires_grad = False class BasicBlock(nn.Sequential): def __init__( self, conv, in_channels, out_channels, kernel_size, stride=1, bias=True, bn=False, act=nn.PReLU()): m = [conv(in_channels, out_channels, kernel_size, bias=bias)] if bn: m.append(nn.BatchNorm2d(out_channels)) if act is not None: m.append(act) super(BasicBlock, self).__init__(*m) class ResBlock(nn.Module): def __init__( self, conv, n_feats, kernel_size, bias=True, bn=False, act=nn.PReLU(), res_scale=1): super(ResBlock, self).__init__() m = [] for i in range(2): m.append(conv(n_feats, n_feats, kernel_size, bias=bias)) if bn: m.append(nn.BatchNorm2d(n_feats)) if i == 0: m.append(act) self.body = nn.Sequential(*m) self.res_scale = res_scale def forward(self, x): res = self.body(x).mul(self.res_scale) res += x return res class Upsampler(nn.Sequential): def __init__(self, conv, scale, n_feats, bn=False, act=False, bias=True): m = [] if (scale & (scale - 1)) == 0: # Is scale = 2^n? for _ in range(int(math.log(scale, 2))): m.append(conv(n_feats, 4 * n_feats, 3, bias)) m.append(nn.PixelShuffle(2)) if bn: m.append(nn.BatchNorm2d(n_feats)) if act == 'relu': m.append(nn.ReLU(True)) elif act == 'prelu': m.append(nn.PReLU(n_feats)) elif scale == 3: m.append(conv(n_feats, 9 * n_feats, 3, bias)) m.append(nn.PixelShuffle(3)) if bn: m.append(nn.BatchNorm2d(n_feats)) if act == 'relu': m.append(nn.ReLU(True)) elif act == 'prelu': m.append(nn.PReLU(n_feats)) else: raise NotImplementedError super(Upsampler, self).__init__(*m)
2,799
30.460674
80
py
Pyramid-Attention-Networks
Pyramid-Attention-Networks-master/Demosaic/code/model/__init__.py
import os from importlib import import_module import torch import torch.nn as nn from torch.autograd import Variable class Model(nn.Module): def __init__(self, args, ckp): super(Model, self).__init__() print('Making model...') self.scale = args.scale self.idx_scale = 0 self.self_ensemble = args.self_ensemble self.chop = args.chop self.precision = args.precision self.cpu = args.cpu self.device = torch.device('cpu' if args.cpu else 'cuda') self.n_GPUs = args.n_GPUs self.save_models = args.save_models module = import_module('model.' + args.model.lower()) self.model = module.make_model(args).to(self.device) if args.precision == 'half': self.model.half() if not args.cpu and args.n_GPUs > 1: self.model = nn.DataParallel(self.model, range(args.n_GPUs)) self.load( ckp.dir, pre_train=args.pre_train, resume=args.resume, cpu=args.cpu ) print(self.model, file=ckp.log_file) def forward(self, x, idx_scale): self.idx_scale = idx_scale target = self.get_model() if hasattr(target, 'set_scale'): target.set_scale(idx_scale) if self.self_ensemble and not self.training: if self.chop: forward_function = self.forward_chop else: forward_function = self.model.forward return self.forward_x8(x, forward_function) elif self.chop and not self.training: return self.forward_chop(x) else: return self.model(x) def get_model(self): if self.n_GPUs == 1: return self.model else: return self.model.module def state_dict(self, **kwargs): target = self.get_model() return target.state_dict(**kwargs) def save(self, apath, epoch, is_best=False): target = self.get_model() torch.save( target.state_dict(), os.path.join(apath, 'model_latest.pt') ) if is_best: torch.save( target.state_dict(), os.path.join(apath, 'model_best.pt') ) if self.save_models: torch.save( target.state_dict(), os.path.join(apath, 'model_{}.pt'.format(epoch)) ) def load(self, apath, pre_train='.', resume=-1, cpu=False): if cpu: kwargs = {'map_location': lambda storage, loc: storage} else: kwargs = {} if resume == -1: self.get_model().load_state_dict( torch.load( os.path.join(apath, 'model_latest.pt'), **kwargs ), strict=False ) elif resume == 0: if pre_train != '.': print('Loading model from {}'.format(pre_train)) self.get_model().load_state_dict( torch.load(pre_train, **kwargs), strict=False ) else: self.get_model().load_state_dict( torch.load( os.path.join(apath, 'model', 'model_{}.pt'.format(resume)), **kwargs ), strict=False ) def forward_chop(self, x, shave=10, min_size=6400): scale = self.scale[self.idx_scale] scale = 1 n_GPUs = min(self.n_GPUs, 4) b, c, h, w = x.size() h_half, w_half = h // 2, w // 2 h_size, w_size = h_half + shave, w_half + shave lr_list = [ x[:, :, 0:h_size, 0:w_size], x[:, :, 0:h_size, (w - w_size):w], x[:, :, (h - h_size):h, 0:w_size], x[:, :, (h - h_size):h, (w - w_size):w]] if w_size * h_size < min_size: sr_list = [] for i in range(0, 4, n_GPUs): lr_batch = torch.cat(lr_list[i:(i + n_GPUs)], dim=0) sr_batch = self.model(lr_batch) sr_list.extend(sr_batch.chunk(n_GPUs, dim=0)) else: sr_list = [ self.forward_chop(patch, shave=shave, min_size=min_size) \ for patch in lr_list ] h, w = scale * h, scale * w h_half, w_half = scale * h_half, scale * w_half h_size, w_size = scale * h_size, scale * w_size shave *= scale output = x.new(b, c, h, w) output[:, :, 0:h_half, 0:w_half] \ = sr_list[0][:, :, 0:h_half, 0:w_half] output[:, :, 0:h_half, w_half:w] \ = sr_list[1][:, :, 0:h_half, (w_size - w + w_half):w_size] output[:, :, h_half:h, 0:w_half] \ = sr_list[2][:, :, (h_size - h + h_half):h_size, 0:w_half] output[:, :, h_half:h, w_half:w] \ = sr_list[3][:, :, (h_size - h + h_half):h_size, (w_size - w + w_half):w_size] return output def forward_x8(self, x, forward_function): def _transform(v, op): if self.precision != 'single': v = v.float() v2np = v.data.cpu().numpy() if op == 'v': tfnp = v2np[:, :, :, ::-1].copy() elif op == 'h': tfnp = v2np[:, :, ::-1, :].copy() elif op == 't': tfnp = v2np.transpose((0, 1, 3, 2)).copy() ret = torch.Tensor(tfnp).to(self.device) if self.precision == 'half': ret = ret.half() return ret lr_list = [x] for tf in 'v', 'h', 't': lr_list.extend([_transform(t, tf) for t in lr_list]) sr_list = [forward_function(aug) for aug in lr_list] for i in range(len(sr_list)): if i > 3: sr_list[i] = _transform(sr_list[i], 't') if i % 4 > 1: sr_list[i] = _transform(sr_list[i], 'h') if (i % 4) % 2 == 1: sr_list[i] = _transform(sr_list[i], 'v') output_cat = torch.cat(sr_list, dim=0) output = output_cat.mean(dim=0, keepdim=True) return output
6,200
31.465969
90
py
Pyramid-Attention-Networks
Pyramid-Attention-Networks-master/Demosaic/code/model/panet.py
from model import common from model import attention import torch.nn as nn def make_model(args, parent=False): return PANET(args) class PANET(nn.Module): def __init__(self, args, conv=common.default_conv): super(PANET, self).__init__() n_resblocks = args.n_resblocks n_feats = args.n_feats kernel_size = 3 scale = args.scale[0] rgb_mean = (0.4488, 0.4371, 0.4040) rgb_std = (1.0, 1.0, 1.0) self.sub_mean = common.MeanShift(args.rgb_range, rgb_mean, rgb_std) msa = attention.PyramidAttention() # define head module m_head = [conv(args.n_colors, n_feats, kernel_size)] # define body module m_body = [ common.ResBlock( conv, n_feats, kernel_size, nn.PReLU(), res_scale=args.res_scale ) for _ in range(n_resblocks//2) ] m_body.append(msa) for i in range(n_resblocks//2): m_body.append(common.ResBlock(conv,n_feats,kernel_size,nn.PReLU(),res_scale=args.res_scale)) m_body.append(conv(n_feats, n_feats, kernel_size)) # define tail module #m_tail = [ # common.Upsampler(conv, scale, n_feats, act=False), # conv(n_feats, args.n_colors, kernel_size) #] m_tail = [ conv(n_feats, args.n_colors, kernel_size) ] self.add_mean = common.MeanShift(args.rgb_range, rgb_mean, rgb_std, 1) self.head = nn.Sequential(*m_head) self.body = nn.Sequential(*m_body) self.tail = nn.Sequential(*m_tail) def forward(self, x): #x = self.sub_mean(x) x = self.head(x) res = self.body(x) res += x x = self.tail(res) #x = self.add_mean(x) return x def load_state_dict(self, state_dict, strict=True): own_state = self.state_dict() for name, param in state_dict.items(): if name in own_state: if isinstance(param, nn.Parameter): param = param.data try: own_state[name].copy_(param) except Exception: if name.find('tail') == -1: raise RuntimeError('While copying the parameter named {}, ' 'whose dimensions in the model are {} and ' 'whose dimensions in the checkpoint are {}.' .format(name, own_state[name].size(), param.size())) elif strict: if name.find('tail') == -1: raise KeyError('unexpected key "{}" in state_dict' .format(name))
2,779
32.493976
104
py
Pyramid-Attention-Networks
Pyramid-Attention-Networks-master/Demosaic/code/model/attention.py
import torch import torch.nn as nn import torch.nn.functional as F from torchvision import transforms from torchvision import utils as vutils from model import common from utils.tools import extract_image_patches,\ reduce_mean, reduce_sum, same_padding class PyramidAttention(nn.Module): def __init__(self, level=5, res_scale=1, channel=64, reduction=2, ksize=3, stride=1, softmax_scale=10, average=True, conv=common.default_conv): super(PyramidAttention, self).__init__() self.ksize = ksize self.stride = stride self.res_scale = res_scale self.softmax_scale = softmax_scale self.scale = [1-i/10 for i in range(level)] self.average = average escape_NaN = torch.FloatTensor([1e-4]) self.register_buffer('escape_NaN', escape_NaN) self.conv_match_L_base = common.BasicBlock(conv,channel,channel//reduction, 1, bn=False, act=nn.PReLU()) self.conv_match = common.BasicBlock(conv,channel, channel//reduction, 1, bn=False, act=nn.PReLU()) self.conv_assembly = common.BasicBlock(conv,channel, channel,1,bn=False, act=nn.PReLU()) def forward(self, input): res = input #theta match_base = self.conv_match_L_base(input) shape_base = list(res.size()) input_groups = torch.split(match_base,1,dim=0) # patch size for matching kernel = self.ksize # raw_w is for reconstruction raw_w = [] # w is for matching w = [] #build feature pyramid for i in range(len(self.scale)): ref = input if self.scale[i]!=1: ref = F.interpolate(input, scale_factor=self.scale[i], mode='bicubic') #feature transformation function f base = self.conv_assembly(ref) shape_input = base.shape #sampling raw_w_i = extract_image_patches(base, ksizes=[kernel, kernel], strides=[self.stride,self.stride], rates=[1, 1], padding='same') # [N, C*k*k, L] raw_w_i = raw_w_i.view(shape_input[0], shape_input[1], kernel, kernel, -1) raw_w_i = raw_w_i.permute(0, 4, 1, 2, 3) # raw_shape: [N, L, C, k, k] raw_w_i_groups = torch.split(raw_w_i, 1, dim=0) raw_w.append(raw_w_i_groups) #feature transformation function g ref_i = self.conv_match(ref) shape_ref = ref_i.shape #sampling w_i = extract_image_patches(ref_i, ksizes=[self.ksize, self.ksize], strides=[self.stride, self.stride], rates=[1, 1], padding='same') w_i = w_i.view(shape_ref[0], shape_ref[1], self.ksize, self.ksize, -1) w_i = w_i.permute(0, 4, 1, 2, 3) # w shape: [N, L, C, k, k] w_i_groups = torch.split(w_i, 1, dim=0) w.append(w_i_groups) y = [] for idx, xi in enumerate(input_groups): #group in a filter wi = torch.cat([w[i][idx][0] for i in range(len(self.scale))],dim=0) # [L, C, k, k] #normalize max_wi = torch.max(torch.sqrt(reduce_sum(torch.pow(wi, 2), axis=[1, 2, 3], keepdim=True)), self.escape_NaN) wi_normed = wi/ max_wi #matching xi = same_padding(xi, [self.ksize, self.ksize], [1, 1], [1, 1]) # xi: 1*c*H*W yi = F.conv2d(xi, wi_normed, stride=1) # [1, L, H, W] L = shape_ref[2]*shape_ref[3] yi = yi.view(1,wi.shape[0], shape_base[2], shape_base[3]) # (B=1, C=32*32, H=32, W=32) # softmax matching score yi = F.softmax(yi*self.softmax_scale, dim=1) if self.average == False: yi = (yi == yi.max(dim=1,keepdim=True)[0]).float() # deconv for patch pasting raw_wi = torch.cat([raw_w[i][idx][0] for i in range(len(self.scale))],dim=0) yi = F.conv_transpose2d(yi, raw_wi, stride=self.stride,padding=1)/4. y.append(yi) y = torch.cat(y, dim=0)+res*self.res_scale # back to the mini-batch return y
4,427
46.106383
147
py
Pyramid-Attention-Networks
Pyramid-Attention-Networks-master/Demosaic/code/model/vdsr.py
from model import common import torch.nn as nn import torch.nn.init as init url = { 'r20f64': '' } def make_model(args, parent=False): return VDSR(args) class VDSR(nn.Module): def __init__(self, args, conv=common.default_conv): super(VDSR, self).__init__() n_resblocks = args.n_resblocks n_feats = args.n_feats kernel_size = 3 self.url = url['r{}f{}'.format(n_resblocks, n_feats)] self.sub_mean = common.MeanShift(args.rgb_range) self.add_mean = common.MeanShift(args.rgb_range, sign=1) def basic_block(in_channels, out_channels, act): return common.BasicBlock( conv, in_channels, out_channels, kernel_size, bias=True, bn=False, act=act ) # define body module m_body = [] m_body.append(basic_block(args.n_colors, n_feats, nn.ReLU(True))) for _ in range(n_resblocks - 2): m_body.append(basic_block(n_feats, n_feats, nn.ReLU(True))) m_body.append(basic_block(n_feats, args.n_colors, None)) self.body = nn.Sequential(*m_body) def forward(self, x): x = self.sub_mean(x) res = self.body(x) res += x x = self.add_mean(res) return x
1,275
26.148936
73
py
Pyramid-Attention-Networks
Pyramid-Attention-Networks-master/Demosaic/code/model/utils/tools.py
import os import torch import numpy as np from PIL import Image import torch.nn.functional as F def normalize(x): return x.mul_(2).add_(-1) def same_padding(images, ksizes, strides, rates): assert len(images.size()) == 4 batch_size, channel, rows, cols = images.size() out_rows = (rows + strides[0] - 1) // strides[0] out_cols = (cols + strides[1] - 1) // strides[1] effective_k_row = (ksizes[0] - 1) * rates[0] + 1 effective_k_col = (ksizes[1] - 1) * rates[1] + 1 padding_rows = max(0, (out_rows-1)*strides[0]+effective_k_row-rows) padding_cols = max(0, (out_cols-1)*strides[1]+effective_k_col-cols) # Pad the input padding_top = int(padding_rows / 2.) padding_left = int(padding_cols / 2.) padding_bottom = padding_rows - padding_top padding_right = padding_cols - padding_left paddings = (padding_left, padding_right, padding_top, padding_bottom) images = torch.nn.ZeroPad2d(paddings)(images) return images def extract_image_patches(images, ksizes, strides, rates, padding='same'): """ Extract patches from images and put them in the C output dimension. :param padding: :param images: [batch, channels, in_rows, in_cols]. A 4-D Tensor with shape :param ksizes: [ksize_rows, ksize_cols]. The size of the sliding window for each dimension of images :param strides: [stride_rows, stride_cols] :param rates: [dilation_rows, dilation_cols] :return: A Tensor """ assert len(images.size()) == 4 assert padding in ['same', 'valid'] batch_size, channel, height, width = images.size() if padding == 'same': images = same_padding(images, ksizes, strides, rates) elif padding == 'valid': pass else: raise NotImplementedError('Unsupported padding type: {}.\ Only "same" or "valid" are supported.'.format(padding)) unfold = torch.nn.Unfold(kernel_size=ksizes, dilation=rates, padding=0, stride=strides) patches = unfold(images) return patches # [N, C*k*k, L], L is the total number of such blocks def reduce_mean(x, axis=None, keepdim=False): if not axis: axis = range(len(x.shape)) for i in sorted(axis, reverse=True): x = torch.mean(x, dim=i, keepdim=keepdim) return x def reduce_std(x, axis=None, keepdim=False): if not axis: axis = range(len(x.shape)) for i in sorted(axis, reverse=True): x = torch.std(x, dim=i, keepdim=keepdim) return x def reduce_sum(x, axis=None, keepdim=False): if not axis: axis = range(len(x.shape)) for i in sorted(axis, reverse=True): x = torch.sum(x, dim=i, keepdim=keepdim) return x
2,777
32.878049
79
py