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
DiffPure
DiffPure-master/score_sde/models/layers.py
# coding=utf-8 # Copyright 2020 The Google Research Authors. # # Licensed under the Apache License, Version 2.0 (the "License"); # you may not use this file except in compliance with the License. # You may obtain a copy of the License at # # http://www.apache.org/licenses/LICENSE-2.0 # # Unless required by applicable law or agreed to in writing, software # distributed under the License is distributed on an "AS IS" BASIS, # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. # pylint: skip-file """Common layers for defining score networks. """ import math import string from functools import partial import torch.nn as nn import torch import torch.nn.functional as F import numpy as np from .normalization import ConditionalInstanceNorm2dPlus def get_act(config): """Get activation functions from the config file.""" if config.model.nonlinearity.lower() == 'elu': return nn.ELU() elif config.model.nonlinearity.lower() == 'relu': return nn.ReLU() elif config.model.nonlinearity.lower() == 'lrelu': return nn.LeakyReLU(negative_slope=0.2) elif config.model.nonlinearity.lower() == 'swish': return nn.SiLU() else: raise NotImplementedError('activation function does not exist!') def ncsn_conv1x1(in_planes, out_planes, stride=1, bias=True, dilation=1, init_scale=1., padding=0): """1x1 convolution. Same as NCSNv1/v2.""" conv = nn.Conv2d(in_planes, out_planes, kernel_size=1, stride=stride, bias=bias, dilation=dilation, padding=padding) init_scale = 1e-10 if init_scale == 0 else init_scale conv.weight.data *= init_scale conv.bias.data *= init_scale return conv def variance_scaling(scale, mode, distribution, in_axis=1, out_axis=0, dtype=torch.float32, device='cpu'): """Ported from JAX. """ def _compute_fans(shape, in_axis=1, out_axis=0): receptive_field_size = np.prod(shape) / shape[in_axis] / shape[out_axis] fan_in = shape[in_axis] * receptive_field_size fan_out = shape[out_axis] * receptive_field_size return fan_in, fan_out def init(shape, dtype=dtype, device=device): fan_in, fan_out = _compute_fans(shape, in_axis, out_axis) if mode == "fan_in": denominator = fan_in elif mode == "fan_out": denominator = fan_out elif mode == "fan_avg": denominator = (fan_in + fan_out) / 2 else: raise ValueError( "invalid mode for variance scaling initializer: {}".format(mode)) variance = scale / denominator if distribution == "normal": return torch.randn(*shape, dtype=dtype, device=device) * np.sqrt(variance) elif distribution == "uniform": return (torch.rand(*shape, dtype=dtype, device=device) * 2. - 1.) * np.sqrt(3 * variance) else: raise ValueError("invalid distribution for variance scaling initializer") return init def default_init(scale=1.): """The same initialization used in DDPM.""" scale = 1e-10 if scale == 0 else scale return variance_scaling(scale, 'fan_avg', 'uniform') class Dense(nn.Module): """Linear layer with `default_init`.""" def __init__(self): super().__init__() def ddpm_conv1x1(in_planes, out_planes, stride=1, bias=True, init_scale=1., padding=0): """1x1 convolution with DDPM initialization.""" conv = nn.Conv2d(in_planes, out_planes, kernel_size=1, stride=stride, padding=padding, bias=bias) conv.weight.data = default_init(init_scale)(conv.weight.data.shape) nn.init.zeros_(conv.bias) return conv def ncsn_conv3x3(in_planes, out_planes, stride=1, bias=True, dilation=1, init_scale=1., padding=1): """3x3 convolution with PyTorch initialization. Same as NCSNv1/NCSNv2.""" init_scale = 1e-10 if init_scale == 0 else init_scale conv = nn.Conv2d(in_planes, out_planes, stride=stride, bias=bias, dilation=dilation, padding=padding, kernel_size=3) conv.weight.data *= init_scale conv.bias.data *= init_scale return conv def ddpm_conv3x3(in_planes, out_planes, stride=1, bias=True, dilation=1, init_scale=1., padding=1): """3x3 convolution with DDPM initialization.""" conv = nn.Conv2d(in_planes, out_planes, kernel_size=3, stride=stride, padding=padding, dilation=dilation, bias=bias) conv.weight.data = default_init(init_scale)(conv.weight.data.shape) nn.init.zeros_(conv.bias) return conv ########################################################################### # Functions below are ported over from the NCSNv1/NCSNv2 codebase: # https://github.com/ermongroup/ncsn # https://github.com/ermongroup/ncsnv2 ########################################################################### class CRPBlock(nn.Module): def __init__(self, features, n_stages, act=nn.ReLU(), maxpool=True): super().__init__() self.convs = nn.ModuleList() for i in range(n_stages): self.convs.append(ncsn_conv3x3(features, features, stride=1, bias=False)) self.n_stages = n_stages if maxpool: self.pool = nn.MaxPool2d(kernel_size=5, stride=1, padding=2) else: self.pool = nn.AvgPool2d(kernel_size=5, stride=1, padding=2) self.act = act def forward(self, x): x = self.act(x) path = x for i in range(self.n_stages): path = self.pool(path) path = self.convs[i](path) x = path + x return x class CondCRPBlock(nn.Module): def __init__(self, features, n_stages, num_classes, normalizer, act=nn.ReLU()): super().__init__() self.convs = nn.ModuleList() self.norms = nn.ModuleList() self.normalizer = normalizer for i in range(n_stages): self.norms.append(normalizer(features, num_classes, bias=True)) self.convs.append(ncsn_conv3x3(features, features, stride=1, bias=False)) self.n_stages = n_stages self.pool = nn.AvgPool2d(kernel_size=5, stride=1, padding=2) self.act = act def forward(self, x, y): x = self.act(x) path = x for i in range(self.n_stages): path = self.norms[i](path, y) path = self.pool(path) path = self.convs[i](path) x = path + x return x class RCUBlock(nn.Module): def __init__(self, features, n_blocks, n_stages, act=nn.ReLU()): super().__init__() for i in range(n_blocks): for j in range(n_stages): setattr(self, '{}_{}_conv'.format(i + 1, j + 1), ncsn_conv3x3(features, features, stride=1, bias=False)) self.stride = 1 self.n_blocks = n_blocks self.n_stages = n_stages self.act = act def forward(self, x): for i in range(self.n_blocks): residual = x for j in range(self.n_stages): x = self.act(x) x = getattr(self, '{}_{}_conv'.format(i + 1, j + 1))(x) x += residual return x class CondRCUBlock(nn.Module): def __init__(self, features, n_blocks, n_stages, num_classes, normalizer, act=nn.ReLU()): super().__init__() for i in range(n_blocks): for j in range(n_stages): setattr(self, '{}_{}_norm'.format(i + 1, j + 1), normalizer(features, num_classes, bias=True)) setattr(self, '{}_{}_conv'.format(i + 1, j + 1), ncsn_conv3x3(features, features, stride=1, bias=False)) self.stride = 1 self.n_blocks = n_blocks self.n_stages = n_stages self.act = act self.normalizer = normalizer def forward(self, x, y): for i in range(self.n_blocks): residual = x for j in range(self.n_stages): x = getattr(self, '{}_{}_norm'.format(i + 1, j + 1))(x, y) x = self.act(x) x = getattr(self, '{}_{}_conv'.format(i + 1, j + 1))(x) x += residual return x class MSFBlock(nn.Module): def __init__(self, in_planes, features): super().__init__() assert isinstance(in_planes, list) or isinstance(in_planes, tuple) self.convs = nn.ModuleList() self.features = features for i in range(len(in_planes)): self.convs.append(ncsn_conv3x3(in_planes[i], features, stride=1, bias=True)) def forward(self, xs, shape): sums = torch.zeros(xs[0].shape[0], self.features, *shape, device=xs[0].device) for i in range(len(self.convs)): h = self.convs[i](xs[i]) h = F.interpolate(h, size=shape, mode='bilinear', align_corners=True) sums += h return sums class CondMSFBlock(nn.Module): def __init__(self, in_planes, features, num_classes, normalizer): super().__init__() assert isinstance(in_planes, list) or isinstance(in_planes, tuple) self.convs = nn.ModuleList() self.norms = nn.ModuleList() self.features = features self.normalizer = normalizer for i in range(len(in_planes)): self.convs.append(ncsn_conv3x3(in_planes[i], features, stride=1, bias=True)) self.norms.append(normalizer(in_planes[i], num_classes, bias=True)) def forward(self, xs, y, shape): sums = torch.zeros(xs[0].shape[0], self.features, *shape, device=xs[0].device) for i in range(len(self.convs)): h = self.norms[i](xs[i], y) h = self.convs[i](h) h = F.interpolate(h, size=shape, mode='bilinear', align_corners=True) sums += h return sums class RefineBlock(nn.Module): def __init__(self, in_planes, features, act=nn.ReLU(), start=False, end=False, maxpool=True): super().__init__() assert isinstance(in_planes, tuple) or isinstance(in_planes, list) self.n_blocks = n_blocks = len(in_planes) self.adapt_convs = nn.ModuleList() for i in range(n_blocks): self.adapt_convs.append(RCUBlock(in_planes[i], 2, 2, act)) self.output_convs = RCUBlock(features, 3 if end else 1, 2, act) if not start: self.msf = MSFBlock(in_planes, features) self.crp = CRPBlock(features, 2, act, maxpool=maxpool) def forward(self, xs, output_shape): assert isinstance(xs, tuple) or isinstance(xs, list) hs = [] for i in range(len(xs)): h = self.adapt_convs[i](xs[i]) hs.append(h) if self.n_blocks > 1: h = self.msf(hs, output_shape) else: h = hs[0] h = self.crp(h) h = self.output_convs(h) return h class CondRefineBlock(nn.Module): def __init__(self, in_planes, features, num_classes, normalizer, act=nn.ReLU(), start=False, end=False): super().__init__() assert isinstance(in_planes, tuple) or isinstance(in_planes, list) self.n_blocks = n_blocks = len(in_planes) self.adapt_convs = nn.ModuleList() for i in range(n_blocks): self.adapt_convs.append( CondRCUBlock(in_planes[i], 2, 2, num_classes, normalizer, act) ) self.output_convs = CondRCUBlock(features, 3 if end else 1, 2, num_classes, normalizer, act) if not start: self.msf = CondMSFBlock(in_planes, features, num_classes, normalizer) self.crp = CondCRPBlock(features, 2, num_classes, normalizer, act) def forward(self, xs, y, output_shape): assert isinstance(xs, tuple) or isinstance(xs, list) hs = [] for i in range(len(xs)): h = self.adapt_convs[i](xs[i], y) hs.append(h) if self.n_blocks > 1: h = self.msf(hs, y, output_shape) else: h = hs[0] h = self.crp(h, y) h = self.output_convs(h, y) return h class ConvMeanPool(nn.Module): def __init__(self, input_dim, output_dim, kernel_size=3, biases=True, adjust_padding=False): super().__init__() if not adjust_padding: conv = nn.Conv2d(input_dim, output_dim, kernel_size, stride=1, padding=kernel_size // 2, bias=biases) self.conv = conv else: conv = nn.Conv2d(input_dim, output_dim, kernel_size, stride=1, padding=kernel_size // 2, bias=biases) self.conv = nn.Sequential( nn.ZeroPad2d((1, 0, 1, 0)), conv ) def forward(self, inputs): output = self.conv(inputs) output = sum([output[:, :, ::2, ::2], output[:, :, 1::2, ::2], output[:, :, ::2, 1::2], output[:, :, 1::2, 1::2]]) / 4. return output class MeanPoolConv(nn.Module): def __init__(self, input_dim, output_dim, kernel_size=3, biases=True): super().__init__() self.conv = nn.Conv2d(input_dim, output_dim, kernel_size, stride=1, padding=kernel_size // 2, bias=biases) def forward(self, inputs): output = inputs output = sum([output[:, :, ::2, ::2], output[:, :, 1::2, ::2], output[:, :, ::2, 1::2], output[:, :, 1::2, 1::2]]) / 4. return self.conv(output) class UpsampleConv(nn.Module): def __init__(self, input_dim, output_dim, kernel_size=3, biases=True): super().__init__() self.conv = nn.Conv2d(input_dim, output_dim, kernel_size, stride=1, padding=kernel_size // 2, bias=biases) self.pixelshuffle = nn.PixelShuffle(upscale_factor=2) def forward(self, inputs): output = inputs output = torch.cat([output, output, output, output], dim=1) output = self.pixelshuffle(output) return self.conv(output) class ConditionalResidualBlock(nn.Module): def __init__(self, input_dim, output_dim, num_classes, resample=1, act=nn.ELU(), normalization=ConditionalInstanceNorm2dPlus, adjust_padding=False, dilation=None): super().__init__() self.non_linearity = act self.input_dim = input_dim self.output_dim = output_dim self.resample = resample self.normalization = normalization if resample == 'down': if dilation > 1: self.conv1 = ncsn_conv3x3(input_dim, input_dim, dilation=dilation) self.normalize2 = normalization(input_dim, num_classes) self.conv2 = ncsn_conv3x3(input_dim, output_dim, dilation=dilation) conv_shortcut = partial(ncsn_conv3x3, dilation=dilation) else: self.conv1 = ncsn_conv3x3(input_dim, input_dim) self.normalize2 = normalization(input_dim, num_classes) self.conv2 = ConvMeanPool(input_dim, output_dim, 3, adjust_padding=adjust_padding) conv_shortcut = partial(ConvMeanPool, kernel_size=1, adjust_padding=adjust_padding) elif resample is None: if dilation > 1: conv_shortcut = partial(ncsn_conv3x3, dilation=dilation) self.conv1 = ncsn_conv3x3(input_dim, output_dim, dilation=dilation) self.normalize2 = normalization(output_dim, num_classes) self.conv2 = ncsn_conv3x3(output_dim, output_dim, dilation=dilation) else: conv_shortcut = nn.Conv2d self.conv1 = ncsn_conv3x3(input_dim, output_dim) self.normalize2 = normalization(output_dim, num_classes) self.conv2 = ncsn_conv3x3(output_dim, output_dim) else: raise Exception('invalid resample value') if output_dim != input_dim or resample is not None: self.shortcut = conv_shortcut(input_dim, output_dim) self.normalize1 = normalization(input_dim, num_classes) def forward(self, x, y): output = self.normalize1(x, y) output = self.non_linearity(output) output = self.conv1(output) output = self.normalize2(output, y) output = self.non_linearity(output) output = self.conv2(output) if self.output_dim == self.input_dim and self.resample is None: shortcut = x else: shortcut = self.shortcut(x) return shortcut + output class ResidualBlock(nn.Module): def __init__(self, input_dim, output_dim, resample=None, act=nn.ELU(), normalization=nn.InstanceNorm2d, adjust_padding=False, dilation=1): super().__init__() self.non_linearity = act self.input_dim = input_dim self.output_dim = output_dim self.resample = resample self.normalization = normalization if resample == 'down': if dilation > 1: self.conv1 = ncsn_conv3x3(input_dim, input_dim, dilation=dilation) self.normalize2 = normalization(input_dim) self.conv2 = ncsn_conv3x3(input_dim, output_dim, dilation=dilation) conv_shortcut = partial(ncsn_conv3x3, dilation=dilation) else: self.conv1 = ncsn_conv3x3(input_dim, input_dim) self.normalize2 = normalization(input_dim) self.conv2 = ConvMeanPool(input_dim, output_dim, 3, adjust_padding=adjust_padding) conv_shortcut = partial(ConvMeanPool, kernel_size=1, adjust_padding=adjust_padding) elif resample is None: if dilation > 1: conv_shortcut = partial(ncsn_conv3x3, dilation=dilation) self.conv1 = ncsn_conv3x3(input_dim, output_dim, dilation=dilation) self.normalize2 = normalization(output_dim) self.conv2 = ncsn_conv3x3(output_dim, output_dim, dilation=dilation) else: # conv_shortcut = nn.Conv2d ### Something wierd here. conv_shortcut = partial(ncsn_conv1x1) self.conv1 = ncsn_conv3x3(input_dim, output_dim) self.normalize2 = normalization(output_dim) self.conv2 = ncsn_conv3x3(output_dim, output_dim) else: raise Exception('invalid resample value') if output_dim != input_dim or resample is not None: self.shortcut = conv_shortcut(input_dim, output_dim) self.normalize1 = normalization(input_dim) def forward(self, x): output = self.normalize1(x) output = self.non_linearity(output) output = self.conv1(output) output = self.normalize2(output) output = self.non_linearity(output) output = self.conv2(output) if self.output_dim == self.input_dim and self.resample is None: shortcut = x else: shortcut = self.shortcut(x) return shortcut + output ########################################################################### # Functions below are ported over from the DDPM codebase: # https://github.com/hojonathanho/diffusion/blob/master/diffusion_tf/nn.py ########################################################################### def get_timestep_embedding(timesteps, embedding_dim, max_positions=10000): assert len(timesteps.shape) == 1 # and timesteps.dtype == tf.int32 half_dim = embedding_dim // 2 # magic number 10000 is from transformers emb = math.log(max_positions) / (half_dim - 1) # emb = math.log(2.) / (half_dim - 1) emb = torch.exp(torch.arange(half_dim, dtype=torch.float32, device=timesteps.device) * -emb) # emb = tf.range(num_embeddings, dtype=jnp.float32)[:, None] * emb[None, :] # emb = tf.cast(timesteps, dtype=jnp.float32)[:, None] * emb[None, :] emb = timesteps.float()[:, None] * emb[None, :] emb = torch.cat([torch.sin(emb), torch.cos(emb)], dim=1) if embedding_dim % 2 == 1: # zero pad emb = F.pad(emb, (0, 1), mode='constant') assert emb.shape == (timesteps.shape[0], embedding_dim) return emb def _einsum(a, b, c, x, y): einsum_str = '{},{}->{}'.format(''.join(a), ''.join(b), ''.join(c)) return torch.einsum(einsum_str, x, y) def contract_inner(x, y): """tensordot(x, y, 1).""" x_chars = list(string.ascii_lowercase[:len(x.shape)]) y_chars = list(string.ascii_lowercase[len(x.shape):len(y.shape) + len(x.shape)]) y_chars[0] = x_chars[-1] # first axis of y and last of x get summed out_chars = x_chars[:-1] + y_chars[1:] return _einsum(x_chars, y_chars, out_chars, x, y) class NIN(nn.Module): def __init__(self, in_dim, num_units, init_scale=0.1): super().__init__() self.W = nn.Parameter(default_init(scale=init_scale)((in_dim, num_units)), requires_grad=True) self.b = nn.Parameter(torch.zeros(num_units), requires_grad=True) def forward(self, x): x = x.permute(0, 2, 3, 1) y = contract_inner(x, self.W) + self.b return y.permute(0, 3, 1, 2) class AttnBlock(nn.Module): """Channel-wise self-attention block.""" def __init__(self, channels): super().__init__() self.GroupNorm_0 = nn.GroupNorm(num_groups=32, num_channels=channels, eps=1e-6) self.NIN_0 = NIN(channels, channels) self.NIN_1 = NIN(channels, channels) self.NIN_2 = NIN(channels, channels) self.NIN_3 = NIN(channels, channels, init_scale=0.) def forward(self, x): B, C, H, W = x.shape h = self.GroupNorm_0(x) q = self.NIN_0(h) k = self.NIN_1(h) v = self.NIN_2(h) w = torch.einsum('bchw,bcij->bhwij', q, k) * (int(C) ** (-0.5)) w = torch.reshape(w, (B, H, W, H * W)) w = F.softmax(w, dim=-1) w = torch.reshape(w, (B, H, W, H, W)) h = torch.einsum('bhwij,bcij->bchw', w, v) h = self.NIN_3(h) return x + h class Upsample(nn.Module): def __init__(self, channels, with_conv=False): super().__init__() if with_conv: self.Conv_0 = ddpm_conv3x3(channels, channels) self.with_conv = with_conv def forward(self, x): B, C, H, W = x.shape h = F.interpolate(x, (H * 2, W * 2), mode='nearest') if self.with_conv: h = self.Conv_0(h) return h class Downsample(nn.Module): def __init__(self, channels, with_conv=False): super().__init__() if with_conv: self.Conv_0 = ddpm_conv3x3(channels, channels, stride=2, padding=0) self.with_conv = with_conv def forward(self, x): B, C, H, W = x.shape # Emulate 'SAME' padding if self.with_conv: x = F.pad(x, (0, 1, 0, 1)) x = self.Conv_0(x) else: x = F.avg_pool2d(x, kernel_size=2, stride=2, padding=0) assert x.shape == (B, C, H // 2, W // 2) return x class ResnetBlockDDPM(nn.Module): """The ResNet Blocks used in DDPM.""" def __init__(self, act, in_ch, out_ch=None, temb_dim=None, conv_shortcut=False, dropout=0.1): super().__init__() if out_ch is None: out_ch = in_ch self.GroupNorm_0 = nn.GroupNorm(num_groups=32, num_channels=in_ch, eps=1e-6) self.act = act self.Conv_0 = ddpm_conv3x3(in_ch, out_ch) if temb_dim is not None: self.Dense_0 = nn.Linear(temb_dim, out_ch) self.Dense_0.weight.data = default_init()(self.Dense_0.weight.data.shape) nn.init.zeros_(self.Dense_0.bias) self.GroupNorm_1 = nn.GroupNorm(num_groups=32, num_channels=out_ch, eps=1e-6) self.Dropout_0 = nn.Dropout(dropout) self.Conv_1 = ddpm_conv3x3(out_ch, out_ch, init_scale=0.) if in_ch != out_ch: if conv_shortcut: self.Conv_2 = ddpm_conv3x3(in_ch, out_ch) else: self.NIN_0 = NIN(in_ch, out_ch) self.out_ch = out_ch self.in_ch = in_ch self.conv_shortcut = conv_shortcut def forward(self, x, temb=None): B, C, H, W = x.shape assert C == self.in_ch out_ch = self.out_ch if self.out_ch else self.in_ch h = self.act(self.GroupNorm_0(x)) h = self.Conv_0(h) # Add bias to each feature map conditioned on the time embedding if temb is not None: h += self.Dense_0(self.act(temb))[:, :, None, None] h = self.act(self.GroupNorm_1(h)) h = self.Dropout_0(h) h = self.Conv_1(h) if C != out_ch: if self.conv_shortcut: x = self.Conv_2(x) else: x = self.NIN_0(x) return x + h
22,687
33.271903
112
py
DiffPure
DiffPure-master/score_sde/models/ddpm.py
# coding=utf-8 # Copyright 2020 The Google Research Authors. # # Licensed under the Apache License, Version 2.0 (the "License"); # you may not use this file except in compliance with the License. # You may obtain a copy of the License at # # http://www.apache.org/licenses/LICENSE-2.0 # # Unless required by applicable law or agreed to in writing, software # distributed under the License is distributed on an "AS IS" BASIS, # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. # pylint: skip-file """DDPM model. This code is the pytorch equivalent of: https://github.com/hojonathanho/diffusion/blob/master/diffusion_tf/models/unet.py """ import torch import torch.nn as nn import functools from . import utils, layers, normalization RefineBlock = layers.RefineBlock ResidualBlock = layers.ResidualBlock ResnetBlockDDPM = layers.ResnetBlockDDPM Upsample = layers.Upsample Downsample = layers.Downsample conv3x3 = layers.ddpm_conv3x3 get_act = layers.get_act get_normalization = normalization.get_normalization default_initializer = layers.default_init @utils.register_model(name='ddpm') class DDPM(nn.Module): def __init__(self, config): super().__init__() self.act = act = get_act(config) self.register_buffer('sigmas', torch.tensor(utils.get_sigmas(config))) self.nf = nf = config.model.nf ch_mult = config.model.ch_mult self.num_res_blocks = num_res_blocks = config.model.num_res_blocks self.attn_resolutions = attn_resolutions = config.model.attn_resolutions dropout = config.model.dropout resamp_with_conv = config.model.resamp_with_conv self.num_resolutions = num_resolutions = len(ch_mult) self.all_resolutions = all_resolutions = [config.data.image_size // (2 ** i) for i in range(num_resolutions)] AttnBlock = functools.partial(layers.AttnBlock) self.conditional = conditional = config.model.conditional ResnetBlock = functools.partial(ResnetBlockDDPM, act=act, temb_dim=4 * nf, dropout=dropout) if conditional: # Condition on noise levels. modules = [nn.Linear(nf, nf * 4)] modules[0].weight.data = default_initializer()(modules[0].weight.data.shape) nn.init.zeros_(modules[0].bias) modules.append(nn.Linear(nf * 4, nf * 4)) modules[1].weight.data = default_initializer()(modules[1].weight.data.shape) nn.init.zeros_(modules[1].bias) self.centered = config.data.centered channels = config.data.num_channels # Downsampling block modules.append(conv3x3(channels, nf)) hs_c = [nf] in_ch = nf for i_level in range(num_resolutions): # Residual blocks for this resolution for i_block in range(num_res_blocks): out_ch = nf * ch_mult[i_level] modules.append(ResnetBlock(in_ch=in_ch, out_ch=out_ch)) in_ch = out_ch if all_resolutions[i_level] in attn_resolutions: modules.append(AttnBlock(channels=in_ch)) hs_c.append(in_ch) if i_level != num_resolutions - 1: modules.append(Downsample(channels=in_ch, with_conv=resamp_with_conv)) hs_c.append(in_ch) in_ch = hs_c[-1] modules.append(ResnetBlock(in_ch=in_ch)) modules.append(AttnBlock(channels=in_ch)) modules.append(ResnetBlock(in_ch=in_ch)) # Upsampling block for i_level in reversed(range(num_resolutions)): for i_block in range(num_res_blocks + 1): out_ch = nf * ch_mult[i_level] modules.append(ResnetBlock(in_ch=in_ch + hs_c.pop(), out_ch=out_ch)) in_ch = out_ch if all_resolutions[i_level] in attn_resolutions: modules.append(AttnBlock(channels=in_ch)) if i_level != 0: modules.append(Upsample(channels=in_ch, with_conv=resamp_with_conv)) assert not hs_c modules.append(nn.GroupNorm(num_channels=in_ch, num_groups=32, eps=1e-6)) modules.append(conv3x3(in_ch, channels, init_scale=0.)) self.all_modules = nn.ModuleList(modules) self.scale_by_sigma = config.model.scale_by_sigma def forward(self, x, labels): modules = self.all_modules m_idx = 0 if self.conditional: # timestep/scale embedding timesteps = labels temb = layers.get_timestep_embedding(timesteps, self.nf) temb = modules[m_idx](temb) m_idx += 1 temb = modules[m_idx](self.act(temb)) m_idx += 1 else: temb = None if self.centered: # Input is in [-1, 1] h = x else: # Input is in [0, 1] h = 2 * x - 1. # Downsampling block hs = [modules[m_idx](h)] m_idx += 1 for i_level in range(self.num_resolutions): # Residual blocks for this resolution for i_block in range(self.num_res_blocks): h = modules[m_idx](hs[-1], temb) m_idx += 1 if h.shape[-1] in self.attn_resolutions: h = modules[m_idx](h) m_idx += 1 hs.append(h) if i_level != self.num_resolutions - 1: hs.append(modules[m_idx](hs[-1])) m_idx += 1 h = hs[-1] h = modules[m_idx](h, temb) m_idx += 1 h = modules[m_idx](h) m_idx += 1 h = modules[m_idx](h, temb) m_idx += 1 # Upsampling block for i_level in reversed(range(self.num_resolutions)): for i_block in range(self.num_res_blocks + 1): h = modules[m_idx](torch.cat([h, hs.pop()], dim=1), temb) m_idx += 1 if h.shape[-1] in self.attn_resolutions: h = modules[m_idx](h) m_idx += 1 if i_level != 0: h = modules[m_idx](h) m_idx += 1 assert not hs h = self.act(modules[m_idx](h)) m_idx += 1 h = modules[m_idx](h) m_idx += 1 assert m_idx == len(modules) if self.scale_by_sigma: # Divide the output by sigmas. Useful for training with the NCSN loss. # The DDPM loss scales the network output by sigma in the loss function, # so no need of doing it here. used_sigmas = self.sigmas[labels, None, None, None] h = h / used_sigmas return h
6,082
32.423077
113
py
DiffPure
DiffPure-master/score_sde/models/ncsnv2.py
# coding=utf-8 # Copyright 2020 The Google Research Authors. # # Licensed under the Apache License, Version 2.0 (the "License"); # you may not use this file except in compliance with the License. # You may obtain a copy of the License at # # http://www.apache.org/licenses/LICENSE-2.0 # # Unless required by applicable law or agreed to in writing, software # distributed under the License is distributed on an "AS IS" BASIS, # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. # pylint: skip-file """The NCSNv2 model.""" import torch import torch.nn as nn import functools from .utils import get_sigmas, register_model from .layers import (CondRefineBlock, RefineBlock, ResidualBlock, ncsn_conv3x3, ConditionalResidualBlock, get_act) from .normalization import get_normalization CondResidualBlock = ConditionalResidualBlock conv3x3 = ncsn_conv3x3 def get_network(config): if config.data.image_size < 96: return functools.partial(NCSNv2, config=config) elif 96 <= config.data.image_size <= 128: return functools.partial(NCSNv2_128, config=config) elif 128 < config.data.image_size <= 256: return functools.partial(NCSNv2_256, config=config) else: raise NotImplementedError( f'No network suitable for {config.data.image_size}px implemented yet.') @register_model(name='ncsnv2_64') class NCSNv2(nn.Module): def __init__(self, config): super().__init__() self.centered = config.data.centered self.norm = get_normalization(config) self.nf = nf = config.model.nf self.act = act = get_act(config) self.register_buffer('sigmas', torch.tensor(get_sigmas(config))) self.config = config self.begin_conv = nn.Conv2d(config.data.channels, nf, 3, stride=1, padding=1) self.normalizer = self.norm(nf, config.model.num_scales) self.end_conv = nn.Conv2d(nf, config.data.channels, 3, stride=1, padding=1) self.res1 = nn.ModuleList([ ResidualBlock(self.nf, self.nf, resample=None, act=act, normalization=self.norm), ResidualBlock(self.nf, self.nf, resample=None, act=act, normalization=self.norm)] ) self.res2 = nn.ModuleList([ ResidualBlock(self.nf, 2 * self.nf, resample='down', act=act, normalization=self.norm), ResidualBlock(2 * self.nf, 2 * self.nf, resample=None, act=act, normalization=self.norm)] ) self.res3 = nn.ModuleList([ ResidualBlock(2 * self.nf, 2 * self.nf, resample='down', act=act, normalization=self.norm, dilation=2), ResidualBlock(2 * self.nf, 2 * self.nf, resample=None, act=act, normalization=self.norm, dilation=2)] ) if config.data.image_size == 28: self.res4 = nn.ModuleList([ ResidualBlock(2 * self.nf, 2 * self.nf, resample='down', act=act, normalization=self.norm, adjust_padding=True, dilation=4), ResidualBlock(2 * self.nf, 2 * self.nf, resample=None, act=act, normalization=self.norm, dilation=4)] ) else: self.res4 = nn.ModuleList([ ResidualBlock(2 * self.nf, 2 * self.nf, resample='down', act=act, normalization=self.norm, adjust_padding=False, dilation=4), ResidualBlock(2 * self.nf, 2 * self.nf, resample=None, act=act, normalization=self.norm, dilation=4)] ) self.refine1 = RefineBlock([2 * self.nf], 2 * self.nf, act=act, start=True) self.refine2 = RefineBlock([2 * self.nf, 2 * self.nf], 2 * self.nf, act=act) self.refine3 = RefineBlock([2 * self.nf, 2 * self.nf], self.nf, act=act) self.refine4 = RefineBlock([self.nf, self.nf], self.nf, act=act, end=True) def _compute_cond_module(self, module, x): for m in module: x = m(x) return x def forward(self, x, y): if not self.centered: h = 2 * x - 1. else: h = x output = self.begin_conv(h) layer1 = self._compute_cond_module(self.res1, output) layer2 = self._compute_cond_module(self.res2, layer1) layer3 = self._compute_cond_module(self.res3, layer2) layer4 = self._compute_cond_module(self.res4, layer3) ref1 = self.refine1([layer4], layer4.shape[2:]) ref2 = self.refine2([layer3, ref1], layer3.shape[2:]) ref3 = self.refine3([layer2, ref2], layer2.shape[2:]) output = self.refine4([layer1, ref3], layer1.shape[2:]) output = self.normalizer(output) output = self.act(output) output = self.end_conv(output) used_sigmas = self.sigmas[y].view(x.shape[0], *([1] * len(x.shape[1:]))) output = output / used_sigmas return output @register_model(name='ncsn') class NCSN(nn.Module): def __init__(self, config): super().__init__() self.centered = config.data.centered self.norm = get_normalization(config) self.nf = nf = config.model.nf self.act = act = get_act(config) self.config = config self.begin_conv = nn.Conv2d(config.data.channels, nf, 3, stride=1, padding=1) self.normalizer = self.norm(nf, config.model.num_scales) self.end_conv = nn.Conv2d(nf, config.data.channels, 3, stride=1, padding=1) self.res1 = nn.ModuleList([ ConditionalResidualBlock(self.nf, self.nf, config.model.num_scales, resample=None, act=act, normalization=self.norm), ConditionalResidualBlock(self.nf, self.nf, config.model.num_scales, resample=None, act=act, normalization=self.norm)] ) self.res2 = nn.ModuleList([ ConditionalResidualBlock(self.nf, 2 * self.nf, config.model.num_scales, resample='down', act=act, normalization=self.norm), ConditionalResidualBlock(2 * self.nf, 2 * self.nf, config.model.num_scales, resample=None, act=act, normalization=self.norm)] ) self.res3 = nn.ModuleList([ ConditionalResidualBlock(2 * self.nf, 2 * self.nf, config.model.num_scales, resample='down', act=act, normalization=self.norm, dilation=2), ConditionalResidualBlock(2 * self.nf, 2 * self.nf, config.model.num_scales, resample=None, act=act, normalization=self.norm, dilation=2)] ) if config.data.image_size == 28: self.res4 = nn.ModuleList([ ConditionalResidualBlock(2 * self.nf, 2 * self.nf, config.model.num_scales, resample='down', act=act, normalization=self.norm, adjust_padding=True, dilation=4), ConditionalResidualBlock(2 * self.nf, 2 * self.nf, config.model.num_scales, resample=None, act=act, normalization=self.norm, dilation=4)] ) else: self.res4 = nn.ModuleList([ ConditionalResidualBlock(2 * self.nf, 2 * self.nf, config.model.num_scales, resample='down', act=act, normalization=self.norm, adjust_padding=False, dilation=4), ConditionalResidualBlock(2 * self.nf, 2 * self.nf, config.model.num_scales, resample=None, act=act, normalization=self.norm, dilation=4)] ) self.refine1 = CondRefineBlock([2 * self.nf], 2 * self.nf, config.model.num_scales, self.norm, act=act, start=True) self.refine2 = CondRefineBlock([2 * self.nf, 2 * self.nf], 2 * self.nf, config.model.num_scales, self.norm, act=act) self.refine3 = CondRefineBlock([2 * self.nf, 2 * self.nf], self.nf, config.model.num_scales, self.norm, act=act) self.refine4 = CondRefineBlock([self.nf, self.nf], self.nf, config.model.num_scales, self.norm, act=act, end=True) def _compute_cond_module(self, module, x, y): for m in module: x = m(x, y) return x def forward(self, x, y): if not self.centered: h = 2 * x - 1. else: h = x output = self.begin_conv(h) layer1 = self._compute_cond_module(self.res1, output, y) layer2 = self._compute_cond_module(self.res2, layer1, y) layer3 = self._compute_cond_module(self.res3, layer2, y) layer4 = self._compute_cond_module(self.res4, layer3, y) ref1 = self.refine1([layer4], y, layer4.shape[2:]) ref2 = self.refine2([layer3, ref1], y, layer3.shape[2:]) ref3 = self.refine3([layer2, ref2], y, layer2.shape[2:]) output = self.refine4([layer1, ref3], y, layer1.shape[2:]) output = self.normalizer(output, y) output = self.act(output) output = self.end_conv(output) return output @register_model(name='ncsnv2_128') class NCSNv2_128(nn.Module): """NCSNv2 model architecture for 128px images.""" def __init__(self, config): super().__init__() self.centered = config.data.centered self.norm = get_normalization(config) self.nf = nf = config.model.nf self.act = act = get_act(config) self.register_buffer('sigmas', torch.tensor(get_sigmas(config))) self.config = config self.begin_conv = nn.Conv2d(config.data.channels, nf, 3, stride=1, padding=1) self.normalizer = self.norm(nf, config.model.num_scales) self.end_conv = nn.Conv2d(nf, config.data.channels, 3, stride=1, padding=1) self.res1 = nn.ModuleList([ ResidualBlock(self.nf, self.nf, resample=None, act=act, normalization=self.norm), ResidualBlock(self.nf, self.nf, resample=None, act=act, normalization=self.norm)] ) self.res2 = nn.ModuleList([ ResidualBlock(self.nf, 2 * self.nf, resample='down', act=act, normalization=self.norm), ResidualBlock(2 * self.nf, 2 * self.nf, resample=None, act=act, normalization=self.norm)] ) self.res3 = nn.ModuleList([ ResidualBlock(2 * self.nf, 2 * self.nf, resample='down', act=act, normalization=self.norm), ResidualBlock(2 * self.nf, 2 * self.nf, resample=None, act=act, normalization=self.norm)] ) self.res4 = nn.ModuleList([ ResidualBlock(2 * self.nf, 4 * self.nf, resample='down', act=act, normalization=self.norm, dilation=2), ResidualBlock(4 * self.nf, 4 * self.nf, resample=None, act=act, normalization=self.norm, dilation=2)] ) self.res5 = nn.ModuleList([ ResidualBlock(4 * self.nf, 4 * self.nf, resample='down', act=act, normalization=self.norm, dilation=4), ResidualBlock(4 * self.nf, 4 * self.nf, resample=None, act=act, normalization=self.norm, dilation=4)] ) self.refine1 = RefineBlock([4 * self.nf], 4 * self.nf, act=act, start=True) self.refine2 = RefineBlock([4 * self.nf, 4 * self.nf], 2 * self.nf, act=act) self.refine3 = RefineBlock([2 * self.nf, 2 * self.nf], 2 * self.nf, act=act) self.refine4 = RefineBlock([2 * self.nf, 2 * self.nf], self.nf, act=act) self.refine5 = RefineBlock([self.nf, self.nf], self.nf, act=act, end=True) def _compute_cond_module(self, module, x): for m in module: x = m(x) return x def forward(self, x, y): if not self.centered: h = 2 * x - 1. else: h = x output = self.begin_conv(h) layer1 = self._compute_cond_module(self.res1, output) layer2 = self._compute_cond_module(self.res2, layer1) layer3 = self._compute_cond_module(self.res3, layer2) layer4 = self._compute_cond_module(self.res4, layer3) layer5 = self._compute_cond_module(self.res5, layer4) ref1 = self.refine1([layer5], layer5.shape[2:]) ref2 = self.refine2([layer4, ref1], layer4.shape[2:]) ref3 = self.refine3([layer3, ref2], layer3.shape[2:]) ref4 = self.refine4([layer2, ref3], layer2.shape[2:]) output = self.refine5([layer1, ref4], layer1.shape[2:]) output = self.normalizer(output) output = self.act(output) output = self.end_conv(output) used_sigmas = self.sigmas[y].view(x.shape[0], *([1] * len(x.shape[1:]))) output = output / used_sigmas return output @register_model(name='ncsnv2_256') class NCSNv2_256(nn.Module): """NCSNv2 model architecture for 256px images.""" def __init__(self, config): super().__init__() self.centered = config.data.centered self.norm = get_normalization(config) self.nf = nf = config.model.nf self.act = act = get_act(config) self.register_buffer('sigmas', torch.tensor(get_sigmas(config))) self.config = config self.begin_conv = nn.Conv2d(config.data.channels, nf, 3, stride=1, padding=1) self.normalizer = self.norm(nf, config.model.num_scales) self.end_conv = nn.Conv2d(nf, config.data.channels, 3, stride=1, padding=1) self.res1 = nn.ModuleList([ ResidualBlock(self.nf, self.nf, resample=None, act=act, normalization=self.norm), ResidualBlock(self.nf, self.nf, resample=None, act=act, normalization=self.norm)] ) self.res2 = nn.ModuleList([ ResidualBlock(self.nf, 2 * self.nf, resample='down', act=act, normalization=self.norm), ResidualBlock(2 * self.nf, 2 * self.nf, resample=None, act=act, normalization=self.norm)] ) self.res3 = nn.ModuleList([ ResidualBlock(2 * self.nf, 2 * self.nf, resample='down', act=act, normalization=self.norm), ResidualBlock(2 * self.nf, 2 * self.nf, resample=None, act=act, normalization=self.norm)] ) self.res31 = nn.ModuleList([ ResidualBlock(2 * self.nf, 2 * self.nf, resample='down', act=act, normalization=self.norm), ResidualBlock(2 * self.nf, 2 * self.nf, resample=None, act=act, normalization=self.norm)] ) self.res4 = nn.ModuleList([ ResidualBlock(2 * self.nf, 4 * self.nf, resample='down', act=act, normalization=self.norm, dilation=2), ResidualBlock(4 * self.nf, 4 * self.nf, resample=None, act=act, normalization=self.norm, dilation=2)] ) self.res5 = nn.ModuleList([ ResidualBlock(4 * self.nf, 4 * self.nf, resample='down', act=act, normalization=self.norm, dilation=4), ResidualBlock(4 * self.nf, 4 * self.nf, resample=None, act=act, normalization=self.norm, dilation=4)] ) self.refine1 = RefineBlock([4 * self.nf], 4 * self.nf, act=act, start=True) self.refine2 = RefineBlock([4 * self.nf, 4 * self.nf], 2 * self.nf, act=act) self.refine3 = RefineBlock([2 * self.nf, 2 * self.nf], 2 * self.nf, act=act) self.refine31 = RefineBlock([2 * self.nf, 2 * self.nf], 2 * self.nf, act=act) self.refine4 = RefineBlock([2 * self.nf, 2 * self.nf], self.nf, act=act) self.refine5 = RefineBlock([self.nf, self.nf], self.nf, act=act, end=True) def _compute_cond_module(self, module, x): for m in module: x = m(x) return x def forward(self, x, y): if not self.centered: h = 2 * x - 1. else: h = x output = self.begin_conv(h) layer1 = self._compute_cond_module(self.res1, output) layer2 = self._compute_cond_module(self.res2, layer1) layer3 = self._compute_cond_module(self.res3, layer2) layer31 = self._compute_cond_module(self.res31, layer3) layer4 = self._compute_cond_module(self.res4, layer31) layer5 = self._compute_cond_module(self.res5, layer4) ref1 = self.refine1([layer5], layer5.shape[2:]) ref2 = self.refine2([layer4, ref1], layer4.shape[2:]) ref31 = self.refine31([layer31, ref2], layer31.shape[2:]) ref3 = self.refine3([layer3, ref31], layer3.shape[2:]) ref4 = self.refine4([layer2, ref3], layer2.shape[2:]) output = self.refine5([layer1, ref4], layer1.shape[2:]) output = self.normalizer(output) output = self.act(output) output = self.end_conv(output) used_sigmas = self.sigmas[y].view(x.shape[0], *([1] * len(x.shape[1:]))) output = output / used_sigmas return output
16,043
37.567308
120
py
DiffPure
DiffPure-master/score_sde/models/normalization.py
# coding=utf-8 # Copyright 2020 The Google Research Authors. # # Licensed under the Apache License, Version 2.0 (the "License"); # you may not use this file except in compliance with the License. # You may obtain a copy of the License at # # http://www.apache.org/licenses/LICENSE-2.0 # # Unless required by applicable law or agreed to in writing, software # distributed under the License is distributed on an "AS IS" BASIS, # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. """Normalization layers.""" import torch.nn as nn import torch import functools def get_normalization(config, conditional=False): """Obtain normalization modules from the config file.""" norm = config.model.normalization if conditional: if norm == 'InstanceNorm++': return functools.partial(ConditionalInstanceNorm2dPlus, num_classes=config.model.num_classes) else: raise NotImplementedError(f'{norm} not implemented yet.') else: if norm == 'InstanceNorm': return nn.InstanceNorm2d elif norm == 'InstanceNorm++': return InstanceNorm2dPlus elif norm == 'VarianceNorm': return VarianceNorm2d elif norm == 'GroupNorm': return nn.GroupNorm else: raise ValueError('Unknown normalization: %s' % norm) class ConditionalBatchNorm2d(nn.Module): def __init__(self, num_features, num_classes, bias=True): super().__init__() self.num_features = num_features self.bias = bias self.bn = nn.BatchNorm2d(num_features, affine=False) if self.bias: self.embed = nn.Embedding(num_classes, num_features * 2) self.embed.weight.data[:, :num_features].uniform_() # Initialise scale at N(1, 0.02) self.embed.weight.data[:, num_features:].zero_() # Initialise bias at 0 else: self.embed = nn.Embedding(num_classes, num_features) self.embed.weight.data.uniform_() def forward(self, x, y): out = self.bn(x) if self.bias: gamma, beta = self.embed(y).chunk(2, dim=1) out = gamma.view(-1, self.num_features, 1, 1) * out + beta.view(-1, self.num_features, 1, 1) else: gamma = self.embed(y) out = gamma.view(-1, self.num_features, 1, 1) * out return out class ConditionalInstanceNorm2d(nn.Module): def __init__(self, num_features, num_classes, bias=True): super().__init__() self.num_features = num_features self.bias = bias self.instance_norm = nn.InstanceNorm2d(num_features, affine=False, track_running_stats=False) if bias: self.embed = nn.Embedding(num_classes, num_features * 2) self.embed.weight.data[:, :num_features].uniform_() # Initialise scale at N(1, 0.02) self.embed.weight.data[:, num_features:].zero_() # Initialise bias at 0 else: self.embed = nn.Embedding(num_classes, num_features) self.embed.weight.data.uniform_() def forward(self, x, y): h = self.instance_norm(x) if self.bias: gamma, beta = self.embed(y).chunk(2, dim=-1) out = gamma.view(-1, self.num_features, 1, 1) * h + beta.view(-1, self.num_features, 1, 1) else: gamma = self.embed(y) out = gamma.view(-1, self.num_features, 1, 1) * h return out class ConditionalVarianceNorm2d(nn.Module): def __init__(self, num_features, num_classes, bias=False): super().__init__() self.num_features = num_features self.bias = bias self.embed = nn.Embedding(num_classes, num_features) self.embed.weight.data.normal_(1, 0.02) def forward(self, x, y): vars = torch.var(x, dim=(2, 3), keepdim=True) h = x / torch.sqrt(vars + 1e-5) gamma = self.embed(y) out = gamma.view(-1, self.num_features, 1, 1) * h return out class VarianceNorm2d(nn.Module): def __init__(self, num_features, bias=False): super().__init__() self.num_features = num_features self.bias = bias self.alpha = nn.Parameter(torch.zeros(num_features)) self.alpha.data.normal_(1, 0.02) def forward(self, x): vars = torch.var(x, dim=(2, 3), keepdim=True) h = x / torch.sqrt(vars + 1e-5) out = self.alpha.view(-1, self.num_features, 1, 1) * h return out class ConditionalNoneNorm2d(nn.Module): def __init__(self, num_features, num_classes, bias=True): super().__init__() self.num_features = num_features self.bias = bias if bias: self.embed = nn.Embedding(num_classes, num_features * 2) self.embed.weight.data[:, :num_features].uniform_() # Initialise scale at N(1, 0.02) self.embed.weight.data[:, num_features:].zero_() # Initialise bias at 0 else: self.embed = nn.Embedding(num_classes, num_features) self.embed.weight.data.uniform_() def forward(self, x, y): if self.bias: gamma, beta = self.embed(y).chunk(2, dim=-1) out = gamma.view(-1, self.num_features, 1, 1) * x + beta.view(-1, self.num_features, 1, 1) else: gamma = self.embed(y) out = gamma.view(-1, self.num_features, 1, 1) * x return out class NoneNorm2d(nn.Module): def __init__(self, num_features, bias=True): super().__init__() def forward(self, x): return x class InstanceNorm2dPlus(nn.Module): def __init__(self, num_features, bias=True): super().__init__() self.num_features = num_features self.bias = bias self.instance_norm = nn.InstanceNorm2d(num_features, affine=False, track_running_stats=False) self.alpha = nn.Parameter(torch.zeros(num_features)) self.gamma = nn.Parameter(torch.zeros(num_features)) self.alpha.data.normal_(1, 0.02) self.gamma.data.normal_(1, 0.02) if bias: self.beta = nn.Parameter(torch.zeros(num_features)) def forward(self, x): means = torch.mean(x, dim=(2, 3)) m = torch.mean(means, dim=-1, keepdim=True) v = torch.var(means, dim=-1, keepdim=True) means = (means - m) / (torch.sqrt(v + 1e-5)) h = self.instance_norm(x) if self.bias: h = h + means[..., None, None] * self.alpha[..., None, None] out = self.gamma.view(-1, self.num_features, 1, 1) * h + self.beta.view(-1, self.num_features, 1, 1) else: h = h + means[..., None, None] * self.alpha[..., None, None] out = self.gamma.view(-1, self.num_features, 1, 1) * h return out class ConditionalInstanceNorm2dPlus(nn.Module): def __init__(self, num_features, num_classes, bias=True): super().__init__() self.num_features = num_features self.bias = bias self.instance_norm = nn.InstanceNorm2d(num_features, affine=False, track_running_stats=False) if bias: self.embed = nn.Embedding(num_classes, num_features * 3) self.embed.weight.data[:, :2 * num_features].normal_(1, 0.02) # Initialise scale at N(1, 0.02) self.embed.weight.data[:, 2 * num_features:].zero_() # Initialise bias at 0 else: self.embed = nn.Embedding(num_classes, 2 * num_features) self.embed.weight.data.normal_(1, 0.02) def forward(self, x, y): means = torch.mean(x, dim=(2, 3)) m = torch.mean(means, dim=-1, keepdim=True) v = torch.var(means, dim=-1, keepdim=True) means = (means - m) / (torch.sqrt(v + 1e-5)) h = self.instance_norm(x) if self.bias: gamma, alpha, beta = self.embed(y).chunk(3, dim=-1) h = h + means[..., None, None] * alpha[..., None, None] out = gamma.view(-1, self.num_features, 1, 1) * h + beta.view(-1, self.num_features, 1, 1) else: gamma, alpha = self.embed(y).chunk(2, dim=-1) h = h + means[..., None, None] * alpha[..., None, None] out = gamma.view(-1, self.num_features, 1, 1) * h return out
7,657
34.453704
106
py
DiffPure
DiffPure-master/score_sde/models/ema.py
# --------------------------------------------------------------- # Taken from the following link as is from: # https://github.com/yang-song/score_sde_pytorch/blob/main/models/ema.py # # The license for the original version of this file can be # found in the `score_sde` directory (LICENSE_SCORE_SDE). # --------------------------------------------------------------- # Modified from https://raw.githubusercontent.com/fadel/pytorch_ema/master/torch_ema/ema.py from __future__ import division from __future__ import unicode_literals import torch # Partially based on: https://github.com/tensorflow/tensorflow/blob/r1.13/tensorflow/python/training/moving_averages.py class ExponentialMovingAverage: """ Maintains (exponential) moving average of a set of parameters. """ def __init__(self, parameters, decay, use_num_updates=True): """ Args: parameters: Iterable of `torch.nn.Parameter`; usually the result of `model.parameters()`. decay: The exponential decay. use_num_updates: Whether to use number of updates when computing averages. """ if decay < 0.0 or decay > 1.0: raise ValueError('Decay must be between 0 and 1') self.decay = decay self.num_updates = 0 if use_num_updates else None self.shadow_params = [p.clone().detach() for p in parameters if p.requires_grad] self.collected_params = [] def update(self, parameters): """ Update currently maintained parameters. Call this every time the parameters are updated, such as the result of the `optimizer.step()` call. Args: parameters: Iterable of `torch.nn.Parameter`; usually the same set of parameters used to initialize this object. """ decay = self.decay if self.num_updates is not None: self.num_updates += 1 decay = min(decay, (1 + self.num_updates) / (10 + self.num_updates)) one_minus_decay = 1.0 - decay with torch.no_grad(): parameters = [p for p in parameters if p.requires_grad] for s_param, param in zip(self.shadow_params, parameters): s_param.sub_(one_minus_decay * (s_param - param)) def copy_to(self, parameters): """ Copy current parameters into given collection of parameters. Args: parameters: Iterable of `torch.nn.Parameter`; the parameters to be updated with the stored moving averages. """ parameters = [p for p in parameters if p.requires_grad] for s_param, param in zip(self.shadow_params, parameters): if param.requires_grad: param.data.copy_(s_param.data) def store(self, parameters): """ Save the current parameters for restoring later. Args: parameters: Iterable of `torch.nn.Parameter`; the parameters to be temporarily stored. """ self.collected_params = [param.clone() for param in parameters] def restore(self, parameters): """ Restore the parameters stored with the `store` method. Useful to validate the model with EMA parameters without affecting the original optimization process. Store the parameters before the `copy_to` method. After validation (or model saving), use this to restore the former parameters. Args: parameters: Iterable of `torch.nn.Parameter`; the parameters to be updated with the stored parameters. """ for c_param, param in zip(self.collected_params, parameters): param.data.copy_(c_param.data) def state_dict(self): return dict(decay=self.decay, num_updates=self.num_updates, shadow_params=self.shadow_params) def load_state_dict(self, state_dict): self.decay = state_dict['decay'] self.num_updates = state_dict['num_updates'] self.shadow_params = state_dict['shadow_params']
3,783
34.698113
119
py
DiffPure
DiffPure-master/score_sde/models/ncsnpp.py
# coding=utf-8 # Copyright 2020 The Google Research Authors. # # Licensed under the Apache License, Version 2.0 (the "License"); # you may not use this file except in compliance with the License. # You may obtain a copy of the License at # # http://www.apache.org/licenses/LICENSE-2.0 # # Unless required by applicable law or agreed to in writing, software # distributed under the License is distributed on an "AS IS" BASIS, # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. # pylint: skip-file from . import utils, layers, layerspp, normalization import torch.nn as nn import functools import torch import numpy as np ResnetBlockDDPM = layerspp.ResnetBlockDDPMpp ResnetBlockBigGAN = layerspp.ResnetBlockBigGANpp Combine = layerspp.Combine conv3x3 = layerspp.conv3x3 conv1x1 = layerspp.conv1x1 get_act = layers.get_act get_normalization = normalization.get_normalization default_initializer = layers.default_init @utils.register_model(name='ncsnpp') class NCSNpp(nn.Module): """NCSN++ model""" def __init__(self, config): super().__init__() self.config = config self.act = act = get_act(config) self.register_buffer('sigmas', torch.tensor(utils.get_sigmas(config))) self.nf = nf = config.model.nf ch_mult = config.model.ch_mult self.num_res_blocks = num_res_blocks = config.model.num_res_blocks self.attn_resolutions = attn_resolutions = config.model.attn_resolutions dropout = config.model.dropout resamp_with_conv = config.model.resamp_with_conv self.num_resolutions = num_resolutions = len(ch_mult) self.all_resolutions = all_resolutions = [config.data.image_size // (2 ** i) for i in range(num_resolutions)] self.conditional = conditional = config.model.conditional # noise-conditional fir = config.model.fir fir_kernel = config.model.fir_kernel self.skip_rescale = skip_rescale = config.model.skip_rescale self.resblock_type = resblock_type = config.model.resblock_type.lower() self.progressive = progressive = config.model.progressive.lower() self.progressive_input = progressive_input = config.model.progressive_input.lower() self.embedding_type = embedding_type = config.model.embedding_type.lower() init_scale = config.model.init_scale assert progressive in ['none', 'output_skip', 'residual'] assert progressive_input in ['none', 'input_skip', 'residual'] assert embedding_type in ['fourier', 'positional'] combine_method = config.model.progressive_combine.lower() combiner = functools.partial(Combine, method=combine_method) modules = [] # timestep/noise_level embedding; only for continuous training if embedding_type == 'fourier': # Gaussian Fourier features embeddings. assert config.training.continuous, "Fourier features are only used for continuous training." modules.append(layerspp.GaussianFourierProjection( embedding_size=nf, scale=config.model.fourier_scale )) embed_dim = 2 * nf elif embedding_type == 'positional': embed_dim = nf else: raise ValueError(f'embedding type {embedding_type} unknown.') if conditional: modules.append(nn.Linear(embed_dim, nf * 4)) modules[-1].weight.data = default_initializer()(modules[-1].weight.shape) nn.init.zeros_(modules[-1].bias) modules.append(nn.Linear(nf * 4, nf * 4)) modules[-1].weight.data = default_initializer()(modules[-1].weight.shape) nn.init.zeros_(modules[-1].bias) AttnBlock = functools.partial(layerspp.AttnBlockpp, init_scale=init_scale, skip_rescale=skip_rescale) Upsample = functools.partial(layerspp.Upsample, with_conv=resamp_with_conv, fir=fir, fir_kernel=fir_kernel) if progressive == 'output_skip': self.pyramid_upsample = layerspp.Upsample(fir=fir, fir_kernel=fir_kernel, with_conv=False) elif progressive == 'residual': pyramid_upsample = functools.partial(layerspp.Upsample, fir=fir, fir_kernel=fir_kernel, with_conv=True) Downsample = functools.partial(layerspp.Downsample, with_conv=resamp_with_conv, fir=fir, fir_kernel=fir_kernel) if progressive_input == 'input_skip': self.pyramid_downsample = layerspp.Downsample(fir=fir, fir_kernel=fir_kernel, with_conv=False) elif progressive_input == 'residual': pyramid_downsample = functools.partial(layerspp.Downsample, fir=fir, fir_kernel=fir_kernel, with_conv=True) if resblock_type == 'ddpm': ResnetBlock = functools.partial(ResnetBlockDDPM, act=act, dropout=dropout, init_scale=init_scale, skip_rescale=skip_rescale, temb_dim=nf * 4) elif resblock_type == 'biggan': ResnetBlock = functools.partial(ResnetBlockBigGAN, act=act, dropout=dropout, fir=fir, fir_kernel=fir_kernel, init_scale=init_scale, skip_rescale=skip_rescale, temb_dim=nf * 4) else: raise ValueError(f'resblock type {resblock_type} unrecognized.') # Downsampling block channels = config.data.num_channels if progressive_input != 'none': input_pyramid_ch = channels modules.append(conv3x3(channels, nf)) hs_c = [nf] in_ch = nf for i_level in range(num_resolutions): # Residual blocks for this resolution for i_block in range(num_res_blocks): out_ch = nf * ch_mult[i_level] modules.append(ResnetBlock(in_ch=in_ch, out_ch=out_ch)) in_ch = out_ch if all_resolutions[i_level] in attn_resolutions: modules.append(AttnBlock(channels=in_ch)) hs_c.append(in_ch) if i_level != num_resolutions - 1: if resblock_type == 'ddpm': modules.append(Downsample(in_ch=in_ch)) else: modules.append(ResnetBlock(down=True, in_ch=in_ch)) if progressive_input == 'input_skip': modules.append(combiner(dim1=input_pyramid_ch, dim2=in_ch)) if combine_method == 'cat': in_ch *= 2 elif progressive_input == 'residual': modules.append(pyramid_downsample(in_ch=input_pyramid_ch, out_ch=in_ch)) input_pyramid_ch = in_ch hs_c.append(in_ch) in_ch = hs_c[-1] modules.append(ResnetBlock(in_ch=in_ch)) modules.append(AttnBlock(channels=in_ch)) modules.append(ResnetBlock(in_ch=in_ch)) pyramid_ch = 0 # Upsampling block for i_level in reversed(range(num_resolutions)): for i_block in range(num_res_blocks + 1): out_ch = nf * ch_mult[i_level] modules.append(ResnetBlock(in_ch=in_ch + hs_c.pop(), out_ch=out_ch)) in_ch = out_ch if all_resolutions[i_level] in attn_resolutions: modules.append(AttnBlock(channels=in_ch)) if progressive != 'none': if i_level == num_resolutions - 1: if progressive == 'output_skip': modules.append(nn.GroupNorm(num_groups=min(in_ch // 4, 32), num_channels=in_ch, eps=1e-6)) modules.append(conv3x3(in_ch, channels, init_scale=init_scale)) pyramid_ch = channels elif progressive == 'residual': modules.append(nn.GroupNorm(num_groups=min(in_ch // 4, 32), num_channels=in_ch, eps=1e-6)) modules.append(conv3x3(in_ch, in_ch, bias=True)) pyramid_ch = in_ch else: raise ValueError(f'{progressive} is not a valid name.') else: if progressive == 'output_skip': modules.append(nn.GroupNorm(num_groups=min(in_ch // 4, 32), num_channels=in_ch, eps=1e-6)) modules.append(conv3x3(in_ch, channels, bias=True, init_scale=init_scale)) pyramid_ch = channels elif progressive == 'residual': modules.append(pyramid_upsample(in_ch=pyramid_ch, out_ch=in_ch)) pyramid_ch = in_ch else: raise ValueError(f'{progressive} is not a valid name') if i_level != 0: if resblock_type == 'ddpm': modules.append(Upsample(in_ch=in_ch)) else: modules.append(ResnetBlock(in_ch=in_ch, up=True)) assert not hs_c if progressive != 'output_skip': modules.append(nn.GroupNorm(num_groups=min(in_ch // 4, 32), num_channels=in_ch, eps=1e-6)) modules.append(conv3x3(in_ch, channels, init_scale=init_scale)) self.all_modules = nn.ModuleList(modules) def forward(self, x, time_cond): # timestep/noise_level embedding; only for continuous training modules = self.all_modules m_idx = 0 if self.embedding_type == 'fourier': # Gaussian Fourier features embeddings. used_sigmas = time_cond temb = modules[m_idx](torch.log(used_sigmas)) m_idx += 1 elif self.embedding_type == 'positional': # Sinusoidal positional embeddings. timesteps = time_cond used_sigmas = self.sigmas[time_cond.long()] temb = layers.get_timestep_embedding(timesteps, self.nf) else: raise ValueError(f'embedding type {self.embedding_type} unknown.') if self.conditional: temb = modules[m_idx](temb) m_idx += 1 temb = modules[m_idx](self.act(temb)) m_idx += 1 else: temb = None if not self.config.data.centered: # If input data is in [0, 1] x = 2 * x - 1. # Downsampling block input_pyramid = None if self.progressive_input != 'none': input_pyramid = x hs = [modules[m_idx](x)] m_idx += 1 for i_level in range(self.num_resolutions): # Residual blocks for this resolution for i_block in range(self.num_res_blocks): h = modules[m_idx](hs[-1], temb) m_idx += 1 if h.shape[-1] in self.attn_resolutions: h = modules[m_idx](h) m_idx += 1 hs.append(h) if i_level != self.num_resolutions - 1: if self.resblock_type == 'ddpm': h = modules[m_idx](hs[-1]) m_idx += 1 else: h = modules[m_idx](hs[-1], temb) m_idx += 1 if self.progressive_input == 'input_skip': input_pyramid = self.pyramid_downsample(input_pyramid) h = modules[m_idx](input_pyramid, h) m_idx += 1 elif self.progressive_input == 'residual': input_pyramid = modules[m_idx](input_pyramid) m_idx += 1 if self.skip_rescale: input_pyramid = (input_pyramid + h) / np.sqrt(2.) else: input_pyramid = input_pyramid + h h = input_pyramid hs.append(h) h = hs[-1] h = modules[m_idx](h, temb) m_idx += 1 h = modules[m_idx](h) m_idx += 1 h = modules[m_idx](h, temb) m_idx += 1 pyramid = None # Upsampling block for i_level in reversed(range(self.num_resolutions)): for i_block in range(self.num_res_blocks + 1): h = modules[m_idx](torch.cat([h, hs.pop()], dim=1), temb) m_idx += 1 if h.shape[-1] in self.attn_resolutions: h = modules[m_idx](h) m_idx += 1 if self.progressive != 'none': if i_level == self.num_resolutions - 1: if self.progressive == 'output_skip': pyramid = self.act(modules[m_idx](h)) m_idx += 1 pyramid = modules[m_idx](pyramid) m_idx += 1 elif self.progressive == 'residual': pyramid = self.act(modules[m_idx](h)) m_idx += 1 pyramid = modules[m_idx](pyramid) m_idx += 1 else: raise ValueError(f'{self.progressive} is not a valid name.') else: if self.progressive == 'output_skip': pyramid = self.pyramid_upsample(pyramid) pyramid_h = self.act(modules[m_idx](h)) m_idx += 1 pyramid_h = modules[m_idx](pyramid_h) m_idx += 1 pyramid = pyramid + pyramid_h elif self.progressive == 'residual': pyramid = modules[m_idx](pyramid) m_idx += 1 if self.skip_rescale: pyramid = (pyramid + h) / np.sqrt(2.) else: pyramid = pyramid + h h = pyramid else: raise ValueError(f'{self.progressive} is not a valid name') if i_level != 0: if self.resblock_type == 'ddpm': h = modules[m_idx](h) m_idx += 1 else: h = modules[m_idx](h, temb) m_idx += 1 assert not hs if self.progressive == 'output_skip': h = pyramid else: h = self.act(modules[m_idx](h)) m_idx += 1 h = modules[m_idx](h) m_idx += 1 assert m_idx == len(modules) if self.config.model.scale_by_sigma: used_sigmas = used_sigmas.reshape((x.shape[0], *([1] * len(x.shape[1:])))) h = h / used_sigmas return h
13,653
34.743455
113
py
DiffPure
DiffPure-master/score_sde/models/layerspp.py
# coding=utf-8 # Copyright 2020 The Google Research Authors. # # Licensed under the Apache License, Version 2.0 (the "License"); # you may not use this file except in compliance with the License. # You may obtain a copy of the License at # # http://www.apache.org/licenses/LICENSE-2.0 # # Unless required by applicable law or agreed to in writing, software # distributed under the License is distributed on an "AS IS" BASIS, # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. # pylint: skip-file """Layers for defining NCSN++. """ from . import layers from . import up_or_down_sampling import torch.nn as nn import torch import torch.nn.functional as F import numpy as np conv1x1 = layers.ddpm_conv1x1 conv3x3 = layers.ddpm_conv3x3 NIN = layers.NIN default_init = layers.default_init class GaussianFourierProjection(nn.Module): """Gaussian Fourier embeddings for noise levels.""" def __init__(self, embedding_size=256, scale=1.0): super().__init__() self.W = nn.Parameter(torch.randn(embedding_size) * scale, requires_grad=False) def forward(self, x): x_proj = x[:, None] * self.W[None, :] * 2 * np.pi return torch.cat([torch.sin(x_proj), torch.cos(x_proj)], dim=-1) class Combine(nn.Module): """Combine information from skip connections.""" def __init__(self, dim1, dim2, method='cat'): super().__init__() self.Conv_0 = conv1x1(dim1, dim2) self.method = method def forward(self, x, y): h = self.Conv_0(x) if self.method == 'cat': return torch.cat([h, y], dim=1) elif self.method == 'sum': return h + y else: raise ValueError(f'Method {self.method} not recognized.') class AttnBlockpp(nn.Module): """Channel-wise self-attention block. Modified from DDPM.""" def __init__(self, channels, skip_rescale=False, init_scale=0.): super().__init__() self.GroupNorm_0 = nn.GroupNorm(num_groups=min(channels // 4, 32), num_channels=channels, eps=1e-6) self.NIN_0 = NIN(channels, channels) self.NIN_1 = NIN(channels, channels) self.NIN_2 = NIN(channels, channels) self.NIN_3 = NIN(channels, channels, init_scale=init_scale) self.skip_rescale = skip_rescale def forward(self, x): B, C, H, W = x.shape h = self.GroupNorm_0(x) q = self.NIN_0(h) k = self.NIN_1(h) v = self.NIN_2(h) w = torch.einsum('bchw,bcij->bhwij', q, k) * (int(C) ** (-0.5)) w = torch.reshape(w, (B, H, W, H * W)) w = F.softmax(w, dim=-1) w = torch.reshape(w, (B, H, W, H, W)) h = torch.einsum('bhwij,bcij->bchw', w, v) h = self.NIN_3(h) if not self.skip_rescale: return x + h else: return (x + h) / np.sqrt(2.) class Upsample(nn.Module): def __init__(self, in_ch=None, out_ch=None, with_conv=False, fir=False, fir_kernel=(1, 3, 3, 1)): super().__init__() out_ch = out_ch if out_ch else in_ch if not fir: if with_conv: self.Conv_0 = conv3x3(in_ch, out_ch) else: if with_conv: self.Conv2d_0 = up_or_down_sampling.Conv2d(in_ch, out_ch, kernel=3, up=True, resample_kernel=fir_kernel, use_bias=True, kernel_init=default_init()) self.fir = fir self.with_conv = with_conv self.fir_kernel = fir_kernel self.out_ch = out_ch def forward(self, x): B, C, H, W = x.shape if not self.fir: h = F.interpolate(x, (H * 2, W * 2), 'nearest') if self.with_conv: h = self.Conv_0(h) else: if not self.with_conv: h = up_or_down_sampling.upsample_2d(x, self.fir_kernel, factor=2) else: h = self.Conv2d_0(x) return h class Downsample(nn.Module): def __init__(self, in_ch=None, out_ch=None, with_conv=False, fir=False, fir_kernel=(1, 3, 3, 1)): super().__init__() out_ch = out_ch if out_ch else in_ch if not fir: if with_conv: self.Conv_0 = conv3x3(in_ch, out_ch, stride=2, padding=0) else: if with_conv: self.Conv2d_0 = up_or_down_sampling.Conv2d(in_ch, out_ch, kernel=3, down=True, resample_kernel=fir_kernel, use_bias=True, kernel_init=default_init()) self.fir = fir self.fir_kernel = fir_kernel self.with_conv = with_conv self.out_ch = out_ch def forward(self, x): B, C, H, W = x.shape if not self.fir: if self.with_conv: x = F.pad(x, (0, 1, 0, 1)) x = self.Conv_0(x) else: x = F.avg_pool2d(x, 2, stride=2) else: if not self.with_conv: x = up_or_down_sampling.downsample_2d(x, self.fir_kernel, factor=2) else: x = self.Conv2d_0(x) return x class ResnetBlockDDPMpp(nn.Module): """ResBlock adapted from DDPM.""" def __init__(self, act, in_ch, out_ch=None, temb_dim=None, conv_shortcut=False, dropout=0.1, skip_rescale=False, init_scale=0.): super().__init__() out_ch = out_ch if out_ch else in_ch self.GroupNorm_0 = nn.GroupNorm(num_groups=min(in_ch // 4, 32), num_channels=in_ch, eps=1e-6) self.Conv_0 = conv3x3(in_ch, out_ch) if temb_dim is not None: self.Dense_0 = nn.Linear(temb_dim, out_ch) self.Dense_0.weight.data = default_init()(self.Dense_0.weight.data.shape) nn.init.zeros_(self.Dense_0.bias) self.GroupNorm_1 = nn.GroupNorm(num_groups=min(out_ch // 4, 32), num_channels=out_ch, eps=1e-6) self.Dropout_0 = nn.Dropout(dropout) self.Conv_1 = conv3x3(out_ch, out_ch, init_scale=init_scale) if in_ch != out_ch: if conv_shortcut: self.Conv_2 = conv3x3(in_ch, out_ch) else: self.NIN_0 = NIN(in_ch, out_ch) self.skip_rescale = skip_rescale self.act = act self.out_ch = out_ch self.conv_shortcut = conv_shortcut def forward(self, x, temb=None): h = self.act(self.GroupNorm_0(x)) h = self.Conv_0(h) if temb is not None: h += self.Dense_0(self.act(temb))[:, :, None, None] h = self.act(self.GroupNorm_1(h)) h = self.Dropout_0(h) h = self.Conv_1(h) if x.shape[1] != self.out_ch: if self.conv_shortcut: x = self.Conv_2(x) else: x = self.NIN_0(x) if not self.skip_rescale: return x + h else: return (x + h) / np.sqrt(2.) class ResnetBlockBigGANpp(nn.Module): def __init__(self, act, in_ch, out_ch=None, temb_dim=None, up=False, down=False, dropout=0.1, fir=False, fir_kernel=(1, 3, 3, 1), skip_rescale=True, init_scale=0.): super().__init__() out_ch = out_ch if out_ch else in_ch self.GroupNorm_0 = nn.GroupNorm(num_groups=min(in_ch // 4, 32), num_channels=in_ch, eps=1e-6) self.up = up self.down = down self.fir = fir self.fir_kernel = fir_kernel self.Conv_0 = conv3x3(in_ch, out_ch) if temb_dim is not None: self.Dense_0 = nn.Linear(temb_dim, out_ch) self.Dense_0.weight.data = default_init()(self.Dense_0.weight.shape) nn.init.zeros_(self.Dense_0.bias) self.GroupNorm_1 = nn.GroupNorm(num_groups=min(out_ch // 4, 32), num_channels=out_ch, eps=1e-6) self.Dropout_0 = nn.Dropout(dropout) self.Conv_1 = conv3x3(out_ch, out_ch, init_scale=init_scale) if in_ch != out_ch or up or down: self.Conv_2 = conv1x1(in_ch, out_ch) self.skip_rescale = skip_rescale self.act = act self.in_ch = in_ch self.out_ch = out_ch def forward(self, x, temb=None): h = self.act(self.GroupNorm_0(x)) if self.up: if self.fir: h = up_or_down_sampling.upsample_2d(h, self.fir_kernel, factor=2) x = up_or_down_sampling.upsample_2d(x, self.fir_kernel, factor=2) else: h = up_or_down_sampling.naive_upsample_2d(h, factor=2) x = up_or_down_sampling.naive_upsample_2d(x, factor=2) elif self.down: if self.fir: h = up_or_down_sampling.downsample_2d(h, self.fir_kernel, factor=2) x = up_or_down_sampling.downsample_2d(x, self.fir_kernel, factor=2) else: h = up_or_down_sampling.naive_downsample_2d(h, factor=2) x = up_or_down_sampling.naive_downsample_2d(x, factor=2) h = self.Conv_0(h) # Add bias to each feature map conditioned on the time embedding if temb is not None: h += self.Dense_0(self.act(temb))[:, :, None, None] h = self.act(self.GroupNorm_1(h)) h = self.Dropout_0(h) h = self.Conv_1(h) if self.in_ch != self.out_ch or self.up or self.down: x = self.Conv_2(x) if not self.skip_rescale: return x + h else: return (x + h) / np.sqrt(2.)
9,001
31.734545
99
py
DiffPure
DiffPure-master/score_sde/op/upfirdn2d.py
# --------------------------------------------------------------- # Taken from the following link as is from: # https://github.com/yang-song/score_sde_pytorch/blob/main/op/upfirdn2d.py # # The license for the original version of this file can be # found in the `score_sde` directory (LICENSE_SCORE_SDE). # --------------------------------------------------------------- import os import torch from torch.nn import functional as F from torch.autograd import Function from torch.utils.cpp_extension import load module_path = os.path.dirname(__file__) upfirdn2d_op = load( "upfirdn2d", sources=[ os.path.join(module_path, "upfirdn2d.cpp"), os.path.join(module_path, "upfirdn2d_kernel.cu"), ], ) class UpFirDn2dBackward(Function): @staticmethod def forward( ctx, grad_output, kernel, grad_kernel, up, down, pad, g_pad, in_size, out_size ): up_x, up_y = up down_x, down_y = down g_pad_x0, g_pad_x1, g_pad_y0, g_pad_y1 = g_pad grad_output = grad_output.reshape(-1, out_size[0], out_size[1], 1) grad_input = upfirdn2d_op.upfirdn2d( grad_output, grad_kernel, down_x, down_y, up_x, up_y, g_pad_x0, g_pad_x1, g_pad_y0, g_pad_y1, ) grad_input = grad_input.view(in_size[0], in_size[1], in_size[2], in_size[3]) ctx.save_for_backward(kernel) pad_x0, pad_x1, pad_y0, pad_y1 = pad ctx.up_x = up_x ctx.up_y = up_y ctx.down_x = down_x ctx.down_y = down_y ctx.pad_x0 = pad_x0 ctx.pad_x1 = pad_x1 ctx.pad_y0 = pad_y0 ctx.pad_y1 = pad_y1 ctx.in_size = in_size ctx.out_size = out_size return grad_input @staticmethod def backward(ctx, gradgrad_input): kernel, = ctx.saved_tensors gradgrad_input = gradgrad_input.reshape(-1, ctx.in_size[2], ctx.in_size[3], 1) gradgrad_out = upfirdn2d_op.upfirdn2d( gradgrad_input, kernel, ctx.up_x, ctx.up_y, ctx.down_x, ctx.down_y, ctx.pad_x0, ctx.pad_x1, ctx.pad_y0, ctx.pad_y1, ) # gradgrad_out = gradgrad_out.view(ctx.in_size[0], ctx.out_size[0], ctx.out_size[1], ctx.in_size[3]) gradgrad_out = gradgrad_out.view( ctx.in_size[0], ctx.in_size[1], ctx.out_size[0], ctx.out_size[1] ) return gradgrad_out, None, None, None, None, None, None, None, None class UpFirDn2d(Function): @staticmethod def forward(ctx, input, kernel, up, down, pad): up_x, up_y = up down_x, down_y = down pad_x0, pad_x1, pad_y0, pad_y1 = pad kernel_h, kernel_w = kernel.shape batch, channel, in_h, in_w = input.shape ctx.in_size = input.shape input = input.reshape(-1, in_h, in_w, 1) ctx.save_for_backward(kernel, torch.flip(kernel, [0, 1])) out_h = (in_h * up_y + pad_y0 + pad_y1 - kernel_h) // down_y + 1 out_w = (in_w * up_x + pad_x0 + pad_x1 - kernel_w) // down_x + 1 ctx.out_size = (out_h, out_w) ctx.up = (up_x, up_y) ctx.down = (down_x, down_y) ctx.pad = (pad_x0, pad_x1, pad_y0, pad_y1) g_pad_x0 = kernel_w - pad_x0 - 1 g_pad_y0 = kernel_h - pad_y0 - 1 g_pad_x1 = in_w * up_x - out_w * down_x + pad_x0 - up_x + 1 g_pad_y1 = in_h * up_y - out_h * down_y + pad_y0 - up_y + 1 ctx.g_pad = (g_pad_x0, g_pad_x1, g_pad_y0, g_pad_y1) out = upfirdn2d_op.upfirdn2d( input, kernel, up_x, up_y, down_x, down_y, pad_x0, pad_x1, pad_y0, pad_y1 ) # out = out.view(major, out_h, out_w, minor) out = out.view(-1, channel, out_h, out_w) return out @staticmethod def backward(ctx, grad_output): kernel, grad_kernel = ctx.saved_tensors grad_input = UpFirDn2dBackward.apply( grad_output, kernel, grad_kernel, ctx.up, ctx.down, ctx.pad, ctx.g_pad, ctx.in_size, ctx.out_size, ) return grad_input, None, None, None, None def upfirdn2d(input, kernel, up=1, down=1, pad=(0, 0)): if input.device.type == "cpu": out = upfirdn2d_native( input, kernel, up, up, down, down, pad[0], pad[1], pad[0], pad[1] ) else: out = UpFirDn2d.apply( input, kernel, (up, up), (down, down), (pad[0], pad[1], pad[0], pad[1]) ) return out def upfirdn2d_native( input, kernel, up_x, up_y, down_x, down_y, pad_x0, pad_x1, pad_y0, pad_y1 ): _, channel, in_h, in_w = input.shape input = input.reshape(-1, in_h, in_w, 1) _, in_h, in_w, minor = input.shape kernel_h, kernel_w = kernel.shape out = input.view(-1, in_h, 1, in_w, 1, minor) out = F.pad(out, [0, 0, 0, up_x - 1, 0, 0, 0, up_y - 1]) out = out.view(-1, in_h * up_y, in_w * up_x, minor) out = F.pad( out, [0, 0, max(pad_x0, 0), max(pad_x1, 0), max(pad_y0, 0), max(pad_y1, 0)] ) out = out[ :, max(-pad_y0, 0) : out.shape[1] - max(-pad_y1, 0), max(-pad_x0, 0) : out.shape[2] - max(-pad_x1, 0), :, ] out = out.permute(0, 3, 1, 2) out = out.reshape( [-1, 1, in_h * up_y + pad_y0 + pad_y1, in_w * up_x + pad_x0 + pad_x1] ) w = torch.flip(kernel, [0, 1]).view(1, 1, kernel_h, kernel_w) out = F.conv2d(out, w) out = out.reshape( -1, minor, in_h * up_y + pad_y0 + pad_y1 - kernel_h + 1, in_w * up_x + pad_x0 + pad_x1 - kernel_w + 1, ) out = out.permute(0, 2, 3, 1) out = out[:, ::down_y, ::down_x, :] out_h = (in_h * up_y + pad_y0 + pad_y1 - kernel_h) // down_y + 1 out_w = (in_w * up_x + pad_x0 + pad_x1 - kernel_w) // down_x + 1 return out.view(-1, channel, out_h, out_w)
6,043
27.91866
108
py
DiffPure
DiffPure-master/score_sde/op/__init__.py
# --------------------------------------------------------------- # Taken from the following link as is from: # https://github.com/yang-song/score_sde_pytorch/blob/main/op/__init__.py # # The license for the original version of this file can be # found in the `score_sde` directory (LICENSE_SCORE_SDE). # --------------------------------------------------------------- from .fused_act import FusedLeakyReLU, fused_leaky_relu from .upfirdn2d import upfirdn2d
459
40.818182
73
py
DiffPure
DiffPure-master/score_sde/op/fused_act.py
# --------------------------------------------------------------- # Taken from the following link as is from: # https://github.com/yang-song/score_sde_pytorch/blob/main/op/fused_act.py # # The license for the original version of this file can be # found in the `score_sde` directory (LICENSE_SCORE_SDE). # --------------------------------------------------------------- import os import torch from torch import nn from torch.nn import functional as F from torch.autograd import Function from torch.utils.cpp_extension import load module_path = os.path.dirname(__file__) fused = load( "fused", sources=[ os.path.join(module_path, "fused_bias_act.cpp"), os.path.join(module_path, "fused_bias_act_kernel.cu"), ], ) class FusedLeakyReLUFunctionBackward(Function): @staticmethod def forward(ctx, grad_output, out, negative_slope, scale): ctx.save_for_backward(out) ctx.negative_slope = negative_slope ctx.scale = scale empty = grad_output.new_empty(0) grad_input = fused.fused_bias_act( grad_output, empty, out, 3, 1, negative_slope, scale ) dim = [0] if grad_input.ndim > 2: dim += list(range(2, grad_input.ndim)) grad_bias = grad_input.sum(dim).detach() return grad_input, grad_bias @staticmethod def backward(ctx, gradgrad_input, gradgrad_bias): out, = ctx.saved_tensors gradgrad_out = fused.fused_bias_act( gradgrad_input, gradgrad_bias, out, 3, 1, ctx.negative_slope, ctx.scale ) return gradgrad_out, None, None, None class FusedLeakyReLUFunction(Function): @staticmethod def forward(ctx, input, bias, negative_slope, scale): empty = input.new_empty(0) out = fused.fused_bias_act(input, bias, empty, 3, 0, negative_slope, scale) ctx.save_for_backward(out) ctx.negative_slope = negative_slope ctx.scale = scale return out @staticmethod def backward(ctx, grad_output): out, = ctx.saved_tensors grad_input, grad_bias = FusedLeakyReLUFunctionBackward.apply( grad_output, out, ctx.negative_slope, ctx.scale ) return grad_input, grad_bias, None, None class FusedLeakyReLU(nn.Module): def __init__(self, channel, negative_slope=0.2, scale=2 ** 0.5): super().__init__() self.bias = nn.Parameter(torch.zeros(channel)) self.negative_slope = negative_slope self.scale = scale def forward(self, input): return fused_leaky_relu(input, self.bias, self.negative_slope, self.scale) def fused_leaky_relu(input, bias, negative_slope=0.2, scale=2 ** 0.5): if input.device.type == "cpu": rest_dim = [1] * (input.ndim - bias.ndim - 1) return ( F.leaky_relu( input + bias.view(1, bias.shape[0], *rest_dim), negative_slope=0.2 ) * scale ) else: return FusedLeakyReLUFunction.apply(input, bias, negative_slope, scale)
3,061
27.886792
83
py
DiffPure
DiffPure-master/runners/diffpure_guided.py
# --------------------------------------------------------------- # Copyright (c) 2022, NVIDIA CORPORATION. All rights reserved. # # This work is licensed under the NVIDIA Source Code License # for DiffPure. To view a copy of this license, see the LICENSE file. # --------------------------------------------------------------- import os import random import torch import torchvision.utils as tvu from guided_diffusion.script_util import create_model_and_diffusion, model_and_diffusion_defaults class GuidedDiffusion(torch.nn.Module): def __init__(self, args, config, device=None, model_dir='pretrained/guided_diffusion'): super().__init__() self.args = args self.config = config if device is None: device = torch.device("cuda") if torch.cuda.is_available() else torch.device("cpu") self.device = device # load model model_config = model_and_diffusion_defaults() model_config.update(vars(self.config.model)) print(f'model_config: {model_config}') model, diffusion = create_model_and_diffusion(**model_config) model.load_state_dict(torch.load(f'{model_dir}/256x256_diffusion_uncond.pt', map_location='cpu')) model.requires_grad_(False).eval().to(self.device) if model_config['use_fp16']: model.convert_to_fp16() self.model = model self.diffusion = diffusion self.betas = torch.from_numpy(diffusion.betas).float().to(self.device) def image_editing_sample(self, img, bs_id=0, tag=None): with torch.no_grad(): assert isinstance(img, torch.Tensor) batch_size = img.shape[0] if tag is None: tag = 'rnd' + str(random.randint(0, 10000)) out_dir = os.path.join(self.args.log_dir, 'bs' + str(bs_id) + '_' + tag) assert img.ndim == 4, img.ndim img = img.to(self.device) x0 = img if bs_id < 2: os.makedirs(out_dir, exist_ok=True) tvu.save_image((x0 + 1) * 0.5, os.path.join(out_dir, f'original_input.png')) xs = [] for it in range(self.args.sample_step): e = torch.randn_like(x0) total_noise_levels = self.args.t a = (1 - self.betas).cumprod(dim=0) x = x0 * a[total_noise_levels - 1].sqrt() + e * (1.0 - a[total_noise_levels - 1]).sqrt() if bs_id < 2: tvu.save_image((x + 1) * 0.5, os.path.join(out_dir, f'init_{it}.png')) for i in reversed(range(total_noise_levels)): t = torch.tensor([i] * batch_size, device=self.device) x = self.diffusion.p_sample(self.model, x, t, clip_denoised=True, denoised_fn=None, cond_fn=None, model_kwargs=None)["sample"] # added intermediate step vis if (i - 99) % 100 == 0 and bs_id < 2: tvu.save_image((x + 1) * 0.5, os.path.join(out_dir, f'noise_t_{i}_{it}.png')) x0 = x if bs_id < 2: torch.save(x0, os.path.join(out_dir, f'samples_{it}.pth')) tvu.save_image((x0 + 1) * 0.5, os.path.join(out_dir, f'samples_{it}.png')) xs.append(x0) return torch.cat(xs, dim=0)
3,551
38.466667
105
py
DiffPure
DiffPure-master/runners/diffpure_ode.py
# --------------------------------------------------------------- # Copyright (c) 2022, NVIDIA CORPORATION. All rights reserved. # # This work is licensed under the NVIDIA Source Code License # for DiffPure. To view a copy of this license, see the LICENSE file. # --------------------------------------------------------------- import os import random import numpy as np import torch import torchvision.utils as tvu from torchdiffeq import odeint_adjoint from guided_diffusion.script_util import create_model_and_diffusion, model_and_diffusion_defaults from score_sde.losses import get_optimizer from score_sde.models import utils as mutils from score_sde.models.ema import ExponentialMovingAverage from score_sde import sde_lib def _extract_into_tensor(arr_or_func, timesteps, broadcast_shape): """ Extract values from a 1-D numpy array for a batch of indices. :param arr: the 1-D numpy array or a func. :param timesteps: a tensor of indices into the array to extract. :param broadcast_shape: a larger shape of K dimensions with the batch dimension equal to the length of timesteps. :return: a tensor of shape [batch_size, 1, ...] where the shape has K dims. """ if callable(arr_or_func): res = arr_or_func(timesteps).float() else: res = arr_or_func.to(device=timesteps.device)[timesteps].float() while len(res.shape) < len(broadcast_shape): res = res[..., None] return res.expand(broadcast_shape) def restore_checkpoint(ckpt_dir, state, device): loaded_state = torch.load(ckpt_dir, map_location=device) state['optimizer'].load_state_dict(loaded_state['optimizer']) state['model'].load_state_dict(loaded_state['model'], strict=False) state['ema'].load_state_dict(loaded_state['ema']) state['step'] = loaded_state['step'] class VPODE(torch.nn.Module): def __init__(self, model, score_type='guided_diffusion', beta_min=0.1, beta_max=20, N=1000, img_shape=(3, 256, 256), model_kwargs=None): """Construct a Variance Preserving SDE. Args: model: diffusion model score_type: [guided_diffusion, score_sde, ddpm] beta_min: value of beta(0) beta_max: value of beta(1) """ super().__init__() self.model = model self.score_type = score_type self.model_kwargs = model_kwargs self.img_shape = img_shape self.beta_0 = beta_min self.beta_1 = beta_max self.N = N self.discrete_betas = torch.linspace(beta_min / N, beta_max / N, N) self.alphas = 1. - self.discrete_betas self.alphas_cumprod = torch.cumprod(self.alphas, dim=0) self.sqrt_alphas_cumprod = torch.sqrt(self.alphas_cumprod) self.sqrt_1m_alphas_cumprod = torch.sqrt(1. - self.alphas_cumprod) self.alphas_cumprod_cont = lambda t: torch.exp(-0.5 * (beta_max - beta_min) * t**2 - beta_min * t) self.sqrt_1m_alphas_cumprod_neg_recip_cont = lambda t: -1. / torch.sqrt(1. - self.alphas_cumprod_cont(t)) def _scale_timesteps(self, t): assert torch.all(t <= 1) and torch.all(t >= 0), f't has to be in [0, 1], but get {t} with shape {t.shape}' return (t.float() * self.N).long() def vpsde_fn(self, t, x): beta_t = self.beta_0 + t * (self.beta_1 - self.beta_0) drift = -0.5 * beta_t[:, None] * x diffusion = torch.sqrt(beta_t) return drift, diffusion def ode_fn(self, t, x): """Create the drift and diffusion functions for the reverse SDE""" drift, diffusion = self.vpsde_fn(t, x) assert x.ndim == 2 and np.prod(self.img_shape) == x.shape[1], x.shape x_img = x.view(-1, *self.img_shape) if self.score_type == 'guided_diffusion': # model output is epsilon if self.model_kwargs is None: self.model_kwargs = {} disc_steps = self._scale_timesteps(t) # (batch_size, ), from float in [0,1] to int in [0, 1000] model_output = self.model(x_img, disc_steps, **self.model_kwargs) # with learned sigma, so model_output contains (mean, val) model_output, _ = torch.split(model_output, self.img_shape[0], dim=1) assert x_img.shape == model_output.shape, f'{x_img.shape}, {model_output.shape}' model_output = model_output.view(x.shape[0], -1) score = _extract_into_tensor(self.sqrt_1m_alphas_cumprod_neg_recip_cont, t, x.shape) * model_output elif self.score_type == 'score_sde': # model output is epsilon sde = sde_lib.VPSDE(beta_min=self.beta_0, beta_max=self.beta_1, N=self.N) score_fn = mutils.get_score_fn(sde, self.model, train=False, continuous=True) score = score_fn(x_img, t) assert x_img.shape == score.shape, f'{x_img.shape}, {score.shape}' score = score.view(x.shape[0], -1) else: raise NotImplementedError(f'Unknown score type in RevVPSDE: {self.score_type}!') ode_coef = drift - 0.5 * diffusion[:, None] ** 2 * score return ode_coef def forward(self, t, states): x = states[0] t = t.expand(x.shape[0]) # (batch_size, ) dx_dt = self.ode_fn(t, x) assert dx_dt.shape == x.shape return dx_dt, class OdeGuidedDiffusion(torch.nn.Module): def __init__(self, args, config, device=None): super().__init__() self.args = args self.config = config if device is None: device = torch.device("cuda") if torch.cuda.is_available() else torch.device("cpu") self.device = device # load model if config.data.dataset == 'ImageNet': img_shape = (3, 256, 256) model_dir = 'pretrained/guided_diffusion' model_config = model_and_diffusion_defaults() model_config.update(vars(self.config.model)) print(f'model_config: {model_config}') model, _ = create_model_and_diffusion(**model_config) model.load_state_dict(torch.load(f'{model_dir}/256x256_diffusion_uncond.pt', map_location='cpu')) if model_config['use_fp16']: model.convert_to_fp16() elif config.data.dataset == 'CIFAR10': img_shape = (3, 32, 32) model_dir = 'pretrained/score_sde' print(f'model_config: {config}') model = mutils.create_model(config) optimizer = get_optimizer(config, model.parameters()) ema = ExponentialMovingAverage(model.parameters(), decay=config.model.ema_rate) state = dict(step=0, optimizer=optimizer, model=model, ema=ema) restore_checkpoint(f'{model_dir}/checkpoint_8.pth', state, device) ema.copy_to(model.parameters()) else: raise NotImplementedError(f'Unknown dataset {config.data.dataset}!') model.eval().to(self.device) self.model = model self.vpode = VPODE(model=model, score_type=args.score_type, img_shape=img_shape, model_kwargs=None).to(self.device) self.betas = self.vpode.discrete_betas.float().to(self.device) self.atol, self.rtol = 1e-3, 1e-3 self.method = 'euler' print(f'method: {self.method}, atol: {self.atol}, rtol: {self.rtol}, step_size: {self.args.step_size}') def image_editing_sample(self, img, bs_id=0, tag=None): assert isinstance(img, torch.Tensor) batch_size = img.shape[0] if tag is None: tag = 'rnd' + str(random.randint(0, 10000)) out_dir = os.path.join(self.args.log_dir, 'bs' + str(bs_id) + '_' + tag) assert img.ndim == 4, img.ndim img = img.to(self.device) x0 = img if bs_id < 2: os.makedirs(out_dir, exist_ok=True) tvu.save_image((x0 + 1) * 0.5, os.path.join(out_dir, f'original_input.png')) xs = [] for it in range(self.args.sample_step): if self.args.fix_rand: # fix initial randomness noise_fixed = torch.FloatTensor(1, *x0.shape[1:]).\ normal_(0, 1, generator=torch.manual_seed(self.args.seed)).to(self.device) print(f'noise_fixed: {noise_fixed[0, 0, 0, :3]}') e = noise_fixed.repeat(x0.shape[0], 1, 1, 1) else: e = torch.randn_like(x0).to(self.device) assert e.shape == x0.shape total_noise_levels = self.args.t a = (1 - self.betas).cumprod(dim=0).to(self.device) x = x0 * a[total_noise_levels - 1].sqrt() + e * (1.0 - a[total_noise_levels - 1]).sqrt() if bs_id < 2: tvu.save_image((x + 1) * 0.5, os.path.join(out_dir, f'init_{it}.png')) epsilon_dt0, epsilon_dt1 = 0, 1e-5 t0, t1 = self.args.t * 1. / 1000 - epsilon_dt0, epsilon_dt1 t_size = 2 ts = torch.linspace(t0, t1, t_size).to(self.device) x_ = x.view(batch_size, -1) # (batch_size, state_size) states = (x_, ) # ODE solver odeint = odeint_adjoint state_t = odeint( self.vpode, states, ts, atol=self.atol, rtol=self.rtol, method=self.method, options=None if self.method != 'euler' else dict(step_size=self.args.step_size) # only used for fixed-point method ) # 'euler', 'dopri5' x0_ = state_t[0][-1] x0 = x0_.view(x.shape) # (batch_size, c, h, w) if bs_id < 2: torch.save(x0, os.path.join(out_dir, f'samples_{it}.pth')) tvu.save_image((x0 + 1) * 0.5, os.path.join(out_dir, f'samples_{it}.png')) xs.append(x0) return torch.cat(xs, dim=0)
9,933
38.736
131
py
DiffPure
DiffPure-master/runners/diffpure_ldsde.py
# --------------------------------------------------------------- # Copyright (c) 2022, NVIDIA CORPORATION. All rights reserved. # # This work is licensed under the NVIDIA Source Code License # for DiffPure. To view a copy of this license, see the LICENSE file. # --------------------------------------------------------------- import os import random import numpy as np import torch import torchvision.utils as tvu import torchsde from guided_diffusion.script_util import create_model_and_diffusion, model_and_diffusion_defaults from score_sde.losses import get_optimizer from score_sde.models import utils as mutils from score_sde.models.ema import ExponentialMovingAverage from score_sde import sde_lib def _extract_into_tensor(arr_or_func, timesteps, broadcast_shape): """ Extract values from a 1-D numpy array for a batch of indices. :param arr: the 1-D numpy array or a func. :param timesteps: a tensor of indices into the array to extract. :param broadcast_shape: a larger shape of K dimensions with the batch dimension equal to the length of timesteps. :return: a tensor of shape [batch_size, 1, ...] where the shape has K dims. """ if callable(arr_or_func): res = arr_or_func(timesteps).float() else: res = arr_or_func.to(device=timesteps.device)[timesteps].float() while len(res.shape) < len(broadcast_shape): res = res[..., None] return res.expand(broadcast_shape) def restore_checkpoint(ckpt_dir, state, device): loaded_state = torch.load(ckpt_dir, map_location=device) state['optimizer'].load_state_dict(loaded_state['optimizer']) state['model'].load_state_dict(loaded_state['model'], strict=False) state['ema'].load_state_dict(loaded_state['ema']) state['step'] = loaded_state['step'] class LDSDE(torch.nn.Module): def __init__(self, model, x_init, score_type='guided_diffusion', beta_min=0.1, beta_max=20, N=1000, img_shape=(3, 256, 256), sigma2=0.001, lambda_ld=0.01, eta=5, model_kwargs=None): """Construct a Variance Preserving SDE. Args: model: diffusion model score_type: [guided_diffusion, score_sde, ddpm] beta_min: value of beta(0) beta_max: value of beta(1) """ super().__init__() self.model = model self.x_init = x_init self.sigma2 = sigma2 self.eta = eta self.lambda_ld = lambda_ld # damping coefficient self.score_type = score_type self.model_kwargs = model_kwargs self.img_shape = img_shape self.beta_0 = beta_min self.beta_1 = beta_max self.N = N self.discrete_betas = torch.linspace(beta_min / N, beta_max / N, N) self.alphas = 1. - self.discrete_betas self.alphas_cumprod = torch.cumprod(self.alphas, dim=0) self.sqrt_alphas_cumprod = torch.sqrt(self.alphas_cumprod) self.sqrt_1m_alphas_cumprod = torch.sqrt(1. - self.alphas_cumprod) self.alphas_cumprod_cont = lambda t: torch.exp(-0.5 * (beta_max - beta_min) * t**2 - beta_min * t) self.sqrt_1m_alphas_cumprod_neg_recip_cont = lambda t: -1. / torch.sqrt(1. - self.alphas_cumprod_cont(t)) self.noise_type = "diagonal" self.sde_type = "ito" print(f'sigma2: {self.sigma2}, lambda_ld: {self.lambda_ld}, eta: {self.eta}') def _scale_timesteps(self, t): assert torch.all(t <= 1) and torch.all(t >= 0), f't has to be in [0, 1], but get {t} with shape {t.shape}' return (t.float() * self.N).long() def ldsde_fn(self, t, x, return_type='drift'): """Create the drift and diffusion functions for the reverse SDE""" t = torch.zeros_like(t, dtype=torch.float, device=t.device) + 1e-2 if return_type == 'drift': assert x.ndim == 2 and np.prod(self.img_shape) == x.shape[1], x.shape x_img = x.view(-1, *self.img_shape) if self.score_type == 'guided_diffusion': # model output is epsilon if self.model_kwargs is None: self.model_kwargs = {} disc_steps = self._scale_timesteps(t) # (batch_size, ), from float in [0,1] to int in [0, 1000] model_output = self.model(x_img, disc_steps, **self.model_kwargs) # with learned sigma, so model_output contains (mean, val) model_output, _ = torch.split(model_output, self.img_shape[0], dim=1) assert x_img.shape == model_output.shape, f'{x_img.shape}, {model_output.shape}' model_output = model_output.view(x.shape[0], -1) score = _extract_into_tensor(self.sqrt_1m_alphas_cumprod_neg_recip_cont, t, x.shape) * model_output elif self.score_type == 'score_sde': # model output is epsilon sde = sde_lib.VPSDE(beta_min=self.beta_0, beta_max=self.beta_1, N=self.N) score_fn = mutils.get_score_fn(sde, self.model, train=False, continuous=True) score = score_fn(x_img, t) assert x_img.shape == score.shape, f'{x_img.shape}, {score.shape}' score = score.view(x.shape[0], -1) else: raise NotImplementedError(f'Unknown score type in RevVPSDE: {self.score_type}!') drift = -0.5 * (-score + (x - self.x_init) / self.sigma2) * self.lambda_ld # TODO return drift else: diffusion_coef = np.sqrt(self.lambda_ld) * self.eta return torch.tensor([diffusion_coef], dtype=torch.float).expand(x.shape[0]).to(x.device) def f(self, t, x): """Create the drift function f(x, t) sdeint only support a 2D tensor (batch_size, c*h*w) """ t = t.expand(x.shape[0]) # (batch_size, ) drift = self.ldsde_fn(t, x, return_type='drift') assert drift.shape == x.shape return drift def g(self, t, x): """Create the diffusion function g(t) sdeint only support a 2D tensor (batch_size, c*h*w) """ t = t.expand(x.shape[0]) # (batch_size, ) diffusion = self.ldsde_fn(t, x, return_type='diffusion') assert diffusion.shape == (x.shape[0], ) return diffusion[:, None].expand(x.shape) class LDGuidedDiffusion(torch.nn.Module): def __init__(self, args, config, device=None): super().__init__() self.args = args self.config = config if device is None: device = torch.device("cuda") if torch.cuda.is_available() else torch.device("cpu") self.device = device # load model if config.data.dataset == 'ImageNet': img_shape = (3, 256, 256) model_dir = 'pretrained/guided_diffusion' model_config = model_and_diffusion_defaults() model_config.update(vars(self.config.model)) print(f'model_config: {model_config}') model, _ = create_model_and_diffusion(**model_config) model.load_state_dict(torch.load(f'{model_dir}/256x256_diffusion_uncond.pt', map_location='cpu')) if model_config['use_fp16']: model.convert_to_fp16() elif config.data.dataset == 'CIFAR10': img_shape = (3, 32, 32) model_dir = 'pretrained/score_sde' print(f'model_config: {config}') model = mutils.create_model(config) optimizer = get_optimizer(config, model.parameters()) ema = ExponentialMovingAverage(model.parameters(), decay=config.model.ema_rate) state = dict(step=0, optimizer=optimizer, model=model, ema=ema) restore_checkpoint(f'{model_dir}/checkpoint_8.pth', state, device) ema.copy_to(model.parameters()) else: raise NotImplementedError(f'Unknown dataset {config.data.dataset}!') model.eval().to(self.device) self.model = model self.img_shape = img_shape print(f'use_bm: {args.use_bm}') self.args_dict = { 'method': 'euler', # ["srk", "euler", None] 'adaptive': False, 'dt': 1e-2, } print(f'args_dict: {self.args_dict}') def image_editing_sample(self, img, bs_id=0, tag=None): assert isinstance(img, torch.Tensor) batch_size = img.shape[0] state_size = int(np.prod(img.shape[1:])) # c*h*w if tag is None: tag = 'rnd' + str(random.randint(0, 10000)) out_dir = os.path.join(self.args.log_dir, 'bs' + str(bs_id) + '_' + tag) assert img.ndim == 4, img.ndim img = img.to(self.device) x0 = img x0_ = x0.view(batch_size, -1) # (batch_size, state_size) self.ldsde = LDSDE(model=self.model, x_init=x0_, score_type=self.args.score_type, img_shape=self.img_shape, sigma2=self.args.sigma2, lambda_ld=self.args.lambda_ld, eta=self.args.eta, model_kwargs=None).to(self.device) self.betas = self.ldsde.discrete_betas.float().to(self.device) if bs_id < 2: os.makedirs(out_dir, exist_ok=True) tvu.save_image((x0 + 1) * 0.5, os.path.join(out_dir, f'original_input.png')) xs = [] for it in range(self.args.sample_step): x = x0 if bs_id < 2: tvu.save_image((x + 1) * 0.5, os.path.join(out_dir, f'init_{it}.png')) epsilon_dt0, epsilon_dt1 = 0, 1e-5 t0, t1 = 1 - self.args.t * 1. / 1000 + epsilon_dt0, 1 - epsilon_dt1 t_size = 2 ts = torch.linspace(t0, t1, t_size).to(self.device) x_ = x.view(batch_size, -1) # (batch_size, state_size) if self.args.use_bm: bm = torchsde.BrownianInterval(t0=t0, t1=t1, size=(batch_size, state_size), device=self.device) xs_ = torchsde.sdeint_adjoint(self.ldsde, x_, ts, bm=bm, **self.args_dict) else: xs_ = torchsde.sdeint_adjoint(self.ldsde, x_, ts, **self.args_dict) x0 = xs_[-1].view(x.shape) # (batch_size, c, h, w) if bs_id < 2: torch.save(x0, os.path.join(out_dir, f'samples_{it}.pth')) tvu.save_image((x0 + 1) * 0.5, os.path.join(out_dir, f'samples_{it}.png')) xs.append(x0) return torch.cat(xs, dim=0)
10,418
40.181818
115
py
DiffPure
DiffPure-master/runners/diffpure_ddpm.py
# --------------------------------------------------------------- # Copyright (c) 2022, NVIDIA CORPORATION. All rights reserved. # # This work is licensed under the NVIDIA Source Code License # for DiffPure. To view a copy of this license, see the LICENSE file. # --------------------------------------------------------------- import os import random import numpy as np import torch import torchvision.utils as tvu from ddpm.unet_ddpm import Model def get_beta_schedule(*, beta_start, beta_end, num_diffusion_timesteps): betas = np.linspace(beta_start, beta_end, num_diffusion_timesteps, dtype=np.float64) assert betas.shape == (num_diffusion_timesteps,) return betas def extract(a, t, x_shape): """Extract coefficients from a based on t and reshape to make it broadcastable with x_shape.""" bs, = t.shape assert x_shape[0] == bs out = torch.gather(torch.tensor(a, dtype=torch.float, device=t.device), 0, t.long()) assert out.shape == (bs,) out = out.reshape((bs,) + (1,) * (len(x_shape) - 1)) return out def image_editing_denoising_step_flexible_mask(x, t, *, model, logvar, betas): """ Sample from p(x_{t-1} | x_t) """ alphas = 1.0 - betas alphas_cumprod = alphas.cumprod(dim=0) model_output = model(x, t) weighted_score = betas / torch.sqrt(1 - alphas_cumprod) mean = extract(1 / torch.sqrt(alphas), t, x.shape) * (x - extract(weighted_score, t, x.shape) * model_output) logvar = extract(logvar, t, x.shape) noise = torch.randn_like(x) mask = 1 - (t == 0).float() mask = mask.reshape((x.shape[0],) + (1,) * (len(x.shape) - 1)) sample = mean + mask * torch.exp(0.5 * logvar) * noise sample = sample.float() return sample class Diffusion(torch.nn.Module): def __init__(self, args, config, device=None): super().__init__() self.args = args self.config = config if device is None: device = torch.device("cuda") if torch.cuda.is_available() else torch.device("cpu") self.device = device print("Loading model") if self.config.data.dataset == "CelebA_HQ": url = "https://image-editing-test-12345.s3-us-west-2.amazonaws.com/checkpoints/celeba_hq.ckpt" else: raise ValueError model = Model(self.config) ckpt = torch.hub.load_state_dict_from_url(url, map_location='cpu') model.load_state_dict(ckpt) model.eval() self.model = model self.model_var_type = config.model.var_type betas = get_beta_schedule( beta_start=config.diffusion.beta_start, beta_end=config.diffusion.beta_end, num_diffusion_timesteps=config.diffusion.num_diffusion_timesteps ) self.betas = torch.from_numpy(betas).float() self.num_timesteps = betas.shape[0] alphas = 1.0 - betas alphas_cumprod = np.cumprod(alphas, axis=0) alphas_cumprod_prev = np.append(1.0, alphas_cumprod[:-1]) posterior_variance = betas * \ (1.0 - alphas_cumprod_prev) / (1.0 - alphas_cumprod) if self.model_var_type == "fixedlarge": self.logvar = np.log(np.append(posterior_variance[1], betas[1:])) elif self.model_var_type == 'fixedsmall': self.logvar = np.log(np.maximum(posterior_variance, 1e-20)) def image_editing_sample(self, img=None, bs_id=0, tag=None): assert isinstance(img, torch.Tensor) batch_size = img.shape[0] with torch.no_grad(): if tag is None: tag = 'rnd' + str(random.randint(0, 10000)) out_dir = os.path.join(self.args.log_dir, 'bs' + str(bs_id) + '_' + tag) assert img.ndim == 4, img.ndim x0 = img if bs_id < 2: os.makedirs(out_dir, exist_ok=True) tvu.save_image((x0 + 1) * 0.5, os.path.join(out_dir, f'original_input.png')) xs = [] for it in range(self.args.sample_step): e = torch.randn_like(x0) total_noise_levels = self.args.t a = (1 - self.betas).cumprod(dim=0).to(x0.device) x = x0 * a[total_noise_levels - 1].sqrt() + e * (1.0 - a[total_noise_levels - 1]).sqrt() if bs_id < 2: tvu.save_image((x + 1) * 0.5, os.path.join(out_dir, f'init_{it}.png')) for i in reversed(range(total_noise_levels)): t = torch.tensor([i] * batch_size, device=img.device) x = image_editing_denoising_step_flexible_mask(x, t=t, model=self.model, logvar=self.logvar, betas=self.betas.to(img.device)) # added intermediate step vis if (i - 49) % 50 == 0 and bs_id < 2: tvu.save_image((x + 1) * 0.5, os.path.join(out_dir, f'noise_t_{i}_{it}.png')) x0 = x if bs_id < 2: torch.save(x0, os.path.join(out_dir, f'samples_{it}.pth')) tvu.save_image((x0 + 1) * 0.5, os.path.join(out_dir, f'samples_{it}.png')) xs.append(x0) return torch.cat(xs, dim=0)
5,358
36.475524
113
py
DiffPure
DiffPure-master/runners/diffpure_sde.py
# --------------------------------------------------------------- # Copyright (c) 2022, NVIDIA CORPORATION. All rights reserved. # # This work is licensed under the NVIDIA Source Code License # for DiffPure. To view a copy of this license, see the LICENSE file. # --------------------------------------------------------------- import os import random import numpy as np import torch import torchvision.utils as tvu import torchsde from guided_diffusion.script_util import create_model_and_diffusion, model_and_diffusion_defaults from score_sde.losses import get_optimizer from score_sde.models import utils as mutils from score_sde.models.ema import ExponentialMovingAverage from score_sde import sde_lib def _extract_into_tensor(arr_or_func, timesteps, broadcast_shape): """ Extract values from a 1-D numpy array for a batch of indices. :param arr: the 1-D numpy array or a func. :param timesteps: a tensor of indices into the array to extract. :param broadcast_shape: a larger shape of K dimensions with the batch dimension equal to the length of timesteps. :return: a tensor of shape [batch_size, 1, ...] where the shape has K dims. """ if callable(arr_or_func): res = arr_or_func(timesteps).float() else: res = arr_or_func.to(device=timesteps.device)[timesteps].float() while len(res.shape) < len(broadcast_shape): res = res[..., None] return res.expand(broadcast_shape) def restore_checkpoint(ckpt_dir, state, device): loaded_state = torch.load(ckpt_dir, map_location=device) state['optimizer'].load_state_dict(loaded_state['optimizer']) state['model'].load_state_dict(loaded_state['model'], strict=False) state['ema'].load_state_dict(loaded_state['ema']) state['step'] = loaded_state['step'] class RevVPSDE(torch.nn.Module): def __init__(self, model, score_type='guided_diffusion', beta_min=0.1, beta_max=20, N=1000, img_shape=(3, 256, 256), model_kwargs=None): """Construct a Variance Preserving SDE. Args: model: diffusion model score_type: [guided_diffusion, score_sde, ddpm] beta_min: value of beta(0) beta_max: value of beta(1) """ super().__init__() self.model = model self.score_type = score_type self.model_kwargs = model_kwargs self.img_shape = img_shape self.beta_0 = beta_min self.beta_1 = beta_max self.N = N self.discrete_betas = torch.linspace(beta_min / N, beta_max / N, N) self.alphas = 1. - self.discrete_betas self.alphas_cumprod = torch.cumprod(self.alphas, dim=0) self.sqrt_alphas_cumprod = torch.sqrt(self.alphas_cumprod) self.sqrt_1m_alphas_cumprod = torch.sqrt(1. - self.alphas_cumprod) self.alphas_cumprod_cont = lambda t: torch.exp(-0.5 * (beta_max - beta_min) * t**2 - beta_min * t) self.sqrt_1m_alphas_cumprod_neg_recip_cont = lambda t: -1. / torch.sqrt(1. - self.alphas_cumprod_cont(t)) self.noise_type = "diagonal" self.sde_type = "ito" def _scale_timesteps(self, t): assert torch.all(t <= 1) and torch.all(t >= 0), f't has to be in [0, 1], but get {t} with shape {t.shape}' return (t.float() * self.N).long() def vpsde_fn(self, t, x): beta_t = self.beta_0 + t * (self.beta_1 - self.beta_0) drift = -0.5 * beta_t[:, None] * x diffusion = torch.sqrt(beta_t) return drift, diffusion def rvpsde_fn(self, t, x, return_type='drift'): """Create the drift and diffusion functions for the reverse SDE""" drift, diffusion = self.vpsde_fn(t, x) if return_type == 'drift': assert x.ndim == 2 and np.prod(self.img_shape) == x.shape[1], x.shape x_img = x.view(-1, *self.img_shape) if self.score_type == 'guided_diffusion': # model output is epsilon if self.model_kwargs is None: self.model_kwargs = {} disc_steps = self._scale_timesteps(t) # (batch_size, ), from float in [0,1] to int in [0, 1000] model_output = self.model(x_img, disc_steps, **self.model_kwargs) # with learned sigma, so model_output contains (mean, val) model_output, _ = torch.split(model_output, self.img_shape[0], dim=1) assert x_img.shape == model_output.shape, f'{x_img.shape}, {model_output.shape}' model_output = model_output.view(x.shape[0], -1) score = _extract_into_tensor(self.sqrt_1m_alphas_cumprod_neg_recip_cont, t, x.shape) * model_output elif self.score_type == 'score_sde': # model output is epsilon sde = sde_lib.VPSDE(beta_min=self.beta_0, beta_max=self.beta_1, N=self.N) score_fn = mutils.get_score_fn(sde, self.model, train=False, continuous=True) score = score_fn(x_img, t) assert x_img.shape == score.shape, f'{x_img.shape}, {score.shape}' score = score.view(x.shape[0], -1) else: raise NotImplementedError(f'Unknown score type in RevVPSDE: {self.score_type}!') drift = drift - diffusion[:, None] ** 2 * score return drift else: return diffusion def f(self, t, x): """Create the drift function -f(x, 1-t) (by t' = 1 - t) sdeint only support a 2D tensor (batch_size, c*h*w) """ t = t.expand(x.shape[0]) # (batch_size, ) drift = self.rvpsde_fn(1 - t, x, return_type='drift') assert drift.shape == x.shape return -drift def g(self, t, x): """Create the diffusion function g(1-t) (by t' = 1 - t) sdeint only support a 2D tensor (batch_size, c*h*w) """ t = t.expand(x.shape[0]) # (batch_size, ) diffusion = self.rvpsde_fn(1 - t, x, return_type='diffusion') assert diffusion.shape == (x.shape[0], ) return diffusion[:, None].expand(x.shape) class RevGuidedDiffusion(torch.nn.Module): def __init__(self, args, config, device=None): super().__init__() self.args = args self.config = config if device is None: device = torch.device("cuda") if torch.cuda.is_available() else torch.device("cpu") self.device = device # load model if config.data.dataset == 'ImageNet': img_shape = (3, 256, 256) model_dir = 'pretrained/guided_diffusion' model_config = model_and_diffusion_defaults() model_config.update(vars(self.config.model)) print(f'model_config: {model_config}') model, _ = create_model_and_diffusion(**model_config) model.load_state_dict(torch.load(f'{model_dir}/256x256_diffusion_uncond.pt', map_location='cpu')) if model_config['use_fp16']: model.convert_to_fp16() elif config.data.dataset == 'CIFAR10': img_shape = (3, 32, 32) model_dir = 'pretrained/score_sde' print(f'model_config: {config}') model = mutils.create_model(config) optimizer = get_optimizer(config, model.parameters()) ema = ExponentialMovingAverage(model.parameters(), decay=config.model.ema_rate) state = dict(step=0, optimizer=optimizer, model=model, ema=ema) restore_checkpoint(f'{model_dir}/checkpoint_8.pth', state, device) ema.copy_to(model.parameters()) else: raise NotImplementedError(f'Unknown dataset {config.data.dataset}!') model.eval().to(self.device) self.model = model self.rev_vpsde = RevVPSDE(model=model, score_type=args.score_type, img_shape=img_shape, model_kwargs=None).to(self.device) self.betas = self.rev_vpsde.discrete_betas.float().to(self.device) print(f't: {args.t}, rand_t: {args.rand_t}, t_delta: {args.t_delta}') print(f'use_bm: {args.use_bm}') def image_editing_sample(self, img, bs_id=0, tag=None): assert isinstance(img, torch.Tensor) batch_size = img.shape[0] state_size = int(np.prod(img.shape[1:])) # c*h*w if tag is None: tag = 'rnd' + str(random.randint(0, 10000)) out_dir = os.path.join(self.args.log_dir, 'bs' + str(bs_id) + '_' + tag) assert img.ndim == 4, img.ndim img = img.to(self.device) x0 = img if bs_id < 2: os.makedirs(out_dir, exist_ok=True) tvu.save_image((x0 + 1) * 0.5, os.path.join(out_dir, f'original_input.png')) xs = [] for it in range(self.args.sample_step): e = torch.randn_like(x0).to(self.device) total_noise_levels = self.args.t if self.args.rand_t: total_noise_levels = self.args.t + np.random.randint(-self.args.t_delta, self.args.t_delta) print(f'total_noise_levels: {total_noise_levels}') a = (1 - self.betas).cumprod(dim=0).to(self.device) x = x0 * a[total_noise_levels - 1].sqrt() + e * (1.0 - a[total_noise_levels - 1]).sqrt() if bs_id < 2: tvu.save_image((x + 1) * 0.5, os.path.join(out_dir, f'init_{it}.png')) epsilon_dt0, epsilon_dt1 = 0, 1e-5 t0, t1 = 1 - self.args.t * 1. / 1000 + epsilon_dt0, 1 - epsilon_dt1 t_size = 2 ts = torch.linspace(t0, t1, t_size).to(self.device) x_ = x.view(batch_size, -1) # (batch_size, state_size) if self.args.use_bm: bm = torchsde.BrownianInterval(t0=t0, t1=t1, size=(batch_size, state_size), device=self.device) xs_ = torchsde.sdeint_adjoint(self.rev_vpsde, x_, ts, method='euler', bm=bm) else: xs_ = torchsde.sdeint_adjoint(self.rev_vpsde, x_, ts, method='euler') x0 = xs_[-1].view(x.shape) # (batch_size, c, h, w) if bs_id < 2: torch.save(x0, os.path.join(out_dir, f'samples_{it}.pth')) tvu.save_image((x0 + 1) * 0.5, os.path.join(out_dir, f'samples_{it}.png')) xs.append(x0) return torch.cat(xs, dim=0)
10,334
40.673387
115
py
DiffPure
DiffPure-master/bpda_eot/bpda_eot_attack.py
# --------------------------------------------------------------- # Copyright (c) 2022, NVIDIA CORPORATION. All rights reserved. # # This file has been modified from ebm-defense. # # Source: # https://github.com/point0bar1/ebm-defense/blob/master/bpda_eot_attack.py # # The license for the original version of this file can be # found in this directory (LICENSE_BPDA). # The modifications to this file are subject to the same license. # --------------------------------------------------------------- import torch import torch.nn.functional as F criterion = torch.nn.CrossEntropyLoss() class BPDA_EOT_Attack(): def __init__(self, model, adv_eps=8.0/255, eot_defense_reps=150, eot_attack_reps=15): self.model = model self.config = { 'eot_defense_ave': 'logits', 'eot_attack_ave': 'logits', 'eot_defense_reps': eot_defense_reps, 'eot_attack_reps': eot_attack_reps, 'adv_steps': 50, 'adv_norm': 'l_inf', 'adv_eps': adv_eps, 'adv_eta': 2.0 / 255, 'log_freq': 10 } print(f'BPDA_EOT config: {self.config}') def purify(self, x): return self.model(x, mode='purify') def eot_defense_prediction(seslf, logits, reps=1, eot_defense_ave=None): if eot_defense_ave == 'logits': logits_pred = logits.view([reps, int(logits.shape[0]/reps), logits.shape[1]]).mean(0) elif eot_defense_ave == 'softmax': logits_pred = F.softmax(logits, dim=1).view([reps, int(logits.shape[0]/reps), logits.shape[1]]).mean(0) elif eot_defense_ave == 'logsoftmax': logits_pred = F.log_softmax(logits, dim=1).view([reps, int(logits.shape[0] / reps), logits.shape[1]]).mean(0) elif reps == 1: logits_pred = logits else: raise RuntimeError('Invalid ave_method_pred (use "logits" or "softmax" or "logsoftmax")') _, y_pred = torch.max(logits_pred, 1) return y_pred def eot_attack_loss(self, logits, y, reps=1, eot_attack_ave='loss'): if eot_attack_ave == 'logits': logits_loss = logits.view([reps, int(logits.shape[0] / reps), logits.shape[1]]).mean(0) y_loss = y elif eot_attack_ave == 'softmax': logits_loss = torch.log(F.softmax(logits, dim=1).view([reps, int(logits.shape[0] / reps), logits.shape[1]]).mean(0)) y_loss = y elif eot_attack_ave == 'logsoftmax': logits_loss = F.log_softmax(logits, dim=1).view([reps, int(logits.shape[0] / reps), logits.shape[1]]).mean(0) y_loss = y elif eot_attack_ave == 'loss': logits_loss = logits y_loss = y.repeat(reps) else: raise RuntimeError('Invalid ave_method_eot ("logits", "softmax", "logsoftmax", "loss")') loss = criterion(logits_loss, y_loss) return loss def predict(self, X, y, requires_grad=True, reps=1, eot_defense_ave=None, eot_attack_ave='loss'): if requires_grad: logits = self.model(X, mode='classify') else: with torch.no_grad(): logits = self.model(X.data, mode='classify') y_pred = self.eot_defense_prediction(logits.detach(), reps, eot_defense_ave) correct = torch.eq(y_pred, y) loss = self.eot_attack_loss(logits, y, reps, eot_attack_ave) return correct.detach(), loss def pgd_update(self, X_adv, grad, X, adv_norm, adv_eps, adv_eta, eps=1e-10): if adv_norm == 'l_inf': X_adv.data += adv_eta * torch.sign(grad) X_adv = torch.clamp(torch.min(X + adv_eps, torch.max(X - adv_eps, X_adv)), min=0, max=1) elif adv_norm == 'l_2': X_adv.data += adv_eta * grad / grad.view(X.shape[0], -1).norm(p=2, dim=1).view(X.shape[0], 1, 1, 1) dists = (X_adv - X).view(X.shape[0], -1).norm(dim=1, p=2).view(X.shape[0], 1, 1, 1) X_adv = torch.clamp(X + torch.min(dists, adv_eps*torch.ones_like(dists))*(X_adv-X)/(dists+eps), min=0, max=1) else: raise RuntimeError('Invalid adv_norm ("l_inf" or "l_2"') return X_adv def purify_and_predict(self, X, y, purify_reps=1, requires_grad=True): X_repeat = X.repeat([purify_reps, 1, 1, 1]) X_repeat_purified = self.purify(X_repeat).detach().clone() X_repeat_purified.requires_grad_() correct, loss = self.predict(X_repeat_purified, y, requires_grad, purify_reps, self.config['eot_defense_ave'], self.config['eot_attack_ave']) if requires_grad: X_grads = torch.autograd.grad(loss, [X_repeat_purified])[0] # average gradients over parallel samples for EOT attack attack_grad = X_grads.view([purify_reps]+list(X.shape)).mean(dim=0) return correct, attack_grad else: return correct, None def eot_defense_verification(self, X_adv, y, correct, defended): for verify_ind in range(correct.nelement()): if correct[verify_ind] == 0 and defended[verify_ind] == 1: defended[verify_ind] = self.purify_and_predict(X_adv[verify_ind].unsqueeze(0), y[verify_ind].view([1]), self.config['eot_defense_reps'], requires_grad=False)[0] return defended def eval_and_bpda_eot_grad(self, X_adv, y, defended, requires_grad=True): correct, attack_grad = self.purify_and_predict(X_adv, y, self.config['eot_attack_reps'], requires_grad) if self.config['eot_defense_reps'] > 0: defended = self.eot_defense_verification(X_adv, y, correct, defended) else: defended *= correct return defended, attack_grad def attack_batch(self, X, y): # get baseline accuracy for natural images defended = self.eval_and_bpda_eot_grad(X, y, torch.ones_like(y).bool(), False)[0] print('Baseline: {} of {}'.format(defended.sum(), len(defended))) class_batch = torch.zeros([self.config['adv_steps'] + 2, X.shape[0]]).bool() class_batch[0] = defended.cpu() ims_adv_batch = torch.zeros(X.shape) for ind in range(defended.nelement()): if defended[ind] == 0: ims_adv_batch[ind] = X[ind].cpu() X_adv = X.clone() # adversarial attacks on a single batch of images for step in range(self.config['adv_steps'] + 1): defended, attack_grad = self.eval_and_bpda_eot_grad(X_adv, y, defended) class_batch[step+1] = defended.cpu() for ind in range(defended.nelement()): if class_batch[step, ind] == 1 and defended[ind] == 0: ims_adv_batch[ind] = X_adv[ind].cpu() # update adversarial images (except on final iteration so final adv images match final eval) if step < self.config['adv_steps']: X_adv = self.pgd_update(X_adv, attack_grad, X, self.config['adv_norm'], self.config['adv_eps'], self.config['adv_eta']) X_adv = X_adv.detach().clone() if step == 1 or step % self.config['log_freq'] == 0 or step == self.config['adv_steps']: print('Attack {} of {} Batch defended: {} of {}'. format(step, self.config['adv_steps'], int(torch.sum(defended).cpu().numpy()), X_adv.shape[0])) if int(torch.sum(defended).cpu().numpy()) == 0: print('Attack successfully to the batch!') break for ind in range(defended.nelement()): if defended[ind] == 1: ims_adv_batch[ind] = X_adv[ind].cpu() return class_batch, ims_adv_batch def attack_all(self, X, y, batch_size): class_path = torch.zeros([self.config['adv_steps'] + 2, 0]).bool() ims_adv = torch.zeros(0) n_batches = X.shape[0] // batch_size if n_batches == 0 and X.shape[0] > 0: n_batches = 1 for counter in range(n_batches): X_batch = X[counter * batch_size:min((counter + 1) * batch_size, X.shape[0])].clone().to(X.device) y_batch = y[counter * batch_size:min((counter + 1) * batch_size, X.shape[0])].clone().to(X.device) class_batch, ims_adv_batch = self.attack_batch(X_batch.contiguous(), y_batch.contiguous()) class_path = torch.cat((class_path, class_batch), dim=1) ims_adv = torch.cat((ims_adv, ims_adv_batch), dim=0) print(f'finished {counter}-th batch in attack_all') return class_path, ims_adv
8,620
45.349462
135
py
DiffPure
DiffPure-master/guided_diffusion/resample.py
# --------------------------------------------------------------- # Taken from the following link as is from: # https://github.com/openai/guided-diffusion/blob/main/guided_diffusion/resample.py # # The license for the original version of this file can be # found in this directory (LICENSE_GUIDED_DIFFUSION). # --------------------------------------------------------------- from abc import ABC, abstractmethod import numpy as np import torch as th import torch.distributed as dist def create_named_schedule_sampler(name, diffusion): """ Create a ScheduleSampler from a library of pre-defined samplers. :param name: the name of the sampler. :param diffusion: the diffusion object to sample for. """ if name == "uniform": return UniformSampler(diffusion) elif name == "loss-second-moment": return LossSecondMomentResampler(diffusion) else: raise NotImplementedError(f"unknown schedule sampler: {name}") class ScheduleSampler(ABC): """ A distribution over timesteps in the diffusion process, intended to reduce variance of the objective. By default, samplers perform unbiased importance sampling, in which the objective's mean is unchanged. However, subclasses may override sample() to change how the resampled terms are reweighted, allowing for actual changes in the objective. """ @abstractmethod def weights(self): """ Get a numpy array of weights, one per diffusion step. The weights needn't be normalized, but must be positive. """ def sample(self, batch_size, device): """ Importance-sample timesteps for a batch. :param batch_size: the number of timesteps. :param device: the torch device to save to. :return: a tuple (timesteps, weights): - timesteps: a tensor of timestep indices. - weights: a tensor of weights to scale the resulting losses. """ w = self.weights() p = w / np.sum(w) indices_np = np.random.choice(len(p), size=(batch_size,), p=p) indices = th.from_numpy(indices_np).long().to(device) weights_np = 1 / (len(p) * p[indices_np]) weights = th.from_numpy(weights_np).float().to(device) return indices, weights class UniformSampler(ScheduleSampler): def __init__(self, diffusion): self.diffusion = diffusion self._weights = np.ones([diffusion.num_timesteps]) def weights(self): return self._weights class LossAwareSampler(ScheduleSampler): def update_with_local_losses(self, local_ts, local_losses): """ Update the reweighting using losses from a model. Call this method from each rank with a batch of timesteps and the corresponding losses for each of those timesteps. This method will perform synchronization to make sure all of the ranks maintain the exact same reweighting. :param local_ts: an integer Tensor of timesteps. :param local_losses: a 1D Tensor of losses. """ batch_sizes = [ th.tensor([0], dtype=th.int32, device=local_ts.device) for _ in range(dist.get_world_size()) ] dist.all_gather( batch_sizes, th.tensor([len(local_ts)], dtype=th.int32, device=local_ts.device), ) # Pad all_gather batches to be the maximum batch size. batch_sizes = [x.item() for x in batch_sizes] max_bs = max(batch_sizes) timestep_batches = [th.zeros(max_bs).to(local_ts) for bs in batch_sizes] loss_batches = [th.zeros(max_bs).to(local_losses) for bs in batch_sizes] dist.all_gather(timestep_batches, local_ts) dist.all_gather(loss_batches, local_losses) timesteps = [ x.item() for y, bs in zip(timestep_batches, batch_sizes) for x in y[:bs] ] losses = [x.item() for y, bs in zip(loss_batches, batch_sizes) for x in y[:bs]] self.update_with_all_losses(timesteps, losses) @abstractmethod def update_with_all_losses(self, ts, losses): """ Update the reweighting using losses from a model. Sub-classes should override this method to update the reweighting using losses from the model. This method directly updates the reweighting without synchronizing between workers. It is called by update_with_local_losses from all ranks with identical arguments. Thus, it should have deterministic behavior to maintain state across workers. :param ts: a list of int timesteps. :param losses: a list of float losses, one per timestep. """ class LossSecondMomentResampler(LossAwareSampler): def __init__(self, diffusion, history_per_term=10, uniform_prob=0.001): self.diffusion = diffusion self.history_per_term = history_per_term self.uniform_prob = uniform_prob self._loss_history = np.zeros( [diffusion.num_timesteps, history_per_term], dtype=np.float64 ) self._loss_counts = np.zeros([diffusion.num_timesteps], dtype=np.int) def weights(self): if not self._warmed_up(): return np.ones([self.diffusion.num_timesteps], dtype=np.float64) weights = np.sqrt(np.mean(self._loss_history ** 2, axis=-1)) weights /= np.sum(weights) weights *= 1 - self.uniform_prob weights += self.uniform_prob / len(weights) return weights def update_with_all_losses(self, ts, losses): for t, loss in zip(ts, losses): if self._loss_counts[t] == self.history_per_term: # Shift out the oldest loss term. self._loss_history[t, :-1] = self._loss_history[t, 1:] self._loss_history[t, -1] = loss else: self._loss_history[t, self._loss_counts[t]] = loss self._loss_counts[t] += 1 def _warmed_up(self): return (self._loss_counts == self.history_per_term).all()
6,065
36.214724
87
py
DiffPure
DiffPure-master/guided_diffusion/losses.py
# --------------------------------------------------------------- # Taken from the following link as is from: # https://github.com/openai/guided-diffusion/blob/main/guided_diffusion/losses.py # # The license for the original version of this file can be # found in this directory (LICENSE_GUIDED_DIFFUSION). # --------------------------------------------------------------- """ Helpers for various likelihood-based losses. These are ported from the original Ho et al. diffusion models codebase: https://github.com/hojonathanho/diffusion/blob/1e0dceb3b3495bbe19116a5e1b3596cd0706c543/diffusion_tf/utils.py """ import numpy as np import torch as th def normal_kl(mean1, logvar1, mean2, logvar2): """ Compute the KL divergence between two gaussians. Shapes are automatically broadcasted, so batches can be compared to scalars, among other use cases. """ tensor = None for obj in (mean1, logvar1, mean2, logvar2): if isinstance(obj, th.Tensor): tensor = obj break assert tensor is not None, "at least one argument must be a Tensor" # Force variances to be Tensors. Broadcasting helps convert scalars to # Tensors, but it does not work for th.exp(). logvar1, logvar2 = [ x if isinstance(x, th.Tensor) else th.tensor(x).to(tensor) for x in (logvar1, logvar2) ] return 0.5 * ( -1.0 + logvar2 - logvar1 + th.exp(logvar1 - logvar2) + ((mean1 - mean2) ** 2) * th.exp(-logvar2) ) def approx_standard_normal_cdf(x): """ A fast approximation of the cumulative distribution function of the standard normal. """ return 0.5 * (1.0 + th.tanh(np.sqrt(2.0 / np.pi) * (x + 0.044715 * th.pow(x, 3)))) def discretized_gaussian_log_likelihood(x, *, means, log_scales): """ Compute the log-likelihood of a Gaussian distribution discretizing to a given image. :param x: the target images. It is assumed that this was uint8 values, rescaled to the range [-1, 1]. :param means: the Gaussian mean Tensor. :param log_scales: the Gaussian log stddev Tensor. :return: a tensor like x of log probabilities (in nats). """ assert x.shape == means.shape == log_scales.shape centered_x = x - means inv_stdv = th.exp(-log_scales) plus_in = inv_stdv * (centered_x + 1.0 / 255.0) cdf_plus = approx_standard_normal_cdf(plus_in) min_in = inv_stdv * (centered_x - 1.0 / 255.0) cdf_min = approx_standard_normal_cdf(min_in) log_cdf_plus = th.log(cdf_plus.clamp(min=1e-12)) log_one_minus_cdf_min = th.log((1.0 - cdf_min).clamp(min=1e-12)) cdf_delta = cdf_plus - cdf_min log_probs = th.where( x < -0.999, log_cdf_plus, th.where(x > 0.999, log_one_minus_cdf_min, th.log(cdf_delta.clamp(min=1e-12))), ) assert log_probs.shape == x.shape return log_probs
2,908
32.825581
109
py
DiffPure
DiffPure-master/guided_diffusion/image_datasets.py
# --------------------------------------------------------------- # Taken from the following link as is from: # https://github.com/openai/guided-diffusion/blob/main/guided_diffusion/image_datasets.py # # The license for the original version of this file can be # found in this directory (LICENSE_GUIDED_DIFFUSION). # --------------------------------------------------------------- import math import random from PIL import Image import blobfile as bf from mpi4py import MPI import numpy as np from torch.utils.data import DataLoader, Dataset def load_data( *, data_dir, batch_size, image_size, class_cond=False, deterministic=False, random_crop=False, random_flip=True, ): """ For a dataset, create a generator over (images, kwargs) pairs. Each images is an NCHW float tensor, and the kwargs dict contains zero or more keys, each of which map to a batched Tensor of their own. The kwargs dict can be used for class labels, in which case the key is "y" and the values are integer tensors of class labels. :param data_dir: a dataset directory. :param batch_size: the batch size of each returned pair. :param image_size: the size to which images are resized. :param class_cond: if True, include a "y" key in returned dicts for class label. If classes are not available and this is true, an exception will be raised. :param deterministic: if True, yield results in a deterministic order. :param random_crop: if True, randomly crop the images for augmentation. :param random_flip: if True, randomly flip the images for augmentation. """ if not data_dir: raise ValueError("unspecified data directory") all_files = _list_image_files_recursively(data_dir) classes = None if class_cond: # Assume classes are the first part of the filename, # before an underscore. class_names = [bf.basename(path).split("_")[0] for path in all_files] sorted_classes = {x: i for i, x in enumerate(sorted(set(class_names)))} classes = [sorted_classes[x] for x in class_names] dataset = ImageDataset( image_size, all_files, classes=classes, shard=MPI.COMM_WORLD.Get_rank(), num_shards=MPI.COMM_WORLD.Get_size(), random_crop=random_crop, random_flip=random_flip, ) if deterministic: loader = DataLoader( dataset, batch_size=batch_size, shuffle=False, num_workers=1, drop_last=True ) else: loader = DataLoader( dataset, batch_size=batch_size, shuffle=True, num_workers=1, drop_last=True ) while True: yield from loader def _list_image_files_recursively(data_dir): results = [] for entry in sorted(bf.listdir(data_dir)): full_path = bf.join(data_dir, entry) ext = entry.split(".")[-1] if "." in entry and ext.lower() in ["jpg", "jpeg", "png", "gif"]: results.append(full_path) elif bf.isdir(full_path): results.extend(_list_image_files_recursively(full_path)) return results class ImageDataset(Dataset): def __init__( self, resolution, image_paths, classes=None, shard=0, num_shards=1, random_crop=False, random_flip=True, ): super().__init__() self.resolution = resolution self.local_images = image_paths[shard:][::num_shards] self.local_classes = None if classes is None else classes[shard:][::num_shards] self.random_crop = random_crop self.random_flip = random_flip def __len__(self): return len(self.local_images) def __getitem__(self, idx): path = self.local_images[idx] with bf.BlobFile(path, "rb") as f: pil_image = Image.open(f) pil_image.load() pil_image = pil_image.convert("RGB") if self.random_crop: arr = random_crop_arr(pil_image, self.resolution) else: arr = center_crop_arr(pil_image, self.resolution) if self.random_flip and random.random() < 0.5: arr = arr[:, ::-1] arr = arr.astype(np.float32) / 127.5 - 1 out_dict = {} if self.local_classes is not None: out_dict["y"] = np.array(self.local_classes[idx], dtype=np.int64) return np.transpose(arr, [2, 0, 1]), out_dict def center_crop_arr(pil_image, image_size): # We are not on a new enough PIL to support the `reducing_gap` # argument, which uses BOX downsampling at powers of two first. # Thus, we do it by hand to improve downsample quality. while min(*pil_image.size) >= 2 * image_size: pil_image = pil_image.resize( tuple(x // 2 for x in pil_image.size), resample=Image.BOX ) scale = image_size / min(*pil_image.size) pil_image = pil_image.resize( tuple(round(x * scale) for x in pil_image.size), resample=Image.BICUBIC ) arr = np.array(pil_image) crop_y = (arr.shape[0] - image_size) // 2 crop_x = (arr.shape[1] - image_size) // 2 return arr[crop_y : crop_y + image_size, crop_x : crop_x + image_size] def random_crop_arr(pil_image, image_size, min_crop_frac=0.8, max_crop_frac=1.0): min_smaller_dim_size = math.ceil(image_size / max_crop_frac) max_smaller_dim_size = math.ceil(image_size / min_crop_frac) smaller_dim_size = random.randrange(min_smaller_dim_size, max_smaller_dim_size + 1) # We are not on a new enough PIL to support the `reducing_gap` # argument, which uses BOX downsampling at powers of two first. # Thus, we do it by hand to improve downsample quality. while min(*pil_image.size) >= 2 * smaller_dim_size: pil_image = pil_image.resize( tuple(x // 2 for x in pil_image.size), resample=Image.BOX ) scale = smaller_dim_size / min(*pil_image.size) pil_image = pil_image.resize( tuple(round(x * scale) for x in pil_image.size), resample=Image.BICUBIC ) arr = np.array(pil_image) crop_y = random.randrange(arr.shape[0] - image_size + 1) crop_x = random.randrange(arr.shape[1] - image_size + 1) return arr[crop_y : crop_y + image_size, crop_x : crop_x + image_size]
6,312
34.869318
89
py
DiffPure
DiffPure-master/guided_diffusion/nn.py
# --------------------------------------------------------------- # Taken from the following link as is from: # https://github.com/openai/guided-diffusion/blob/main/guided_diffusion/nn.py # # The license for the original version of this file can be # found in this directory (LICENSE_GUIDED_DIFFUSION). # --------------------------------------------------------------- """ Various utilities for neural networks. """ import math import torch as th import torch.nn as nn # PyTorch 1.7 has SiLU, but we support PyTorch 1.5. class SiLU(nn.Module): def forward(self, x): return x * th.sigmoid(x) class GroupNorm32(nn.GroupNorm): def forward(self, x): return super().forward(x.float()).type(x.dtype) def conv_nd(dims, *args, **kwargs): """ Create a 1D, 2D, or 3D convolution module. """ if dims == 1: return nn.Conv1d(*args, **kwargs) elif dims == 2: return nn.Conv2d(*args, **kwargs) elif dims == 3: return nn.Conv3d(*args, **kwargs) raise ValueError(f"unsupported dimensions: {dims}") def linear(*args, **kwargs): """ Create a linear module. """ return nn.Linear(*args, **kwargs) def avg_pool_nd(dims, *args, **kwargs): """ Create a 1D, 2D, or 3D average pooling module. """ if dims == 1: return nn.AvgPool1d(*args, **kwargs) elif dims == 2: return nn.AvgPool2d(*args, **kwargs) elif dims == 3: return nn.AvgPool3d(*args, **kwargs) raise ValueError(f"unsupported dimensions: {dims}") def update_ema(target_params, source_params, rate=0.99): """ Update target parameters to be closer to those of source parameters using an exponential moving average. :param target_params: the target parameter sequence. :param source_params: the source parameter sequence. :param rate: the EMA rate (closer to 1 means slower). """ for targ, src in zip(target_params, source_params): targ.detach().mul_(rate).add_(src, alpha=1 - rate) def zero_module(module): """ Zero out the parameters of a module and return it. """ for p in module.parameters(): p.detach().zero_() return module def scale_module(module, scale): """ Scale the parameters of a module and return it. """ for p in module.parameters(): p.detach().mul_(scale) return module def mean_flat(tensor): """ Take the mean over all non-batch dimensions. """ return tensor.mean(dim=list(range(1, len(tensor.shape)))) def normalization(channels): """ Make a standard normalization layer. :param channels: number of input channels. :return: an nn.Module for normalization. """ return GroupNorm32(32, channels) def timestep_embedding(timesteps, dim, max_period=10000): """ Create sinusoidal timestep embeddings. :param timesteps: a 1-D Tensor of N indices, one per batch element. These may be fractional. :param dim: the dimension of the output. :param max_period: controls the minimum frequency of the embeddings. :return: an [N x dim] Tensor of positional embeddings. """ half = dim // 2 freqs = th.exp( -math.log(max_period) * th.arange(start=0, end=half, dtype=th.float32) / half ).to(device=timesteps.device) args = timesteps[:, None].float() * freqs[None] embedding = th.cat([th.cos(args), th.sin(args)], dim=-1) if dim % 2: embedding = th.cat([embedding, th.zeros_like(embedding[:, :1])], dim=-1) return embedding def checkpoint(func, inputs, params, flag): """ Evaluate a function without caching intermediate activations, allowing for reduced memory at the expense of extra compute in the backward pass. :param func: the function to evaluate. :param inputs: the argument sequence to pass to `func`. :param params: a sequence of parameters `func` depends on but does not explicitly take as arguments. :param flag: if False, disable gradient checkpointing. """ if flag: args = tuple(inputs) + tuple(params) return CheckpointFunction.apply(func, len(inputs), *args) else: return func(*inputs) class CheckpointFunction(th.autograd.Function): @staticmethod def forward(ctx, run_function, length, *args): ctx.run_function = run_function ctx.input_tensors = list(args[:length]) ctx.input_params = list(args[length:]) with th.no_grad(): output_tensors = ctx.run_function(*ctx.input_tensors) return output_tensors @staticmethod def backward(ctx, *output_grads): ctx.input_tensors = [x.detach().requires_grad_(True) for x in ctx.input_tensors] with th.enable_grad(): # Fixes a bug where the first op in run_function modifies the # Tensor storage in place, which is not allowed for detach()'d # Tensors. shallow_copies = [x.view_as(x) for x in ctx.input_tensors] output_tensors = ctx.run_function(*shallow_copies) input_grads = th.autograd.grad( output_tensors, ctx.input_tensors + ctx.input_params, output_grads, allow_unused=True, ) del ctx.input_tensors del ctx.input_params del output_tensors return (None, None) + input_grads
5,390
29.117318
88
py
DiffPure
DiffPure-master/guided_diffusion/fp16_util.py
# --------------------------------------------------------------- # Taken from the following link as is from: # https://github.com/openai/guided-diffusion/blob/main/guided_diffusion/fp16_util.py # # The license for the original version of this file can be # found in this directory (LICENSE_GUIDED_DIFFUSION). # --------------------------------------------------------------- """ Helpers to train with 16-bit precision. """ import numpy as np import torch as th import torch.nn as nn from torch._utils import _flatten_dense_tensors, _unflatten_dense_tensors from . import logger INITIAL_LOG_LOSS_SCALE = 20.0 def convert_module_to_f16(l): """ Convert primitive modules to float16. """ if isinstance(l, (nn.Conv1d, nn.Conv2d, nn.Conv3d)): l.weight.data = l.weight.data.half() if l.bias is not None: l.bias.data = l.bias.data.half() def convert_module_to_f32(l): """ Convert primitive modules to float32, undoing convert_module_to_f16(). """ if isinstance(l, (nn.Conv1d, nn.Conv2d, nn.Conv3d)): l.weight.data = l.weight.data.float() if l.bias is not None: l.bias.data = l.bias.data.float() def make_master_params(param_groups_and_shapes): """ Copy model parameters into a (differently-shaped) list of full-precision parameters. """ master_params = [] for param_group, shape in param_groups_and_shapes: master_param = nn.Parameter( _flatten_dense_tensors( [param.detach().float() for (_, param) in param_group] ).view(shape) ) master_param.requires_grad = True master_params.append(master_param) return master_params def model_grads_to_master_grads(param_groups_and_shapes, master_params): """ Copy the gradients from the model parameters into the master parameters from make_master_params(). """ for master_param, (param_group, shape) in zip( master_params, param_groups_and_shapes ): master_param.grad = _flatten_dense_tensors( [param_grad_or_zeros(param) for (_, param) in param_group] ).view(shape) def master_params_to_model_params(param_groups_and_shapes, master_params): """ Copy the master parameter data back into the model parameters. """ # Without copying to a list, if a generator is passed, this will # silently not copy any parameters. for master_param, (param_group, _) in zip(master_params, param_groups_and_shapes): for (_, param), unflat_master_param in zip( param_group, unflatten_master_params(param_group, master_param.view(-1)) ): param.detach().copy_(unflat_master_param) def unflatten_master_params(param_group, master_param): return _unflatten_dense_tensors(master_param, [param for (_, param) in param_group]) def get_param_groups_and_shapes(named_model_params): named_model_params = list(named_model_params) scalar_vector_named_params = ( [(n, p) for (n, p) in named_model_params if p.ndim <= 1], (-1), ) matrix_named_params = ( [(n, p) for (n, p) in named_model_params if p.ndim > 1], (1, -1), ) return [scalar_vector_named_params, matrix_named_params] def master_params_to_state_dict( model, param_groups_and_shapes, master_params, use_fp16 ): if use_fp16: state_dict = model.state_dict() for master_param, (param_group, _) in zip( master_params, param_groups_and_shapes ): for (name, _), unflat_master_param in zip( param_group, unflatten_master_params(param_group, master_param.view(-1)) ): assert name in state_dict state_dict[name] = unflat_master_param else: state_dict = model.state_dict() for i, (name, _value) in enumerate(model.named_parameters()): assert name in state_dict state_dict[name] = master_params[i] return state_dict def state_dict_to_master_params(model, state_dict, use_fp16): if use_fp16: named_model_params = [ (name, state_dict[name]) for name, _ in model.named_parameters() ] param_groups_and_shapes = get_param_groups_and_shapes(named_model_params) master_params = make_master_params(param_groups_and_shapes) else: master_params = [state_dict[name] for name, _ in model.named_parameters()] return master_params def zero_master_grads(master_params): for param in master_params: param.grad = None def zero_grad(model_params): for param in model_params: # Taken from https://pytorch.org/docs/stable/_modules/torch/optim/optimizer.html#Optimizer.add_param_group if param.grad is not None: param.grad.detach_() param.grad.zero_() def param_grad_or_zeros(param): if param.grad is not None: return param.grad.data.detach() else: return th.zeros_like(param) class MixedPrecisionTrainer: def __init__( self, *, model, use_fp16=False, fp16_scale_growth=1e-3, initial_lg_loss_scale=INITIAL_LOG_LOSS_SCALE, ): self.model = model self.use_fp16 = use_fp16 self.fp16_scale_growth = fp16_scale_growth self.model_params = list(self.model.parameters()) self.master_params = self.model_params self.param_groups_and_shapes = None self.lg_loss_scale = initial_lg_loss_scale if self.use_fp16: self.param_groups_and_shapes = get_param_groups_and_shapes( self.model.named_parameters() ) self.master_params = make_master_params(self.param_groups_and_shapes) self.model.convert_to_fp16() def zero_grad(self): zero_grad(self.model_params) def backward(self, loss: th.Tensor): if self.use_fp16: loss_scale = 2 ** self.lg_loss_scale (loss * loss_scale).backward() else: loss.backward() def optimize(self, opt: th.optim.Optimizer): if self.use_fp16: return self._optimize_fp16(opt) else: return self._optimize_normal(opt) def _optimize_fp16(self, opt: th.optim.Optimizer): logger.logkv_mean("lg_loss_scale", self.lg_loss_scale) model_grads_to_master_grads(self.param_groups_and_shapes, self.master_params) grad_norm, param_norm = self._compute_norms(grad_scale=2 ** self.lg_loss_scale) if check_overflow(grad_norm): self.lg_loss_scale -= 1 logger.log(f"Found NaN, decreased lg_loss_scale to {self.lg_loss_scale}") zero_master_grads(self.master_params) return False logger.logkv_mean("grad_norm", grad_norm) logger.logkv_mean("param_norm", param_norm) self.master_params[0].grad.mul_(1.0 / (2 ** self.lg_loss_scale)) opt.step() zero_master_grads(self.master_params) master_params_to_model_params(self.param_groups_and_shapes, self.master_params) self.lg_loss_scale += self.fp16_scale_growth return True def _optimize_normal(self, opt: th.optim.Optimizer): grad_norm, param_norm = self._compute_norms() logger.logkv_mean("grad_norm", grad_norm) logger.logkv_mean("param_norm", param_norm) opt.step() return True def _compute_norms(self, grad_scale=1.0): grad_norm = 0.0 param_norm = 0.0 for p in self.master_params: with th.no_grad(): param_norm += th.norm(p, p=2, dtype=th.float32).item() ** 2 if p.grad is not None: grad_norm += th.norm(p.grad, p=2, dtype=th.float32).item() ** 2 return np.sqrt(grad_norm) / grad_scale, np.sqrt(param_norm) def master_params_to_state_dict(self, master_params): return master_params_to_state_dict( self.model, self.param_groups_and_shapes, master_params, self.use_fp16 ) def state_dict_to_master_params(self, state_dict): return state_dict_to_master_params(self.model, state_dict, self.use_fp16) def check_overflow(value): return (value == float("inf")) or (value == -float("inf")) or (value != value)
8,318
32.955102
114
py
DiffPure
DiffPure-master/guided_diffusion/unet.py
# --------------------------------------------------------------- # Taken from the following link as is from: # https://github.com/openai/guided-diffusion/blob/main/guided_diffusion/unet.py # # The license for the original version of this file can be # found in this directory (LICENSE_GUIDED_DIFFUSION). # --------------------------------------------------------------- from abc import abstractmethod import math import numpy as np import torch as th import torch.nn as nn import torch.nn.functional as F from .fp16_util import convert_module_to_f16, convert_module_to_f32 from .nn import ( checkpoint, conv_nd, linear, avg_pool_nd, zero_module, normalization, timestep_embedding, ) class AttentionPool2d(nn.Module): """ Adapted from CLIP: https://github.com/openai/CLIP/blob/main/clip/model.py """ def __init__( self, spacial_dim: int, embed_dim: int, num_heads_channels: int, output_dim: int = None, ): super().__init__() self.positional_embedding = nn.Parameter( th.randn(embed_dim, spacial_dim ** 2 + 1) / embed_dim ** 0.5 ) self.qkv_proj = conv_nd(1, embed_dim, 3 * embed_dim, 1) self.c_proj = conv_nd(1, embed_dim, output_dim or embed_dim, 1) self.num_heads = embed_dim // num_heads_channels self.attention = QKVAttention(self.num_heads) def forward(self, x): b, c, *_spatial = x.shape x = x.reshape(b, c, -1) # NC(HW) x = th.cat([x.mean(dim=-1, keepdim=True), x], dim=-1) # NC(HW+1) x = x + self.positional_embedding[None, :, :].to(x.dtype) # NC(HW+1) x = self.qkv_proj(x) x = self.attention(x) x = self.c_proj(x) return x[:, :, 0] class TimestepBlock(nn.Module): """ Any module where forward() takes timestep embeddings as a second argument. """ @abstractmethod def forward(self, x, emb): """ Apply the module to `x` given `emb` timestep embeddings. """ class TimestepEmbedSequential(nn.Sequential, TimestepBlock): """ A sequential module that passes timestep embeddings to the children that support it as an extra input. """ def forward(self, x, emb): for layer in self: if isinstance(layer, TimestepBlock): x = layer(x, emb) else: x = layer(x) return x class Upsample(nn.Module): """ An upsampling layer with an optional convolution. :param channels: channels in the inputs and outputs. :param use_conv: a bool determining if a convolution is applied. :param dims: determines if the signal is 1D, 2D, or 3D. If 3D, then upsampling occurs in the inner-two dimensions. """ def __init__(self, channels, use_conv, dims=2, out_channels=None): super().__init__() self.channels = channels self.out_channels = out_channels or channels self.use_conv = use_conv self.dims = dims if use_conv: self.conv = conv_nd(dims, self.channels, self.out_channels, 3, padding=1) def forward(self, x): assert x.shape[1] == self.channels if self.dims == 3: x = F.interpolate( x, (x.shape[2], x.shape[3] * 2, x.shape[4] * 2), mode="nearest" ) else: x = F.interpolate(x, scale_factor=2, mode="nearest") if self.use_conv: x = self.conv(x) return x class Downsample(nn.Module): """ A downsampling layer with an optional convolution. :param channels: channels in the inputs and outputs. :param use_conv: a bool determining if a convolution is applied. :param dims: determines if the signal is 1D, 2D, or 3D. If 3D, then downsampling occurs in the inner-two dimensions. """ def __init__(self, channels, use_conv, dims=2, out_channels=None): super().__init__() self.channels = channels self.out_channels = out_channels or channels self.use_conv = use_conv self.dims = dims stride = 2 if dims != 3 else (1, 2, 2) if use_conv: self.op = conv_nd( dims, self.channels, self.out_channels, 3, stride=stride, padding=1 ) else: assert self.channels == self.out_channels self.op = avg_pool_nd(dims, kernel_size=stride, stride=stride) def forward(self, x): assert x.shape[1] == self.channels return self.op(x) class ResBlock(TimestepBlock): """ A residual block that can optionally change the number of channels. :param channels: the number of input channels. :param emb_channels: the number of timestep embedding channels. :param dropout: the rate of dropout. :param out_channels: if specified, the number of out channels. :param use_conv: if True and out_channels is specified, use a spatial convolution instead of a smaller 1x1 convolution to change the channels in the skip connection. :param dims: determines if the signal is 1D, 2D, or 3D. :param use_checkpoint: if True, use gradient checkpointing on this module. :param up: if True, use this block for upsampling. :param down: if True, use this block for downsampling. """ def __init__( self, channels, emb_channels, dropout, out_channels=None, use_conv=False, use_scale_shift_norm=False, dims=2, use_checkpoint=False, up=False, down=False, ): super().__init__() self.channels = channels self.emb_channels = emb_channels self.dropout = dropout self.out_channels = out_channels or channels self.use_conv = use_conv self.use_checkpoint = use_checkpoint self.use_scale_shift_norm = use_scale_shift_norm self.in_layers = nn.Sequential( normalization(channels), nn.SiLU(), conv_nd(dims, channels, self.out_channels, 3, padding=1), ) self.updown = up or down if up: self.h_upd = Upsample(channels, False, dims) self.x_upd = Upsample(channels, False, dims) elif down: self.h_upd = Downsample(channels, False, dims) self.x_upd = Downsample(channels, False, dims) else: self.h_upd = self.x_upd = nn.Identity() self.emb_layers = nn.Sequential( nn.SiLU(), linear( emb_channels, 2 * self.out_channels if use_scale_shift_norm else self.out_channels, ), ) self.out_layers = nn.Sequential( normalization(self.out_channels), nn.SiLU(), nn.Dropout(p=dropout), zero_module( conv_nd(dims, self.out_channels, self.out_channels, 3, padding=1) ), ) if self.out_channels == channels: self.skip_connection = nn.Identity() elif use_conv: self.skip_connection = conv_nd( dims, channels, self.out_channels, 3, padding=1 ) else: self.skip_connection = conv_nd(dims, channels, self.out_channels, 1) def forward(self, x, emb): """ Apply the block to a Tensor, conditioned on a timestep embedding. :param x: an [N x C x ...] Tensor of features. :param emb: an [N x emb_channels] Tensor of timestep embeddings. :return: an [N x C x ...] Tensor of outputs. """ return checkpoint( self._forward, (x, emb), self.parameters(), self.use_checkpoint ) def _forward(self, x, emb): if self.updown: in_rest, in_conv = self.in_layers[:-1], self.in_layers[-1] h = in_rest(x) h = self.h_upd(h) x = self.x_upd(x) h = in_conv(h) else: h = self.in_layers(x) emb_out = self.emb_layers(emb).type(h.dtype) while len(emb_out.shape) < len(h.shape): emb_out = emb_out[..., None] if self.use_scale_shift_norm: out_norm, out_rest = self.out_layers[0], self.out_layers[1:] scale, shift = th.chunk(emb_out, 2, dim=1) h = out_norm(h) * (1 + scale) + shift h = out_rest(h) else: h = h + emb_out h = self.out_layers(h) return self.skip_connection(x) + h class AttentionBlock(nn.Module): """ An attention block that allows spatial positions to attend to each other. Originally ported from here, but adapted to the N-d case. https://github.com/hojonathanho/diffusion/blob/1e0dceb3b3495bbe19116a5e1b3596cd0706c543/diffusion_tf/models/unet.py#L66. """ def __init__( self, channels, num_heads=1, num_head_channels=-1, use_checkpoint=False, use_new_attention_order=False, ): super().__init__() self.channels = channels if num_head_channels == -1: self.num_heads = num_heads else: assert ( channels % num_head_channels == 0 ), f"q,k,v channels {channels} is not divisible by num_head_channels {num_head_channels}" self.num_heads = channels // num_head_channels self.use_checkpoint = use_checkpoint self.norm = normalization(channels) self.qkv = conv_nd(1, channels, channels * 3, 1) if use_new_attention_order: # split qkv before split heads self.attention = QKVAttention(self.num_heads) else: # split heads before split qkv self.attention = QKVAttentionLegacy(self.num_heads) self.proj_out = zero_module(conv_nd(1, channels, channels, 1)) def forward(self, x): return checkpoint(self._forward, (x,), self.parameters(), True) def _forward(self, x): b, c, *spatial = x.shape x = x.reshape(b, c, -1) qkv = self.qkv(self.norm(x)) h = self.attention(qkv) h = self.proj_out(h) return (x + h).reshape(b, c, *spatial) def count_flops_attn(model, _x, y): """ A counter for the `thop` package to count the operations in an attention operation. Meant to be used like: macs, params = thop.profile( model, inputs=(inputs, timestamps), custom_ops={QKVAttention: QKVAttention.count_flops}, ) """ b, c, *spatial = y[0].shape num_spatial = int(np.prod(spatial)) # We perform two matmuls with the same number of ops. # The first computes the weight matrix, the second computes # the combination of the value vectors. matmul_ops = 2 * b * (num_spatial ** 2) * c model.total_ops += th.DoubleTensor([matmul_ops]) class QKVAttentionLegacy(nn.Module): """ A module which performs QKV attention. Matches legacy QKVAttention + input/ouput heads shaping """ def __init__(self, n_heads): super().__init__() self.n_heads = n_heads def forward(self, qkv): """ Apply QKV attention. :param qkv: an [N x (H * 3 * C) x T] tensor of Qs, Ks, and Vs. :return: an [N x (H * C) x T] tensor after attention. """ bs, width, length = qkv.shape assert width % (3 * self.n_heads) == 0 ch = width // (3 * self.n_heads) q, k, v = qkv.reshape(bs * self.n_heads, ch * 3, length).split(ch, dim=1) scale = 1 / math.sqrt(math.sqrt(ch)) weight = th.einsum( "bct,bcs->bts", q * scale, k * scale ) # More stable with f16 than dividing afterwards weight = th.softmax(weight.float(), dim=-1).type(weight.dtype) a = th.einsum("bts,bcs->bct", weight, v) return a.reshape(bs, -1, length) @staticmethod def count_flops(model, _x, y): return count_flops_attn(model, _x, y) class QKVAttention(nn.Module): """ A module which performs QKV attention and splits in a different order. """ def __init__(self, n_heads): super().__init__() self.n_heads = n_heads def forward(self, qkv): """ Apply QKV attention. :param qkv: an [N x (3 * H * C) x T] tensor of Qs, Ks, and Vs. :return: an [N x (H * C) x T] tensor after attention. """ bs, width, length = qkv.shape assert width % (3 * self.n_heads) == 0 ch = width // (3 * self.n_heads) q, k, v = qkv.chunk(3, dim=1) scale = 1 / math.sqrt(math.sqrt(ch)) weight = th.einsum( "bct,bcs->bts", (q * scale).view(bs * self.n_heads, ch, length), (k * scale).view(bs * self.n_heads, ch, length), ) # More stable with f16 than dividing afterwards weight = th.softmax(weight.float(), dim=-1).type(weight.dtype) a = th.einsum("bts,bcs->bct", weight, v.reshape(bs * self.n_heads, ch, length)) return a.reshape(bs, -1, length) @staticmethod def count_flops(model, _x, y): return count_flops_attn(model, _x, y) class UNetModel(nn.Module): """ The full UNet model with attention and timestep embedding. :param in_channels: channels in the input Tensor. :param model_channels: base channel count for the model. :param out_channels: channels in the output Tensor. :param num_res_blocks: number of residual blocks per downsample. :param attention_resolutions: a collection of downsample rates at which attention will take place. May be a set, list, or tuple. For example, if this contains 4, then at 4x downsampling, attention will be used. :param dropout: the dropout probability. :param channel_mult: channel multiplier for each level of the UNet. :param conv_resample: if True, use learned convolutions for upsampling and downsampling. :param dims: determines if the signal is 1D, 2D, or 3D. :param num_classes: if specified (as an int), then this model will be class-conditional with `num_classes` classes. :param use_checkpoint: use gradient checkpointing to reduce memory usage. :param num_heads: the number of attention heads in each attention layer. :param num_heads_channels: if specified, ignore num_heads and instead use a fixed channel width per attention head. :param num_heads_upsample: works with num_heads to set a different number of heads for upsampling. Deprecated. :param use_scale_shift_norm: use a FiLM-like conditioning mechanism. :param resblock_updown: use residual blocks for up/downsampling. :param use_new_attention_order: use a different attention pattern for potentially increased efficiency. """ def __init__( self, image_size, in_channels, model_channels, out_channels, num_res_blocks, attention_resolutions, dropout=0, channel_mult=(1, 2, 4, 8), conv_resample=True, dims=2, num_classes=None, use_checkpoint=False, use_fp16=False, num_heads=1, num_head_channels=-1, num_heads_upsample=-1, use_scale_shift_norm=False, resblock_updown=False, use_new_attention_order=False, ): super().__init__() if num_heads_upsample == -1: num_heads_upsample = num_heads self.image_size = image_size self.in_channels = in_channels self.model_channels = model_channels self.out_channels = out_channels self.num_res_blocks = num_res_blocks self.attention_resolutions = attention_resolutions self.dropout = dropout self.channel_mult = channel_mult self.conv_resample = conv_resample self.num_classes = num_classes self.use_checkpoint = use_checkpoint self.dtype = th.float16 if use_fp16 else th.float32 self.num_heads = num_heads self.num_head_channels = num_head_channels self.num_heads_upsample = num_heads_upsample time_embed_dim = model_channels * 4 self.time_embed = nn.Sequential( linear(model_channels, time_embed_dim), nn.SiLU(), linear(time_embed_dim, time_embed_dim), ) if self.num_classes is not None: self.label_emb = nn.Embedding(num_classes, time_embed_dim) ch = input_ch = int(channel_mult[0] * model_channels) self.input_blocks = nn.ModuleList( [TimestepEmbedSequential(conv_nd(dims, in_channels, ch, 3, padding=1))] ) self._feature_size = ch input_block_chans = [ch] ds = 1 for level, mult in enumerate(channel_mult): for _ in range(num_res_blocks): layers = [ ResBlock( ch, time_embed_dim, dropout, out_channels=int(mult * model_channels), dims=dims, use_checkpoint=use_checkpoint, use_scale_shift_norm=use_scale_shift_norm, ) ] ch = int(mult * model_channels) if ds in attention_resolutions: layers.append( AttentionBlock( ch, use_checkpoint=use_checkpoint, num_heads=num_heads, num_head_channels=num_head_channels, use_new_attention_order=use_new_attention_order, ) ) self.input_blocks.append(TimestepEmbedSequential(*layers)) self._feature_size += ch input_block_chans.append(ch) if level != len(channel_mult) - 1: out_ch = ch self.input_blocks.append( TimestepEmbedSequential( ResBlock( ch, time_embed_dim, dropout, out_channels=out_ch, dims=dims, use_checkpoint=use_checkpoint, use_scale_shift_norm=use_scale_shift_norm, down=True, ) if resblock_updown else Downsample( ch, conv_resample, dims=dims, out_channels=out_ch ) ) ) ch = out_ch input_block_chans.append(ch) ds *= 2 self._feature_size += ch self.middle_block = TimestepEmbedSequential( ResBlock( ch, time_embed_dim, dropout, dims=dims, use_checkpoint=use_checkpoint, use_scale_shift_norm=use_scale_shift_norm, ), AttentionBlock( ch, use_checkpoint=use_checkpoint, num_heads=num_heads, num_head_channels=num_head_channels, use_new_attention_order=use_new_attention_order, ), ResBlock( ch, time_embed_dim, dropout, dims=dims, use_checkpoint=use_checkpoint, use_scale_shift_norm=use_scale_shift_norm, ), ) self._feature_size += ch self.output_blocks = nn.ModuleList([]) for level, mult in list(enumerate(channel_mult))[::-1]: for i in range(num_res_blocks + 1): ich = input_block_chans.pop() layers = [ ResBlock( ch + ich, time_embed_dim, dropout, out_channels=int(model_channels * mult), dims=dims, use_checkpoint=use_checkpoint, use_scale_shift_norm=use_scale_shift_norm, ) ] ch = int(model_channels * mult) if ds in attention_resolutions: layers.append( AttentionBlock( ch, use_checkpoint=use_checkpoint, num_heads=num_heads_upsample, num_head_channels=num_head_channels, use_new_attention_order=use_new_attention_order, ) ) if level and i == num_res_blocks: out_ch = ch layers.append( ResBlock( ch, time_embed_dim, dropout, out_channels=out_ch, dims=dims, use_checkpoint=use_checkpoint, use_scale_shift_norm=use_scale_shift_norm, up=True, ) if resblock_updown else Upsample(ch, conv_resample, dims=dims, out_channels=out_ch) ) ds //= 2 self.output_blocks.append(TimestepEmbedSequential(*layers)) self._feature_size += ch self.out = nn.Sequential( normalization(ch), nn.SiLU(), zero_module(conv_nd(dims, input_ch, out_channels, 3, padding=1)), ) def convert_to_fp16(self): """ Convert the torso of the model to float16. """ self.input_blocks.apply(convert_module_to_f16) self.middle_block.apply(convert_module_to_f16) self.output_blocks.apply(convert_module_to_f16) def convert_to_fp32(self): """ Convert the torso of the model to float32. """ self.input_blocks.apply(convert_module_to_f32) self.middle_block.apply(convert_module_to_f32) self.output_blocks.apply(convert_module_to_f32) def forward(self, x, timesteps, y=None): """ Apply the model to an input batch. :param x: an [N x C x ...] Tensor of inputs. :param timesteps: a 1-D batch of timesteps. :param y: an [N] Tensor of labels, if class-conditional. :return: an [N x C x ...] Tensor of outputs. """ assert (y is not None) == ( self.num_classes is not None ), "must specify y if and only if the model is class-conditional" hs = [] emb = self.time_embed(timestep_embedding(timesteps, self.model_channels)) if self.num_classes is not None: assert y.shape == (x.shape[0],) emb = emb + self.label_emb(y) h = x.type(self.dtype) for module in self.input_blocks: h = module(h, emb) hs.append(h) h = self.middle_block(h, emb) for module in self.output_blocks: h = th.cat([h, hs.pop()], dim=1) h = module(h, emb) h = h.type(x.dtype) return self.out(h) class SuperResModel(UNetModel): """ A UNetModel that performs super-resolution. Expects an extra kwarg `low_res` to condition on a low-resolution image. """ def __init__(self, image_size, in_channels, *args, **kwargs): super().__init__(image_size, in_channels * 2, *args, **kwargs) def forward(self, x, timesteps, low_res=None, **kwargs): _, _, new_height, new_width = x.shape upsampled = F.interpolate(low_res, (new_height, new_width), mode="bilinear") x = th.cat([x, upsampled], dim=1) return super().forward(x, timesteps, **kwargs) class EncoderUNetModel(nn.Module): """ The half UNet model with attention and timestep embedding. For usage, see UNet. """ def __init__( self, image_size, in_channels, model_channels, out_channels, num_res_blocks, attention_resolutions, dropout=0, channel_mult=(1, 2, 4, 8), conv_resample=True, dims=2, use_checkpoint=False, use_fp16=False, num_heads=1, num_head_channels=-1, num_heads_upsample=-1, use_scale_shift_norm=False, resblock_updown=False, use_new_attention_order=False, pool="adaptive", ): super().__init__() if num_heads_upsample == -1: num_heads_upsample = num_heads self.in_channels = in_channels self.model_channels = model_channels self.out_channels = out_channels self.num_res_blocks = num_res_blocks self.attention_resolutions = attention_resolutions self.dropout = dropout self.channel_mult = channel_mult self.conv_resample = conv_resample self.use_checkpoint = use_checkpoint self.dtype = th.float16 if use_fp16 else th.float32 self.num_heads = num_heads self.num_head_channels = num_head_channels self.num_heads_upsample = num_heads_upsample time_embed_dim = model_channels * 4 self.time_embed = nn.Sequential( linear(model_channels, time_embed_dim), nn.SiLU(), linear(time_embed_dim, time_embed_dim), ) ch = int(channel_mult[0] * model_channels) self.input_blocks = nn.ModuleList( [TimestepEmbedSequential(conv_nd(dims, in_channels, ch, 3, padding=1))] ) self._feature_size = ch input_block_chans = [ch] ds = 1 for level, mult in enumerate(channel_mult): for _ in range(num_res_blocks): layers = [ ResBlock( ch, time_embed_dim, dropout, out_channels=int(mult * model_channels), dims=dims, use_checkpoint=use_checkpoint, use_scale_shift_norm=use_scale_shift_norm, ) ] ch = int(mult * model_channels) if ds in attention_resolutions: layers.append( AttentionBlock( ch, use_checkpoint=use_checkpoint, num_heads=num_heads, num_head_channels=num_head_channels, use_new_attention_order=use_new_attention_order, ) ) self.input_blocks.append(TimestepEmbedSequential(*layers)) self._feature_size += ch input_block_chans.append(ch) if level != len(channel_mult) - 1: out_ch = ch self.input_blocks.append( TimestepEmbedSequential( ResBlock( ch, time_embed_dim, dropout, out_channels=out_ch, dims=dims, use_checkpoint=use_checkpoint, use_scale_shift_norm=use_scale_shift_norm, down=True, ) if resblock_updown else Downsample( ch, conv_resample, dims=dims, out_channels=out_ch ) ) ) ch = out_ch input_block_chans.append(ch) ds *= 2 self._feature_size += ch self.middle_block = TimestepEmbedSequential( ResBlock( ch, time_embed_dim, dropout, dims=dims, use_checkpoint=use_checkpoint, use_scale_shift_norm=use_scale_shift_norm, ), AttentionBlock( ch, use_checkpoint=use_checkpoint, num_heads=num_heads, num_head_channels=num_head_channels, use_new_attention_order=use_new_attention_order, ), ResBlock( ch, time_embed_dim, dropout, dims=dims, use_checkpoint=use_checkpoint, use_scale_shift_norm=use_scale_shift_norm, ), ) self._feature_size += ch self.pool = pool if pool == "adaptive": self.out = nn.Sequential( normalization(ch), nn.SiLU(), nn.AdaptiveAvgPool2d((1, 1)), zero_module(conv_nd(dims, ch, out_channels, 1)), nn.Flatten(), ) elif pool == "attention": assert num_head_channels != -1 self.out = nn.Sequential( normalization(ch), nn.SiLU(), AttentionPool2d( (image_size // ds), ch, num_head_channels, out_channels ), ) elif pool == "spatial": self.out = nn.Sequential( nn.Linear(self._feature_size, 2048), nn.ReLU(), nn.Linear(2048, self.out_channels), ) elif pool == "spatial_v2": self.out = nn.Sequential( nn.Linear(self._feature_size, 2048), normalization(2048), nn.SiLU(), nn.Linear(2048, self.out_channels), ) else: raise NotImplementedError(f"Unexpected {pool} pooling") def convert_to_fp16(self): """ Convert the torso of the model to float16. """ self.input_blocks.apply(convert_module_to_f16) self.middle_block.apply(convert_module_to_f16) def convert_to_fp32(self): """ Convert the torso of the model to float32. """ self.input_blocks.apply(convert_module_to_f32) self.middle_block.apply(convert_module_to_f32) def forward(self, x, timesteps): """ Apply the model to an input batch. :param x: an [N x C x ...] Tensor of inputs. :param timesteps: a 1-D batch of timesteps. :return: an [N x K] Tensor of outputs. """ emb = self.time_embed(timestep_embedding(timesteps, self.model_channels)) results = [] h = x.type(self.dtype) for module in self.input_blocks: h = module(h, emb) if self.pool.startswith("spatial"): results.append(h.type(x.dtype).mean(dim=(2, 3))) h = self.middle_block(h, emb) if self.pool.startswith("spatial"): results.append(h.type(x.dtype).mean(dim=(2, 3))) h = th.cat(results, axis=-1) return self.out(h) else: h = h.type(x.dtype) return self.out(h)
31,605
34.001107
124
py
DiffPure
DiffPure-master/guided_diffusion/gaussian_diffusion.py
# --------------------------------------------------------------- # Taken from the following link as is from: # https://github.com/openai/guided-diffusion/blob/main/guided_diffusion/gaussian_diffusion.py # # The license for the original version of this file can be # found in this directory (LICENSE_GUIDED_DIFFUSION). # --------------------------------------------------------------- """ This code started out as a PyTorch port of Ho et al's diffusion models: https://github.com/hojonathanho/diffusion/blob/1e0dceb3b3495bbe19116a5e1b3596cd0706c543/diffusion_tf/diffusion_utils_2.py Docstrings have been added, as well as DDIM sampling and a new collection of beta schedules. """ import enum import math import numpy as np import torch as th from .nn import mean_flat from .losses import normal_kl, discretized_gaussian_log_likelihood def get_named_beta_schedule(schedule_name, num_diffusion_timesteps): """ Get a pre-defined beta schedule for the given name. The beta schedule library consists of beta schedules which remain similar in the limit of num_diffusion_timesteps. Beta schedules may be added, but should not be removed or changed once they are committed to maintain backwards compatibility. """ if schedule_name == "linear": # Linear schedule from Ho et al, extended to work for any number of # diffusion steps. scale = 1000 / num_diffusion_timesteps beta_start = scale * 0.0001 beta_end = scale * 0.02 return np.linspace( beta_start, beta_end, num_diffusion_timesteps, dtype=np.float64 ) elif schedule_name == "cosine": return betas_for_alpha_bar( num_diffusion_timesteps, lambda t: math.cos((t + 0.008) / 1.008 * math.pi / 2) ** 2, ) else: raise NotImplementedError(f"unknown beta schedule: {schedule_name}") def betas_for_alpha_bar(num_diffusion_timesteps, alpha_bar, max_beta=0.999): """ Create a beta schedule that discretizes the given alpha_t_bar function, which defines the cumulative product of (1-beta) over time from t = [0,1]. :param num_diffusion_timesteps: the number of betas to produce. :param alpha_bar: a lambda that takes an argument t from 0 to 1 and produces the cumulative product of (1-beta) up to that part of the diffusion process. :param max_beta: the maximum beta to use; use values lower than 1 to prevent singularities. """ betas = [] for i in range(num_diffusion_timesteps): t1 = i / num_diffusion_timesteps t2 = (i + 1) / num_diffusion_timesteps betas.append(min(1 - alpha_bar(t2) / alpha_bar(t1), max_beta)) return np.array(betas) class ModelMeanType(enum.Enum): """ Which type of output the model predicts. """ PREVIOUS_X = enum.auto() # the model predicts x_{t-1} START_X = enum.auto() # the model predicts x_0 EPSILON = enum.auto() # the model predicts epsilon class ModelVarType(enum.Enum): """ What is used as the model's output variance. The LEARNED_RANGE option has been added to allow the model to predict values between FIXED_SMALL and FIXED_LARGE, making its job easier. """ LEARNED = enum.auto() FIXED_SMALL = enum.auto() FIXED_LARGE = enum.auto() LEARNED_RANGE = enum.auto() class LossType(enum.Enum): MSE = enum.auto() # use raw MSE loss (and KL when learning variances) RESCALED_MSE = ( enum.auto() ) # use raw MSE loss (with RESCALED_KL when learning variances) KL = enum.auto() # use the variational lower-bound RESCALED_KL = enum.auto() # like KL, but rescale to estimate the full VLB def is_vb(self): return self == LossType.KL or self == LossType.RESCALED_KL class GaussianDiffusion: """ Utilities for training and sampling diffusion models. Ported directly from here, and then adapted over time to further experimentation. https://github.com/hojonathanho/diffusion/blob/1e0dceb3b3495bbe19116a5e1b3596cd0706c543/diffusion_tf/diffusion_utils_2.py#L42 :param betas: a 1-D numpy array of betas for each diffusion timestep, starting at T and going to 1. :param model_mean_type: a ModelMeanType determining what the model outputs. :param model_var_type: a ModelVarType determining how variance is output. :param loss_type: a LossType determining the loss function to use. :param rescale_timesteps: if True, pass floating point timesteps into the model so that they are always scaled like in the original paper (0 to 1000). """ def __init__( self, *, betas, model_mean_type, model_var_type, loss_type, rescale_timesteps=False, ): self.model_mean_type = model_mean_type self.model_var_type = model_var_type self.loss_type = loss_type self.rescale_timesteps = rescale_timesteps # Use float64 for accuracy. betas = np.array(betas, dtype=np.float64) self.betas = betas assert len(betas.shape) == 1, "betas must be 1-D" assert (betas > 0).all() and (betas <= 1).all() self.num_timesteps = int(betas.shape[0]) alphas = 1.0 - betas self.alphas_cumprod = np.cumprod(alphas, axis=0) self.alphas_cumprod_prev = np.append(1.0, self.alphas_cumprod[:-1]) self.alphas_cumprod_next = np.append(self.alphas_cumprod[1:], 0.0) assert self.alphas_cumprod_prev.shape == (self.num_timesteps,) # calculations for diffusion q(x_t | x_{t-1}) and others self.sqrt_alphas_cumprod = np.sqrt(self.alphas_cumprod) self.sqrt_one_minus_alphas_cumprod = np.sqrt(1.0 - self.alphas_cumprod) self.log_one_minus_alphas_cumprod = np.log(1.0 - self.alphas_cumprod) self.sqrt_recip_alphas_cumprod = np.sqrt(1.0 / self.alphas_cumprod) self.sqrt_recipm1_alphas_cumprod = np.sqrt(1.0 / self.alphas_cumprod - 1) # calculations for posterior q(x_{t-1} | x_t, x_0) self.posterior_variance = ( betas * (1.0 - self.alphas_cumprod_prev) / (1.0 - self.alphas_cumprod) ) # log calculation clipped because the posterior variance is 0 at the # beginning of the diffusion chain. self.posterior_log_variance_clipped = np.log( np.append(self.posterior_variance[1], self.posterior_variance[1:]) ) self.posterior_mean_coef1 = ( betas * np.sqrt(self.alphas_cumprod_prev) / (1.0 - self.alphas_cumprod) ) self.posterior_mean_coef2 = ( (1.0 - self.alphas_cumprod_prev) * np.sqrt(alphas) / (1.0 - self.alphas_cumprod) ) def q_mean_variance(self, x_start, t): """ Get the distribution q(x_t | x_0). :param x_start: the [N x C x ...] tensor of noiseless inputs. :param t: the number of diffusion steps (minus 1). Here, 0 means one step. :return: A tuple (mean, variance, log_variance), all of x_start's shape. """ mean = ( _extract_into_tensor(self.sqrt_alphas_cumprod, t, x_start.shape) * x_start ) variance = _extract_into_tensor(1.0 - self.alphas_cumprod, t, x_start.shape) log_variance = _extract_into_tensor( self.log_one_minus_alphas_cumprod, t, x_start.shape ) return mean, variance, log_variance def q_sample(self, x_start, t, noise=None): """ Diffuse the data for a given number of diffusion steps. In other words, sample from q(x_t | x_0). :param x_start: the initial data batch. :param t: the number of diffusion steps (minus 1). Here, 0 means one step. :param noise: if specified, the split-out normal noise. :return: A noisy version of x_start. """ if noise is None: noise = th.randn_like(x_start) assert noise.shape == x_start.shape return ( _extract_into_tensor(self.sqrt_alphas_cumprod, t, x_start.shape) * x_start + _extract_into_tensor(self.sqrt_one_minus_alphas_cumprod, t, x_start.shape) * noise ) def q_posterior_mean_variance(self, x_start, x_t, t): """ Compute the mean and variance of the diffusion posterior: q(x_{t-1} | x_t, x_0) """ assert x_start.shape == x_t.shape posterior_mean = ( _extract_into_tensor(self.posterior_mean_coef1, t, x_t.shape) * x_start + _extract_into_tensor(self.posterior_mean_coef2, t, x_t.shape) * x_t ) posterior_variance = _extract_into_tensor(self.posterior_variance, t, x_t.shape) posterior_log_variance_clipped = _extract_into_tensor( self.posterior_log_variance_clipped, t, x_t.shape ) assert ( posterior_mean.shape[0] == posterior_variance.shape[0] == posterior_log_variance_clipped.shape[0] == x_start.shape[0] ) return posterior_mean, posterior_variance, posterior_log_variance_clipped def p_mean_variance( self, model, x, t, clip_denoised=True, denoised_fn=None, model_kwargs=None ): """ Apply the model to get p(x_{t-1} | x_t), as well as a prediction of the initial x, x_0. :param model: the model, which takes a signal and a batch of timesteps as input. :param x: the [N x C x ...] tensor at time t. :param t: a 1-D Tensor of timesteps. :param clip_denoised: if True, clip the denoised signal into [-1, 1]. :param denoised_fn: if not None, a function which applies to the x_start prediction before it is used to sample. Applies before clip_denoised. :param model_kwargs: if not None, a dict of extra keyword arguments to pass to the model. This can be used for conditioning. :return: a dict with the following keys: - 'mean': the model mean output. - 'variance': the model variance output. - 'log_variance': the log of 'variance'. - 'pred_xstart': the prediction for x_0. """ if model_kwargs is None: model_kwargs = {} B, C = x.shape[:2] assert t.shape == (B,) model_output = model(x, self._scale_timesteps(t), **model_kwargs) if self.model_var_type in [ModelVarType.LEARNED, ModelVarType.LEARNED_RANGE]: assert model_output.shape == (B, C * 2, *x.shape[2:]) model_output, model_var_values = th.split(model_output, C, dim=1) if self.model_var_type == ModelVarType.LEARNED: model_log_variance = model_var_values model_variance = th.exp(model_log_variance) else: min_log = _extract_into_tensor( self.posterior_log_variance_clipped, t, x.shape ) max_log = _extract_into_tensor(np.log(self.betas), t, x.shape) # The model_var_values is [-1, 1] for [min_var, max_var]. frac = (model_var_values + 1) / 2 model_log_variance = frac * max_log + (1 - frac) * min_log model_variance = th.exp(model_log_variance) else: model_variance, model_log_variance = { # for fixedlarge, we set the initial (log-)variance like so # to get a better decoder log likelihood. ModelVarType.FIXED_LARGE: ( np.append(self.posterior_variance[1], self.betas[1:]), np.log(np.append(self.posterior_variance[1], self.betas[1:])), ), ModelVarType.FIXED_SMALL: ( self.posterior_variance, self.posterior_log_variance_clipped, ), }[self.model_var_type] model_variance = _extract_into_tensor(model_variance, t, x.shape) model_log_variance = _extract_into_tensor(model_log_variance, t, x.shape) def process_xstart(x): if denoised_fn is not None: x = denoised_fn(x) if clip_denoised: return x.clamp(-1, 1) return x if self.model_mean_type == ModelMeanType.PREVIOUS_X: pred_xstart = process_xstart( self._predict_xstart_from_xprev(x_t=x, t=t, xprev=model_output) ) model_mean = model_output elif self.model_mean_type in [ModelMeanType.START_X, ModelMeanType.EPSILON]: if self.model_mean_type == ModelMeanType.START_X: pred_xstart = process_xstart(model_output) else: pred_xstart = process_xstart( self._predict_xstart_from_eps(x_t=x, t=t, eps=model_output) ) model_mean, _, _ = self.q_posterior_mean_variance( x_start=pred_xstart, x_t=x, t=t ) else: raise NotImplementedError(self.model_mean_type) assert ( model_mean.shape == model_log_variance.shape == pred_xstart.shape == x.shape ) return { "mean": model_mean, "variance": model_variance, "log_variance": model_log_variance, "pred_xstart": pred_xstart, } def _predict_xstart_from_eps(self, x_t, t, eps): assert x_t.shape == eps.shape return ( _extract_into_tensor(self.sqrt_recip_alphas_cumprod, t, x_t.shape) * x_t - _extract_into_tensor(self.sqrt_recipm1_alphas_cumprod, t, x_t.shape) * eps ) def _predict_xstart_from_xprev(self, x_t, t, xprev): assert x_t.shape == xprev.shape return ( # (xprev - coef2*x_t) / coef1 _extract_into_tensor(1.0 / self.posterior_mean_coef1, t, x_t.shape) * xprev - _extract_into_tensor( self.posterior_mean_coef2 / self.posterior_mean_coef1, t, x_t.shape ) * x_t ) def _predict_eps_from_xstart(self, x_t, t, pred_xstart): return ( _extract_into_tensor(self.sqrt_recip_alphas_cumprod, t, x_t.shape) * x_t - pred_xstart ) / _extract_into_tensor(self.sqrt_recipm1_alphas_cumprod, t, x_t.shape) def _scale_timesteps(self, t): if self.rescale_timesteps: return t.float() * (1000.0 / self.num_timesteps) return t def condition_mean(self, cond_fn, p_mean_var, x, t, model_kwargs=None): """ Compute the mean for the previous step, given a function cond_fn that computes the gradient of a conditional log probability with respect to x. In particular, cond_fn computes grad(log(p(y|x))), and we want to condition on y. This uses the conditioning strategy from Sohl-Dickstein et al. (2015). """ gradient = cond_fn(x, self._scale_timesteps(t), **model_kwargs) new_mean = ( p_mean_var["mean"].float() + p_mean_var["variance"] * gradient.float() ) return new_mean def condition_score(self, cond_fn, p_mean_var, x, t, model_kwargs=None): """ Compute what the p_mean_variance output would have been, should the model's score function be conditioned by cond_fn. See condition_mean() for details on cond_fn. Unlike condition_mean(), this instead uses the conditioning strategy from Song et al (2020). """ alpha_bar = _extract_into_tensor(self.alphas_cumprod, t, x.shape) eps = self._predict_eps_from_xstart(x, t, p_mean_var["pred_xstart"]) eps = eps - (1 - alpha_bar).sqrt() * cond_fn( x, self._scale_timesteps(t), **model_kwargs ) out = p_mean_var.copy() out["pred_xstart"] = self._predict_xstart_from_eps(x, t, eps) out["mean"], _, _ = self.q_posterior_mean_variance( x_start=out["pred_xstart"], x_t=x, t=t ) return out def p_sample( self, model, x, t, clip_denoised=True, denoised_fn=None, cond_fn=None, model_kwargs=None, ): """ Sample x_{t-1} from the model at the given timestep. :param model: the model to sample from. :param x: the current tensor at x_{t-1}. :param t: the value of t, starting at 0 for the first diffusion step. :param clip_denoised: if True, clip the x_start prediction to [-1, 1]. :param denoised_fn: if not None, a function which applies to the x_start prediction before it is used to sample. :param cond_fn: if not None, this is a gradient function that acts similarly to the model. :param model_kwargs: if not None, a dict of extra keyword arguments to pass to the model. This can be used for conditioning. :return: a dict containing the following keys: - 'sample': a random sample from the model. - 'pred_xstart': a prediction of x_0. """ out = self.p_mean_variance( model, x, t, clip_denoised=clip_denoised, denoised_fn=denoised_fn, model_kwargs=model_kwargs, ) noise = th.randn_like(x) nonzero_mask = ( (t != 0).float().view(-1, *([1] * (len(x.shape) - 1))) ) # no noise when t == 0 if cond_fn is not None: out["mean"] = self.condition_mean( cond_fn, out, x, t, model_kwargs=model_kwargs ) sample = out["mean"] + nonzero_mask * th.exp(0.5 * out["log_variance"]) * noise return {"sample": sample, "pred_xstart": out["pred_xstart"]} def p_sample_loop( self, model, shape, noise=None, clip_denoised=True, denoised_fn=None, cond_fn=None, model_kwargs=None, device=None, progress=False, ): """ Generate samples from the model. :param model: the model module. :param shape: the shape of the samples, (N, C, H, W). :param noise: if specified, the noise from the encoder to sample. Should be of the same shape as `shape`. :param clip_denoised: if True, clip x_start predictions to [-1, 1]. :param denoised_fn: if not None, a function which applies to the x_start prediction before it is used to sample. :param cond_fn: if not None, this is a gradient function that acts similarly to the model. :param model_kwargs: if not None, a dict of extra keyword arguments to pass to the model. This can be used for conditioning. :param device: if specified, the device to create the samples on. If not specified, use a model parameter's device. :param progress: if True, show a tqdm progress bar. :return: a non-differentiable batch of samples. """ final = None for sample in self.p_sample_loop_progressive( model, shape, noise=noise, clip_denoised=clip_denoised, denoised_fn=denoised_fn, cond_fn=cond_fn, model_kwargs=model_kwargs, device=device, progress=progress, ): final = sample return final["sample"] def p_sample_loop_progressive( self, model, shape, noise=None, clip_denoised=True, denoised_fn=None, cond_fn=None, model_kwargs=None, device=None, progress=False, ): """ Generate samples from the model and yield intermediate samples from each timestep of diffusion. Arguments are the same as p_sample_loop(). Returns a generator over dicts, where each dict is the return value of p_sample(). """ if device is None: device = next(model.parameters()).device assert isinstance(shape, (tuple, list)) if noise is not None: img = noise else: img = th.randn(*shape, device=device) indices = list(range(self.num_timesteps))[::-1] if progress: # Lazy import so that we don't depend on tqdm. from tqdm.auto import tqdm indices = tqdm(indices) for i in indices: t = th.tensor([i] * shape[0], device=device) with th.no_grad(): out = self.p_sample( model, img, t, clip_denoised=clip_denoised, denoised_fn=denoised_fn, cond_fn=cond_fn, model_kwargs=model_kwargs, ) yield out img = out["sample"] def ddim_sample( self, model, x, t, clip_denoised=True, denoised_fn=None, cond_fn=None, model_kwargs=None, eta=0.0, ): """ Sample x_{t-1} from the model using DDIM. Same usage as p_sample(). """ out = self.p_mean_variance( model, x, t, clip_denoised=clip_denoised, denoised_fn=denoised_fn, model_kwargs=model_kwargs, ) if cond_fn is not None: out = self.condition_score(cond_fn, out, x, t, model_kwargs=model_kwargs) # Usually our model outputs epsilon, but we re-derive it # in case we used x_start or x_prev prediction. eps = self._predict_eps_from_xstart(x, t, out["pred_xstart"]) alpha_bar = _extract_into_tensor(self.alphas_cumprod, t, x.shape) alpha_bar_prev = _extract_into_tensor(self.alphas_cumprod_prev, t, x.shape) sigma = ( eta * th.sqrt((1 - alpha_bar_prev) / (1 - alpha_bar)) * th.sqrt(1 - alpha_bar / alpha_bar_prev) ) # Equation 12. noise = th.randn_like(x) mean_pred = ( out["pred_xstart"] * th.sqrt(alpha_bar_prev) + th.sqrt(1 - alpha_bar_prev - sigma ** 2) * eps ) nonzero_mask = ( (t != 0).float().view(-1, *([1] * (len(x.shape) - 1))) ) # no noise when t == 0 sample = mean_pred + nonzero_mask * sigma * noise return {"sample": sample, "pred_xstart": out["pred_xstart"]} def ddim_reverse_sample( self, model, x, t, clip_denoised=True, denoised_fn=None, model_kwargs=None, eta=0.0, ): """ Sample x_{t+1} from the model using DDIM reverse ODE. """ assert eta == 0.0, "Reverse ODE only for deterministic path" out = self.p_mean_variance( model, x, t, clip_denoised=clip_denoised, denoised_fn=denoised_fn, model_kwargs=model_kwargs, ) # Usually our model outputs epsilon, but we re-derive it # in case we used x_start or x_prev prediction. eps = ( _extract_into_tensor(self.sqrt_recip_alphas_cumprod, t, x.shape) * x - out["pred_xstart"] ) / _extract_into_tensor(self.sqrt_recipm1_alphas_cumprod, t, x.shape) alpha_bar_next = _extract_into_tensor(self.alphas_cumprod_next, t, x.shape) # Equation 12. reversed mean_pred = ( out["pred_xstart"] * th.sqrt(alpha_bar_next) + th.sqrt(1 - alpha_bar_next) * eps ) return {"sample": mean_pred, "pred_xstart": out["pred_xstart"]} def ddim_sample_loop( self, model, shape, noise=None, clip_denoised=True, denoised_fn=None, cond_fn=None, model_kwargs=None, device=None, progress=False, eta=0.0, ): """ Generate samples from the model using DDIM. Same usage as p_sample_loop(). """ final = None for sample in self.ddim_sample_loop_progressive( model, shape, noise=noise, clip_denoised=clip_denoised, denoised_fn=denoised_fn, cond_fn=cond_fn, model_kwargs=model_kwargs, device=device, progress=progress, eta=eta, ): final = sample return final["sample"] def ddim_sample_loop_progressive( self, model, shape, noise=None, clip_denoised=True, denoised_fn=None, cond_fn=None, model_kwargs=None, device=None, progress=False, eta=0.0, ): """ Use DDIM to sample from the model and yield intermediate samples from each timestep of DDIM. Same usage as p_sample_loop_progressive(). """ if device is None: device = next(model.parameters()).device assert isinstance(shape, (tuple, list)) if noise is not None: img = noise else: img = th.randn(*shape, device=device) indices = list(range(self.num_timesteps))[::-1] if progress: # Lazy import so that we don't depend on tqdm. from tqdm.auto import tqdm indices = tqdm(indices) for i in indices: t = th.tensor([i] * shape[0], device=device) with th.no_grad(): out = self.ddim_sample( model, img, t, clip_denoised=clip_denoised, denoised_fn=denoised_fn, cond_fn=cond_fn, model_kwargs=model_kwargs, eta=eta, ) yield out img = out["sample"] def _vb_terms_bpd( self, model, x_start, x_t, t, clip_denoised=True, model_kwargs=None ): """ Get a term for the variational lower-bound. The resulting units are bits (rather than nats, as one might expect). This allows for comparison to other papers. :return: a dict with the following keys: - 'output': a shape [N] tensor of NLLs or KLs. - 'pred_xstart': the x_0 predictions. """ true_mean, _, true_log_variance_clipped = self.q_posterior_mean_variance( x_start=x_start, x_t=x_t, t=t ) out = self.p_mean_variance( model, x_t, t, clip_denoised=clip_denoised, model_kwargs=model_kwargs ) kl = normal_kl( true_mean, true_log_variance_clipped, out["mean"], out["log_variance"] ) kl = mean_flat(kl) / np.log(2.0) decoder_nll = -discretized_gaussian_log_likelihood( x_start, means=out["mean"], log_scales=0.5 * out["log_variance"] ) assert decoder_nll.shape == x_start.shape decoder_nll = mean_flat(decoder_nll) / np.log(2.0) # At the first timestep return the decoder NLL, # otherwise return KL(q(x_{t-1}|x_t,x_0) || p(x_{t-1}|x_t)) output = th.where((t == 0), decoder_nll, kl) return {"output": output, "pred_xstart": out["pred_xstart"]} def training_losses(self, model, x_start, t, model_kwargs=None, noise=None): """ Compute training losses for a single timestep. :param model: the model to evaluate loss on. :param x_start: the [N x C x ...] tensor of inputs. :param t: a batch of timestep indices. :param model_kwargs: if not None, a dict of extra keyword arguments to pass to the model. This can be used for conditioning. :param noise: if specified, the specific Gaussian noise to try to remove. :return: a dict with the key "loss" containing a tensor of shape [N]. Some mean or variance settings may also have other keys. """ if model_kwargs is None: model_kwargs = {} if noise is None: noise = th.randn_like(x_start) x_t = self.q_sample(x_start, t, noise=noise) terms = {} if self.loss_type == LossType.KL or self.loss_type == LossType.RESCALED_KL: terms["loss"] = self._vb_terms_bpd( model=model, x_start=x_start, x_t=x_t, t=t, clip_denoised=False, model_kwargs=model_kwargs, )["output"] if self.loss_type == LossType.RESCALED_KL: terms["loss"] *= self.num_timesteps elif self.loss_type == LossType.MSE or self.loss_type == LossType.RESCALED_MSE: model_output = model(x_t, self._scale_timesteps(t), **model_kwargs) if self.model_var_type in [ ModelVarType.LEARNED, ModelVarType.LEARNED_RANGE, ]: B, C = x_t.shape[:2] assert model_output.shape == (B, C * 2, *x_t.shape[2:]) model_output, model_var_values = th.split(model_output, C, dim=1) # Learn the variance using the variational bound, but don't let # it affect our mean prediction. frozen_out = th.cat([model_output.detach(), model_var_values], dim=1) terms["vb"] = self._vb_terms_bpd( model=lambda *args, r=frozen_out: r, x_start=x_start, x_t=x_t, t=t, clip_denoised=False, )["output"] if self.loss_type == LossType.RESCALED_MSE: # Divide by 1000 for equivalence with initial implementation. # Without a factor of 1/1000, the VB term hurts the MSE term. terms["vb"] *= self.num_timesteps / 1000.0 target = { ModelMeanType.PREVIOUS_X: self.q_posterior_mean_variance( x_start=x_start, x_t=x_t, t=t )[0], ModelMeanType.START_X: x_start, ModelMeanType.EPSILON: noise, }[self.model_mean_type] assert model_output.shape == target.shape == x_start.shape terms["mse"] = mean_flat((target - model_output) ** 2) if "vb" in terms: terms["loss"] = terms["mse"] + terms["vb"] else: terms["loss"] = terms["mse"] else: raise NotImplementedError(self.loss_type) return terms def _prior_bpd(self, x_start): """ Get the prior KL term for the variational lower-bound, measured in bits-per-dim. This term can't be optimized, as it only depends on the encoder. :param x_start: the [N x C x ...] tensor of inputs. :return: a batch of [N] KL values (in bits), one per batch element. """ batch_size = x_start.shape[0] t = th.tensor([self.num_timesteps - 1] * batch_size, device=x_start.device) qt_mean, _, qt_log_variance = self.q_mean_variance(x_start, t) kl_prior = normal_kl( mean1=qt_mean, logvar1=qt_log_variance, mean2=0.0, logvar2=0.0 ) return mean_flat(kl_prior) / np.log(2.0) def calc_bpd_loop(self, model, x_start, clip_denoised=True, model_kwargs=None): """ Compute the entire variational lower-bound, measured in bits-per-dim, as well as other related quantities. :param model: the model to evaluate loss on. :param x_start: the [N x C x ...] tensor of inputs. :param clip_denoised: if True, clip denoised samples. :param model_kwargs: if not None, a dict of extra keyword arguments to pass to the model. This can be used for conditioning. :return: a dict containing the following keys: - total_bpd: the total variational lower-bound, per batch element. - prior_bpd: the prior term in the lower-bound. - vb: an [N x T] tensor of terms in the lower-bound. - xstart_mse: an [N x T] tensor of x_0 MSEs for each timestep. - mse: an [N x T] tensor of epsilon MSEs for each timestep. """ device = x_start.device batch_size = x_start.shape[0] vb = [] xstart_mse = [] mse = [] for t in list(range(self.num_timesteps))[::-1]: t_batch = th.tensor([t] * batch_size, device=device) noise = th.randn_like(x_start) x_t = self.q_sample(x_start=x_start, t=t_batch, noise=noise) # Calculate VLB term at the current timestep with th.no_grad(): out = self._vb_terms_bpd( model, x_start=x_start, x_t=x_t, t=t_batch, clip_denoised=clip_denoised, model_kwargs=model_kwargs, ) vb.append(out["output"]) xstart_mse.append(mean_flat((out["pred_xstart"] - x_start) ** 2)) eps = self._predict_eps_from_xstart(x_t, t_batch, out["pred_xstart"]) mse.append(mean_flat((eps - noise) ** 2)) vb = th.stack(vb, dim=1) xstart_mse = th.stack(xstart_mse, dim=1) mse = th.stack(mse, dim=1) prior_bpd = self._prior_bpd(x_start) total_bpd = vb.sum(dim=1) + prior_bpd return { "total_bpd": total_bpd, "prior_bpd": prior_bpd, "vb": vb, "xstart_mse": xstart_mse, "mse": mse, } def _extract_into_tensor(arr, timesteps, broadcast_shape): """ Extract values from a 1-D numpy array for a batch of indices. :param arr: the 1-D numpy array. :param timesteps: a tensor of indices into the array to extract. :param broadcast_shape: a larger shape of K dimensions with the batch dimension equal to the length of timesteps. :return: a tensor of shape [batch_size, 1, ...] where the shape has K dims. """ res = th.from_numpy(arr).to(device=timesteps.device)[timesteps].float() while len(res.shape) < len(broadcast_shape): res = res[..., None] return res.expand(broadcast_shape)
34,721
36.864776
129
py
DiffPure
DiffPure-master/guided_diffusion/train_util.py
# --------------------------------------------------------------- # Taken from the following link as is from: # https://github.com/openai/guided-diffusion/blob/main/guided_diffusion/train_util.py # # The license for the original version of this file can be # found in this directory (LICENSE_GUIDED_DIFFUSION). # --------------------------------------------------------------- import copy import functools import os import blobfile as bf import torch as th import torch.distributed as dist from torch.nn.parallel.distributed import DistributedDataParallel as DDP from torch.optim import AdamW from . import dist_util, logger from .fp16_util import MixedPrecisionTrainer from .nn import update_ema from .resample import LossAwareSampler, UniformSampler # For ImageNet experiments, this was a good default value. # We found that the lg_loss_scale quickly climbed to # 20-21 within the first ~1K steps of training. INITIAL_LOG_LOSS_SCALE = 20.0 class TrainLoop: def __init__( self, *, model, diffusion, data, batch_size, microbatch, lr, ema_rate, log_interval, save_interval, resume_checkpoint, use_fp16=False, fp16_scale_growth=1e-3, schedule_sampler=None, weight_decay=0.0, lr_anneal_steps=0, ): self.model = model self.diffusion = diffusion self.data = data self.batch_size = batch_size self.microbatch = microbatch if microbatch > 0 else batch_size self.lr = lr self.ema_rate = ( [ema_rate] if isinstance(ema_rate, float) else [float(x) for x in ema_rate.split(",")] ) self.log_interval = log_interval self.save_interval = save_interval self.resume_checkpoint = resume_checkpoint self.use_fp16 = use_fp16 self.fp16_scale_growth = fp16_scale_growth self.schedule_sampler = schedule_sampler or UniformSampler(diffusion) self.weight_decay = weight_decay self.lr_anneal_steps = lr_anneal_steps self.step = 0 self.resume_step = 0 self.global_batch = self.batch_size * dist.get_world_size() self.sync_cuda = th.cuda.is_available() self._load_and_sync_parameters() self.mp_trainer = MixedPrecisionTrainer( model=self.model, use_fp16=self.use_fp16, fp16_scale_growth=fp16_scale_growth, ) self.opt = AdamW( self.mp_trainer.master_params, lr=self.lr, weight_decay=self.weight_decay ) if self.resume_step: self._load_optimizer_state() # Model was resumed, either due to a restart or a checkpoint # being specified at the command line. self.ema_params = [ self._load_ema_parameters(rate) for rate in self.ema_rate ] else: self.ema_params = [ copy.deepcopy(self.mp_trainer.master_params) for _ in range(len(self.ema_rate)) ] if th.cuda.is_available(): self.use_ddp = True self.ddp_model = DDP( self.model, device_ids=[dist_util.dev()], output_device=dist_util.dev(), broadcast_buffers=False, bucket_cap_mb=128, find_unused_parameters=False, ) else: if dist.get_world_size() > 1: logger.warn( "Distributed training requires CUDA. " "Gradients will not be synchronized properly!" ) self.use_ddp = False self.ddp_model = self.model def _load_and_sync_parameters(self): resume_checkpoint = find_resume_checkpoint() or self.resume_checkpoint if resume_checkpoint: self.resume_step = parse_resume_step_from_filename(resume_checkpoint) if dist.get_rank() == 0: logger.log(f"loading model from checkpoint: {resume_checkpoint}...") self.model.load_state_dict( dist_util.load_state_dict( resume_checkpoint, map_location=dist_util.dev() ) ) dist_util.sync_params(self.model.parameters()) def _load_ema_parameters(self, rate): ema_params = copy.deepcopy(self.mp_trainer.master_params) main_checkpoint = find_resume_checkpoint() or self.resume_checkpoint ema_checkpoint = find_ema_checkpoint(main_checkpoint, self.resume_step, rate) if ema_checkpoint: if dist.get_rank() == 0: logger.log(f"loading EMA from checkpoint: {ema_checkpoint}...") state_dict = dist_util.load_state_dict( ema_checkpoint, map_location=dist_util.dev() ) ema_params = self.mp_trainer.state_dict_to_master_params(state_dict) dist_util.sync_params(ema_params) return ema_params def _load_optimizer_state(self): main_checkpoint = find_resume_checkpoint() or self.resume_checkpoint opt_checkpoint = bf.join( bf.dirname(main_checkpoint), f"opt{self.resume_step:06}.pt" ) if bf.exists(opt_checkpoint): logger.log(f"loading optimizer state from checkpoint: {opt_checkpoint}") state_dict = dist_util.load_state_dict( opt_checkpoint, map_location=dist_util.dev() ) self.opt.load_state_dict(state_dict) def run_loop(self): while ( not self.lr_anneal_steps or self.step + self.resume_step < self.lr_anneal_steps ): batch, cond = next(self.data) self.run_step(batch, cond) if self.step % self.log_interval == 0: logger.dumpkvs() if self.step % self.save_interval == 0: self.save() # Run for a finite amount of time in integration tests. if os.environ.get("DIFFUSION_TRAINING_TEST", "") and self.step > 0: return self.step += 1 # Save the last checkpoint if it wasn't already saved. if (self.step - 1) % self.save_interval != 0: self.save() def run_step(self, batch, cond): self.forward_backward(batch, cond) took_step = self.mp_trainer.optimize(self.opt) if took_step: self._update_ema() self._anneal_lr() self.log_step() def forward_backward(self, batch, cond): self.mp_trainer.zero_grad() for i in range(0, batch.shape[0], self.microbatch): micro = batch[i : i + self.microbatch].to(dist_util.dev()) micro_cond = { k: v[i : i + self.microbatch].to(dist_util.dev()) for k, v in cond.items() } last_batch = (i + self.microbatch) >= batch.shape[0] t, weights = self.schedule_sampler.sample(micro.shape[0], dist_util.dev()) compute_losses = functools.partial( self.diffusion.training_losses, self.ddp_model, micro, t, model_kwargs=micro_cond, ) if last_batch or not self.use_ddp: losses = compute_losses() else: with self.ddp_model.no_sync(): losses = compute_losses() if isinstance(self.schedule_sampler, LossAwareSampler): self.schedule_sampler.update_with_local_losses( t, losses["loss"].detach() ) loss = (losses["loss"] * weights).mean() log_loss_dict( self.diffusion, t, {k: v * weights for k, v in losses.items()} ) self.mp_trainer.backward(loss) def _update_ema(self): for rate, params in zip(self.ema_rate, self.ema_params): update_ema(params, self.mp_trainer.master_params, rate=rate) def _anneal_lr(self): if not self.lr_anneal_steps: return frac_done = (self.step + self.resume_step) / self.lr_anneal_steps lr = self.lr * (1 - frac_done) for param_group in self.opt.param_groups: param_group["lr"] = lr def log_step(self): logger.logkv("step", self.step + self.resume_step) logger.logkv("samples", (self.step + self.resume_step + 1) * self.global_batch) def save(self): def save_checkpoint(rate, params): state_dict = self.mp_trainer.master_params_to_state_dict(params) if dist.get_rank() == 0: logger.log(f"saving model {rate}...") if not rate: filename = f"model{(self.step+self.resume_step):06d}.pt" else: filename = f"ema_{rate}_{(self.step+self.resume_step):06d}.pt" with bf.BlobFile(bf.join(get_blob_logdir(), filename), "wb") as f: th.save(state_dict, f) save_checkpoint(0, self.mp_trainer.master_params) for rate, params in zip(self.ema_rate, self.ema_params): save_checkpoint(rate, params) if dist.get_rank() == 0: with bf.BlobFile( bf.join(get_blob_logdir(), f"opt{(self.step+self.resume_step):06d}.pt"), "wb", ) as f: th.save(self.opt.state_dict(), f) dist.barrier() def parse_resume_step_from_filename(filename): """ Parse filenames of the form path/to/modelNNNNNN.pt, where NNNNNN is the checkpoint's number of steps. """ split = filename.split("model") if len(split) < 2: return 0 split1 = split[-1].split(".")[0] try: return int(split1) except ValueError: return 0 def get_blob_logdir(): # You can change this to be a separate path to save checkpoints to # a blobstore or some external drive. return logger.get_dir() def find_resume_checkpoint(): # On your infrastructure, you may want to override this to automatically # discover the latest checkpoint on your blob storage, etc. return None def find_ema_checkpoint(main_checkpoint, step, rate): if main_checkpoint is None: return None filename = f"ema_{rate}_{(step):06d}.pt" path = bf.join(bf.dirname(main_checkpoint), filename) if bf.exists(path): return path return None def log_loss_dict(diffusion, ts, losses): for key, values in losses.items(): logger.logkv_mean(key, values.mean().item()) # Log the quantiles (four quartiles, in particular). for sub_t, sub_loss in zip(ts.cpu().numpy(), values.detach().cpu().numpy()): quartile = int(4 * sub_t / diffusion.num_timesteps) logger.logkv_mean(f"{key}_q{quartile}", sub_loss)
10,982
34.429032
88
py
DiffPure
DiffPure-master/guided_diffusion/respace.py
# --------------------------------------------------------------- # Taken from the following link as is from: # https://github.com/openai/guided-diffusion/blob/main/guided_diffusion/respace.py # # The license for the original version of this file can be # found in this directory (LICENSE_GUIDED_DIFFUSION). # --------------------------------------------------------------- import numpy as np import torch as th from .gaussian_diffusion import GaussianDiffusion def space_timesteps(num_timesteps, section_counts): """ Create a list of timesteps to use from an original diffusion process, given the number of timesteps we want to take from equally-sized portions of the original process. For example, if there's 300 timesteps and the section counts are [10,15,20] then the first 100 timesteps are strided to be 10 timesteps, the second 100 are strided to be 15 timesteps, and the final 100 are strided to be 20. If the stride is a string starting with "ddim", then the fixed striding from the DDIM paper is used, and only one section is allowed. :param num_timesteps: the number of diffusion steps in the original process to divide up. :param section_counts: either a list of numbers, or a string containing comma-separated numbers, indicating the step count per section. As a special case, use "ddimN" where N is a number of steps to use the striding from the DDIM paper. :return: a set of diffusion steps from the original process to use. """ if isinstance(section_counts, str): if section_counts.startswith("ddim"): desired_count = int(section_counts[len("ddim") :]) for i in range(1, num_timesteps): if len(range(0, num_timesteps, i)) == desired_count: return set(range(0, num_timesteps, i)) raise ValueError( f"cannot create exactly {num_timesteps} steps with an integer stride" ) section_counts = [int(x) for x in section_counts.split(",")] size_per = num_timesteps // len(section_counts) extra = num_timesteps % len(section_counts) start_idx = 0 all_steps = [] for i, section_count in enumerate(section_counts): size = size_per + (1 if i < extra else 0) if size < section_count: raise ValueError( f"cannot divide section of {size} steps into {section_count}" ) if section_count <= 1: frac_stride = 1 else: frac_stride = (size - 1) / (section_count - 1) cur_idx = 0.0 taken_steps = [] for _ in range(section_count): taken_steps.append(start_idx + round(cur_idx)) cur_idx += frac_stride all_steps += taken_steps start_idx += size return set(all_steps) class SpacedDiffusion(GaussianDiffusion): """ A diffusion process which can skip steps in a base diffusion process. :param use_timesteps: a collection (sequence or set) of timesteps from the original diffusion process to retain. :param kwargs: the kwargs to create the base diffusion process. """ def __init__(self, use_timesteps, **kwargs): self.use_timesteps = set(use_timesteps) self.timestep_map = [] self.original_num_steps = len(kwargs["betas"]) base_diffusion = GaussianDiffusion(**kwargs) # pylint: disable=missing-kwoa last_alpha_cumprod = 1.0 new_betas = [] for i, alpha_cumprod in enumerate(base_diffusion.alphas_cumprod): if i in self.use_timesteps: new_betas.append(1 - alpha_cumprod / last_alpha_cumprod) last_alpha_cumprod = alpha_cumprod self.timestep_map.append(i) kwargs["betas"] = np.array(new_betas) super().__init__(**kwargs) def p_mean_variance( self, model, *args, **kwargs ): # pylint: disable=signature-differs return super().p_mean_variance(self._wrap_model(model), *args, **kwargs) def training_losses( self, model, *args, **kwargs ): # pylint: disable=signature-differs return super().training_losses(self._wrap_model(model), *args, **kwargs) def condition_mean(self, cond_fn, *args, **kwargs): return super().condition_mean(self._wrap_model(cond_fn), *args, **kwargs) def condition_score(self, cond_fn, *args, **kwargs): return super().condition_score(self._wrap_model(cond_fn), *args, **kwargs) def _wrap_model(self, model): if isinstance(model, _WrappedModel): return model return _WrappedModel( model, self.timestep_map, self.rescale_timesteps, self.original_num_steps ) def _scale_timesteps(self, t): # Scaling is done by the wrapped model. return t class _WrappedModel: def __init__(self, model, timestep_map, rescale_timesteps, original_num_steps): self.model = model self.timestep_map = timestep_map self.rescale_timesteps = rescale_timesteps self.original_num_steps = original_num_steps def __call__(self, x, ts, **kwargs): map_tensor = th.tensor(self.timestep_map, device=ts.device, dtype=ts.dtype) new_ts = map_tensor[ts] if self.rescale_timesteps: new_ts = new_ts.float() * (1000.0 / self.original_num_steps) return self.model(x, new_ts, **kwargs)
5,568
39.649635
85
py
DiffPure
DiffPure-master/guided_diffusion/dist_util.py
# --------------------------------------------------------------- # Taken from the following link as is from: # https://github.com/openai/guided-diffusion/blob/main/guided_diffusion/dist_util.py # # The license for the original version of this file can be # found in this directory (LICENSE_GUIDED_DIFFUSION). # --------------------------------------------------------------- """ Helpers for distributed training. """ import io import os import socket import blobfile as bf from mpi4py import MPI import torch as th import torch.distributed as dist # Change this to reflect your cluster layout. # The GPU for a given rank is (rank % GPUS_PER_NODE). GPUS_PER_NODE = 8 SETUP_RETRY_COUNT = 3 def setup_dist(): """ Setup a distributed process group. """ if dist.is_initialized(): return os.environ["CUDA_VISIBLE_DEVICES"] = f"{MPI.COMM_WORLD.Get_rank() % GPUS_PER_NODE}" comm = MPI.COMM_WORLD backend = "gloo" if not th.cuda.is_available() else "nccl" if backend == "gloo": hostname = "localhost" else: hostname = socket.gethostbyname(socket.getfqdn()) os.environ["MASTER_ADDR"] = comm.bcast(hostname, root=0) os.environ["RANK"] = str(comm.rank) os.environ["WORLD_SIZE"] = str(comm.size) port = comm.bcast(_find_free_port(), root=0) os.environ["MASTER_PORT"] = str(port) dist.init_process_group(backend=backend, init_method="env://") def dev(): """ Get the device to use for torch.distributed. """ if th.cuda.is_available(): return th.device(f"cuda") return th.device("cpu") def load_state_dict(path, **kwargs): """ Load a PyTorch file without redundant fetches across MPI ranks. """ chunk_size = 2 ** 30 # MPI has a relatively small size limit if MPI.COMM_WORLD.Get_rank() == 0: with bf.BlobFile(path, "rb") as f: data = f.read() num_chunks = len(data) // chunk_size if len(data) % chunk_size: num_chunks += 1 MPI.COMM_WORLD.bcast(num_chunks) for i in range(0, len(data), chunk_size): MPI.COMM_WORLD.bcast(data[i : i + chunk_size]) else: num_chunks = MPI.COMM_WORLD.bcast(None) data = bytes() for _ in range(num_chunks): data += MPI.COMM_WORLD.bcast(None) return th.load(io.BytesIO(data), **kwargs) def sync_params(params): """ Synchronize a sequence of Tensors across ranks from rank 0. """ for p in params: with th.no_grad(): dist.broadcast(p, 0) def _find_free_port(): try: s = socket.socket(socket.AF_INET, socket.SOCK_STREAM) s.bind(("", 0)) s.setsockopt(socket.SOL_SOCKET, socket.SO_REUSEADDR, 1) return s.getsockname()[1] finally: s.close()
2,801
26.470588
87
py
DiffPure
DiffPure-master/classifiers/cifar10_resnet.py
# --------------------------------------------------------------- # Copyright (c) 2022, NVIDIA CORPORATION. All rights reserved. # # This work is licensed under the NVIDIA Source Code License # for DiffPure. To view a copy of this license, see the LICENSE file. # --------------------------------------------------------------- import math import torch import torch.nn.functional as F import torch.nn as nn # ---------------------------- ResNet ---------------------------- class Bottleneck(nn.Module): expansion = 4 def __init__(self, in_planes, planes, stride=1): super(Bottleneck, self).__init__() self.conv1 = nn.Conv2d(in_planes, planes, kernel_size=1, bias=False) self.bn1 = nn.BatchNorm2d(planes) self.conv2 = nn.Conv2d(planes, planes, kernel_size=3, stride=stride, padding=1, bias=False) self.bn2 = nn.BatchNorm2d(planes) self.conv3 = nn.Conv2d(planes, self.expansion * planes, kernel_size=1, bias=False) self.bn3 = nn.BatchNorm2d(self.expansion * planes) self.shortcut = nn.Sequential() if stride != 1 or in_planes != self.expansion * planes: self.shortcut = nn.Sequential( nn.Conv2d(in_planes, self.expansion * planes, kernel_size=1, stride=stride, bias=False), nn.BatchNorm2d(self.expansion * planes) ) def forward(self, x): out = F.relu(self.bn1(self.conv1(x))) out = F.relu(self.bn2(self.conv2(out))) out = self.bn3(self.conv3(out)) out += self.shortcut(x) out = F.relu(out) return out class ResNet(nn.Module): def __init__(self, block, num_blocks, num_classes=10): super(ResNet, self).__init__() self.in_planes = 64 num_input_channels = 3 mean = (0.4914, 0.4822, 0.4465) std = (0.2471, 0.2435, 0.2616) self.mean = torch.tensor(mean).view(num_input_channels, 1, 1) self.std = torch.tensor(std).view(num_input_channels, 1, 1) self.conv1 = nn.Conv2d(3, 64, kernel_size=3, stride=1, padding=1, bias=False) self.bn1 = nn.BatchNorm2d(64) self.layer1 = self._make_layer(block, 64, num_blocks[0], stride=1) self.layer2 = self._make_layer(block, 128, num_blocks[1], stride=2) self.layer3 = self._make_layer(block, 256, num_blocks[2], stride=2) self.layer4 = self._make_layer(block, 512, num_blocks[3], stride=2) self.linear = nn.Linear(512 * block.expansion, num_classes) def _make_layer(self, block, planes, num_blocks, stride): strides = [stride] + [1] * (num_blocks - 1) layers = [] for stride in strides: layers.append(block(self.in_planes, planes, stride)) self.in_planes = planes * block.expansion return nn.Sequential(*layers) def forward(self, x): out = (x - self.mean.to(x.device)) / self.std.to(x.device) out = F.relu(self.bn1(self.conv1(out))) out = self.layer1(out) out = self.layer2(out) out = self.layer3(out) out = self.layer4(out) out = F.avg_pool2d(out, 4) out = out.view(out.size(0), -1) out = self.linear(out) return out def ResNet50(): return ResNet(Bottleneck, [3, 4, 6, 3]) # ---------------------------- ResNet ---------------------------- # ---------------------------- WideResNet ---------------------------- class BasicBlock(nn.Module): def __init__(self, in_planes, out_planes, stride, dropRate=0.0): super(BasicBlock, self).__init__() self.bn1 = nn.BatchNorm2d(in_planes) self.relu1 = nn.ReLU(inplace=True) self.conv1 = nn.Conv2d(in_planes, out_planes, kernel_size=3, stride=stride, padding=1, bias=False) self.bn2 = nn.BatchNorm2d(out_planes) self.relu2 = nn.ReLU(inplace=True) self.conv2 = nn.Conv2d(out_planes, out_planes, kernel_size=3, stride=1, padding=1, bias=False) self.droprate = dropRate self.equalInOut = (in_planes == out_planes) self.convShortcut = (not self.equalInOut) and nn.Conv2d(in_planes, out_planes, kernel_size=1, stride=stride, padding=0, bias=False) or None def forward(self, x): if not self.equalInOut: x = self.relu1(self.bn1(x)) else: out = self.relu1(self.bn1(x)) out = self.relu2(self.bn2(self.conv1(out if self.equalInOut else x))) if self.droprate > 0: out = F.dropout(out, p=self.droprate, training=self.training) out = self.conv2(out) return torch.add(x if self.equalInOut else self.convShortcut(x), out) class NetworkBlock(nn.Module): def __init__(self, nb_layers, in_planes, out_planes, block, stride, dropRate=0.0): super(NetworkBlock, self).__init__() self.layer = self._make_layer(block, in_planes, out_planes, nb_layers, stride, dropRate) def _make_layer(self, block, in_planes, out_planes, nb_layers, stride, dropRate): layers = [] for i in range(int(nb_layers)): layers.append(block(i == 0 and in_planes or out_planes, out_planes, i == 0 and stride or 1, dropRate)) return nn.Sequential(*layers) def forward(self, x): return self.layer(x) class WideResNet(nn.Module): """ Based on code from https://github.com/yaodongyu/TRADES """ def __init__(self, depth=28, num_classes=10, widen_factor=10, sub_block1=False, dropRate=0.0, bias_last=True): super(WideResNet, self).__init__() num_input_channels = 3 mean = (0.4914, 0.4822, 0.4465) std = (0.2471, 0.2435, 0.2616) self.mean = torch.tensor(mean).view(num_input_channels, 1, 1) self.std = torch.tensor(std).view(num_input_channels, 1, 1) nChannels = [16, 16 * widen_factor, 32 * widen_factor, 64 * widen_factor] assert ((depth - 4) % 6 == 0) n = (depth - 4) / 6 block = BasicBlock # 1st conv before any network block self.conv1 = nn.Conv2d(3, nChannels[0], kernel_size=3, stride=1, padding=1, bias=False) # 1st block self.block1 = NetworkBlock(n, nChannels[0], nChannels[1], block, 1, dropRate) if sub_block1: # 1st sub-block self.sub_block1 = NetworkBlock(n, nChannels[0], nChannels[1], block, 1, dropRate) # 2nd block self.block2 = NetworkBlock(n, nChannels[1], nChannels[2], block, 2, dropRate) # 3rd block self.block3 = NetworkBlock(n, nChannels[2], nChannels[3], block, 2, dropRate) # global average pooling and classifier self.bn1 = nn.BatchNorm2d(nChannels[3]) self.relu = nn.ReLU(inplace=True) self.fc = nn.Linear(nChannels[3], num_classes, bias=bias_last) self.nChannels = nChannels[3] for m in self.modules(): if isinstance(m, nn.Conv2d): n = m.kernel_size[0] * m.kernel_size[1] * m.out_channels m.weight.data.normal_(0, math.sqrt(2. / n)) elif isinstance(m, nn.BatchNorm2d): m.weight.data.fill_(1) m.bias.data.zero_() elif isinstance(m, nn.Linear) and not m.bias is None: m.bias.data.zero_() def forward(self, x): out = (x - self.mean.to(x.device)) / self.std.to(x.device) out = self.conv1(out) out = self.block1(out) out = self.block2(out) out = self.block3(out) out = self.relu(self.bn1(out)) out = F.avg_pool2d(out, 8) out = out.view(-1, self.nChannels) return self.fc(out) def WideResNet_70_16(): return WideResNet(depth=70, widen_factor=16, dropRate=0.0) def WideResNet_70_16_dropout(): return WideResNet(depth=70, widen_factor=16, dropRate=0.3) # ---------------------------- WideResNet ----------------------------
7,977
38.89
116
py
DiffPure
DiffPure-master/classifiers/attribute_classifier.py
# --------------------------------------------------------------- # Copyright (c) 2022, NVIDIA CORPORATION. All rights reserved. # # This work is licensed under the NVIDIA Source Code License # for DiffPure. To view a copy of this license, see the LICENSE file. # --------------------------------------------------------------- import torch import os from . import attribute_net softmax = torch.nn.Softmax(dim=1) def downsample(images, size=256): # Downsample to 256x256. The attribute classifiers were built for 256x256. # follows https://github.com/NVlabs/stylegan/blob/master/metrics/linear_separability.py#L127 if images.shape[2] > size: factor = images.shape[2] // size assert (factor * size == images.shape[2]) images = images.view( [-1, images.shape[1], images.shape[2] // factor, factor, images.shape[3] // factor, factor]) images = images.mean(dim=[3, 5]) return images else: assert (images.shape[-1] == 256) return images def get_logit(net, im): im_256 = downsample(im) logit = net(im_256) return logit def get_softmaxed(net, im): logit = get_logit(net, im) logits = torch.cat([logit, -logit], dim=1) softmaxed = softmax(torch.cat([logit, -logit], dim=1))[:, 1] return logits, softmaxed def load_attribute_classifier(attribute, ckpt_path=None): if ckpt_path is None: base_path = 'pretrained/celebahq' attribute_pkl = os.path.join(base_path, attribute, 'net_best.pth') ckpt = torch.load(attribute_pkl) else: ckpt = torch.load(ckpt_path) print("Using classifier at epoch: %d" % ckpt['epoch']) if 'valacc' in ckpt.keys(): print("Validation acc on raw images: %0.5f" % ckpt['valacc']) detector = attribute_net.from_state_dict( ckpt['state_dict'], fixed_size=True, use_mbstd=False).cuda().eval() return detector class ClassifierWrapper(torch.nn.Module): def __init__(self, classifier_name, ckpt_path=None, device='cuda'): super(ClassifierWrapper, self).__init__() self.net = load_attribute_classifier(classifier_name, ckpt_path).eval().to(device) def forward(self, ims): out = (ims - 0.5) / 0.5 return get_softmaxed(self.net, out)[0]
2,276
33.5
104
py
DiffPure
DiffPure-master/classifiers/attribute_net.py
# --------------------------------------------------------------- # Copyright (c) 2022, NVIDIA CORPORATION. All rights reserved. # # This work is licensed under the NVIDIA Source Code License # for DiffPure. To view a copy of this license, see the LICENSE file. # --------------------------------------------------------------- import torch import torch.nn as nn import numpy as np def lerp_clip(a, b, t): return a + (b - a) * torch.clamp(t, 0.0, 1.0) class WScaleLayer(nn.Module): def __init__(self, size, fan_in, gain=np.sqrt(2), bias=True): super(WScaleLayer, self).__init__() self.scale = gain / np.sqrt(fan_in) # No longer a parameter if bias: self.b = nn.Parameter(torch.randn(size)) else: self.b = 0 self.size = size def forward(self, x): x_size = x.size() x = x * self.scale # modified to remove warning if type(self.b) == nn.Parameter and len(x_size) == 4: x = x + self.b.view(1, -1, 1, 1).expand( x_size[0], self.size, x_size[2], x_size[3]) if type(self.b) == nn.Parameter and len(x_size) == 2: x = x + self.b.view(1, -1).expand( x_size[0], self.size) return x class WScaleConv2d(nn.Module): def __init__(self, in_channels, out_channels, kernel_size, padding=0, bias=True, gain=np.sqrt(2)): super().__init__() self.conv = nn.Conv2d(in_channels, out_channels, kernel_size=kernel_size, padding=padding, bias=False) fan_in = in_channels * kernel_size * kernel_size self.wscale = WScaleLayer(out_channels, fan_in, gain=gain, bias=bias) def forward(self, x): return self.wscale(self.conv(x)) class WScaleLinear(nn.Module): def __init__(self, in_channels, out_channels, bias=True, gain=np.sqrt(2)): super().__init__() self.linear = nn.Linear(in_channels, out_channels, bias=False) self.wscale = WScaleLayer(out_channels, in_channels, gain=gain, bias=bias) def forward(self, x): return self.wscale(self.linear(x)) class FromRGB(nn.Module): def __init__(self, in_channels, out_channels, kernel_size, act=nn.LeakyReLU(0.2), bias=True): super().__init__() self.conv = WScaleConv2d(in_channels, out_channels, kernel_size, padding=0, bias=bias) self.act = act def forward(self, x): return self.act(self.conv(x)) class Downscale2d(nn.Module): def __init__(self, factor=2): super().__init__() self.downsample = nn.AvgPool2d(kernel_size=factor, stride=factor) def forward(self, x): return self.downsample(x) class DownscaleConvBlock(nn.Module): def __init__(self, in_channels, conv0_channels, conv1_channels, kernel_size, padding, bias=True, act=nn.LeakyReLU(0.2)): super().__init__() self.downscale = Downscale2d() self.conv0 = WScaleConv2d(in_channels, conv0_channels, kernel_size=kernel_size, padding=padding, bias=bias) self.conv1 = WScaleConv2d(conv0_channels, conv1_channels, kernel_size=kernel_size, padding=padding, bias=bias) self.act = act def forward(self, x): x = self.act(self.conv0(x)) # conv2d_downscale2d applies downscaling before activation # the order matters here! has to be conv -> bias -> downscale -> act x = self.conv1(x) x = self.downscale(x) x = self.act(x) return x class MinibatchStdLayer(nn.Module): def __init__(self, group_size=4): super().__init__() self.group_size = group_size def forward(self, x): group_size = min(self.group_size, x.shape[0]) s = x.shape y = x.view([group_size, -1, s[1], s[2], s[3]]) y = y.float() y = y - torch.mean(y, dim=0, keepdim=True) y = torch.mean(y * y, dim=0) y = torch.sqrt(y + 1e-8) y = torch.mean(torch.mean(torch.mean(y, dim=3, keepdim=True), dim=2, keepdim=True), dim=1, keepdim=True) y = y.type(x.type()) y = y.repeat(group_size, 1, s[2], s[3]) return torch.cat([x, y], dim=1) class PredictionBlock(nn.Module): def __init__(self, in_channels, dense0_feat, dense1_feat, out_feat, pool_size=2, act=nn.LeakyReLU(0.2), use_mbstd=True): super().__init__() self.use_mbstd = use_mbstd # attribute classifiers don't have this if self.use_mbstd: self.mbstd_layer = MinibatchStdLayer() # MinibatchStdLayer adds an additional feature dimension self.conv = WScaleConv2d(in_channels + int(self.use_mbstd), dense0_feat, kernel_size=3, padding=1) self.dense0 = WScaleLinear(dense0_feat * pool_size * pool_size, dense1_feat) self.dense1 = WScaleLinear(dense1_feat, out_feat, gain=1) self.act = act def forward(self, x): if self.use_mbstd: x = self.mbstd_layer(x) x = self.act(self.conv(x)) x = x.view([x.shape[0], -1]) x = self.act(self.dense0(x)) x = self.dense1(x) return x class D(nn.Module): def __init__( self, num_channels=3, # Number of input color channels. Overridden based on dataset. resolution=128, # Input resolution. Overridden based on dataset. fmap_base=8192, # Overall multiplier for the number of feature maps. fmap_decay=1.0, # log2 feature map reduction when doubling the resolution. fmap_max=512, # Maximum number of feature maps in any layer. fixed_size=False, # True = load fromrgb_lod0 weights only use_mbstd=True, # False = no mbstd layer in PredictionBlock **kwargs): # Ignore unrecognized keyword args. super().__init__() self.resolution_log2 = resolution_log2 = int(np.log2(resolution)) assert resolution == 2 ** resolution_log2 and resolution >= 4 def nf(stage): return min(int(fmap_base / (2.0 ** (stage * fmap_decay))), fmap_max) self.register_buffer('lod_in', torch.from_numpy(np.array(0.0))) res = resolution_log2 setattr(self, 'fromrgb_lod0', FromRGB(num_channels, nf(res - 1), 1)) for i, res in enumerate(range(resolution_log2, 2, -1), 1): lod = resolution_log2 - res block = DownscaleConvBlock(nf(res - 1), nf(res - 1), nf(res - 2), kernel_size=3, padding=1) setattr(self, '%dx%d' % (2 ** res, 2 ** res), block) fromrgb = FromRGB(3, nf(res - 2), 1) if not fixed_size: setattr(self, 'fromrgb_lod%d' % i, fromrgb) res = 2 pool_size = 2 ** res block = PredictionBlock(nf(res + 1 - 2), nf(res - 1), nf(res - 2), 1, pool_size, use_mbstd=use_mbstd) setattr(self, '%dx%d' % (pool_size, pool_size), block) self.downscale = Downscale2d() self.fixed_size = fixed_size def forward(self, img): x = self.fromrgb_lod0(img) for i, res in enumerate(range(self.resolution_log2, 2, -1), 1): lod = self.resolution_log2 - res x = getattr(self, '%dx%d' % (2 ** res, 2 ** res))(x) if not self.fixed_size: img = self.downscale(img) y = getattr(self, 'fromrgb_lod%d' % i)(img) x = lerp_clip(x, y, self.lod_in - lod) res = 2 pool_size = 2 ** res out = getattr(self, '%dx%d' % (pool_size, pool_size))(x) return out def max_res_from_state_dict(state_dict): for i in range(3, 12): if '%dx%d.conv0.conv.weight' % (2 ** i, 2 ** i) not in state_dict: break return 2 ** (i - 1) def from_state_dict(state_dict, fixed_size=False, use_mbstd=True): res = max_res_from_state_dict(state_dict) print(f'res: {res}') d = D(num_channels=3, resolution=res, fixed_size=fixed_size, use_mbstd=use_mbstd) d.load_state_dict(state_dict) return d
8,507
36.315789
91
py
DiffPure
DiffPure-master/data/datasets.py
# --------------------------------------------------------------- # Copyright (c) 2022, NVIDIA CORPORATION. All rights reserved. # # This work is licensed under the NVIDIA Source Code License # for DiffPure. To view a copy of this license, see the LICENSE file. # --------------------------------------------------------------- import os, sys import io import lmdb import pandas as pd import numpy as np from PIL import Image import torch import torchvision from torch.utils.data import Dataset, Subset import torchvision.transforms as transforms from torchvision.datasets.vision import VisionDataset from torchvision.datasets import folder, ImageFolder # --------------------------------------------------------------------------------------------------- def remove_prefix(s, prefix): if s.startswith(prefix): s = s[len(prefix):] return s class ImageDataset(VisionDataset): """ modified from: https://pytorch.org/docs/stable/_modules/torchvision/datasets/folder.html#ImageFolder uses cached directory listing if available rather than walking directory Attributes: classes (list): List of the class names. class_to_idx (dict): Dict with items (class_name, class_index). samples (list): List of (sample path, class_index) tuples targets (list): The class_index value for each image in the dataset """ def __init__(self, root, loader=folder.default_loader, extensions=folder.IMG_EXTENSIONS, transform=None, target_transform=None, is_valid_file=None, return_path=False): super(ImageDataset, self).__init__(root, transform=transform, target_transform=target_transform) classes, class_to_idx = self._find_classes(self.root) cache = self.root.rstrip('/') + '.txt' if os.path.isfile(cache): print("Using directory list at: %s" % cache) with open(cache) as f: samples = [] for line in f: (path, idx) = line.strip().split(';') samples.append((os.path.join(self.root, path), int(idx))) else: print("Walking directory: %s" % self.root) samples = folder.make_dataset(self.root, class_to_idx, extensions, is_valid_file) with open(cache, 'w') as f: for line in samples: path, label = line f.write('%s;%d\n' % (remove_prefix(path, self.root).lstrip('/'), label)) if len(samples) == 0: raise (RuntimeError( "Found 0 files in subfolders of: " + self.root + "\nSupported extensions are: " + ",".join(extensions))) self.loader = loader self.classes = classes self.class_to_idx = class_to_idx self.samples = samples self.return_path = return_path def _find_classes(self, dir): """ Finds the class folders in a dataset. Ensures: No class is a subdirectory of another. """ if sys.version_info >= (3, 5): # Faster and available in Python 3.5 and above classes = [d.name for d in os.scandir(dir) if d.is_dir()] else: classes = [d for d in os.listdir(dir) if os.path.isdir(os.path.join(dir, d))] classes.sort() class_to_idx = {classes[i]: i for i in range(len(classes))} return classes, class_to_idx def __getitem__(self, index): path, target = self.samples[index] sample = self.loader(path) if self.transform is not None: sample = self.transform(sample) if self.target_transform is not None: target = self.target_transform(target) if self.return_path: return sample, target, path return sample, target def __len__(self): return len(self.samples) # --------------------------------------------------------------------------------------------------- # get the attributes from celebahq subset def make_table(root): filenames = sorted(os.listdir(f'{root}/images')) # filter out non-png files, rename it to jpg to match entries in list_attr_celeba.txt celebahq = [os.path.basename(f).replace('png', 'jpg') if f.endswith('png') else os.path.basename(f) for f in filenames] attr_gt = pd.read_csv(f'{root}/list_attr_celeba.txt', skiprows=1, delim_whitespace=True, index_col=0) attr_celebahq = attr_gt.reindex(index=celebahq).replace(-1, 0) # get the train/test/val partitions partitions = {} with open(f'{root}/list_eval_partition.txt') as f: for line in f: filename, part = line.strip().split(' ') partitions[filename] = int(part) partitions_list = [partitions[fname] for fname in attr_celebahq.index] attr_celebahq['partition'] = partitions_list return attr_celebahq ###### dataset functions ###### class CelebAHQDataset(Dataset): def __init__(self, partition, attribute, root=None, fraction=None, data_seed=1, chunk_length=None, chunk_idx=-1, **kwargs): if root is None: root = './dataset/celebahq' self.fraction = fraction self.dset = ImageDataset(root, **kwargs) # make table attr_celebahq = make_table(root) # convert from train/val/test to partition numbers part_to_int = dict(train=0, val=1, test=2) def get_partition_indices(part): return np.where(attr_celebahq['partition'] == part_to_int[part])[0] partition_idx = get_partition_indices(partition) # if we want to further subsample the dataset, just subsample # partition_idx and Subset() once if fraction is not None: print("Using a fraction of the original dataset") print("The original dataset has length %d" % len(partition_idx)) new_length = int(fraction / 100 * len(partition_idx)) rng = np.random.RandomState(data_seed) new_indices = rng.choice(partition_idx, new_length, replace=False) partition_idx = new_indices print("The subsetted dataset has length %d" % len(partition_idx)) elif chunk_length is not None and chunk_idx > 0: print(f"Using a fraction of the original dataset with chunk_length: {chunk_length}, chunk_idx: {chunk_idx}") print("The original dataset has length %d" % len(partition_idx)) new_indices = partition_idx[chunk_length * chunk_idx: chunk_length * (chunk_idx + 1)] partition_idx = new_indices print("The subsetted dataset has length %d" % len(partition_idx)) self.dset = Subset(self.dset, partition_idx) attr_subset = attr_celebahq.iloc[partition_idx] self.attr_subset = attr_subset[attribute] print('attribute freq: %0.4f (%d / %d)' % (self.attr_subset.mean(), self.attr_subset.sum(), len(self.attr_subset))) def __len__(self): return len(self.dset) def __getitem__(self, idx): data = self.dset[idx] # first element is the class, replace it label = self.attr_subset[idx] return (data[0], label, *data[2:]) ###### transformation functions ###### def get_transform(dataset, transform_type, base_size=256): if dataset.lower() == "celebahq": assert base_size == 256, base_size if transform_type == 'imtrain': return transforms.Compose([ transforms.Resize(base_size), transforms.RandomHorizontalFlip(p=0.5), transforms.ToTensor(), transforms.Normalize((0.5, 0.5, 0.5), (0.5, 0.5, 0.5)) ]) elif transform_type == 'imval': return transforms.Compose([ transforms.Resize(base_size), # no horizontal flip for standard validation transforms.ToTensor(), # transforms.Normalize((0.5, 0.5, 0.5), (0.5, 0.5, 0.5)) ]) elif transform_type == 'imcolor': return transforms.Compose([ transforms.Resize(base_size), transforms.RandomHorizontalFlip(p=0.5), transforms.ColorJitter(brightness=.05, contrast=.05, saturation=.05, hue=.05), transforms.ToTensor(), transforms.Normalize((0.5, 0.5, 0.5), (0.5, 0.5, 0.5)), ]) elif transform_type == 'imcrop': return transforms.Compose([ # 1024 + 32, or 256 + 8 transforms.Resize(int(1.03125 * base_size)), transforms.RandomCrop(base_size), transforms.RandomHorizontalFlip(p=0.5), transforms.ToTensor(), transforms.Normalize((0.5, 0.5, 0.5), (0.5, 0.5, 0.5)), ]) elif transform_type == 'tensorbase': # dummy transform for compatibility with other datasets return transforms.Lambda(lambda x: x) else: raise NotImplementedError elif "imagenet" in dataset.lower(): assert base_size == 224, base_size if transform_type == 'imtrain': return transforms.Compose([ transforms.Resize(256), transforms.CenterCrop(base_size), transforms.RandomHorizontalFlip(p=0.5), transforms.ToTensor(), # transforms.Normalize((0.485, 0.456, 0.406), (0.229, 0.224, 0.225)) ]) elif transform_type == 'imval': return transforms.Compose([ transforms.Resize(256), transforms.CenterCrop(base_size), # no horizontal flip for standard validation transforms.ToTensor(), # transforms.Normalize((0.485, 0.456, 0.406), (0.229, 0.224, 0.225)) ]) else: raise NotImplementedError else: raise NotImplementedError ################################################################################ # ImageNet - LMDB ############################################################################### def lmdb_loader(path, lmdb_data): # In-memory binary streams with lmdb_data.begin(write=False, buffers=True) as txn: bytedata = txn.get(path.encode('ascii')) img = Image.open(io.BytesIO(bytedata)) return img.convert('RGB') def imagenet_lmdb_dataset( root, transform=None, target_transform=None, loader=lmdb_loader): """ You can create this dataloader using: train_data = imagenet_lmdb_dataset(traindir, transform=train_transform) valid_data = imagenet_lmdb_dataset(validdir, transform=val_transform) """ if root.endswith('/'): root = root[:-1] pt_path = os.path.join( root + '_faster_imagefolder.lmdb.pt') lmdb_path = os.path.join( root + '_faster_imagefolder.lmdb') if os.path.isfile(pt_path) and os.path.isdir(lmdb_path): print('Loading pt {} and lmdb {}'.format(pt_path, lmdb_path)) data_set = torch.load(pt_path) else: data_set = ImageFolder( root, None, None, None) torch.save(data_set, pt_path, pickle_protocol=4) print('Saving pt to {}'.format(pt_path)) print('Building lmdb to {}'.format(lmdb_path)) env = lmdb.open(lmdb_path, map_size=1e12) with env.begin(write=True) as txn: for path, class_index in data_set.imgs: with open(path, 'rb') as f: data = f.read() txn.put(path.encode('ascii'), data) data_set.lmdb_data = lmdb.open( lmdb_path, readonly=True, max_readers=1, lock=False, readahead=False, meminit=False) # reset transform and target_transform data_set.samples = data_set.imgs data_set.transform = transform data_set.target_transform = target_transform data_set.loader = lambda path: loader(path, data_set.lmdb_data) return data_set def imagenet_lmdb_dataset_sub( root, transform=None, target_transform=None, loader=lmdb_loader, num_sub=-1, data_seed=0): data_set = imagenet_lmdb_dataset( root, transform=transform, target_transform=target_transform, loader=loader) if num_sub > 0: partition_idx = np.random.RandomState(data_seed).choice(len(data_set), num_sub, replace=False) data_set = Subset(data_set, partition_idx) return data_set ################################################################################ # CIFAR-10 ############################################################################### def cifar10_dataset_sub(root, transform=None, num_sub=-1, data_seed=0): val_data = torchvision.datasets.CIFAR10(root=root, transform=transform, download=True, train=False) if num_sub > 0: partition_idx = np.random.RandomState(data_seed).choice(len(val_data), num_sub, replace=False) val_data = Subset(val_data, partition_idx) return val_data
13,203
38.181009
120
py
Tencent_wsdm_cup2023
Tencent_wsdm_cup2023-main/pytorch_unbias/pretrain/dataset.py
# -*- coding: utf-8 -*- import sys,os import random import collections from models.utils import SPECIAL_TOKENS import logging import numpy as np import torch from torch.utils.data import Dataset, IterableDataset from dataclasses import dataclass from typing import List, Dict, Any logger = logging.getLogger(__name__) MaskedLmInstance = collections.namedtuple("MaskedLmInstance", ["index", "label"]) class MaskingOp(): def __init__(self, masked_lm_prob, max_predictions_per_seq, vocab_list): self.masked_lm_prob = masked_lm_prob self.max_predictions_per_seq = max_predictions_per_seq self.vocab_list = vocab_list def create_masked_lm_predictions(self, tokens): """Creates the predictions for the masked LM objective. This is mostly copied from the Google BERT repo, but with several refactors to clean it up and remove a lot of unnecessary variables.""" cand_indices = [] START_FROM_DOC = False for (i, token) in enumerate(tokens): # token_ids if token == SPECIAL_TOKENS['SEP']: # SEP START_FROM_DOC = True continue if token == SPECIAL_TOKENS['CLS']: # CLS # if token in SPECIAL_TOKENS.values(): continue if not START_FROM_DOC: continue cand_indices.append([i]) num_to_mask = min(self.max_predictions_per_seq, max(1, int(round(len(cand_indices) * self.masked_lm_prob)))) random.shuffle(cand_indices) masked_lms = [] covered_indexes = set() for index_set in cand_indices: if len(masked_lms) >= num_to_mask: break # If adding a whole-word mask would exceed the maximum number of # predictions, then just skip this candidate. if len(masked_lms) + len(index_set) > num_to_mask: continue is_any_index_covered = False for index in index_set: if index in covered_indexes: is_any_index_covered = True break if is_any_index_covered: continue for index in index_set: covered_indexes.add(index) # 80% of the time, replace with [MASK] if random.random() < 0.8: masked_token = SPECIAL_TOKENS['MASK'] #103 else: # 10% of the time, keep original if random.random() < 0.5: masked_token = tokens[index] # 10% of the time, replace with random word else: masked_token = random.choice(self.vocab_list) masked_lms.append(MaskedLmInstance(index=index, label=tokens[index])) tokens[index] = masked_token assert len(masked_lms) <= num_to_mask masked_lms = sorted(masked_lms, key=lambda x: x.index) mask_indices = [p.index for p in masked_lms] masked_token_labels = [p.label for p in masked_lms] return tokens, masked_token_labels, mask_indices def process_data(query, title, content, max_seq_len, masking_obj=None): """ process [query, title, content] into a tensor [CLS] + query + [SEP] + title + [SEP] + content + [SEP] + [PAD] """ data = [SPECIAL_TOKENS['CLS']] segment = [0] data = data + [int(item) + 10 for item in query.split(b'\x01')] # query data = data + [SPECIAL_TOKENS['SEP']] segment = segment + [0] * (len(query.split(b'\x01')) + 1) data = data + [int(item) + 10 for item in title.split(b'\x01')] # content data = data + [SPECIAL_TOKENS['SEP']] # sep defined as 1 segment = segment + [1] * (len(title.split(b'\x01')) + 1) data = data + [int(item) + 10 for item in content.split(b'\x01')] # content data = data + [SPECIAL_TOKENS['SEP']] segment = segment + [1] * (len(content.split(b'\x01')) + 1) #padding_mask = [False] * len(data) if len(data) < max_seq_len: #padding_mask += [True] * (max_seq_len - len(data)) data += [SPECIAL_TOKENS['PAD']] * (max_seq_len - len(data)) else: #padding_mask = padding_mask[:max_seq_len] data = data[:max_seq_len] # segment id if len(segment) < max_seq_len: segment += [1] * (max_seq_len - len(segment)) else: segment = segment[:max_seq_len] #padding_mask = torch.BoolTensor(padding_mask) #data = torch.LongTensor(data) #segment = torch.LongTensor(segment) if masking_obj is not None: token_ids, masked_lm_ids, masked_lm_positions = masking_obj.create_masked_lm_predictions(data) lm_label_array = np.full(max_seq_len, dtype=np.int, fill_value=-1) lm_label_array[masked_lm_positions] = masked_lm_ids return token_ids, segment, lm_label_array else: return data, segment class TrainDatasetBase(IterableDataset): def __init__(self, directory_path, fea_d, fea_c, args): self.directory_path = directory_path self.files = [f for f in os.listdir(self.directory_path) if f.startswith('part')] random.shuffle(self.files) self.cur_query = "#" self.max_seq_len = args.max_seq_len#128 self.num_candidates = args.num_candidates # -1 for no debias self.vocab_size = args.vocab_size#22000 self.vocab_list = list(range(self.vocab_size)) self.masking_obj = MaskingOp(args.masked_lm_prob, args.max_predictions_per_seq, self.vocab_list) self.fea_d = fea_d self.fea_c = fea_c if self.num_candidates > -1: assert "rank_pos" in self.fea_d def get_normal_feature(self, partition_num, partition_l, min_v, max_v, value): # TODO if value < min_v: return 0 elif value >= max_v: return partition_num - 1 else: return int((value - min_v) / partition_l) + 1 def __iter__(self): buffer_per_query = [] #sample_count = 0 info = torch.utils.data.get_worker_info() if info is None: worker_num = 1 worker_id = 0 else: worker_num = info.num_workers worker_id = info.id ## each worker parses one file local_files = [f for i, f in enumerate(self.files) if i % worker_num == worker_id] for i, file in enumerate(local_files): logger.info(f'load file: {file}') if file == 'part-00000': # part-00000.gz is for evaluation continue for line in open(os.path.join(self.directory_path, file), 'rb'): line_list = line.strip(b'\n').split(b'\t') if len(line_list) == 3: # new query self.cur_query = line_list[1] if len(buffer_per_query) > 0: if self.num_candidates > 0: buffer_per_query = buffer_per_query[:self.num_candidates] data = self.yield_data(buffer_per_query, self.num_candidates) if data is not None: yield data buffer_per_query = [] elif len(line_list) > 6: # urls position, title, content, click_label = line_list[0], line_list[2], line_list[3], line_list[5] if self.num_candidates > 0: if int(position) > 10: continue try: src_input, segment, masked_lm_labels = process_data(self.cur_query, title, content, self.max_seq_len, self.masking_obj) sample = {'src_input': src_input, 'segment': segment, 'masked_lm_labels': masked_lm_labels, 'click_label': float(click_label), 'rank_pos': int(position) - 1, } except: continue for fea_name in self.fea_d: if fea_name == 'rank_pos': continue fea = line_list[self.fea_d[fea_name][-1]] emb_idx = self.fea_d[fea_name][1][fea] sample.update({fea_name: emb_idx}) for fea_name in self.fea_c: fea = float(line_list[self.fea_c[fea_name][-1]]) min_v = self.fea_c[fea_name][1] max_v = self.fea_c[fea_name][2] partition_l = self.fea_c[fea_name][3] partition_num = self.fea_c[fea_name][0] emb_idx = self.get_normal_feature(partition_num, partition_l, min_v, max_v, fea) sample.update({fea_name: emb_idx}) buffer_per_query.append(sample) def yield_data(self, buffer_per_query, num_candidates): pass class PreTrainDatasetGroupwise(TrainDatasetBase): def __init__(self, directory_path, train_group_size, fea_d, fea_c, args): super(PreTrainDatasetGroupwise, self).__init__(directory_path, fea_d, fea_c, args) self.train_group_size = train_group_size def yield_data(self, buffer_per_query, num_candidates): # pad item to num candidates buffer_out = buffer_per_query.copy() random.shuffle(buffer_per_query) pos_buffer, neg_buffer = [], [] for record in buffer_per_query: if record['click_label'] > 0: pos_buffer.append(record) else: neg_buffer.append(record) if len(pos_buffer) == 0 or len(neg_buffer) == 0: return None pos_record = random.choice(pos_buffer) if len(neg_buffer) < self.train_group_size - 1: negs = random.choices(neg_buffer, k=self.train_group_size - 1) else: negs = random.sample(neg_buffer, k=self.train_group_size - 1) group = [pos_record] + negs pad_size = num_candidates - len(buffer_out) for _ in range(pad_size): sample_pad = {} for name in self.fea_d: sample_pad.update({name: self.fea_d[name][0]}) for name in self.fea_c: sample_pad.update({name: self.fea_c[name][0]}) buffer_out.append(sample_pad) return group, buffer_out @dataclass class DataCollator: def __call__(self, features) -> Dict[str, Any]: if isinstance(features[0], tuple): relative_fea = [f[0] for f in features] relative_fea = sum(relative_fea, []) token_ids = torch.LongTensor([f['src_input'] for f in relative_fea]) segment = torch.LongTensor([f['segment'] for f in relative_fea]) batch_data = { 'input_ids': token_ids, 'attention_mask': (token_ids > 0).float(), # torch.FloatTensor, 0 is PAD 'token_type_ids': segment } for name in relative_fea[0]: if name in ["qid", "label", "freq", "src_input", "segment"]: continue else: fea = torch.LongTensor([f[name] for f in relative_fea]) batch_data.update({name: fea}) debias_fea = [f[1] for f in features] debias_fea = sum(debias_fea, []) debias_dict = collections.OrderedDict() for name in debias_fea[0]: if name in ["qid", "label", "freq", "src_input", "segment", "src_input", "segment", "masked_lm_labels", "click_label"]: continue else: fea = torch.LongTensor([f[name] for f in debias_fea]) debias_dict.update({name: fea}) batch_data.update({"debias_fea": debias_dict}) else: token_ids = torch.LongTensor([f['src_input'] for f in features]) segment = torch.LongTensor([f['segment'] for f in features]) batch_data = { 'input_ids': token_ids, 'attention_mask': (token_ids > 0).float(), # torch.FloatTensor, 0 is PAD 'token_type_ids': segment } for name in features[0]: if name in ["qid", "label", "freq", "src_input", "segment"]: continue else: fea = torch.LongTensor([f[name] for f in features]) batch_data.update({name: fea}) if 'qid' in features[0]: qids = torch.LongTensor([f['qid'] for f in features]) batch_data.update({'qids': qids}) if 'label' in features[0]: labels = torch.LongTensor([f['label'] for f in features]) batch_data.update({'labels': labels}) if 'freq' in features[0]: freqs = torch.LongTensor([f['freq'] for f in features]) batch_data.update({'freqs': freqs}) return batch_data ###################Test dataset################################ class TestDataset(Dataset): def __init__(self, fpath, max_seq_len, data_type, buffer_size=300000): self.buffer_size = buffer_size self.max_seq_len = max_seq_len self.data_type = data_type if data_type == 'annotate': self.buffer = self.load_annotate_data(fpath) #self.buffer = self.buffer[:10000] elif data_type == 'click': self.buffer = self.load_click_data(fpath) #self.total_freqs = None def __len__(self): return len(self.buffer) def __getitem__(self, index): #return self.buffer[index] if len(self.buffer[index]) == 4: src_input, segment, qid, label= self.buffer[index] sample = {'src_input': src_input, 'segment': segment, 'qid': qid, 'label': label} elif len(self.buffer[index]) == 5: src_input, segment, qid, label, freq = self.buffer[index] sample = {'src_input': src_input, 'segment': segment, 'qid': qid, 'label': label, 'freq': freq} return sample def load_annotate_data(self, fpath): logger.info(f'load annotated data from {fpath}') buffer = [] for line in open(fpath, 'rb'): line_list = line.strip(b'\n').split(b'\t') qid, query, title, content, label, freq = line_list if 0 <= int(freq) <= 2: # high freq freq = 0 elif 3 <= int(freq) <= 6: # mid freq freq = 1 elif 7 <= int(freq): # tail freq = 2 src_input, src_segment = process_data(query, title, content, self.max_seq_len) buffer.append([src_input, src_segment, int(qid), int(label), freq]) return buffer#, total_qids, total_labels, total_freqs def load_click_data(self, fpath): logger.info(f'load logged click data from {fpath}') buffer = [] cur_qids = 0 for line in open(fpath, 'rb'): line_list = line.strip(b'\n').split(b'\t') if len(line_list) == 3: # new query self.cur_query = line_list[1] cur_qids += 1 elif len(line_list) > 6: # urls position, title, content, click_label = line_list[0], line_list[2], line_list[3], line_list[5] try: src_input, src_segment = process_data(self.cur_query, title, content, self.max_seq_len) buffer.append([src_input, src_segment, cur_qids, int(click_label)]) except: pass if len(buffer) >= self.buffer_size: # we use 300,000 click records for test break return buffer#, total_qids, total_labels
15,976
41.155673
143
py
Tencent_wsdm_cup2023
Tencent_wsdm_cup2023-main/pytorch_unbias/pretrain/trainer.py
# -*- coding: utf-8 -*- import os from typing import Dict, List, Tuple, Optional, Any, Union import torch from transformers.trainer import Trainer from transformers.trainer_pt_utils import nested_detach import logging logger = logging.getLogger(__name__) class Pretrainer(Trainer): def compute_loss(self, model, inputs): outputs, mlm_loss, ctr_loss = model(**inputs) if self.state.global_step % self.args.logging_steps == 0: self.log({'MLM Loss':mlm_loss.item(), 'CTR Loss': ctr_loss.item()}) # loss = mlm_loss + ctr_loss loss = mlm_loss + ctr_loss return loss def _maybe_log_save_evaluate(self, tr_loss, model, trial, epoch, ignore_keys_for_eval): if self.control.should_log: ''' if is_torch_tpu_available(): xm.mark_step() ''' logs: Dict[str, float] = {} # all_gather + mean() to get average loss over all processes tr_loss_scalar = self._nested_gather(tr_loss).mean().item() # reset tr_loss to zero tr_loss -= tr_loss logs["loss"] = round(tr_loss_scalar / (self.state.global_step - self._globalstep_last_logged), 4) logs["learning_rate"] = self._get_learning_rate() self._total_loss_scalar += tr_loss_scalar self._globalstep_last_logged = self.state.global_step self.store_flos() self.log(logs) metrics = {} if self.control.should_evaluate: if isinstance(self.eval_dataset, dict): for eval_dataset_name, eval_dataset in self.eval_dataset.items(): if eval_dataset_name == 'click': self.label_names = ['qids', 'labels'] else: self.label_names = ['qids', 'labels', 'freqs'] local_metrics = self.evaluate( eval_dataset=eval_dataset, ignore_keys=ignore_keys_for_eval, metric_key_prefix=f"eval_{eval_dataset_name}", ) #logger.info(local_metrics) metrics.update(local_metrics) else: metrics = self.evaluate(ignore_keys=ignore_keys_for_eval) #logger.info(metrics) self._report_to_hp_search(trial, epoch, metrics)#self.state.global_step # ''' if self.control.should_evaluate: metrics = self.evaluate(ignore_keys=ignore_keys_for_eval) self._report_to_hp_search(trial, epoch, metrics) ''' if self.control.should_save: self._save_checkpoint(model, trial, metrics=metrics) self.control = self.callback_handler.on_save(self.args, self.state, self.control) def prediction_step( self, model, inputs: Dict[str, Union[torch.Tensor, Any]], prediction_loss_only: bool, ignore_keys: Optional[List[str]] = None, ) -> Tuple[Optional[torch.Tensor], Optional[torch.Tensor], Optional[torch.Tensor]]: assert prediction_loss_only == False assert ignore_keys is None #inputs = self._prepare_inputs(inputs) has_labels = all(inputs.get(k) is not None for k in self.label_names) inputs = self._prepare_inputs(inputs) # labels may be popped when computing the loss (label smoothing for instance) so we grab them first. if has_labels: labels = nested_detach(tuple(inputs.get(name) for name in self.label_names)) if len(labels) == 1: labels = labels[0] else: labels = None with torch.no_grad(): loss = None with self.autocast_smart_context_manager(): logits = model(**inputs).detach() return (loss, logits, labels)
3,918
36.32381
109
py
Tencent_wsdm_cup2023
Tencent_wsdm_cup2023-main/pytorch_unbias/models/modeling.py
# -*- coding: utf-8 -*- import torch from torch import nn import torch.distributed as dist import torch.nn.functional as F from transformers import BertModel, BertPreTrainedModel from torch.nn import CrossEntropyLoss import logging logger = logging.getLogger(__name__) from transformers.activations import ACT2FN from models.debias_model import DenoisingNetMultiFeature class BertPredictionHeadTransform(nn.Module): def __init__(self, config): super().__init__() self.dense = nn.Linear(config.hidden_size, config.hidden_size) if isinstance(config.hidden_act, str): self.transform_act_fn = ACT2FN[config.hidden_act] else: self.transform_act_fn = config.hidden_act self.LayerNorm = nn.LayerNorm(config.hidden_size, eps=config.layer_norm_eps) def forward(self, hidden_states): hidden_states = self.dense(hidden_states) hidden_states = self.transform_act_fn(hidden_states) hidden_states = self.LayerNorm(hidden_states) return hidden_states class BertLMPredictionHead(nn.Module): def __init__(self, config): super().__init__() self.transform = BertPredictionHeadTransform(config) # The output weights are the same as the input embeddings, but there is # an output-only bias for each token. self.decoder = nn.Linear(config.hidden_size, config.vocab_size, bias=False) self.bias = nn.Parameter(torch.zeros(config.vocab_size)) # Need a link between the two variables so that the bias is correctly resized with `resize_token_embeddings` self.decoder.bias = self.bias def forward(self, hidden_states): hidden_states = self.transform(hidden_states) hidden_states = self.decoder(hidden_states) return hidden_states class CTRPretrainingModel(BertPreTrainedModel): def __init__( self, config, model_args, data_args, training_args, fea_d=None, fea_c=None ): super(CTRPretrainingModel, self).__init__(config) self.bert = BertModel(config) self.num_candidates = data_args.num_candidates if self.num_candidates >= 0: fea_list = data_args.feature_list.strip().split(",") self.fea_name = [fea.strip().split("-") for fea in fea_list] self.emb_size = training_args.emb_size self.propensity_net = DenoisingNetMultiFeature(self.fea_name, self.emb_size, self.num_candidates, training_args.per_device_train_batch_size, training_args.train_group_size, fea_d, fea_c) self.dropout = nn.Dropout(config.hidden_dropout_prob) self.domlp = model_args.do_maxmeanmlp self.cls = nn.Linear(config.hidden_size, 1) self.predictions = BertLMPredictionHead(config) # self.bert.embeddings.word_embeddings.weight self.config = config self.model_args, self.data_args, self.training_args = model_args, data_args, training_args self.mlm_loss_fct = CrossEntropyLoss(ignore_index=-1) self.cross_entropy = nn.CrossEntropyLoss(reduction='mean') self.logits_to_prob = nn.Softmax(dim=-1) self.init_weights() if training_args.local_rank >= 0: self.world_size = dist.get_world_size() def get_group_data(self, inputs): inputs.view( self.training_args.per_device_train_batch_size, self.training_args.train_group_size ) def forward(self, input_ids, attention_mask, token_type_ids, masked_lm_labels=None, click_labels=None, **kwargs): outputs = self.bert(input_ids, attention_mask, token_type_ids, return_dict=True) sequence_output = outputs.last_hidden_state # (bs, seq_len, dim) if self.domlp: out = self.cls(sequence_output).squeeze() prediction_scores = out.max(dim=1).values + out.mean(dim=1) else: pooler_output = outputs.pooler_output # (bs, dim) prediction_scores = self.cls(pooler_output) # (bs, 1) mlm_loss = 0 if masked_lm_labels is not None: lm_prediction_scores = self.predictions(sequence_output) mlm_loss += self.mlm_loss_fct(lm_prediction_scores.view(-1, self.config.vocab_size), masked_lm_labels.view(-1)) # if config.MLM else 0. if click_labels is not None: # pointwise ctr_loss = self.pointwise_ctr_loss(prediction_scores, click_labels) else: # pairwise if self.num_candidates >= 0: # use propensity score select_pos = kwargs['rank_pos'] propensity_scores = self.propensity_net(kwargs["debias_fea"], select_pos) # (bs, 1) ctr_loss = self.groupwise_ctr_loss_with_dla(prediction_scores, propensity_scores) else: ctr_loss = self.groupwise_ctr_loss(prediction_scores) #loss = ctr_loss + mlm_loss return prediction_scores, mlm_loss, ctr_loss else: return prediction_scores def pointwise_ctr_loss(self, prediction_scores, click_labels): # TODO return None def groupwise_ctr_loss(self, prediction_scores): logits = prediction_scores.view(-1, 2) if self.training_args.temperature is not None: assert self.training_args.temperature > 0 logits = logits / self.training_args.temperature logits = logits.view( self.training_args.per_device_train_batch_size, self.training_args.train_group_size ) target_label = torch.zeros(self.training_args.per_device_train_batch_size, dtype=torch.long, device=logits.device) loss = self.cross_entropy(logits, target_label) return loss def groupwise_ctr_loss_with_dla(self, prediction_scores, propensity_scores_pos): logits = prediction_scores.view(-1, 2) if self.training_args.temperature is not None: assert self.training_args.temperature > 0 logits = logits / self.training_args.temperature logits = logits.view( self.training_args.per_device_train_batch_size, self.training_args.train_group_size ) propensity_scores_pos = propensity_scores_pos.view( self.training_args.per_device_train_batch_size, self.training_args.train_group_size ) propensity_scores_pos[:, 1:] = 0.1 propensity_scores = propensity_scores_pos logits = logits * propensity_scores target_label = torch.zeros(self.training_args.per_device_train_batch_size, dtype=torch.long, device=logits.device) loss = self.cross_entropy(logits, target_label) return loss
7,146
42.054217
116
py
Tencent_wsdm_cup2023
Tencent_wsdm_cup2023-main/pytorch_unbias/models/debias_model.py
# -*- coding: utf-8 -*- import torch from torch import nn from torch.nn import BatchNorm1d class DenoisingNetMultiFeature(nn.Module): def __init__(self, fea_name, emb_size, num_candidates, per_device_train_batch_size, train_group_size, fea_d, fea_c): super(DenoisingNetMultiFeature, self).__init__() input_vec_size = len(fea_name) * emb_size self.num_candidates = num_candidates self.per_device_train_batch_size = per_device_train_batch_size self.train_group_size = train_group_size ######################### Create MLP model ######################### self.relu_layer = nn.ReLU() self.norm = BatchNorm1d(input_vec_size) self.dropout = nn.Dropout(0.1) self.linear_layer_1 = nn.Linear(input_vec_size, input_vec_size) self.linear_layer_2 = nn.Linear(input_vec_size, input_vec_size) self.linear_layer_3 = nn.Linear(input_vec_size, int(input_vec_size / 2)) self.linear_layer_4 = nn.Linear(int(input_vec_size / 2), int(input_vec_size / 2)) self.linear_layer_5 = nn.Linear(int(input_vec_size / 2), 1) self.propensity_net = nn.Sequential( self.linear_layer_1, BatchNorm1d(input_vec_size), self.relu_layer, self.linear_layer_2, BatchNorm1d(input_vec_size), self.relu_layer, self.linear_layer_3, BatchNorm1d(int(input_vec_size / 2)), self.relu_layer, self.linear_layer_4, BatchNorm1d(int(input_vec_size / 2)), self.relu_layer, self.linear_layer_5, ).cuda() ############# Create Embedding for discrete feature ################ self.fea_emb = {} for name in fea_d: self.fea_emb[name] = torch.nn.Embedding(fea_d[name][0]+1, emb_size).cuda() for name in fea_c: self.fea_emb[name] = torch.nn.Embedding(fea_c[name][0]+1, emb_size).cuda() self.logits_to_prob = nn.Softmax(dim=1) def forward(self, debias_fea, select_pos): # get embedding for discrete feature fea = [] for name in debias_fea: fea.append(self.fea_emb[name](debias_fea[name])) # concat all features fea = torch.cat(fea, dim=1) fea = self.norm(fea) # cal bias score bias_scores = self.propensity_net(fea) bias_scores = bias_scores.view( self.per_device_train_batch_size, self.num_candidates ) bias_scores = self.logits_to_prob(bias_scores) select_pos = select_pos.view( self.per_device_train_batch_size, self.train_group_size ) select_bias_scores = [] for bs in range(self.per_device_train_batch_size): select_bias_scores.append(bias_scores[bs].index_select(dim=0, index=select_pos[bs])) select_bias_scores = torch.cat(select_bias_scores) return select_bias_scores
2,884
44.078125
120
py
Tencent_wsdm_cup2023
Tencent_wsdm_cup2023-main/paddle_pretrain/convert/convert-onnx.py
# -*- coding: utf-8 -*- # @Time : 2023/1/3 23:32 # @Author : Xiangsheng Li # @File : convert-onnx.py import sys import numpy as np sys.path.append('../../pytorch_pretrain') from transformers import AutoConfig from models.modeling import CTRPretrainingModel input_names = ["input_ids","attention_mask","token_type_ids"] output_names = ["output_0"] input_ids = torch.tensor(np.random.rand(1, 128).astype("int64")) attention_mask = torch.tensor(np.random.rand(1, 128).astype("float32")) token_type_ids = torch.tensor(np.random.rand(1, 128).astype("int64")) #Path to torch checkpoint bert_name_path = 'DATA_ROOT/outputs/finetune/large_group2_wwm/checkpoint-590000/checkpoint-4180' torch_config = AutoConfig.from_pretrained(bert_name_path) bert_model = CTRPretrainingModel.from_pretrained(bert_name_path,config=torch_config,model_args=None,data_args=None,training_args=None) torch.onnx.export(bert_model, (input_ids, attention_mask, token_type_ids), 'model.onnx', opset_version=11, input_names=input_names, output_names=output_names, dynamic_axes={'input_ids': [0],'attention_mask': [0],'token_type_ids': [0], 'output_0': [0]}) #python -m onnxsim model.onnx model_sim.onnx #x2paddle --framework=onnx --model=model_sim.onnx --save_dir=pd_model
1,335
40.75
138
py
Tencent_wsdm_cup2023
Tencent_wsdm_cup2023-main/paddle_pretrain/finetune/dataset.py
# -*- coding: utf-8 -*- # @Time : 2022/12/28 23:20 # @Author : Xiangsheng Li # @File : dataset.py import sys,os import random import collections from models.utils import SPECIAL_TOKENS import logging import numpy as np import paddle from paddle.io import Dataset, IterableDataset from dataclasses import dataclass from typing import List, Dict, Any from collections import defaultdict from tqdm.auto import tqdm from pretrain.dataset import ( process_data, TestDataset ) class PairwiseFinetuneDataset(Dataset): def __init__(self, fpath): self.queries = {} self.docs = dict() self.q_rels = defaultdict(list) self.q_irrels = defaultdict(list) self._load_annotated_data(fpath) self.idx2qids = {i:qid for i,qid in enumerate(self.q_rels.keys())} self.all_docids = list(self.docs.keys()) def _load_annotated_data(self, fpath): #buffer = [] doc2docid = {} for line in tqdm(open(fpath, 'rb'), desc='load annotated data'): line_list = line.strip(b'\n').split(b'\t') qid, query, title, content, label, freq = line_list if qid not in self.queries: self.queries[qid] = query if (title, content) not in doc2docid: doc2docid[(title, content)] = len(doc2docid) docid = doc2docid[(title, content)] if docid not in self.docs: self.docs[docid] = (title, content) label = int(label) if label >= 2: #2,1 self.q_rels[qid].append((docid, label)) else: self.q_irrels[qid].append((docid, label)) def __len__(self): return len(self.q_rels) def __getitem__(self, index): qid = self.idx2qids[index] query = self.queries[qid] rel_doc_id, rel_label = random.choice(self.q_rels[qid]) all_doc_list = self.q_rels[qid] + self.q_irrels[qid] rest_docids = [docid for (docid, label) in all_doc_list if (rel_doc_id != docid and rel_label > label)] if len(rest_docids) == 0: irrel_doc_id = random.choice(self.all_docids) else: irrel_doc_id = random.choice(rest_docids) return {'query':query, 'rel_doc': self.docs[rel_doc_id], 'irrel_doc': self.docs[irrel_doc_id]} @dataclass class PairwiseDataCollator: def __init__(self, max_seq_len): self.max_seq_len = max_seq_len def __call__(self, features: Dict[str, Any]) -> Dict[str, Any]: batch_src_input, batch_segment = [], [] for sample in features: src_input, segment = process_data(sample['query'], sample['rel_doc'][0], sample['rel_doc'][1], self.max_seq_len) batch_src_input.append(src_input) batch_segment.append(segment) src_input, segment = process_data(sample['query'], sample['irrel_doc'][0], sample['irrel_doc'][1], self.max_seq_len) batch_src_input.append(src_input) batch_segment.append(segment) batch_src_input = np.array(batch_src_input,dtype=np.int64) #torch.LongTensor(batch_src_input) batch_segment = np.array(batch_segment,dtype=np.int64) #torch.LongTensor(batch_segment) batch_data = { 'input_ids': paddle.to_tensor(batch_src_input, dtype="int64"), 'attention_mask': paddle.to_tensor((batch_src_input > 0).astype('float32'), dtype='float32'), # torch.FloatTensor, 0 is PAD 'token_type_ids': paddle.to_tensor(batch_segment, dtype="int64") } return batch_data
3,711
35.038835
136
py
Tencent_wsdm_cup2023
Tencent_wsdm_cup2023-main/paddle_pretrain/finetune/trainer.py
# -*- coding: utf-8 -*- # @Time : 2022/12/28 23:27 # @Author : Xiangsheng Li # @File : trainer.py import os from typing import Dict, List, Tuple, Optional, Any, Union ''' import torch from torch.utils.data import DataLoader from torch.nn import Softmax, MarginRankingLoss ''' import paddle from paddle.io import DataLoader from paddle.nn import Softmax, MarginRankingLoss from pretrain.trainer import Pretrainer as Trainer from pretrain.dataset import DataCollator import logging logger = logging.getLogger(__name__) class Finetuner(Trainer): def __init__(self, *args, **kwargs): super(Finetuner, self).__init__(*args, **kwargs) self.softmax = Softmax(axis=1) def compute_loss(self, model, inputs): return self.compute_pair_loss(model, inputs) def compute_pair_loss(self, model, inputs): prediction_scores = model(**inputs) logits = prediction_scores.reshape((-1, 2)) if self.args.temperature is not None: assert self.args.temperature > 0 logits = logits / self.args.temperature logits = self.softmax(logits) pos_logits = logits[:, 0] neg_logits = logits[:, 1] marginloss = MarginRankingLoss(margin=1.0) pair_label = paddle.ones_like(pos_logits) pair_loss = marginloss(pos_logits, neg_logits, pair_label) return pair_loss def get_eval_dataloader(self, eval_dataset) -> DataLoader: # use DataCollator for eval dataset eval_dataset = eval_dataset if eval_dataset is not None else self.eval_dataset data_collator = DataCollator() eval_sampler = self._get_eval_sampler(eval_dataset) return DataLoader( eval_dataset, batch_sampler=eval_sampler, collate_fn=data_collator, num_workers=self.args.dataloader_num_workers, )
1,876
26.202899
86
py
Tencent_wsdm_cup2023
Tencent_wsdm_cup2023-main/paddle_pretrain/models/modeling.py
# -*- coding: utf-8 -*- # @Time : 2022/10/25 12:02 # @Author : Xiangsheng Li # @File : modeling.py import paddle import paddle.nn as nn import paddle.nn.functional as F from paddle.nn import Layer from paddlenlp.transformers import ( BertPretrainedModel as BertPreTrainedModel, BertModel, ACT2FN ) from paddle.nn import CrossEntropyLoss import logging logger = logging.getLogger(__name__) class BertPredictionHeadTransform(Layer): def __init__(self, config): super().__init__() self.dense = nn.Linear(config.hidden_size, config.hidden_size) if isinstance(config.hidden_act, str): self.transform_act_fn = ACT2FN[config.hidden_act] else: self.transform_act_fn = config.hidden_act self.LayerNorm = nn.LayerNorm(config.hidden_size, epsilon=config.layer_norm_eps) def forward(self, hidden_states): hidden_states = self.dense(hidden_states) hidden_states = self.transform_act_fn(hidden_states) hidden_states = self.LayerNorm(hidden_states) return hidden_states class BertLMPredictionHead(Layer): def __init__(self, config): super().__init__() self.transform = BertPredictionHeadTransform(config) # The output weights are the same as the input embeddings, but there is # an output-only bias for each token. # Need a link between the two variables so that the bias is correctly resized with `resize_token_embeddings` #self.bias = nn.Parameter(torch.zeros(config.vocab_size)) #self.decoder.bias = self.bias bias_attr = paddle.ParamAttr( name="bias", initializer=paddle.nn.initializer.Constant(value=0.0)) self.decoder = nn.Linear(config.hidden_size, config.vocab_size, bias_attr=bias_attr) # , bias=False def forward(self, hidden_states): hidden_states = self.transform(hidden_states) hidden_states = self.decoder(hidden_states) return hidden_states class CTRPretrainingModel(BertPreTrainedModel): def __init__( self, config, model_args, data_args, training_args ): super(CTRPretrainingModel, self).__init__(config) self.bert = BertModel(config) self.dropout = nn.Dropout(config.hidden_dropout_prob) self.cls = nn.Linear(config.hidden_size, 1) self.predictions = BertLMPredictionHead(config)#self.bert.embeddings.word_embeddings.weight self.config = config self.model_args, self.data_args, self.training_args = model_args, data_args, training_args self.mlm_loss_fct = CrossEntropyLoss(ignore_index=-1) self.cross_entropy = nn.CrossEntropyLoss(reduction='mean') def forward(self, input_ids, attention_mask, token_type_ids, masked_lm_labels=None): outputs = self.bert(input_ids, attention_mask=attention_mask, token_type_ids=token_type_ids, return_dict=True) sequence_output = outputs.last_hidden_state #(bs, seq_len, dim),outputs.last_hidden_state pooler_output = outputs.pooler_output #(bs, dim),outputs.pooler_output prediction_scores = self.cls(pooler_output) #(bs, 1) mlm_loss, ctr_loss = None, None if masked_lm_labels is not None: lm_prediction_scores = self.predictions(sequence_output) mlm_loss = self.mlm_loss_fct(lm_prediction_scores.reshape((-1, self.config.vocab_size)), masked_lm_labels.reshape((-1,)) )# if config.MLM else 0. ctr_loss = self.groupwise_ctr_loss(prediction_scores) if mlm_loss and ctr_loss: return prediction_scores, mlm_loss, ctr_loss else: return prediction_scores def groupwise_ctr_loss(self, prediction_scores): logits = prediction_scores if self.training_args.temperature is not None: assert self.training_args.temperature > 0 logits = logits / self.training_args.temperature logits = logits.reshape( (self.training_args.per_device_train_batch_size, self.training_args.train_group_size) ) target_label = paddle.zeros((self.training_args.per_device_train_batch_size,), dtype='int64') loss = self.cross_entropy(logits, target_label) return loss
4,428
36.218487
118
py
Tencent_wsdm_cup2023
Tencent_wsdm_cup2023-main/pytorch_pretrain/mt_pretrain/dataset.py
# -*- coding: utf-8 -*- # @Time : 2022/10/25 21:07 # @Author : Xiangsheng Li # @File : dataset.py import sys,os import random import collections from models.utils import SPECIAL_TOKENS import logging import numpy as np import torch from torch.utils.data import Dataset,IterableDataset from dataclasses import dataclass from typing import List, Dict, Any logger = logging.getLogger(__name__) MaskedLmInstance = collections.namedtuple("MaskedLmInstance", ["index", "label"]) class MaskingOp(): def __init__(self, masked_lm_prob, max_predictions_per_seq, vocab_list, unigram_set=None): self.masked_lm_prob = masked_lm_prob self.max_predictions_per_seq = max_predictions_per_seq self.vocab_list = vocab_list self.unigram_set = unigram_set def create_masked_lm_predictions(self, tokens): """Creates the predictions for the masked LM objective. This is mostly copied from the Google BERT repo, but with several refactors to clean it up and remove a lot of unnecessary variables.""" cand_indices = [] START_FROM_DOC = False start = 0 #for (i, token) in enumerate(tokens): # token_ids while start < len(tokens): token = tokens[start] if token == SPECIAL_TOKENS['SEP']: # SEP START_FROM_DOC = True start += 1 continue if token in SPECIAL_TOKENS.values(): # CLS start += 1 continue if not START_FROM_DOC: start += 1 continue if self.unigram_set is not None: end = len(tokens) while start < end: if tuple(tokens[start:end]) in self.unigram_set or (start + 1 == end): cand_indices.append(list(range(start,end))) break end -= 1 start = end else: cand_indices.append([start]) start += 1 num_to_mask = min(self.max_predictions_per_seq, max(1, int(round(len(cand_indices) * self.masked_lm_prob)))) random.shuffle(cand_indices) masked_lms = [] covered_indexes = set() for index_set in cand_indices: if len(masked_lms) >= num_to_mask: break # If adding a whole-word mask would exceed the maximum number of # predictions, then just skip this candidate. if len(masked_lms) + len(index_set) > num_to_mask: continue is_any_index_covered = False for index in index_set: if index in covered_indexes: is_any_index_covered = True break if is_any_index_covered: continue for index in index_set: covered_indexes.add(index) # 80% of the time, replace with [MASK] if random.random() < 0.8: masked_token = SPECIAL_TOKENS['MASK']#103 else: # 10% of the time, keep original if random.random() < 0.5: masked_token = tokens[index] # 10% of the time, replace with random word else: masked_token = random.choice(self.vocab_list) masked_lms.append(MaskedLmInstance(index=index, label=tokens[index])) tokens[index] = masked_token assert len(masked_lms) <= num_to_mask masked_lms = sorted(masked_lms, key=lambda x: x.index) mask_indices = [p.index for p in masked_lms] masked_token_labels = [p.label for p in masked_lms] return tokens, masked_token_labels, mask_indices def process_data(query, title, content, max_seq_len, masking_obj=None): """ process [query, title, content] into a tensor [CLS] + query + [SEP] + title + [SEP] + content + [SEP] + [PAD] """ data = [SPECIAL_TOKENS['CLS']] segment = [0] query_splits = [item for item in query.split(b'\x01') if len(item.strip()) > 0] data = data + [int(item) + 10 for item in query_splits] # query data = data + [SPECIAL_TOKENS['SEP']] segment = segment + [0] * (len(query_splits) + 1) title_splits = [item for item in title.split(b'\x01') if len(item.strip()) > 0] data = data + [int(item) + 10 for item in title_splits] # content data = data + [SPECIAL_TOKENS['SEP']] # sep defined as 1 segment = segment + [1] * (len(title_splits) + 1) content_splits = [item for item in content.split(b'\x01') if len(item.strip()) > 0] data = data + [int(item) + 10 for item in content_splits] # content data = data + [SPECIAL_TOKENS['SEP']] segment = segment + [1] * (len(content_splits) + 1) #padding_mask = [False] * len(data) if len(data) < max_seq_len: #padding_mask += [True] * (max_seq_len - len(data)) data += [SPECIAL_TOKENS['PAD']] * (max_seq_len - len(data)) else: #padding_mask = padding_mask[:max_seq_len] data = data[:max_seq_len] # segment id if len(segment) < max_seq_len: segment += [1] * (max_seq_len - len(segment)) else: segment = segment[:max_seq_len] #padding_mask = torch.BoolTensor(padding_mask) #data = torch.LongTensor(data) #segment = torch.LongTensor(segment) #print("data", data) if masking_obj is not None: token_ids, masked_lm_ids, masked_lm_positions = masking_obj.create_masked_lm_predictions(data) lm_label_array = np.full(max_seq_len, dtype=np.int, fill_value=-1) lm_label_array[masked_lm_positions] = masked_lm_ids return token_ids, segment, lm_label_array else: return data, segment class TrainDatasetBase(IterableDataset): def __init__(self, directory_path, args): self.directory_path = directory_path self.files = [f for f in os.listdir(self.directory_path) if f.startswith('part')] random.shuffle(self.files) self.cur_query = "#" self.max_seq_len = args.max_seq_len#128 self.vocab_size = args.vocab_size#22000 self.vocab_list = list(range(self.vocab_size)) #self.unigram_set = set([tuple(map(lambda t:int(t) + 10, line.strip().split('\x01'))) for line in open(args.unigram_dict_addr) if len(line.strip()) > 0]) if args.unigram_dict_addr is not None: self.unigram_set = set( [tuple(map(lambda t:int(t) + 10, line.strip().split('\x01'))) for line in open(args.unigram_dict_addr) if len(line.strip()) > 0] ) else: self.unigram_set = None self.masking_obj = MaskingOp(args.masked_lm_prob, args.max_predictions_per_seq, self.vocab_list, self.unigram_set) def __iter__(self): buffer_per_query = [] #sample_count = 0 info = torch.utils.data.get_worker_info() if info is None: worker_num = 1 worker_id = 0 else: worker_num = info.num_workers worker_id = info.id ## each worker parses one file local_files = [f for i, f in enumerate(self.files) if i % worker_num == worker_id] for i, file in enumerate(local_files): logger.info(f'load file: {file}') if file == 'part-00000': # part-00000.gz is for evaluation continue for line in open(os.path.join(self.directory_path, file), 'rb'): line_list = line.strip(b'\n').split(b'\t') if len(line_list) == 3: # new query self.cur_query = line_list[1] #print("line_list: ", buffer_per_query) if len(buffer_per_query) > 0: (norm_pair, pos_pair) = self.yield_data(buffer_per_query) if norm_pair is not None: yield norm_pair if pos_pair is not None: yield pos_pair buffer_per_query = [] elif len(line_list) > 6: # urls position, title, content, click_label = line_list[0], line_list[2], line_list[3], line_list[5] dwell_time = float(line_list[16]) try: src_input, segment, masked_lm_labels = process_data(self.cur_query, title, content, self.max_seq_len, self.masking_obj) sample = {'src_input': src_input, 'segment': segment, 'masked_lm_labels': masked_lm_labels, 'click_label': float(click_label), 'dwell_time':dwell_time} buffer_per_query.append(sample) except Exception as e: #print(e) pass def yield_data(self,buffer_per_query): pass class PreTrainDatasetGroupwise(TrainDatasetBase): def __init__(self, directory_path, train_group_size, args): super(PreTrainDatasetGroupwise, self).__init__(directory_path, args) self.train_group_size = train_group_size def yield_data(self, buffer_per_query): random.shuffle(buffer_per_query) pos_buffer, neg_buffer = [], [] for record in buffer_per_query: if record['click_label'] > 0: pos_buffer.append(record) else: neg_buffer.append(record) if len(pos_buffer) == 0 or len(neg_buffer) == 0: return None, None pos_record = random.choice(pos_buffer) if len(neg_buffer) < self.train_group_size - 1: negs = random.choices(neg_buffer, k=self.train_group_size - 1) else: negs = random.sample(neg_buffer, k=self.train_group_size - 1) group = [pos_record] + negs pos_pair = None if len(pos_buffer) >= 2: pos_pair = random.choices(pos_buffer, k=2) diff = pos_pair[0]['dwell_time'] - pos_pair[1]['dwell_time'] if abs(diff) < 8: #diff time larger than 8 pos_pair = None elif diff < 0: pos_pair = [pos_pair[1], pos_pair[0]] return (group, pos_pair) @dataclass class DataCollator: def __call__(self, features) -> Dict[str, Any]: if isinstance(features[0], list): features = sum(features, [])#groupwise dataset token_ids = torch.LongTensor([f['src_input'] for f in features]) segment = torch.LongTensor([f['segment'] for f in features]) #For test data batch_data = { 'input_ids':token_ids, 'attention_mask':(token_ids > 0).float(),#torch.FloatTensor, 0 is PAD 'token_type_ids':segment } #For training data, pointwise or groupwise if 'masked_lm_labels' in features[0]: masked_lm_labels = torch.LongTensor([f['masked_lm_labels'] for f in features]) batch_data.update({'masked_lm_labels': masked_lm_labels}) if 'click_label' in features[0]: click_labels = torch.LongTensor([f['click_label'] for f in features]) batch_data.update({'click_labels': click_labels}) return batch_data @dataclass class TestDataCollator: def __call__(self, features) -> Dict[str, Any]: token_ids = torch.LongTensor([f['src_input'] for f in features]) segment = torch.LongTensor([f['segment'] for f in features]) #For test data batch_data = { 'input_ids':token_ids, 'attention_mask':(token_ids > 0).float(),#torch.FloatTensor, 0 is PAD 'token_type_ids':segment } #For testing data if 'qid' in features[0]: qids = torch.LongTensor([f['qid'] for f in features]) batch_data.update({'qids': qids}) if 'label' in features[0]: labels = torch.LongTensor([f['label'] for f in features]) batch_data.update({'labels': labels}) if 'freq' in features[0]: freqs = torch.LongTensor([f['freq'] for f in features]) batch_data.update({'freqs': freqs}) return batch_data ###################Test dataset################################ class TestDataset(Dataset): def __init__(self, fpath, max_seq_len, data_type, buffer_size=300000): self.buffer_size = buffer_size self.max_seq_len = max_seq_len self.data_type = data_type if data_type == 'annotate': self.buffer = self.load_annotate_data(fpath) #self.buffer = self.buffer[:10000] elif data_type == 'click': self.buffer = self.load_click_data(fpath) #self.total_freqs = None def __len__(self): return len(self.buffer) def __getitem__(self, index): #return self.buffer[index] if len(self.buffer[index]) == 4: src_input, segment, qid, label = self.buffer[index] sample = {'src_input': src_input, 'segment': segment, 'qid': qid, 'label': label} else: src_input, segment, qid, label, freq = self.buffer[index] sample = {'src_input': src_input, 'segment': segment, 'qid':qid, 'label':label, 'freq':freq} return sample def load_annotate_data(self, fpath): logger.info(f'load annotated data from {fpath}') #total_qids = [] buffer = [] #total_labels = [] #total_freqs = [] for line in open(fpath, 'rb'): line_list = line.strip(b'\n').split(b'\t') qid, query, title, content, label, freq = line_list if 0 <= int(freq) <= 2: # high freq freq = 0 elif 3 <= int(freq) <= 6: # mid freq freq = 1 elif 7 <= int(freq): # tail freq = 2 #total_qids.append(int(qid)) #total_labels.append(int(label)) #total_freqs.append(freq) src_input, src_segment = process_data(query, title, content, self.max_seq_len) buffer.append([src_input, src_segment, int(qid), int(label), freq]) return buffer#, total_qids, total_labels, total_freqs def load_click_data(self, fpath): logger.info(f'load logged click data from {fpath}') buffer = [] #total_qids = [] #total_labels = [] cur_qids = 0 for line in open(fpath, 'rb'): line_list = line.strip(b'\n').split(b'\t') if len(line_list) == 3: # new query self.cur_query = line_list[1] cur_qids += 1 elif len(line_list) > 6: # urls position, title, content, click_label = line_list[0], line_list[2], line_list[3], line_list[5] try: src_input, src_segment = process_data(self.cur_query, title, content, self.max_seq_len) buffer.append([src_input, src_segment, cur_qids, int(click_label)]) #total_qids.append(cur_qids) #total_labels.append(int(click_label)) except: pass if len(buffer) >= self.buffer_size: # we use 300,000 click records for test break return buffer#, total_qids, total_labels
15,597
39.201031
161
py
Tencent_wsdm_cup2023
Tencent_wsdm_cup2023-main/pytorch_pretrain/mt_pretrain/trainer.py
# -*- coding: utf-8 -*- # @Time : 2022/10/26 16:51 # @Author : Xiangsheng Li # @File : trainer.py import os from typing import Dict, List, Tuple, Optional, Any, Union import torch import torch.distributed as dist from torch import nn, Tensor from torch.cuda.amp import autocast import torch.nn.functional as F from transformers.trainer import Trainer from transformers.trainer_utils import EvalLoopOutput, has_length from transformers.trainer_pt_utils import nested_concat, nested_numpify, nested_detach from torch.utils.data import DataLoader from mt_pretrain.dataset import TestDataCollator import logging from tqdm.auto import tqdm from models.metrics import evaluate_all_metric logger = logging.getLogger(__name__) class Pretrainer(Trainer): def compute_loss(self,model, inputs): outputs, mlm_loss, ctr_loss = model(**inputs) if self.state.global_step % self.args.logging_steps == 0: self.log({'MLM Loss':mlm_loss.item(), 'CTR Loss': ctr_loss.item()}) loss = mlm_loss + ctr_loss return loss def _maybe_log_save_evaluate(self, tr_loss, model, trial, epoch, ignore_keys_for_eval): if self.control.should_log: logs: Dict[str, float] = {} # all_gather + mean() to get average loss over all processes tr_loss_scalar = self._nested_gather(tr_loss).mean().item() # reset tr_loss to zero tr_loss -= tr_loss logs["loss"] = round(tr_loss_scalar / (self.state.global_step - self._globalstep_last_logged), 4) logs["learning_rate"] = self._get_learning_rate() self._total_loss_scalar += tr_loss_scalar self._globalstep_last_logged = self.state.global_step self.store_flos() self.log(logs) metrics = {} if self.control.should_evaluate: if isinstance(self.eval_dataset, dict): for eval_dataset_name, eval_dataset in self.eval_dataset.items(): if eval_dataset_name == 'click': self.label_names = ['qids', 'labels'] else: self.label_names = ['qids', 'labels', 'freqs'] local_metrics = self.evaluate( eval_dataset=eval_dataset, ignore_keys=ignore_keys_for_eval, metric_key_prefix=f"eval_{eval_dataset_name}", ) #logger.info(local_metrics) metrics.update(local_metrics) else: metrics = self.evaluate(ignore_keys=ignore_keys_for_eval) #logger.info(metrics) self._report_to_hp_search(trial, epoch, metrics)#self.state.global_step if self.control.should_save: self._save_checkpoint(model, trial, metrics=metrics) self.control = self.callback_handler.on_save(self.args, self.state, self.control) def prediction_step( self, model, inputs: Dict[str, Union[torch.Tensor, Any]], prediction_loss_only: bool, ignore_keys: Optional[List[str]] = None, ) -> Tuple[Optional[torch.Tensor], Optional[torch.Tensor], Optional[torch.Tensor]]: assert prediction_loss_only == False assert ignore_keys is None #inputs = self._prepare_inputs(inputs) has_labels = all(inputs.get(k) is not None for k in self.label_names) inputs = self._prepare_inputs(inputs) # labels may be popped when computing the loss (label smoothing for instance) so we grab them first. if has_labels: labels = nested_detach(tuple(inputs.get(name) for name in self.label_names)) if len(labels) == 1: labels = labels[0] else: labels = None with torch.no_grad(): loss = None with self.autocast_smart_context_manager(): logits = model(**inputs).detach().contiguous() return (loss, logits, labels) def get_eval_dataloader(self, eval_dataset) -> DataLoader: # use DataCollator for eval dataset eval_dataset = eval_dataset if eval_dataset is not None else self.eval_dataset data_collator = TestDataCollator() eval_sampler = self._get_eval_sampler(eval_dataset) return DataLoader( eval_dataset, sampler=eval_sampler, batch_size=self.args.eval_batch_size, collate_fn=data_collator, drop_last=self.args.dataloader_drop_last, num_workers=self.args.dataloader_num_workers, pin_memory=self.args.dataloader_pin_memory, )
4,701
37.540984
109
py
Tencent_wsdm_cup2023
Tencent_wsdm_cup2023-main/pytorch_pretrain/pretrain/dataset.py
# -*- coding: utf-8 -*- # @Time : 2022/10/25 21:07 # @Author : Xiangsheng Li # @File : dataset.py import sys,os import random import collections from models.utils import SPECIAL_TOKENS import logging import numpy as np import torch from torch.utils.data import Dataset,IterableDataset from dataclasses import dataclass from typing import List, Dict, Any logger = logging.getLogger(__name__) MaskedLmInstance = collections.namedtuple("MaskedLmInstance", ["index", "label"]) class MaskingOp(): def __init__(self, masked_lm_prob, max_predictions_per_seq, vocab_list, unigram_set=None): self.masked_lm_prob = masked_lm_prob self.max_predictions_per_seq = max_predictions_per_seq self.vocab_list = vocab_list self.unigram_set = unigram_set def create_masked_lm_predictions(self, tokens): """Creates the predictions for the masked LM objective. This is mostly copied from the Google BERT repo, but with several refactors to clean it up and remove a lot of unnecessary variables.""" cand_indices = [] START_FROM_DOC = False start = 0 #for (i, token) in enumerate(tokens): # token_ids while start < len(tokens): token = tokens[start] if token == SPECIAL_TOKENS['SEP']: # SEP START_FROM_DOC = True start += 1 continue if token in SPECIAL_TOKENS.values(): # CLS start += 1 continue if not START_FROM_DOC: start += 1 continue if self.unigram_set is not None: end = len(tokens) while start < end: if tuple(tokens[start:end]) in self.unigram_set or (start + 1 == end): cand_indices.append(list(range(start,end))) break end -= 1 start = end else: cand_indices.append([start]) start += 1 num_to_mask = min(self.max_predictions_per_seq, max(1, int(round(len(cand_indices) * self.masked_lm_prob)))) random.shuffle(cand_indices) masked_lms = [] covered_indexes = set() for index_set in cand_indices: if len(masked_lms) >= num_to_mask: break # If adding a whole-word mask would exceed the maximum number of # predictions, then just skip this candidate. if len(masked_lms) + len(index_set) > num_to_mask: continue is_any_index_covered = False for index in index_set: if index in covered_indexes: is_any_index_covered = True break if is_any_index_covered: continue for index in index_set: covered_indexes.add(index) # 80% of the time, replace with [MASK] if random.random() < 0.8: masked_token = SPECIAL_TOKENS['MASK']#103 else: # 10% of the time, keep original if random.random() < 0.5: masked_token = tokens[index] # 10% of the time, replace with random word else: masked_token = random.choice(self.vocab_list) masked_lms.append(MaskedLmInstance(index=index, label=tokens[index])) tokens[index] = masked_token assert len(masked_lms) <= num_to_mask masked_lms = sorted(masked_lms, key=lambda x: x.index) mask_indices = [p.index for p in masked_lms] masked_token_labels = [p.label for p in masked_lms] return tokens, masked_token_labels, mask_indices def process_data(query, title, content, max_seq_len, masking_obj=None): """ process [query, title, content] into a tensor [CLS] + query + [SEP] + title + [SEP] + content + [SEP] + [PAD] """ data = [SPECIAL_TOKENS['CLS']] segment = [0] query_splits = [item for item in query.split(b'\x01') if len(item.strip()) > 0] data = data + [int(item) + 10 for item in query_splits] # query data = data + [SPECIAL_TOKENS['SEP']] segment = segment + [0] * (len(query_splits) + 1) title_splits = [item for item in title.split(b'\x01') if len(item.strip()) > 0] data = data + [int(item) + 10 for item in title_splits] # content data = data + [SPECIAL_TOKENS['SEP']] # sep defined as 1 segment = segment + [1] * (len(title_splits) + 1) content_splits = [item for item in content.split(b'\x01') if len(item.strip()) > 0] data = data + [int(item) + 10 for item in content_splits] # content data = data + [SPECIAL_TOKENS['SEP']] segment = segment + [1] * (len(content_splits) + 1) #padding_mask = [False] * len(data) if len(data) < max_seq_len: #padding_mask += [True] * (max_seq_len - len(data)) data += [SPECIAL_TOKENS['PAD']] * (max_seq_len - len(data)) else: #padding_mask = padding_mask[:max_seq_len] data = data[:max_seq_len] # segment id if len(segment) < max_seq_len: segment += [1] * (max_seq_len - len(segment)) else: segment = segment[:max_seq_len] if masking_obj is not None: token_ids, masked_lm_ids, masked_lm_positions = masking_obj.create_masked_lm_predictions(data) lm_label_array = np.full(max_seq_len, dtype=np.int, fill_value=-1) lm_label_array[masked_lm_positions] = masked_lm_ids return token_ids, segment, lm_label_array else: return data, segment class TrainDatasetBase(IterableDataset): def __init__(self, directory_path, args): self.directory_path = directory_path self.files = [f for f in os.listdir(self.directory_path) if f.startswith('part')] random.shuffle(self.files) self.cur_query = "#" self.max_seq_len = args.max_seq_len#128 self.vocab_size = args.vocab_size#22000 self.vocab_list = list(range(self.vocab_size)) self.unigram_set = set([tuple(map(lambda t:int(t) + 10, line.strip().split('\x01'))) for line in open(args.unigram_dict_addr) if len(line.strip()) > 0]) self.masking_obj = MaskingOp(args.masked_lm_prob, args.max_predictions_per_seq, self.vocab_list, self.unigram_set) def __iter__(self): buffer_per_query = [] #sample_count = 0 info = torch.utils.data.get_worker_info() if info is None: worker_num = 1 worker_id = 0 else: worker_num = info.num_workers worker_id = info.id ## each worker parses one file local_files = [f for i, f in enumerate(self.files) if i % worker_num == worker_id] for i, file in enumerate(local_files): logger.info(f'load file: {file}') if file == 'part-00000': # part-00000.gz is for evaluation continue for line in open(os.path.join(self.directory_path, file), 'rb'): line_list = line.strip(b'\n').split(b'\t') if len(line_list) == 3: # new query self.cur_query = line_list[1] #print("line_list: ", buffer_per_query) if len(buffer_per_query) > 0: data = self.yield_data(buffer_per_query) #print("buffer_per_query: ", buffer_per_query , data) if data is not None: yield data buffer_per_query = [] elif len(line_list) > 6: # urls position, title, content, click_label = line_list[0], line_list[2], line_list[3], line_list[5] try: src_input, segment, masked_lm_labels = process_data(self.cur_query, title, content, self.max_seq_len, self.masking_obj) sample = {'src_input': src_input, 'segment': segment, 'masked_lm_labels': masked_lm_labels, 'click_label': float(click_label)} buffer_per_query.append(sample) except Exception as e: #print(e) pass def yield_data(self,buffer_per_query): pass class PreTrainDatasetGroupwise(TrainDatasetBase): def __init__(self, directory_path, train_group_size, args): super(PreTrainDatasetGroupwise, self).__init__(directory_path, args) self.train_group_size = train_group_size def yield_data(self, buffer_per_query): random.shuffle(buffer_per_query) pos_buffer, neg_buffer = [], [] for record in buffer_per_query: if record['click_label'] > 0: pos_buffer.append(record) else: neg_buffer.append(record) if len(pos_buffer) == 0 or len(neg_buffer) == 0: return None pos_record = random.choice(pos_buffer) if len(neg_buffer) < self.train_group_size - 1: negs = random.choices(neg_buffer, k=self.train_group_size - 1) else: negs = random.sample(neg_buffer, k=self.train_group_size - 1) group = [pos_record] + negs return group @dataclass class DataCollator: def __call__(self, features) -> Dict[str, Any]: if isinstance(features[0], list): features = sum(features, [])#groupwise dataset token_ids = torch.LongTensor([f['src_input'] for f in features]) segment = torch.LongTensor([f['segment'] for f in features]) #For train and test data batch_data = { 'input_ids':token_ids, 'attention_mask':(token_ids > 0).float(),#torch.FloatTensor, 0 is PAD 'token_type_ids':segment } #For training data, pointwise or groupwise if 'masked_lm_labels' in features[0]: masked_lm_labels = torch.LongTensor([f['masked_lm_labels'] for f in features]) batch_data.update({'masked_lm_labels': masked_lm_labels}) #For testing data if 'qid' in features[0]: qids = torch.LongTensor([f['qid'] for f in features]) batch_data.update({'qids': qids}) if 'label' in features[0]: labels = torch.LongTensor([f['label'] for f in features]) batch_data.update({'labels': labels}) if 'freq' in features[0]: freqs = torch.LongTensor([f['freq'] for f in features]) batch_data.update({'freqs': freqs}) return batch_data ###################Test dataset################################ class TestDataset(Dataset): def __init__(self, fpath, max_seq_len, data_type, buffer_size=300000): self.buffer_size = buffer_size self.max_seq_len = max_seq_len self.data_type = data_type if data_type == 'annotate': self.buffer = self.load_annotate_data(fpath) elif data_type == 'click': self.buffer = self.load_click_data(fpath) def __len__(self): return len(self.buffer) def __getitem__(self, index): if len(self.buffer[index]) == 4: src_input, segment, qid, label = self.buffer[index] sample = {'src_input': src_input, 'segment': segment, 'qid': qid, 'label': label} else: src_input, segment, qid, label, freq = self.buffer[index] sample = {'src_input': src_input, 'segment': segment, 'qid':qid, 'label':label, 'freq':freq} return sample def load_annotate_data(self, fpath): logger.info(f'load annotated data from {fpath}') buffer = [] for line in open(fpath, 'rb'): line_list = line.strip(b'\n').split(b'\t') qid, query, title, content, label, freq = line_list if 0 <= int(freq) <= 2: # high freq freq = 0 elif 3 <= int(freq) <= 6: # mid freq freq = 1 elif 7 <= int(freq): # tail freq = 2 src_input, src_segment = process_data(query, title, content, self.max_seq_len) buffer.append([src_input, src_segment, int(qid), int(label), freq]) return buffer def load_click_data(self, fpath): logger.info(f'load logged click data from {fpath}') buffer = [] cur_qids = 0 for line in open(fpath, 'rb'): line_list = line.strip(b'\n').split(b'\t') if len(line_list) == 3: # new query self.cur_query = line_list[1] cur_qids += 1 elif len(line_list) > 6: # urls position, title, content, click_label = line_list[0], line_list[2], line_list[3], line_list[5] try: src_input, src_segment = process_data(self.cur_query, title, content, self.max_seq_len) buffer.append([src_input, src_segment, cur_qids, int(click_label)]) except: pass if len(buffer) >= self.buffer_size: # we use 300,000 click records for test break return buffer
13,414
39.044776
160
py
Tencent_wsdm_cup2023
Tencent_wsdm_cup2023-main/pytorch_pretrain/pretrain/trainer.py
# -*- coding: utf-8 -*- # @Time : 2022/10/26 16:51 # @Author : Xiangsheng Li # @File : trainer.py import os from typing import Dict, List, Tuple, Optional, Any, Union import torch import torch.distributed as dist from torch import nn, Tensor from torch.cuda.amp import autocast import torch.nn.functional as F from transformers.trainer import Trainer from transformers.trainer_utils import EvalLoopOutput, has_length from transformers.trainer_pt_utils import nested_concat, nested_numpify, nested_detach from torch.utils.data import DataLoader import logging from tqdm.auto import tqdm from models.metrics import evaluate_all_metric logger = logging.getLogger(__name__) class Pretrainer(Trainer): def compute_loss(self,model, inputs): outputs, mlm_loss, ctr_loss = model(**inputs) if self.state.global_step % self.args.logging_steps == 0: self.log({'MLM Loss':mlm_loss.item(), 'CTR Loss': ctr_loss.item()}) loss = mlm_loss + ctr_loss return loss def _maybe_log_save_evaluate(self, tr_loss, model, trial, epoch, ignore_keys_for_eval): if self.control.should_log: logs: Dict[str, float] = {} # all_gather + mean() to get average loss over all processes tr_loss_scalar = self._nested_gather(tr_loss).mean().item() # reset tr_loss to zero tr_loss -= tr_loss logs["loss"] = round(tr_loss_scalar / (self.state.global_step - self._globalstep_last_logged), 4) logs["learning_rate"] = self._get_learning_rate() self._total_loss_scalar += tr_loss_scalar self._globalstep_last_logged = self.state.global_step self.store_flos() self.log(logs) metrics = {} if self.control.should_evaluate: if isinstance(self.eval_dataset, dict): for eval_dataset_name, eval_dataset in self.eval_dataset.items(): if eval_dataset_name == 'click': self.label_names = ['qids', 'labels'] else: self.label_names = ['qids', 'labels', 'freqs'] local_metrics = self.evaluate( eval_dataset=eval_dataset, ignore_keys=ignore_keys_for_eval, metric_key_prefix=f"eval_{eval_dataset_name}", ) metrics.update(local_metrics) else: metrics = self.evaluate(ignore_keys=ignore_keys_for_eval) self._report_to_hp_search(trial, epoch, metrics)#self.state.global_step if self.control.should_save: self._save_checkpoint(model, trial, metrics=metrics) self.control = self.callback_handler.on_save(self.args, self.state, self.control) def prediction_step( self, model, inputs: Dict[str, Union[torch.Tensor, Any]], prediction_loss_only: bool, ignore_keys: Optional[List[str]] = None, ) -> Tuple[Optional[torch.Tensor], Optional[torch.Tensor], Optional[torch.Tensor]]: assert prediction_loss_only == False assert ignore_keys is None has_labels = all(inputs.get(k) is not None for k in self.label_names) inputs = self._prepare_inputs(inputs) # labels may be popped when computing the loss (label smoothing for instance) so we grab them first. if has_labels: labels = nested_detach(tuple(inputs.get(name) for name in self.label_names)) if len(labels) == 1: labels = labels[0] else: labels = None eval_input_keys = ['input_ids', 'attention_mask', 'token_type_ids'] eval_inputs = {k: inputs[k] for k in eval_input_keys} with torch.no_grad(): loss = None with self.autocast_smart_context_manager(): logits = model(**eval_inputs).detach().contiguous() return (loss, logits, labels)
4,014
37.605769
109
py
Tencent_wsdm_cup2023
Tencent_wsdm_cup2023-main/pytorch_pretrain/finetune/dataset.py
# -*- coding: utf-8 -*- # @Time : 2022/11/1 16:22 # @Author : Xiangsheng Li # @File : dataset.py import sys,os import random import collections from models.utils import SPECIAL_TOKENS import logging import numpy as np import torch from torch.utils.data import Dataset, IterableDataset from dataclasses import dataclass from typing import List, Dict, Any from collections import defaultdict from tqdm.auto import tqdm from pretrain.dataset import ( process_data, TestDataset ) class PairwiseFinetuneDataset(Dataset): def __init__(self, fpath): self.queries = {} self.docs = dict() self.q_rels = defaultdict(list) self.q_irrels = defaultdict(list) self._load_annotated_data(fpath) self.idx2qids = {i:qid for i,qid in enumerate(self.q_rels.keys())} self.all_docids = list(self.docs.keys()) def _load_annotated_data(self, fpath): doc2docid = {} for line in tqdm(open(fpath, 'rb'), desc='load annotated data'): line_list = line.strip(b'\n').split(b'\t') qid, query, title, content, label, freq = line_list if qid not in self.queries: self.queries[qid] = query if (title, content) not in doc2docid: doc2docid[(title, content)] = len(doc2docid) docid = doc2docid[(title, content)] if docid not in self.docs: self.docs[docid] = (title, content) label = int(label) if label >= 2: #2,1 self.q_rels[qid].append((docid, label)) else: self.q_irrels[qid].append((docid, label)) def __len__(self): return len(self.q_rels) def __getitem__(self, index): qid = self.idx2qids[index] query = self.queries[qid] rel_doc_id, rel_label = random.choice(self.q_rels[qid]) all_doc_list = self.q_rels[qid] + self.q_irrels[qid] rest_docids = [docid for (docid, label) in all_doc_list if (rel_doc_id != docid and rel_label > label)] if len(rest_docids) == 0: irrel_doc_id = random.choice(self.all_docids) else: irrel_doc_id = random.choice(rest_docids) return {'query':query, 'rel_doc': self.docs[rel_doc_id], 'irrel_doc': self.docs[irrel_doc_id]} @dataclass class PairwiseDataCollator: def __init__(self, max_seq_len): self.max_seq_len = max_seq_len def __call__(self, features: Dict[str, Any]) -> Dict[str, Any]: batch_src_input, batch_segment = [], [] for sample in features: src_input, segment = process_data(sample['query'], sample['rel_doc'][0], sample['rel_doc'][1], self.max_seq_len) batch_src_input.append(src_input) batch_segment.append(segment) src_input, segment = process_data(sample['query'], sample['irrel_doc'][0], sample['irrel_doc'][1], self.max_seq_len) batch_src_input.append(src_input) batch_segment.append(segment) batch_src_input = torch.LongTensor(batch_src_input) batch_segment = torch.LongTensor(batch_segment) batch_data = { 'input_ids': batch_src_input, 'attention_mask': (batch_src_input > 0).float(), # torch.FloatTensor, 0 is PAD 'token_type_ids': batch_segment } return batch_data
3,483
34.191919
111
py
Tencent_wsdm_cup2023
Tencent_wsdm_cup2023-main/pytorch_pretrain/finetune/trainer.py
# -*- coding: utf-8 -*- # @Time : 2022/11/2 14:38 # @Author : Xiangsheng Li # @File : trainer.py import os from typing import Dict, List, Tuple, Optional, Any, Union import torch from torch import nn, Tensor import torch.nn.functional as F from torch.utils.data import DataLoader from torch.nn import Softmax, MarginRankingLoss from pretrain.trainer import Pretrainer as Trainer from pretrain.dataset import DataCollator import logging logger = logging.getLogger(__name__) class Finetuner(Trainer): def __init__(self, *args, **kwargs): super(Finetuner, self).__init__(*args, **kwargs) self.softmax = Softmax(dim=1) def compute_loss(self, model, inputs): return self.compute_pair_loss(model, inputs) def compute_pair_loss(self, model, inputs): prediction_scores = model(**inputs) logits = prediction_scores.view(-1, 2) if self.args.temperature is not None: assert self.args.temperature > 0 logits = logits / self.args.temperature logits = self.softmax(logits) pos_logits = logits[:, 0] neg_logits = logits[:, 1] marginloss = MarginRankingLoss(margin=1.0) pair_label = torch.ones_like(pos_logits) pair_loss = marginloss(pos_logits, neg_logits, pair_label) return pair_loss def get_eval_dataloader(self, eval_dataset) -> DataLoader: # use DataCollator for eval dataset eval_dataset = eval_dataset if eval_dataset is not None else self.eval_dataset data_collator = DataCollator() eval_sampler = self._get_eval_sampler(eval_dataset) return DataLoader( eval_dataset, sampler=eval_sampler, batch_size=self.args.eval_batch_size, collate_fn=data_collator, drop_last=self.args.dataloader_drop_last, num_workers=self.args.dataloader_num_workers, pin_memory=self.args.dataloader_pin_memory, )
1,976
28.954545
86
py
Tencent_wsdm_cup2023
Tencent_wsdm_cup2023-main/pytorch_pretrain/models/modeling.py
# -*- coding: utf-8 -*- # @Time : 2022/10/25 12:02 # @Author : Xiangsheng Li # @File : modeling.py import torch from torch import nn, Tensor import torch.distributed as dist import torch.nn.functional as F from transformers import BertModel, BertPreTrainedModel from transformers.modeling_outputs import MaskedLMOutput from torch.nn import CrossEntropyLoss import logging logger = logging.getLogger(__name__) from transformers.activations import ACT2FN class BertPredictionHeadTransform(nn.Module): def __init__(self, config): super().__init__() self.dense = nn.Linear(config.hidden_size, config.hidden_size) if isinstance(config.hidden_act, str): self.transform_act_fn = ACT2FN[config.hidden_act] else: self.transform_act_fn = config.hidden_act self.LayerNorm = nn.LayerNorm(config.hidden_size, eps=config.layer_norm_eps) def forward(self, hidden_states): hidden_states = self.dense(hidden_states) hidden_states = self.transform_act_fn(hidden_states) hidden_states = self.LayerNorm(hidden_states) return hidden_states class BertLMPredictionHead(nn.Module): def __init__(self, config): super().__init__() self.transform = BertPredictionHeadTransform(config) # The output weights are the same as the input embeddings, but there is # an output-only bias for each token. self.decoder = nn.Linear(config.hidden_size, config.vocab_size, bias=False) self.bias = nn.Parameter(torch.zeros(config.vocab_size)) # Need a link between the two variables so that the bias is correctly resized with `resize_token_embeddings` self.decoder.bias = self.bias def forward(self, hidden_states): hidden_states = self.transform(hidden_states) hidden_states = self.decoder(hidden_states) return hidden_states class CTRPretrainingModel(BertPreTrainedModel): def __init__( self, config, model_args, data_args, training_args ): super(CTRPretrainingModel, self).__init__(config) self.bert = BertModel(config) self.cls = nn.Linear(config.hidden_size, 1) self.predictions = BertLMPredictionHead(config)#self.bert.embeddings.word_embeddings.weight self.config = config self.model_args, self.data_args, self.training_args = model_args, data_args, training_args self.mlm_loss_fct = CrossEntropyLoss(ignore_index=-1) self.cross_entropy = nn.CrossEntropyLoss(reduction='mean') self.init_weights() def forward(self, input_ids, attention_mask, token_type_ids,masked_lm_labels=None):#**kwargs outputs = self.bert(input_ids, attention_mask, token_type_ids, return_dict=True) sequence_output = outputs.last_hidden_state #(bs, seq_len, dim) pooler_output = outputs.pooler_output #(bs, dim) prediction_scores = self.cls(pooler_output) #(bs, 1) mlm_loss, ctr_loss = None, None if masked_lm_labels is not None: lm_prediction_scores = self.predictions(sequence_output) mlm_loss = self.mlm_loss_fct(lm_prediction_scores.view(-1, self.config.vocab_size), masked_lm_labels.view(-1))# if config.MLM else 0. ctr_loss = self.groupwise_ctr_loss(prediction_scores) if mlm_loss and ctr_loss: return prediction_scores, mlm_loss, ctr_loss else: return prediction_scores def groupwise_ctr_loss(self, prediction_scores): # only the first column is positive, others are negatives logits = prediction_scores if self.training_args.temperature is not None: assert self.training_args.temperature > 0 logits = logits / self.training_args.temperature logits = logits.view( self.training_args.per_device_train_batch_size, self.training_args.train_group_size ) target_label = torch.zeros(self.training_args.per_device_train_batch_size, dtype=torch.long, device=logits.device) loss = self.cross_entropy(logits, target_label) return loss def pointwise_ctr_loss(self, prediction_scores, click_labels): click_loss_fct = nn.BCEWithLogitsLoss() prediction_scores = torch.squeeze(prediction_scores, dim=-1) point_ctr_loss = click_loss_fct(prediction_scores, click_labels) return point_ctr_loss def pairwise_ctr_loss(self, prediction_scores): # Pairwise loss logits = prediction_scores.view(-1, 2) if self.training_args.temperature is not None: assert self.training_args.temperature > 0 logits = logits / self.training_args.temperature softmax = Softmax(dim=1) logits = softmax(logits) pos_logits = logits[:, 0] neg_logits = logits[:, 1] marginloss = MarginRankingLoss(margin=1.0) pair_label = torch.ones_like(pos_logits) pair_loss = marginloss(pos_logits, neg_logits, pair_label) return pair_loss
5,193
37.474074
116
py
POMO
POMO-master/OLD_ipynb_ver/POMO_TSP/TORCH_OBJECTS.py
""" The MIT License Copyright (c) Yeong-Dae Kwon Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions: The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software. THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. """ import torch use_cuda = torch.cuda.is_available() # use_cuda = False # torch.cuda.set_device(0) FloatTensor = torch.cuda.FloatTensor if use_cuda else torch.FloatTensor LongTensor = torch.cuda.LongTensor if use_cuda else torch.LongTensor ByteTensor = torch.cuda.ByteTensor if use_cuda else torch.ByteTensor BoolTensor = torch.cuda.BoolTensor if use_cuda else torch.BoolTensor Tensor = FloatTensor if use_cuda: device = torch.device("cuda") else: device = torch.device("cpu")
1,570
32.425532
77
py
POMO
POMO-master/OLD_ipynb_ver/POMO_TSP/source/utilities.py
""" The MIT License Copyright (c) 2020 Yeong-Dae Kwon Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions: The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software. THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. """ import logging import os import time import datetime import pytz import re import numpy as np from HYPER_PARAMS import * from TORCH_OBJECTS import * ######################################## # Get_Logger ######################################## tz = pytz.timezone("Asia/Seoul") def timetz(*args): return datetime.datetime.now(tz).timetuple() def Get_Logger(SAVE_FOLDER_NAME): # make_dir ####################################################### prefix = datetime.datetime.now(pytz.timezone("Asia/Seoul")).strftime("%Y%m%d_%H%M__") result_folder_no_postfix = "./result/{}".format(prefix + SAVE_FOLDER_NAME) result_folder_path = result_folder_no_postfix folder_idx = 0 while os.path.exists(result_folder_path): folder_idx += 1 result_folder_path = result_folder_no_postfix + "({})".format(folder_idx) os.makedirs(result_folder_path) # Logger ####################################################### logger = logging.getLogger(result_folder_path) # this already includes streamHandler?? streamHandler = logging.StreamHandler() fileHandler = logging.FileHandler('{}/log.txt'.format(result_folder_path)) formatter = logging.Formatter("[%(asctime)s] %(message)s", "%Y-%m-%d %H:%M:%S") formatter.converter = timetz streamHandler.setFormatter(formatter) fileHandler.setFormatter(formatter) logger.addHandler(streamHandler) logger.addHandler(fileHandler) logger.setLevel(level=logging.INFO) return logger, result_folder_path def Extract_from_LogFile(result_folder_path, variable_name): logfile_path = '{}/log.txt'.format(result_folder_path) with open(logfile_path) as f: datafile = f.readlines() found = False # This isn't really necessary for line in reversed(datafile): if variable_name in line: found = True m = re.search(variable_name + '[^\n]+', line) break exec_command = "Print(No such variable found !!)" if found: return m.group(0) else: return exec_command ######################################## # Average_Meter ######################################## class Average_Meter: def __init__(self): self.sum = None self.count = None self.reset() def reset(self): self.sum = torch.tensor(0.).to(device) self.count = 0 def push(self, some_tensor, n_for_rank_0_tensor=None): assert not some_tensor.requires_grad # You get Memory error, if you keep tensors with grad history rank = len(some_tensor.shape) if rank == 0: # assuming "already averaged" Tensor was pushed self.sum += some_tensor * n_for_rank_0_tensor self.count += n_for_rank_0_tensor else: self.sum += some_tensor.sum() self.count += some_tensor.numel() def peek(self): average = (self.sum / self.count).tolist() return average def result(self): average = (self.sum / self.count).tolist() self.reset() return average ######################################## # View NN Parameters ######################################## def get_n_params1(model): pp = 0 for p in list(model.parameters()): nn_count = 1 for s in list(p.size()): nn_count = nn_count * s pp += nn_count print(nn_count) print(p.shape) print("Total: {:d}".format(pp)) def get_n_params2(model): model_parameters = filter(lambda p: p.requires_grad, model.parameters()) params = sum([np.prod(p.size()) for p in model_parameters]) print(params) def get_n_params3(model): print(sum(p.numel() for p in model.parameters() if p.requires_grad)) def get_structure(model): print(model) ######################################## # Augment xy data ######################################## def augment_xy_data_by_8_fold(xy_data): # xy_data.shape = (batch_s, problem, 2) x = xy_data[:, :, [0]] y = xy_data[:, :, [1]] # x,y shape = (batch, problem, 1) dat1 = torch.cat((x, y), dim=2) dat2 = torch.cat((1-x, y), dim=2) dat3 = torch.cat((x, 1-y), dim=2) dat4 = torch.cat((1-x, 1-y), dim=2) dat5 = torch.cat((y, x), dim=2) dat6 = torch.cat((1-y, x), dim=2) dat7 = torch.cat((y, 1-x), dim=2) dat8 = torch.cat((1-y, 1-x), dim=2) data_augmented = torch.cat((dat1, dat2, dat3, dat4, dat5, dat6, dat7, dat8), dim=0) # shape = (8*batch, problem, 2) return data_augmented
5,647
26.686275
106
py
POMO
POMO-master/OLD_ipynb_ver/POMO_TSP/source/travelling_saleman_problem.py
""" The MIT License Copyright (c) 2020 Yeong-Dae Kwon Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions: The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software. THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. """ #################################### # EXTERNAL LIBRARY #################################### import torch import numpy as np # For debugging from IPython.core.debugger import set_trace from torch.utils.data import Dataset from torch.utils.data import DataLoader #################################### # PROJECT VARIABLES #################################### from HYPER_PARAMS import * from TORCH_OBJECTS import * #################################### # DATA #################################### def TSP_DATA_LOADER__RANDOM(num_sample, num_nodes, batch_size): dataset = TSP_Dataset__Random(num_sample=num_sample, num_nodes=num_nodes) data_loader = DataLoader(dataset=dataset, batch_size=batch_size, shuffle=False, num_workers=0, collate_fn=TSP_collate_fn) return data_loader class TSP_Dataset__Random(Dataset): def __init__(self, num_sample, num_nodes): self.num_sample = num_sample self.num_nodes = num_nodes def __getitem__(self, index): node_xy_data = np.random.rand(self.num_nodes, 2) return node_xy_data def __len__(self): return self.num_sample def TSP_collate_fn(batch): node_xy = Tensor(batch) return node_xy #################################### # STATE #################################### class STATE: def __init__(self, seq): self.seq = seq self.batch_s = seq.size(0) self.current_node = None # History #################################### self.selected_count = 0 self.available_mask = BoolTensor(np.ones((self.batch_s, TSP_SIZE))) self.ninf_mask = Tensor(np.zeros((self.batch_s, TSP_SIZE))) # shape = (batch_s, TSP_SIZE) self.selected_node_list = LongTensor(np.zeros((self.batch_s, 0))) # shape = (batch_s, selected_count) def move_to(self, selected_node_idx): # selected_node_idx.shape = (batch,) self.current_node = selected_node_idx # History #################################### self.selected_count += 1 self.available_mask[torch.arange(self.batch_s), selected_node_idx] = False self.ninf_mask[torch.arange(self.batch_s), selected_node_idx] = -np.inf self.selected_node_list = torch.cat((self.selected_node_list, selected_node_idx[:, None]), dim=1) class GROUP_STATE: def __init__(self, group_size, data): # data.shape = (batch, group, 2) self.batch_s = data.size(0) self.group_s = group_size self.data = data # History #################################### self.selected_count = 0 self.current_node = None # shape = (batch, group) self.selected_node_list = LongTensor(np.zeros((self.batch_s, group_size, 0))) # shape = (batch, group, selected_count) # Status #################################### self.ninf_mask = Tensor(np.zeros((self.batch_s, group_size, TSP_SIZE))) # shape = (batch, group, TSP_SIZE) def move_to(self, selected_idx_mat): # selected_idx_mat.shape = (batch, group) # History #################################### self.selected_count += 1 self.current_node = selected_idx_mat self.selected_node_list = torch.cat((self.selected_node_list, selected_idx_mat[:, :, None]), dim=2) # Status #################################### batch_idx_mat = torch.arange(self.batch_s)[:, None].expand(self.batch_s, self.group_s) group_idx_mat = torch.arange(self.group_s)[None, :].expand(self.batch_s, self.group_s) self.ninf_mask[batch_idx_mat, group_idx_mat, selected_idx_mat] = -np.inf #################################### # ENVIRONMENT #################################### class ENVIRONMENT: def __init__(self, seq): # seq.shape = (batch, TSP_SIZE, 2) self.seq = seq self.batch_s = seq.size(0) self.state = None def reset(self): self.state = STATE(self.seq) reward = None done = False return self.state, reward, done def step(self, selected_node_idx): # selected_node_idx.shape = (batch,) # move state self.state.move_to(selected_node_idx) # returning values done = self.state.selected_count == TSP_SIZE if done: reward = -self._get_travel_distance() # note the minus sign! else: reward = None return self.state, reward, done def _get_travel_distance(self): gathering_index = self.state.selected_node_list.unsqueeze(2).expand(-1, TSP_SIZE, 2) # shape = (batch, TSP_SIZE, 2) ordered_seq = self.seq.gather(dim=1, index=gathering_index) rolled_seq = ordered_seq.roll(dims=1, shifts=-1) segment_lengths = ((ordered_seq-rolled_seq)**2).sum(2).sqrt() # size = (batch, TSP_SIZE) travel_distances = segment_lengths.sum(1) # size = (batch,) return travel_distances class GROUP_ENVIRONMENT: def __init__(self, data): # seq.shape = (batch, TSP_SIZE, 2) self.data = data self.batch_s = data.size(0) self.group_s = None self.group_state = None def reset(self, group_size): self.group_s = group_size self.group_state = GROUP_STATE(group_size=group_size, data=self.data) reward = None done = False return self.group_state, reward, done def step(self, selected_idx_mat): # selected_idx_mat.shape = (batch, group) # move state self.group_state.move_to(selected_idx_mat) # returning values done = (self.group_state.selected_count == TSP_SIZE) if done: reward = -self._get_group_travel_distance() # note the minus sign! else: reward = None return self.group_state, reward, done def _get_group_travel_distance(self): gathering_index = self.group_state.selected_node_list.unsqueeze(3).expand(self.batch_s, -1, TSP_SIZE, 2) # shape = (batch, group, TSP_SIZE, 2) seq_expanded = self.data[:, None, :, :].expand(self.batch_s, self.group_s, TSP_SIZE, 2) ordered_seq = seq_expanded.gather(dim=2, index=gathering_index) # shape = (batch, group, TSP_SIZE, 2) rolled_seq = ordered_seq.roll(dims=2, shifts=-1) segment_lengths = ((ordered_seq-rolled_seq)**2).sum(3).sqrt() # size = (batch, group, TSP_SIZE) group_travel_distances = segment_lengths.sum(2) # size = (batch, group) return group_travel_distances
7,808
30.873469
112
py
POMO
POMO-master/OLD_ipynb_ver/POMO_TSP/source/MODEL__Actor/grouped_actors.py
""" The MIT License Copyright (c) 2020 Yeong-Dae Kwon Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions: The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software. THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. """ import torch import torch.nn as nn import torch.nn.functional as F import numpy as np # For debugging from IPython.core.debugger import set_trace # Hyper Parameters from HYPER_PARAMS import * from TORCH_OBJECTS import * ######################################## # ACTOR ######################################## class ACTOR(nn.Module): def __init__(self): super().__init__() self.box_select_probabilities = None # shape = (batch, group, TSP_SIZE) self.encoder = Encoder() self.node_prob_calculator = Next_Node_Probability_Calculator_for_group() self.batch_s = None self.encoded_nodes = None def reset(self, group_state): self.batch_s = group_state.data.size(0) self.encoded_nodes = self.encoder(group_state.data) # shape = (batch_s, TSP_SIZE, EMBEDDING_DIM) self.node_prob_calculator.reset(self.encoded_nodes, group_ninf_mask=group_state.ninf_mask) def soft_reset(self, group_state): self.node_prob_calculator.reset(self.encoded_nodes, group_ninf_mask=group_state.ninf_mask) def update(self, group_state): encoded_LAST_NODES = pick_nodes_for_each_group(self.encoded_nodes, group_state.current_node) # shape = (batch_s, group, EMBEDDING_DIM) probs = self.node_prob_calculator(encoded_LAST_NODES) # shape = (batch_s, group, TSP_SIZE) self.box_select_probabilities = probs def get_action_probabilities(self): return self.box_select_probabilities ######################################## # ACTOR_SUB_NN : ENCODER ######################################## class Encoder(nn.Module): def __init__(self): super().__init__() self.embedding = nn.Linear(2, EMBEDDING_DIM) self.layers = nn.ModuleList([Encoder_Layer() for _ in range(ENCODER_LAYER_NUM)]) def forward(self, data): # data.shape = (batch_s, TSP_SIZE, 2) embedded_input = self.embedding(data) # shape = (batch_s, TSP_SIZE, EMBEDDING_DIM) out = embedded_input for layer in self.layers: out = layer(out) return out class Encoder_Layer(nn.Module): def __init__(self): super().__init__() self.Wq = nn.Linear(EMBEDDING_DIM, HEAD_NUM * KEY_DIM, bias=False) self.Wk = nn.Linear(EMBEDDING_DIM, HEAD_NUM * KEY_DIM, bias=False) self.Wv = nn.Linear(EMBEDDING_DIM, HEAD_NUM * KEY_DIM, bias=False) self.multi_head_combine = nn.Linear(HEAD_NUM * KEY_DIM, EMBEDDING_DIM) self.addAndNormalization1 = Add_And_Normalization_Module() self.feedForward = Feed_Forward_Module() self.addAndNormalization2 = Add_And_Normalization_Module() def forward(self, input1): # input.shape = (batch_s, TSP_SIZE, EMBEDDING_DIM) q = reshape_by_heads(self.Wq(input1), head_num=HEAD_NUM) k = reshape_by_heads(self.Wk(input1), head_num=HEAD_NUM) v = reshape_by_heads(self.Wv(input1), head_num=HEAD_NUM) # q shape = (batch_s, HEAD_NUM, TSP_SIZE, KEY_DIM) out_concat = multi_head_attention(q, k, v) # shape = (batch_s, TSP_SIZE, HEAD_NUM*KEY_DIM) multi_head_out = self.multi_head_combine(out_concat) # shape = (batch_s, TSP_SIZE, EMBEDDING_DIM) out1 = self.addAndNormalization1(input1, multi_head_out) out2 = self.feedForward(out1) out3 = self.addAndNormalization2(out1, out2) return out3 ######################################## # ACTOR_SUB_NN : Next_Node_Probability_Calculator ######################################## class Next_Node_Probability_Calculator_for_group(nn.Module): def __init__(self): super().__init__() self.Wq_graph = nn.Linear(EMBEDDING_DIM, HEAD_NUM * KEY_DIM, bias=False) self.Wq_first = nn.Linear(EMBEDDING_DIM, HEAD_NUM * KEY_DIM, bias=False) self.Wq_last = nn.Linear(EMBEDDING_DIM, HEAD_NUM * KEY_DIM, bias=False) self.Wk = nn.Linear(EMBEDDING_DIM, HEAD_NUM * KEY_DIM, bias=False) self.Wv = nn.Linear(EMBEDDING_DIM, HEAD_NUM * KEY_DIM, bias=False) self.multi_head_combine = nn.Linear(HEAD_NUM * KEY_DIM, EMBEDDING_DIM) self.q_graph = None # saved q1, for multi-head attention self.q_first = None # saved q2, for multi-head attention self.k = None # saved key, for multi-head attention self.v = None # saved value, for multi-head_attention self.single_head_key = None # saved, for single-head attention self.group_ninf_mask = None # reference to ninf_mask owned by state def reset(self, encoded_nodes, group_ninf_mask): # encoded_nodes.shape = (batch_s, TSP_SIZE, EMBEDDING_DIM) encoded_graph = encoded_nodes.mean(dim=1, keepdim=True) # shape = (batch_s, 1, EMBEDDING_DIM) self.q_graph = reshape_by_heads(self.Wq_graph(encoded_graph), head_num=HEAD_NUM) # shape = (batch_s, HEAD_NUM, 1, KEY_DIM) self.q_first = None # shape = (batch_s, HEAD_NUM, group, KEY_DIM) self.k = reshape_by_heads(self.Wk(encoded_nodes), head_num=HEAD_NUM) self.v = reshape_by_heads(self.Wv(encoded_nodes), head_num=HEAD_NUM) # shape = (batch_s, HEAD_NUM, TSP_SIZE, KEY_DIM) self.single_head_key = encoded_nodes.transpose(1, 2) # shape = (batch_s, EMBEDDING_DIM, TSP_SIZE) self.group_ninf_mask = group_ninf_mask # shape = (batch_s, group, TSP_SIZE) def forward(self, encoded_LAST_NODE): # encoded_LAST_NODE.shape = (batch_s, group, EMBEDDING_DIM) if self.q_first is None: self.q_first = reshape_by_heads(self.Wq_first(encoded_LAST_NODE), head_num=HEAD_NUM) # shape = (batch_s, HEAD_NUM, group, KEY_DIM) # Multi-Head Attention ####################################################### q_last = reshape_by_heads(self.Wq_last(encoded_LAST_NODE), head_num=HEAD_NUM) # shape = (batch_s, HEAD_NUM, group, KEY_DIM) q = self.q_graph + self.q_first + q_last # shape = (batch_s, HEAD_NUM, group, KEY_DIM) out_concat = multi_head_attention(q, self.k, self.v, group_ninf_mask=self.group_ninf_mask) # shape = (batch_s, group, HEAD_NUM*KEY_DIM) mh_atten_out = self.multi_head_combine(out_concat) # shape = (batch_s, group, EMBEDDING_DIM) # Single-Head Attention, for probability calculation ####################################################### score = torch.matmul(mh_atten_out, self.single_head_key) # shape = (batch_s, group, TSP_SIZE) score_scaled = score / np.sqrt(EMBEDDING_DIM) # shape = (batch_s, group, TSP_SIZE) score_clipped = LOGIT_CLIPPING * torch.tanh(score_scaled) score_masked = score_clipped + self.group_ninf_mask.clone() probs = F.softmax(score_masked, dim=2) # shape = (batch_s, group, TSP_SIZE) return probs ######################################## # NN SUB CLASS / FUNCTIONS ######################################## def pick_nodes_for_each_group(encoded_nodes, node_index_to_pick): # encoded_nodes.shape = (batch_s, TSP_SIZE, EMBEDDING_DIM) # node_index_to_pick.shape = (batch_s, group_s) batch_s = node_index_to_pick.size(0) group_s = node_index_to_pick.size(1) gathering_index = node_index_to_pick[:, :, None].expand(batch_s, group_s, EMBEDDING_DIM) # shape = (batch_s, group, EMBEDDING_DIM) picked_nodes = encoded_nodes.gather(dim=1, index=gathering_index) # shape = (batch_s, group, EMBEDDING_DIM) return picked_nodes def reshape_by_heads(qkv, head_num): # q.shape = (batch, C, head_num*key_dim) batch_s = qkv.size(0) C = qkv.size(1) q_reshaped = qkv.reshape(batch_s, C, head_num, -1) # shape = (batch, C, head_num, key_dim) q_transposed = q_reshaped.transpose(1, 2) # shape = (batch, head_num, C, key_dim) return q_transposed def multi_head_attention(q, k, v, ninf_mask=None, group_ninf_mask=None): # q shape = (batch_s, head_num, n, key_dim) : n can be either 1 or TSP_SIZE # k,v shape = (batch_s, head_num, TSP_SIZE, key_dim) # ninf_mask.shape = (batch_s, TSP_SIZE) # group_ninf_mask.shape = (batch_s, group, TSP_SIZE) batch_s = q.size(0) head_num = q.size(1) n = q.size(2) key_dim = q.size(3) score = torch.matmul(q, k.transpose(2, 3)) # shape = (batch_s, head_num, n, TSP_SIZE) score_scaled = score / np.sqrt(key_dim) if ninf_mask is not None: score_scaled = score_scaled + ninf_mask[:, None, None, :].expand(batch_s, head_num, n, TSP_SIZE) if group_ninf_mask is not None: score_scaled = score_scaled + group_ninf_mask[:, None, :, :].expand(batch_s, head_num, n, TSP_SIZE) weights = nn.Softmax(dim=3)(score_scaled) # shape = (batch_s, head_num, n, TSP_SIZE) out = torch.matmul(weights, v) # shape = (batch_s, head_num, n, key_dim) out_transposed = out.transpose(1, 2) # shape = (batch_s, n, head_num, key_dim) out_concat = out_transposed.reshape(batch_s, n, head_num * key_dim) # shape = (batch_s, n, head_num*key_dim) return out_concat class Add_And_Normalization_Module(nn.Module): def __init__(self): super().__init__() self.norm_by_EMB = nn.BatchNorm1d(EMBEDDING_DIM, affine=True) # 'Funny' Batch_Norm, as it will normalized by EMB dim def forward(self, input1, input2): # input.shape = (batch_s, TSP_SIZE, EMBEDDING_DIM) batch_s = input1.size(0) added = input1 + input2 normalized = self.norm_by_EMB(added.reshape(batch_s * TSP_SIZE, EMBEDDING_DIM)) return normalized.reshape(batch_s, TSP_SIZE, EMBEDDING_DIM) class Feed_Forward_Module(nn.Module): def __init__(self): super().__init__() self.W1 = nn.Linear(EMBEDDING_DIM, FF_HIDDEN_DIM) self.W2 = nn.Linear(FF_HIDDEN_DIM, EMBEDDING_DIM) def forward(self, input1): # input.shape = (batch_s, TSP_SIZE, EMBEDDING_DIM) return self.W2(F.relu(self.W1(input1)))
11,175
34.592357
107
py
POMO
POMO-master/OLD_ipynb_ver/POMO_TSP/source/TRAIN_N_EVAL/Train_Grouped_Actors.py
""" The MIT License Copyright (c) 2020 Yeong-Dae Kwon Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions: The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software. THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. """ import numpy as np import time # For debugging from IPython.core.debugger import set_trace # Hyper Parameters from HYPER_PARAMS import * from TORCH_OBJECTS import * from source.utilities import Average_Meter from source.travelling_saleman_problem import TSP_DATA_LOADER__RANDOM, GROUP_ENVIRONMENT ######################################## # TRAIN ######################################## def TRAIN(actor_group, epoch, timer_start, logger): actor_group.train() distance_AM = Average_Meter() actor_loss_AM = Average_Meter() train_loader = TSP_DATA_LOADER__RANDOM(num_sample=TRAIN_DATASET_SIZE, num_nodes=TSP_SIZE, batch_size=TRAIN_BATCH_SIZE) logger_start = time.time() episode = 0 for data in train_loader: # data.shape = (batch_s, TSP_SIZE, 2) batch_s = data.size(0) episode = episode + batch_s # Actor Group Move ############################################### env = GROUP_ENVIRONMENT(data) group_s = TSP_SIZE group_state, reward, done = env.reset(group_size=group_s) actor_group.reset(group_state) # First Move is given first_action = LongTensor(np.arange(group_s))[None, :].expand(batch_s, group_s) group_state, reward, done = env.step(first_action) group_prob_list = Tensor(np.zeros((batch_s, group_s, 0))) while not done: actor_group.update(group_state) action_probs = actor_group.get_action_probabilities() # shape = (batch, group, TSP_SIZE) action = action_probs.reshape(batch_s*group_s, -1).multinomial(1).squeeze(dim=1).reshape(batch_s, group_s) # shape = (batch, group) group_state, reward, done = env.step(action) batch_idx_mat = torch.arange(batch_s)[:, None].expand(batch_s, group_s) group_idx_mat = torch.arange(group_s)[None, :].expand(batch_s, group_s) chosen_action_prob = action_probs[batch_idx_mat, group_idx_mat, action].reshape(batch_s, group_s) # shape = (batch, group) group_prob_list = torch.cat((group_prob_list, chosen_action_prob[:, :, None]), dim=2) # LEARNING - Actor ############################################### group_reward = reward group_log_prob = group_prob_list.log().sum(dim=2) # shape = (batch, group) group_advantage = group_reward - group_reward.mean(dim=1, keepdim=True) group_loss = -group_advantage * group_log_prob # shape = (batch, group) loss = group_loss.mean() actor_group.optimizer.zero_grad() loss.backward() actor_group.optimizer.step() # RECORDING ############################################### max_reward, _ = group_reward.max(dim=1) distance_AM.push(-max_reward) # reward was given as negative dist actor_loss_AM.push(group_loss.detach().reshape(-1)) # LOGGING ############################################### if (time.time()-logger_start > LOG_PERIOD_SEC) or (episode == TRAIN_DATASET_SIZE): timestr = time.strftime("%H:%M:%S", time.gmtime(time.time()-timer_start)) log_str = 'Ep:{:03d}-{:07d}({:5.1f}%) T:{:s} ALoss:{:+5f} CLoss:{:5f} Avg.dist:{:5f}' \ .format(epoch, episode, episode/TRAIN_DATASET_SIZE*100, timestr, actor_loss_AM.result(), 0, distance_AM.result()) logger.info(log_str) logger_start = time.time() # LR STEP, after each epoch actor_group.lr_stepper.step()
4,697
36.584
122
py
POMO
POMO-master/OLD_ipynb_ver/POMO_TSP/source/TRAIN_N_EVAL/Evaluate_Grouped_Actors.py
""" The MIT License Copyright (c) 2020 Yeong-Dae Kwon Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions: The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software. THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. """ import numpy as np # For debugging from IPython.core.debugger import set_trace # Hyper Parameters from HYPER_PARAMS import * from TORCH_OBJECTS import * from source.utilities import Average_Meter from source.travelling_saleman_problem import TSP_DATA_LOADER__RANDOM, GROUP_ENVIRONMENT ######################################## # EVAL ######################################## eval_result = [] def update_eval_result(old_result): global eval_result eval_result = old_result def EVAL(actor_group, epoch, timer_start, logger): global eval_result actor_group.eval() eval_dist_AM = Average_Meter() if TSP_SIZE == 5: raise NotImplementedError elif TSP_SIZE == 10: raise NotImplementedError else: test_loader = TSP_DATA_LOADER__RANDOM(num_sample=TEST_DATASET_SIZE, num_nodes=TSP_SIZE, batch_size=TEST_BATCH_SIZE) for data in test_loader: # data.shape = (batch_s, TSP_SIZE, 2) batch_s = data.size(0) with torch.no_grad(): env = GROUP_ENVIRONMENT(data) group_s = TSP_SIZE group_state, reward, done = env.reset(group_size=group_s) actor_group.reset(group_state) # First Move is given first_action = LongTensor(np.arange(group_s))[None, :].expand(batch_s, group_s) group_state, reward, done = env.step(first_action) while not done: actor_group.update(group_state) action_probs = actor_group.get_action_probabilities() # shape = (batch, group, TSP_SIZE) action = action_probs.argmax(dim=2) # shape = (batch, group) group_state, reward, done = env.step(action) max_reward, _ = reward.max(dim=1) eval_dist_AM.push(-max_reward) # reward was given as negative dist # LOGGING dist_avg = eval_dist_AM.result() eval_result.append(dist_avg) logger.info('--------------------------------------------------------------------------') log_str = ' <<< EVAL after Epoch:{:03d} >>> Avg.dist:{:f}'.format(epoch, dist_avg) logger.info(log_str) logger.info('--------------------------------------------------------------------------') logger.info('eval_result = {}'.format(eval_result)) logger.info('--------------------------------------------------------------------------') logger.info('--------------------------------------------------------------------------') logger.info('--------------------------------------------------------------------------')
3,672
34.317308
123
py
POMO
POMO-master/OLD_ipynb_ver/POMO_KP/TORCH_OBJECTS.py
""" The MIT License Copyright (c) 2020 Yeong-Dae Kwon Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions: The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software. THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. """ import torch use_cuda = torch.cuda.is_available() # use_cuda = False torch.cuda.set_device(0) FloatTensor = torch.cuda.FloatTensor if use_cuda else torch.FloatTensor LongTensor = torch.cuda.LongTensor if use_cuda else torch.LongTensor ByteTensor = torch.cuda.ByteTensor if use_cuda else torch.ByteTensor BoolTensor = torch.cuda.BoolTensor if use_cuda else torch.BoolTensor Tensor = FloatTensor if use_cuda: device = torch.device("cuda") else: device = torch.device("cpu")
1,573
32.489362
77
py
POMO
POMO-master/OLD_ipynb_ver/POMO_KP/source/utilities.py
""" The MIT License Copyright (c) 2020 Yeong-Dae Kwon Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions: The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software. THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. """ import logging import os import time import datetime import pytz import re import numpy as np from HYPER_PARAMS import * from TORCH_OBJECTS import * ######################################## # Get_Logger ######################################## tz = pytz.timezone("Asia/Seoul") def timetz(*args): return datetime.datetime.now(tz).timetuple() def Get_Logger(SAVE_FOLDER_NAME): # make_dir ####################################################### prefix = datetime.datetime.now(pytz.timezone("Asia/Seoul")).strftime("%Y%m%d_%H%M__") result_folder_no_postfix = "./result/{}".format(prefix + SAVE_FOLDER_NAME) result_folder_path = result_folder_no_postfix folder_idx = 0 while os.path.exists(result_folder_path): folder_idx += 1 result_folder_path = result_folder_no_postfix + "({})".format(folder_idx) os.makedirs(result_folder_path) # Logger ####################################################### logger = logging.getLogger(result_folder_path) # this already includes streamHandler?? streamHandler = logging.StreamHandler() fileHandler = logging.FileHandler('{}/log.txt'.format(result_folder_path)) formatter = logging.Formatter("[%(asctime)s] %(message)s", "%Y-%m-%d %H:%M:%S") formatter.converter = timetz streamHandler.setFormatter(formatter) fileHandler.setFormatter(formatter) logger.addHandler(streamHandler) logger.addHandler(fileHandler) logger.setLevel(level=logging.INFO) return logger, result_folder_path def Extract_from_LogFile(result_folder_path, variable_name): logfile_path = '{}/log.txt'.format(result_folder_path) with open(logfile_path) as f: datafile = f.readlines() found = False # This isn't really necessary for line in reversed(datafile): if variable_name in line: found = True m = re.search(variable_name + '[^\n]+', line) break exec_command = "Print(No such variable found !!)" if found: return m.group(0) else: return exec_command ######################################## # Average_Meter ######################################## class Average_Meter: def __init__(self): self.sum = None self.count = None self.reset() def reset(self): self.sum = torch.tensor(0.).to(device) self.count = 0 def push(self, some_tensor, n_for_rank_0_tensor=None): assert not some_tensor.requires_grad # You get Memory error, if you keep tensors with grad history rank = len(some_tensor.shape) if rank == 0: # assuming "already averaged" Tensor was pushed self.sum += some_tensor * n_for_rank_0_tensor self.count += n_for_rank_0_tensor else: self.sum += some_tensor.sum() self.count += some_tensor.numel() def peek(self): average = (self.sum / self.count).tolist() return average def result(self): average = (self.sum / self.count).tolist() self.reset() return average ######################################## # View NN Parameters ######################################## def get_n_params1(model): pp = 0 for p in list(model.parameters()): nn_count = 1 for s in list(p.size()): nn_count = nn_count * s pp += nn_count print(nn_count) print(p.shape) print("Total: {:d}".format(pp)) def get_n_params2(model): model_parameters = filter(lambda p: p.requires_grad, model.parameters()) params = sum([np.prod(p.size()) for p in model_parameters]) print(params) def get_n_params3(model): print(sum(p.numel() for p in model.parameters() if p.requires_grad)) def get_structure(model): print(model)
4,907
27.045714
106
py
POMO
POMO-master/OLD_ipynb_ver/POMO_KP/source/knapsack_problem.py
""" The MIT License Copyright (c) 2020 Yeong-Dae Kwon Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions: The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software. THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. """ #################################### # EXTERNAL LIBRARY #################################### import torch import numpy as np # For debugging from IPython.core.debugger import set_trace from torch.utils.data import Dataset from torch.utils.data import DataLoader #################################### # PROJECT VARIABLES #################################### from HYPER_PARAMS import * from TORCH_OBJECTS import * #################################### # DATA #################################### def KNAPSACK_DATA_LOADER__RANDOM(num_sample, num_items, batch_size): dataset = KnapSack_Dataset__Random(num_sample=num_sample, num_items=num_items) data_loader = DataLoader(dataset=dataset, batch_size=batch_size, shuffle=False, num_workers=0, collate_fn=knapsack_collate_fn) return data_loader class KnapSack_Dataset__Random(Dataset): def __init__(self, num_sample, num_items): self.num_sample = num_sample self.num_items = num_items def __getitem__(self, index): data = np.random.rand(self.num_items, 2) return data def __len__(self): return self.num_sample def knapsack_collate_fn(batch): return Tensor(batch) #################################### # STATE #################################### class STATE: def __init__(self, item_data, capacity): self.batch_s = item_data.size(0) self.items_and_a_dummy = Tensor(np.zeros((self.batch_s, PROBLEM_SIZE+1, 2))) self.items_and_a_dummy[:, :PROBLEM_SIZE, :] = item_data self.item_data = self.items_and_a_dummy[:, :PROBLEM_SIZE, :] # shape = (batch, problem, 2) # History #################################### self.current_node = None self.selected_count = 0 self.selected_node_list = LongTensor(np.zeros((self.batch_s, 0))) # shape = (batch, selected_count) # Status #################################### self.accumulated_value = Tensor(np.zeros((self.batch_s,))) # shape = (batch,) self.capacity = Tensor(np.ones((self.batch_s,))) * capacity # shape = (batch,) self.ninf_mask_w_dummy = Tensor(np.zeros((self.batch_s, PROBLEM_SIZE+1))) self.ninf_mask = self.ninf_mask_w_dummy[:, :PROBLEM_SIZE] # shape = (batch, problem) self.fit_ninf_mask = None self.finished = BoolTensor(np.zeros((self.batch_s,))) # shape = (batch,) def move_to(self, selected_item_idx): # selected_item_idx.shape = (batch,) # History #################################### self.current_node = selected_item_idx self.selected_count += 1 self.selected_node_list = torch.cat((self.selected_node_list, selected_item_idx[:, None]), dim=1) # Status #################################### gathering_index = selected_item_idx[:, None, None].expand(self.batch_s, 1, 2) selected_item = self.items_and_a_dummy.gather(dim=1, index=gathering_index).squeeze(dim=1) # shape = (batch, 2) self.accumulated_value += selected_item[:, 1] self.capacity -= selected_item[:, 0] self.ninf_mask_w_dummy[torch.arange(self.batch_s), selected_item_idx] = -np.inf unfit_bool = (self.capacity[:, None] - self.item_data[:, :, 0]) < 0 # shape = (batch, problem) self.fit_ninf_mask = self.ninf_mask.clone() self.fit_ninf_mask[unfit_bool] = -np.inf self.finished = (self.fit_ninf_mask == -np.inf).all(dim=1) # shape = (batch,) self.fit_ninf_mask[self.finished[:, None].expand(self.batch_s, PROBLEM_SIZE)] = 0 # do not mask finished epi. class GROUP_STATE: def __init__(self, group_size, item_data, capacity): self.batch_s = item_data.size(0) self.group_s = group_size self.items_and_a_dummy = Tensor(np.zeros((self.batch_s, PROBLEM_SIZE+1, 2))) self.items_and_a_dummy[:, :PROBLEM_SIZE, :] = item_data self.item_data = self.items_and_a_dummy[:, :PROBLEM_SIZE, :] # shape = (batch, problem, 2) # History #################################### self.current_node = None # shape = (batch_s, group) self.selected_count = 0 self.selected_node_list = LongTensor(np.zeros((self.batch_s, self.group_s, 0))) # shape = (batch_s, group, selected_count) # Status #################################### self.accumulated_value = Tensor(np.zeros((self.batch_s, self.group_s))) # shape = (batch, group) self.capacity = Tensor(np.ones((self.batch_s, self.group_s))) * capacity # shape = (batch, group) self.ninf_mask_w_dummy = Tensor(np.zeros((self.batch_s, self.group_s, PROBLEM_SIZE+1))) self.ninf_mask = self.ninf_mask_w_dummy[:, :, :PROBLEM_SIZE] # shape = (batch, group, problem) self.fit_ninf_mask = None self.finished = BoolTensor(np.zeros((self.batch_s, self.group_s))) # shape = (batch, group) def move_to(self, selected_idx_mat): # selected_idx_mat.shape = (batch, group) # History #################################### self.current_node = selected_idx_mat self.selected_count += 1 self.selected_node_list = torch.cat((self.selected_node_list, selected_idx_mat[:, :, None]), dim=2) # Status #################################### items_mat = self.items_and_a_dummy[:, None, :, :].expand(self.batch_s, self.group_s, PROBLEM_SIZE+1, 2) gathering_index = selected_idx_mat[:, :, None, None].expand(self.batch_s, self.group_s, 1, 2) selected_item = items_mat.gather(dim=2, index=gathering_index).squeeze(dim=2) # shape = (batch, group, 2) self.accumulated_value += selected_item[:, :, 1] self.capacity -= selected_item[:, :, 0] batch_idx_mat = torch.arange(self.batch_s)[:, None].expand(self.batch_s, self.group_s) group_idx_mat = torch.arange(self.group_s)[None, :].expand(self.batch_s, self.group_s) self.ninf_mask_w_dummy[batch_idx_mat, group_idx_mat, selected_idx_mat] = -np.inf unfit_bool = (self.capacity[:, :, None] - self.item_data[:, None, :, 0]) < 0 # shape = (batch, group, problem) self.fit_ninf_mask = self.ninf_mask.clone() self.fit_ninf_mask[unfit_bool] = -np.inf self.finished = (self.fit_ninf_mask == -np.inf).all(dim=2) # shape = (batch, group) self.fit_ninf_mask[self.finished[:, :, None].expand(self.batch_s, self.group_s, PROBLEM_SIZE)] = 0 # do not mask finished episode #################################### # ENVIRONMENT #################################### class ENVIRONMENT: def __init__(self, item_data): # item_data.shape = (batch, problem, 2) self.item_data = item_data self.batch_s = item_data.size(0) self.state = None def reset(self): if PROBLEM_SIZE == 50: capacity = 12.5 elif PROBLEM_SIZE == 100: capacity = 25 elif PROBLEM_SIZE == 200: capacity = 25 else: raise NotImplementedError self.state = STATE(item_data=self.item_data, capacity=capacity) reward = None done = False return self.state, reward, done def step(self, selected_item_idx): # selected_node_idx.shape = (batch,) # move state self.state.move_to(selected_item_idx) # returning values done = self.state.finished.all() if done: reward = self.state.accumulated_value else: reward = None return self.state, reward, done class GROUP_ENVIRONMENT: def __init__(self, item_data): # seq.shape = (batch, problem, 2) self.item_data = item_data self.batch_s = item_data.size(0) self.group_state = None def reset(self, group_size): if PROBLEM_SIZE == 50: capacity = 12.5 elif PROBLEM_SIZE == 100: capacity = 25 elif PROBLEM_SIZE == 200: capacity = 25 else: raise NotImplementedError self.group_state = GROUP_STATE(group_size=group_size, item_data=self.item_data, capacity=capacity) reward = None done = False return self.group_state, reward, done def step(self, selected_idx_mat): # selected_idx_mat.shape = (batch, group) # move state self.group_state.move_to(selected_idx_mat) # returning values reward = None done = self.group_state.finished.all() # state.finished.shape = (batch, group) if done: reward = self.group_state.accumulated_value else: reward = None return self.group_state, reward, done
10,073
33.033784
118
py
POMO
POMO-master/OLD_ipynb_ver/POMO_KP/source/MODEL__Actor/grouped_actors.py
""" The MIT License Copyright (c) 2020 Yeong-Dae Kwon Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions: The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software. THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. """ import torch import torch.nn as nn import torch.nn.functional as F import numpy as np # For debugging from IPython.core.debugger import set_trace # Hyper Parameters from HYPER_PARAMS import * from TORCH_OBJECTS import * ######################################## # ACTOR ######################################## class ACTOR(nn.Module): def __init__(self): super().__init__() self.encoder = Encoder() self.node_prob_calculator = Next_Node_Probability_Calculator_for_group() self.batch_s = None self.encoded_nodes_and_dummy = None self.encoded_nodes = None self.encoded_graph = None def reset(self, group_state): self.batch_s = group_state.item_data.size(0) self.encoded_nodes_and_dummy = Tensor(np.zeros((self.batch_s, PROBLEM_SIZE+1, EMBEDDING_DIM))) self.encoded_nodes_and_dummy[:, :PROBLEM_SIZE, :] = self.encoder(group_state.item_data) self.encoded_nodes = self.encoded_nodes_and_dummy[:, :PROBLEM_SIZE, :] # shape = (batch, problem, EMBEDDING_DIM) self.encoded_graph = self.encoded_nodes.mean(dim=1, keepdim=True) # shape = (batch, 1, EMBEDDING_DIM) self.node_prob_calculator.reset(self.encoded_nodes) def get_action_probabilities(self, group_state): probs = self.node_prob_calculator(graph=self.encoded_graph, capacity=group_state.capacity, ninf_mask=group_state.fit_ninf_mask) # shape = (batch, group, problem) return probs ######################################## # ACTOR_SUB_NN : ENCODER ######################################## class Encoder(nn.Module): def __init__(self): super().__init__() self.embedding = nn.Linear(2, EMBEDDING_DIM) self.layers = nn.ModuleList([Encoder_Layer() for _ in range(ENCODER_LAYER_NUM)]) def forward(self, item_data): # item_data.shape = (batch, problem, 2) embedded_input = self.embedding(item_data) # shape = (batch, problem, EMBEDDING_DIM) out = embedded_input for layer in self.layers: out = layer(out) return out class Encoder_Layer(nn.Module): def __init__(self): super().__init__() self.Wq = nn.Linear(EMBEDDING_DIM, HEAD_NUM * KEY_DIM, bias=False) self.Wk = nn.Linear(EMBEDDING_DIM, HEAD_NUM * KEY_DIM, bias=False) self.Wv = nn.Linear(EMBEDDING_DIM, HEAD_NUM * KEY_DIM, bias=False) self.multi_head_combine = nn.Linear(HEAD_NUM * KEY_DIM, EMBEDDING_DIM) self.addAndNormalization1 = Add_And_Normalization_Module() self.feedForward = Feed_Forward_Module() self.addAndNormalization2 = Add_And_Normalization_Module() def forward(self, input1): # input.shape = (batch, problem, EMBEDDING_DIM) q = reshape_by_heads(self.Wq(input1), head_num=HEAD_NUM) k = reshape_by_heads(self.Wk(input1), head_num=HEAD_NUM) v = reshape_by_heads(self.Wv(input1), head_num=HEAD_NUM) # q shape = (batch, HEAD_NUM, problem, KEY_DIM) out_concat = multi_head_attention(q, k, v) # shape = (batch, problem, HEAD_NUM*KEY_DIM) multi_head_out = self.multi_head_combine(out_concat) # shape = (batch, problem, EMBEDDING_DIM) out1 = self.addAndNormalization1(input1, multi_head_out) out2 = self.feedForward(out1) out3 = self.addAndNormalization2(out1, out2) return out3 ######################################## # ACTOR_SUB_NN : Next_Node_Probability_Calculator ######################################## class Next_Node_Probability_Calculator_for_group(nn.Module): def __init__(self): super().__init__() self.Wq = nn.Linear(1+EMBEDDING_DIM, HEAD_NUM * KEY_DIM, bias=False) self.Wk = nn.Linear(EMBEDDING_DIM, HEAD_NUM * KEY_DIM, bias=False) self.Wv = nn.Linear(EMBEDDING_DIM, HEAD_NUM * KEY_DIM, bias=False) self.multi_head_combine = nn.Linear(HEAD_NUM * KEY_DIM, EMBEDDING_DIM) self.k = None # saved key, for multi-head attention self.v = None # saved value, for multi-head_attention self.single_head_key = None # saved, for single-head attention def reset(self, encoded_nodes): # encoded_nodes.shape = (batch, problem, EMBEDDING_DIM) self.k = reshape_by_heads(self.Wk(encoded_nodes), head_num=HEAD_NUM) self.v = reshape_by_heads(self.Wv(encoded_nodes), head_num=HEAD_NUM) # shape = (batch, HEAD_NUM, problem, KEY_DIM) self.single_head_key = encoded_nodes.transpose(1, 2) # shape = (batch, EMBEDDING_DIM, problem) def forward(self, graph, capacity, ninf_mask=None): # graph.shape = (batch, 1, EMBEDDING_DIM) # capacity.shape = (batch, group) # ninf_mask.shape = (batch, group, problem) batch_s = capacity.size(0) group_s = capacity.size(1) # Multi-Head Attention ####################################################### input1 = graph.expand(batch_s, group_s, EMBEDDING_DIM) input2 = capacity[:, :, None] input_cat = torch.cat((input1, input2), dim=2) # shape = (batch, group, 1+EMBEDDING_DIM) q = reshape_by_heads(self.Wq(input_cat), head_num=HEAD_NUM) # shape = (batch, HEAD_NUM, group, KEY_DIM) out_concat = multi_head_attention(q, self.k, self.v, ninf_mask=ninf_mask) # shape = (batch, group, HEAD_NUM*KEY_DIM) mh_atten_out = self.multi_head_combine(out_concat) # shape = (batch, group, EMBEDDING_DIM) # Single-Head Attention, for probability calculation ####################################################### score = torch.matmul(mh_atten_out, self.single_head_key) # shape = (batch, group, problem) score_scaled = score / np.sqrt(EMBEDDING_DIM) # shape = (batch, group, problem) score_clipped = LOGIT_CLIPPING * torch.tanh(score_scaled) if ninf_mask is None: score_masked = score_clipped else: score_masked = score_clipped + ninf_mask probs = F.softmax(score_masked, dim=2) # shape = (batch, group, problem) return probs ######################################## # NN SUB CLASS / FUNCTIONS ######################################## def pick_nodes_for_each_group(encoded_nodes, node_index_to_pick): # encoded_nodes.shape = (batch, problem, EMBEDDING_DIM) # node_index_to_pick.shape = (batch, group_s) batch_s = node_index_to_pick.size(0) group_s = node_index_to_pick.size(1) gathering_index = node_index_to_pick[:, :, None].expand(batch_s, group_s, EMBEDDING_DIM) # shape = (batch, group, EMBEDDING_DIM) picked_nodes = encoded_nodes.gather(dim=1, index=gathering_index) # shape = (batch, group, EMBEDDING_DIM) return picked_nodes def reshape_by_heads(qkv, head_num): # q.shape = (batch, C, head_num*key_dim) batch_s = qkv.size(0) C = qkv.size(1) q_reshaped = qkv.reshape(batch_s, C, head_num, -1) # shape = (batch, C, head_num, key_dim) q_transposed = q_reshaped.transpose(1, 2) # shape = (batch, head_num, C, key_dim) return q_transposed def multi_head_attention(q, k, v, ninf_mask=None): # q shape = (batch, head_num, n, key_dim) : n can be either 1 or group # k,v shape = (batch, head_num, problem, key_dim) # ninf_mask.shape = (batch, group, problem) batch_s = q.size(0) head_num = q.size(1) n = q.size(2) key_dim = q.size(3) problem_s = k.size(2) score = torch.matmul(q, k.transpose(2, 3)) # shape = (batch, head_num, n, TSP_SIZE) score_scaled = score / np.sqrt(key_dim) if ninf_mask is not None: score_scaled = score_scaled + ninf_mask[:, None, :, :].expand(batch_s, head_num, n, problem_s) weights = nn.Softmax(dim=3)(score_scaled) # shape = (batch, head_num, n, problem) out = torch.matmul(weights, v) # shape = (batch, head_num, n, key_dim) out_transposed = out.transpose(1, 2) # shape = (batch, n, head_num, key_dim) out_concat = out_transposed.reshape(batch_s, n, head_num * key_dim) # shape = (batch, n, head_num*key_dim) return out_concat class Add_And_Normalization_Module(nn.Module): def __init__(self): super().__init__() self.norm_by_EMB = nn.BatchNorm1d(EMBEDDING_DIM, affine=True) # 'Funny' Batch_Norm, as it will normalized by EMB dim def forward(self, input1, input2): # input.shape = (batch, problem, EMBEDDING_DIM) batch_s = input1.size(0) problem_s = input1.size(1) added = input1 + input2 normalized = self.norm_by_EMB(added.reshape(batch_s * problem_s, EMBEDDING_DIM)) return normalized.reshape(batch_s, problem_s, EMBEDDING_DIM) class Feed_Forward_Module(nn.Module): def __init__(self): super().__init__() self.W1 = nn.Linear(EMBEDDING_DIM, FF_HIDDEN_DIM) self.W2 = nn.Linear(FF_HIDDEN_DIM, EMBEDDING_DIM) def forward(self, input1): # input.shape = (batch, problem, EMBEDDING_DIM) return self.W2(F.relu(self.W1(input1)))
10,279
32.594771
102
py
POMO
POMO-master/OLD_ipynb_ver/POMO_KP/source/TRAIN_N_EVAL/Train_Grouped_Actors.py
""" The MIT License Copyright (c) 2020 Yeong-Dae Kwon Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions: The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software. THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. """ import numpy as np # For Logging import time # For debugging from IPython.core.debugger import set_trace # Hyper Parameters from HYPER_PARAMS import * from TORCH_OBJECTS import * from source.utilities import Average_Meter from source.knapsack_problem import KNAPSACK_DATA_LOADER__RANDOM, GROUP_ENVIRONMENT ######################################## # TRAIN ######################################## def TRAIN(grouped_actor, epoch, timer_start, logger): grouped_actor.train() score_AM = Average_Meter() actor_loss_AM = Average_Meter() train_loader = KNAPSACK_DATA_LOADER__RANDOM(num_sample=TRAIN_DATASET_SIZE, num_items=PROBLEM_SIZE, batch_size=BATCH_SIZE) logger_start = time.time() episode = 0 for item_data in train_loader: # item_data.shape = (batch, problem, 2) batch_s = item_data.size(0) episode = episode + batch_s # Actor Group Move ############################################### env = GROUP_ENVIRONMENT(item_data) group_s = PROBLEM_SIZE group_state, reward, done = env.reset(group_size=group_s) grouped_actor.reset(group_state) # First Move is given first_action = LongTensor(np.arange(group_s))[None, :].expand(batch_s, group_s) group_state, reward, done = env.step(first_action) group_prob_list = Tensor(np.zeros((batch_s, group_s, 0))) while not done: action_probs = grouped_actor.get_action_probabilities(group_state) # shape = (batch, group, problem) action = action_probs.reshape(batch_s*group_s, PROBLEM_SIZE).multinomial(1).squeeze(dim=1).reshape(batch_s, group_s) # shape = (batch, group) action_w_finisehd = action.clone() action_w_finisehd[group_state.finished] = PROBLEM_SIZE # dummy item group_state, reward, done = env.step(action_w_finisehd) batch_idx_mat = torch.arange(batch_s)[:, None].expand(batch_s, group_s) group_idx_mat = torch.arange(group_s)[None, :].expand(batch_s, group_s) chosen_action_prob = action_probs[batch_idx_mat, group_idx_mat, action].reshape(batch_s, group_s) # shape = (batch, group) chosen_action_prob[group_state.finished] = 1 # done episode will gain no more probability group_prob_list = torch.cat((group_prob_list, chosen_action_prob[:, :, None]), dim=2) # shape = (batch, group, x) # LEARNING - Actor ############################################### group_reward = reward # shape = (batch, group) group_log_prob = group_prob_list.log().sum(dim=2) # shape = (batch, group) group_advantage = group_reward - group_reward.mean(dim=1, keepdim=True) group_loss = -group_advantage * group_log_prob # shape = (batch, group) loss = group_loss.mean() grouped_actor.optimizer.zero_grad() loss.backward() grouped_actor.optimizer.step() # RECORDING ############################################### max_reward, _ = group_reward.max(dim=1) score_AM.push(max_reward) actor_loss_AM.push(group_loss.detach()) # LOGGING ############################################### if (time.time()-logger_start > LOG_PERIOD_SEC) or (episode == TRAIN_DATASET_SIZE): timestr = time.strftime("%H:%M:%S", time.gmtime(time.time()-timer_start)) log_str = 'Ep:{:03d}-{:07d}({:5.1f}%) T:{:s} ALoss:{:+5f} Avg.dist:{:5f}' \ .format(epoch, episode, episode/TRAIN_DATASET_SIZE*100, timestr, actor_loss_AM.result(), score_AM.result()) logger.info(log_str) logger_start = time.time() # LR STEP, after each epoch grouped_actor.lr_stepper.step()
5,058
36.753731
128
py
POMO
POMO-master/OLD_ipynb_ver/POMO_KP/source/TRAIN_N_EVAL/Evaluate_Grouped_Actors.py
""" The MIT License Copyright (c) 2020 Yeong-Dae Kwon Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions: The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software. THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. """ import numpy as np # For debugging from IPython.core.debugger import set_trace # Hyper Parameters from HYPER_PARAMS import * from TORCH_OBJECTS import * from source.utilities import Average_Meter from source.knapsack_problem import KNAPSACK_DATA_LOADER__RANDOM, GROUP_ENVIRONMENT ######################################## # EVAL ######################################## eval_result = [] def EVAL(grouped_actor, epoch, timer_start, logger): global eval_result grouped_actor.eval() eval_AM = Average_Meter() test_loader = KNAPSACK_DATA_LOADER__RANDOM(num_sample=TEST_DATASET_SIZE, num_items=PROBLEM_SIZE, batch_size=TEST_BATCH_SIZE) with torch.no_grad(): for item_data in test_loader: # item_data.shape = (batch, problem, 2) batch_s = item_data.size(0) env = GROUP_ENVIRONMENT(item_data) group_s = PROBLEM_SIZE group_state, reward, done = env.reset(group_size=group_s) grouped_actor.reset(group_state) # First Move is given first_action = LongTensor(np.arange(group_s))[None, :].expand(batch_s, group_s) group_state, reward, done = env.step(first_action) while not done: action_probs = grouped_actor.get_action_probabilities(group_state) # shape = (batch, group, problem) action = action_probs.argmax(dim=2) # shape = (batch, group) action_w_finished = action.clone() action_w_finished[group_state.finished] = PROBLEM_SIZE # this is dummy item with 0 size 0 value group_state, reward, done = env.step(action_w_finished) max_reward, _ = group_state.accumulated_value.max(dim=1) eval_AM.push(max_reward) # LOGGING score_avg = eval_AM.result() eval_result.append(score_avg) logger.info('--------------------------------------------------------------------------') log_str = ' <<< EVAL after Epoch:{:03d} >>> Avg.score:{:f}'.format(epoch, score_avg) logger.info(log_str) logger.info('--------------------------------------------------------------------------') logger.info('eval_result = {}'.format(eval_result)) logger.info('--------------------------------------------------------------------------') logger.info('--------------------------------------------------------------------------') logger.info('--------------------------------------------------------------------------')
3,710
36.11
112
py
POMO
POMO-master/OLD_ipynb_ver/POMO_CVRP/TORCH_OBJECTS.py
""" The MIT License Copyright (c) 2020 Yeong-Dae Kwon Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions: The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software. THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. """ import torch use_cuda = torch.cuda.is_available() # use_cuda = False # torch.cuda.set_device(1) FloatTensor = torch.cuda.FloatTensor if use_cuda else torch.FloatTensor LongTensor = torch.cuda.LongTensor if use_cuda else torch.LongTensor ByteTensor = torch.cuda.ByteTensor if use_cuda else torch.ByteTensor BoolTensor = torch.cuda.BoolTensor if use_cuda else torch.BoolTensor Tensor = FloatTensor if use_cuda: device = torch.device("cuda") else: device = torch.device("cpu")
1,575
32.531915
77
py
POMO
POMO-master/OLD_ipynb_ver/POMO_CVRP/source/cvrp.py
""" The MIT License Copyright (c) 2020 Yeong-Dae Kwon Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions: The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software. THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. """ #################################### # EXTERNAL LIBRARY #################################### import torch import numpy as np # For debugging from IPython.core.debugger import set_trace from torch.utils.data import Dataset from torch.utils.data import DataLoader #################################### # PROJECT VARIABLES #################################### from HYPER_PARAMS import * from TORCH_OBJECTS import * #################################### # DATA #################################### def CVRP_DATA_LOADER__RANDOM(num_sample, num_nodes, batch_size): dataset = CVRP_Dataset__Random(num_sample=num_sample, num_nodes=num_nodes) data_loader = DataLoader(dataset=dataset, batch_size=batch_size, shuffle=False, num_workers=0, collate_fn=CVRP_collate_fn) return data_loader class CVRP_Dataset__Random(Dataset): def __init__(self, num_sample, num_nodes): self.num_sample = num_sample self.num_nodes = num_nodes if num_nodes == 20: demand_scaler = 30 elif num_nodes == 50: demand_scaler = 40 elif num_nodes == 100: demand_scaler = 50 else: raise NotImplementedError self.demand_scaler = demand_scaler def __getitem__(self, index): depot_xy_data = np.random.rand(1, 2) node_xy_data = np.random.rand(self.num_nodes, 2) node_demand_data = np.random.randint(1, 10, self.num_nodes) / self.demand_scaler return depot_xy_data, node_xy_data, node_demand_data def __len__(self): return self.num_sample def CVRP_collate_fn(batch): depot_xy_data_tuples, node_xy_data_tuples, node_demand_data_tuples = zip(*batch) depot_xy = Tensor(depot_xy_data_tuples) node_xy = Tensor(node_xy_data_tuples) node_demand = Tensor(node_demand_data_tuples)[:, :, None] # unsqeeeze to match the shape of node_xy return depot_xy, node_xy, node_demand #################################### # STATE #################################### class STATE: def __init__(self, data): # data.shape = (batch, problem+1, 3) self.batch_s = data.size(0) self.data = data # History #################################### self.selected_count = 0 self.current_node = None # shape = (batch,) self.selected_node_list = LongTensor(np.zeros((self.batch_s, 0))) # shape = (batch, selected_count) # Status #################################### self.at_the_depot = None # shape = (batch,) self.loaded = Tensor(np.ones((self.batch_s,))) # shape = (batch,) self.visited_ninf_flag = Tensor(np.zeros((self.batch_s, PROBLEM_SIZE+1))) # shape = (batch, problem+1) self.ninf_mask = Tensor(np.zeros((self.batch_s, PROBLEM_SIZE+1))) # shape = (batch, problem+1) self.finished = BoolTensor(np.zeros((self.batch_s,))) # shape = (batch,) def move_to(self, selected_node_idx): # selected_node_idx.shape = (batch,) # History #################################### self.selected_count += 1 self.current_node = selected_node_idx self.selected_node_list = torch.cat((self.selected_node_list, selected_node_idx[:, None]), dim=1) # Status #################################### self.at_the_depot = (selected_node_idx == 0) demand_list = self.data[:, :, 2] # shape = (batch, problem+1) gathering_index = selected_node_idx[:, None] selected_demand = demand_list.gather(dim=1, index=gathering_index).squeeze(dim=1) # shape = (batch,) self.loaded -= selected_demand self.loaded[self.at_the_depot] = 1 # refill loaded at the depot self.visited_ninf_flag[torch.arange(self.batch_s), selected_node_idx] = -np.inf self.finished = self.finished + (self.visited_ninf_flag == -np.inf).all(dim=1) # shape = (batch,) # Status Edit #################################### self.visited_ninf_flag[:, 0][~self.at_the_depot] = 0 # allow car to visit depot anytime demand_too_large = self.loaded[:, None] < demand_list # shape = (batch, problem+1) self.ninf_mask = self.visited_ninf_flag.clone() self.ninf_mask[demand_too_large] = -np.inf self.ninf_mask[self.finished[:, None].expand(self.batch_s, PROBLEM_SIZE+1)] = 0 # do not mask finished episode class GROUP_STATE: def __init__(self, group_size, data): # data.shape = (batch, problem+1, 3) self.batch_s = data.size(0) self.group_s = group_size self.data = data # History #################################### self.selected_count = 0 self.current_node = None # shape = (batch, group) self.selected_node_list = LongTensor(np.zeros((self.batch_s, self.group_s, 0))) # shape = (batch, group, selected_count) # Status #################################### self.at_the_depot = None # shape = (batch, group) self.loaded = Tensor(np.ones((self.batch_s, self.group_s))) # shape = (batch, group) self.visited_ninf_flag = Tensor(np.zeros((self.batch_s, self.group_s, PROBLEM_SIZE+1))) # shape = (batch, group, problem+1) self.ninf_mask = Tensor(np.zeros((self.batch_s, self.group_s, PROBLEM_SIZE+1))) # shape = (batch, group, problem+1) self.finished = BoolTensor(np.zeros((self.batch_s, self.group_s))) # shape = (batch, group) def move_to(self, selected_idx_mat): # selected_idx_mat.shape = (batch, group) # History #################################### self.selected_count += 1 self.current_node = selected_idx_mat self.selected_node_list = torch.cat((self.selected_node_list, selected_idx_mat[:, :, None]), dim=2) # Status #################################### self.at_the_depot = (selected_idx_mat == 0) demand_list = self.data[:, None, :, 2].expand(self.batch_s, self.group_s, -1) # shape = (batch, group, problem+1) gathering_index = selected_idx_mat[:, :, None] selected_demand = demand_list.gather(dim=2, index=gathering_index).squeeze(dim=2) # shape = (batch, group) self.loaded -= selected_demand self.loaded[self.at_the_depot] = 1 # refill loaded at the depot batch_idx_mat = torch.arange(self.batch_s)[:, None].expand(self.batch_s, self.group_s) group_idx_mat = torch.arange(self.group_s)[None, :].expand(self.batch_s, self.group_s) self.visited_ninf_flag[batch_idx_mat, group_idx_mat, selected_idx_mat] = -np.inf self.finished = self.finished + (self.visited_ninf_flag == -np.inf).all(dim=2) # shape = (batch, group) # Status Edit #################################### self.visited_ninf_flag[:, :, 0][~self.at_the_depot] = 0 # allow car to visit depot anytime round_error_epsilon = 0.000001 demand_too_large = self.loaded[:, :, None] + round_error_epsilon < demand_list # shape = (batch, group, problem+1) self.ninf_mask = self.visited_ninf_flag.clone() self.ninf_mask[demand_too_large] = -np.inf self.ninf_mask[self.finished[:, :, None].expand(self.batch_s, self.group_s, PROBLEM_SIZE+1)] = 0 # do not mask finished episode #################################### # ENVIRONMENT #################################### class ENVIRONMENT: def __init__(self, depot_xy, node_xy, node_demand): # depot_xy.shape = (batch, 1, 2) # node_xy.shape = (batch, problem, 2) # node_demand.shape = (batch, problem, 1) self.batch_s = depot_xy.size(0) self.state = None all_node_xy = torch.cat((depot_xy, node_xy), dim=1) # shape = (batch, problem+1, 2) depot_demand = Tensor(np.zeros((self.batch_s, 1, 1))) all_node_demand = torch.cat((depot_demand, node_demand), dim=1) # shape = (batch, problem+1, 1) self.data = torch.cat((all_node_xy, all_node_demand), dim=2) # shape = (batch, problem+1, 3) def reset(self): self.state = STATE(self.data) reward = None done = False return self.state, reward, done def step(self, selected_node_idx): # selected_node_idx.shape = (batch,) # move state self.state.move_to(selected_node_idx) # returning values done = self.state.finished.all() if done: reward = -self._get_travel_distance() # note the minus sign! else: reward = None return self.state, reward, done def _get_travel_distance(self): all_node_xy = self.data[:, :, 0:2] # shape = (batch, problem+1, 2) gathering_index = self.state.selected_node_list.unsqueeze(2).expand(-1, -1, 2) # shape = (batch, selected_count, 2) ordered_seq = all_node_xy.gather(dim=1, index=gathering_index) # shape = (batch, selected_count, 2) rolled_seq = ordered_seq.roll(dims=1, shifts=-1) segment_lengths = ((ordered_seq-rolled_seq)**2).sum(2).sqrt() # size = (batch, selected_count) travel_distances = segment_lengths.sum(1) # size = (batch,) return travel_distances class GROUP_ENVIRONMENT: def __init__(self, depot_xy, node_xy, node_demand): # depot_xy.shape = (batch, 1, 2) # node_xy.shape = (batch, problem, 2) # node_demand.shape = (batch, problem, 1) self.batch_s = depot_xy.size(0) self.group_s = None self.group_state = None all_node_xy = torch.cat((depot_xy, node_xy), dim=1) # shape = (batch, problem+1, 2) depot_demand = Tensor(np.zeros((self.batch_s, 1, 1))) all_node_demand = torch.cat((depot_demand, node_demand), dim=1) # shape = (batch, problem+1, 1) self.data = torch.cat((all_node_xy, all_node_demand), dim=2) # shape = (batch, problem+1, 3) def reset(self, group_size): self.group_s = group_size self.group_state = GROUP_STATE(group_size=group_size, data=self.data) reward = None done = False return self.group_state, reward, done def step(self, selected_idx_mat): # selected_idx_mat.shape = (batch, group) # move state self.group_state.move_to(selected_idx_mat) # returning values done = self.group_state.finished.all() # state.finished.shape = (batch, group) if done: reward = -self._get_travel_distance() # note the minus sign! else: reward = None return self.group_state, reward, done def _get_travel_distance(self): all_node_xy = self.data[:, None, :, 0:2].expand(self.batch_s, self.group_s, -1, 2) # shape = (batch, group, problem+1, 2) gathering_index = self.group_state.selected_node_list[:, :, :, None].expand(-1, -1, -1, 2) # shape = (batch, group, selected_count, 2) ordered_seq = all_node_xy.gather(dim=2, index=gathering_index) # shape = (batch, group, selected_count, 2) rolled_seq = ordered_seq.roll(dims=2, shifts=-1) segment_lengths = ((ordered_seq-rolled_seq)**2).sum(3).sqrt() # size = (batch, group, selected_count) travel_distances = segment_lengths.sum(2) # size = (batch, group) return travel_distances
12,715
35.645533
119
py
POMO
POMO-master/OLD_ipynb_ver/POMO_CVRP/source/utilities.py
""" The MIT License Copyright (c) 2020 Yeong-Dae Kwon Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions: The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software. THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. """ import logging import os import time import datetime import pytz import re import numpy as np from HYPER_PARAMS import * from TORCH_OBJECTS import * ######################################## # Get_Logger ######################################## tz = pytz.timezone("Asia/Seoul") def timetz(*args): return datetime.datetime.now(tz).timetuple() def Get_Logger(SAVE_FOLDER_NAME): # make_dir ####################################################### prefix = datetime.datetime.now(pytz.timezone("Asia/Seoul")).strftime("%Y%m%d_%H%M__") result_folder_no_postfix = "./result/{}".format(prefix + SAVE_FOLDER_NAME) result_folder_path = result_folder_no_postfix folder_idx = 0 while os.path.exists(result_folder_path): folder_idx += 1 result_folder_path = result_folder_no_postfix + "({})".format(folder_idx) os.makedirs(result_folder_path) # Logger ####################################################### logger = logging.getLogger(result_folder_path) # this already includes streamHandler?? streamHandler = logging.StreamHandler() fileHandler = logging.FileHandler('{}/log.txt'.format(result_folder_path)) formatter = logging.Formatter("[%(asctime)s] %(message)s", "%Y-%m-%d %H:%M:%S") formatter.converter = timetz streamHandler.setFormatter(formatter) fileHandler.setFormatter(formatter) logger.addHandler(streamHandler) logger.addHandler(fileHandler) logger.setLevel(level=logging.INFO) return logger, result_folder_path def Extract_from_LogFile(result_folder_path, variable_name): logfile_path = '{}/log.txt'.format(result_folder_path) with open(logfile_path) as f: datafile = f.readlines() found = False # This isn't really necessary for line in reversed(datafile): if variable_name in line: found = True m = re.search(variable_name + '[^\n]+', line) break exec_command = "Print(No such variable found !!)" if found: return m.group(0) else: return exec_command ######################################## # Average_Meter ######################################## class Average_Meter: def __init__(self): self.sum = None self.count = None self.reset() def reset(self): self.sum = torch.tensor(0.).to(device) self.count = 0 def push(self, some_tensor, n_for_rank_0_tensor=None): assert not some_tensor.requires_grad # You get Memory error, if you keep tensors with grad history rank = len(some_tensor.shape) if rank == 0: # assuming "already averaged" Tensor was pushed self.sum += some_tensor * n_for_rank_0_tensor self.count += n_for_rank_0_tensor else: self.sum += some_tensor.sum() self.count += some_tensor.numel() def peek(self): average = (self.sum / self.count).tolist() return average def result(self): average = (self.sum / self.count).tolist() self.reset() return average ######################################## # View NN Parameters ######################################## def get_n_params1(model): pp = 0 for p in list(model.parameters()): nn_count = 1 for s in list(p.size()): nn_count = nn_count * s pp += nn_count print(nn_count) print(p.shape) print("Total: {:d}".format(pp)) def get_n_params2(model): model_parameters = filter(lambda p: p.requires_grad, model.parameters()) params = sum([np.prod(p.size()) for p in model_parameters]) print(params) def get_n_params3(model): print(sum(p.numel() for p in model.parameters() if p.requires_grad)) def get_structure(model): print(model) ######################################## # Augment xy data ######################################## def augment_xy_data_by_8_fold(xy_data): # xy_data.shape = (batch_s, problem, 2) x = xy_data[:, :, [0]] y = xy_data[:, :, [1]] # x,y shape = (batch, problem, 1) dat1 = torch.cat((x, y), dim=2) dat2 = torch.cat((1-x, y), dim=2) dat3 = torch.cat((x, 1-y), dim=2) dat4 = torch.cat((1-x, 1-y), dim=2) dat5 = torch.cat((y, x), dim=2) dat6 = torch.cat((1-y, x), dim=2) dat7 = torch.cat((y, 1-x), dim=2) dat8 = torch.cat((1-y, 1-x), dim=2) data_augmented = torch.cat((dat1, dat2, dat3, dat4, dat5, dat6, dat7, dat8), dim=0) # shape = (8*batch, problem, 2) return data_augmented
5,645
26.950495
106
py
POMO
POMO-master/OLD_ipynb_ver/POMO_CVRP/source/MODEL__Actor/grouped_actors.py
""" The MIT License Copyright (c) 2020 Yeong-Dae Kwon Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions: The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software. THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. """ import torch import torch.nn as nn import torch.nn.functional as F import numpy as np # For debugging from IPython.core.debugger import set_trace # Hyper Parameters from HYPER_PARAMS import * from TORCH_OBJECTS import * ######################################## # ACTOR ######################################## class ACTOR(nn.Module): def __init__(self): super().__init__() self.encoder = Encoder() self.node_prob_calculator = Next_Node_Probability_Calculator_for_group() self.batch_s = None self.encoded_nodes = None self.encoded_graph = None def reset(self, group_state): self.batch_s = group_state.data.size(0) self.encoded_nodes = self.encoder(group_state.data) # shape = (batch, problem+1, EMBEDDING_DIM) self.encoded_graph = self.encoded_nodes.mean(dim=1, keepdim=True) # shape = (batch, 1, EMBEDDING_DIM) self.node_prob_calculator.reset(self.encoded_nodes) def get_action_probabilities(self, group_state): encoded_LAST_NODES = pick_nodes_for_each_group(self.encoded_nodes, group_state.current_node) # shape = (batch, group, EMBEDDING_DIM) remaining_loaded = group_state.loaded[:, :, None] # shape = (batch, group, 1) item_select_probabilities = self.node_prob_calculator(self.encoded_graph, encoded_LAST_NODES, remaining_loaded, ninf_mask=group_state.ninf_mask) # shape = (batch, group, problem+1) return item_select_probabilities ######################################## # ACTOR_SUB_NN : ENCODER ######################################## class Encoder(nn.Module): def __init__(self): super().__init__() self.embedding_depot = nn.Linear(2, EMBEDDING_DIM) self.embedding_node = nn.Linear(3, EMBEDDING_DIM) self.layers = nn.ModuleList([Encoder_Layer() for _ in range(ENCODER_LAYER_NUM)]) def forward(self, data): # data.shape = (batch, problem+1, 3) depot_xy = data[:, [0], 0:2] # shape = (batch, 1, 2) node_xy_demand = data[:, 1:, 0:3] # shape = (batch, problem, 3) embedded_depot = self.embedding_depot(depot_xy) # shape = (batch, 1, EMBEDDING_DIM) embedded_node = self.embedding_node(node_xy_demand) # shape = (batch, problem, EMBEDDING_DIM) out = torch.cat((embedded_depot, embedded_node), dim=1) # shape = (batch, problem+1, EMBEDDING_DIM) for layer in self.layers: out = layer(out) return out class Encoder_Layer(nn.Module): def __init__(self): super().__init__() self.Wq = nn.Linear(EMBEDDING_DIM, HEAD_NUM * KEY_DIM, bias=False) self.Wk = nn.Linear(EMBEDDING_DIM, HEAD_NUM * KEY_DIM, bias=False) self.Wv = nn.Linear(EMBEDDING_DIM, HEAD_NUM * KEY_DIM, bias=False) self.multi_head_combine = nn.Linear(HEAD_NUM * KEY_DIM, EMBEDDING_DIM) self.addAndNormalization1 = Add_And_Normalization_Module() self.feedForward = Feed_Forward_Module() self.addAndNormalization2 = Add_And_Normalization_Module() def forward(self, input1): # input.shape = (batch, problem, EMBEDDING_DIM) q = reshape_by_heads(self.Wq(input1), head_num=HEAD_NUM) k = reshape_by_heads(self.Wk(input1), head_num=HEAD_NUM) v = reshape_by_heads(self.Wv(input1), head_num=HEAD_NUM) # q shape = (batch, HEAD_NUM, problem, KEY_DIM) out_concat = multi_head_attention(q, k, v) # shape = (batch, problem, HEAD_NUM*KEY_DIM) multi_head_out = self.multi_head_combine(out_concat) # shape = (batch, problem, EMBEDDING_DIM) out1 = self.addAndNormalization1(input1, multi_head_out) out2 = self.feedForward(out1) out3 = self.addAndNormalization2(out1, out2) return out3 ######################################## # ACTOR_SUB_NN : Next_Node_Probability_Calculator ######################################## class Next_Node_Probability_Calculator_for_group(nn.Module): def __init__(self): super().__init__() self.Wq = nn.Linear(2*EMBEDDING_DIM+1, HEAD_NUM * KEY_DIM, bias=False) self.Wk = nn.Linear(EMBEDDING_DIM, HEAD_NUM * KEY_DIM, bias=False) self.Wv = nn.Linear(EMBEDDING_DIM, HEAD_NUM * KEY_DIM, bias=False) self.multi_head_combine = nn.Linear(HEAD_NUM * KEY_DIM, EMBEDDING_DIM) self.k = None # saved key, for multi-head attention self.v = None # saved value, for multi-head_attention self.single_head_key = None # saved, for single-head attention def reset(self, encoded_nodes): # encoded_nodes.shape = (batch, problem+1, EMBEDDING_DIM) self.k = reshape_by_heads(self.Wk(encoded_nodes), head_num=HEAD_NUM) self.v = reshape_by_heads(self.Wv(encoded_nodes), head_num=HEAD_NUM) # shape = (batch, HEAD_NUM, problem+1, KEY_DIM) self.single_head_key = encoded_nodes.transpose(1, 2) # shape = (batch, EMBEDDING_DIM, problem+1) def forward(self, input1, input2, remaining_loaded, ninf_mask=None): # input1.shape = (batch, 1, EMBEDDING_DIM) # input2.shape = (batch, group, EMBEDDING_DIM) # remaining_loaded.shape = (batch, group, 1) # ninf_mask.shape = (batch, group, problem+1) group_s = input2.size(1) # Multi-Head Attention ####################################################### input_cat = torch.cat((input1.expand(-1, group_s, -1), input2, remaining_loaded), dim=2) # shape = (batch, group, 2*EMBEDDING_DIM+1) q = reshape_by_heads(self.Wq(input_cat), head_num=HEAD_NUM) # shape = (batch, HEAD_NUM, group, KEY_DIM) out_concat = multi_head_attention(q, self.k, self.v, ninf_mask=ninf_mask) # shape = (batch, n, HEAD_NUM*KEY_DIM) mh_atten_out = self.multi_head_combine(out_concat) # shape = (batch, n, EMBEDDING_DIM) # Single-Head Attention, for probability calculation ####################################################### score = torch.matmul(mh_atten_out, self.single_head_key) # shape = (batch, n, problem+1) score_scaled = score / np.sqrt(EMBEDDING_DIM) # shape = (batch_s, group, problem+1) score_clipped = LOGIT_CLIPPING * torch.tanh(score_scaled) if ninf_mask is None: score_masked = score_clipped else: score_masked = score_clipped + ninf_mask probs = F.softmax(score_masked, dim=2) # shape = (batch, group, problem+1) return probs ######################################## # NN SUB CLASS / FUNCTIONS ######################################## def pick_nodes_for_each_group(encoded_nodes, node_index_to_pick): # encoded_nodes.shape = (batch, problem, EMBEDDING_DIM) # node_index_to_pick.shape = (batch, group) gathering_index = node_index_to_pick[:, :, None].expand(-1, -1, EMBEDDING_DIM) # shape = (batch, group, EMBEDDING_DIM) picked_nodes = encoded_nodes.gather(dim=1, index=gathering_index) # shape = (batch, group, EMBEDDING_DIM) return picked_nodes def reshape_by_heads(qkv, head_num): # q.shape = (batch, C, head_num*key_dim) batch_s = qkv.size(0) C = qkv.size(1) q_reshaped = qkv.reshape(batch_s, C, head_num, -1) # shape = (batch, C, head_num, key_dim) q_transposed = q_reshaped.transpose(1, 2) # shape = (batch, head_num, C, key_dim) return q_transposed def multi_head_attention(q, k, v, ninf_mask=None): # q shape = (batch, head_num, n, key_dim) : n can be either 1 or group # k,v shape = (batch, head_num, problem, key_dim) # ninf_mask.shape = (batch, group, problem) batch_s = q.size(0) head_num = q.size(1) n = q.size(2) key_dim = q.size(3) problem_s = k.size(2) score = torch.matmul(q, k.transpose(2, 3)) # shape = (batch, head_num, n, problem) score_scaled = score / np.sqrt(key_dim) if ninf_mask is not None: score_scaled = score_scaled + ninf_mask[:, None, :, :].expand(batch_s, head_num, n, problem_s) weights = nn.Softmax(dim=3)(score_scaled) # shape = (batch, head_num, n, problem) out = torch.matmul(weights, v) # shape = (batch, head_num, n, key_dim) out_transposed = out.transpose(1, 2) # shape = (batch, n, head_num, key_dim) out_concat = out_transposed.reshape(batch_s, n, head_num * key_dim) # shape = (batch, n, head_num*key_dim) return out_concat class Add_And_Normalization_Module(nn.Module): def __init__(self): super().__init__() self.norm_by_EMB = nn.BatchNorm1d(EMBEDDING_DIM, affine=True) # 'Funny' Batch_Norm, as it will normalized by EMB dim def forward(self, input1, input2): # input.shape = (batch, problem, EMBEDDING_DIM) batch_s = input1.size(0) problem_s = input1.size(1) added = input1 + input2 normalized = self.norm_by_EMB(added.reshape(batch_s * problem_s, EMBEDDING_DIM)) return normalized.reshape(batch_s, problem_s, EMBEDDING_DIM) class Feed_Forward_Module(nn.Module): def __init__(self): super().__init__() self.W1 = nn.Linear(EMBEDDING_DIM, FF_HIDDEN_DIM) self.W2 = nn.Linear(FF_HIDDEN_DIM, EMBEDDING_DIM) def forward(self, input1): # input.shape = (batch, problem, EMBEDDING_DIM) return self.W2(F.relu(self.W1(input1)))
10,612
33.016026
112
py
POMO
POMO-master/OLD_ipynb_ver/POMO_CVRP/source/TRAIN_N_EVAL/Evaluate__Grouped_Actors.py
""" The MIT License Copyright (c) 2020 Yeong-Dae Kwon Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions: The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software. THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. """ import numpy as np # For debugging from IPython.core.debugger import set_trace # Hyper Parameters from HYPER_PARAMS import * from TORCH_OBJECTS import * from source.utilities import Average_Meter, augment_xy_data_by_8_fold from source.cvrp import CVRP_DATA_LOADER__RANDOM, GROUP_ENVIRONMENT ######################################## # EVAL ######################################## eval_result = [] def EVAL(grouped_actor, epoch, timer_start, logger): global eval_result grouped_actor.eval() eval_AM = Average_Meter() test_loader = CVRP_DATA_LOADER__RANDOM(num_sample=TEST_DATASET_SIZE, num_nodes=PROBLEM_SIZE, batch_size=TEST_BATCH_SIZE) with torch.no_grad(): for depot_xy, node_xy, node_demand in test_loader: # depot_xy.shape = (batch, 1, 2) # node_xy.shape = (batch, problem, 2) # node_demand.shape = (batch, problem, 1) batch_s = depot_xy.size(0) env = GROUP_ENVIRONMENT(depot_xy, node_xy, node_demand) group_s = PROBLEM_SIZE group_state, reward, done = env.reset(group_size=group_s) grouped_actor.reset(group_state) # First Move is given first_action = LongTensor(np.zeros((batch_s, group_s))) # start from node_0-depot group_state, reward, done = env.step(first_action) # Second Move is given second_action = LongTensor(np.arange(group_s))[None, :].expand(batch_s, group_s) group_state, reward, done = env.step(second_action) while not done: action_probs = grouped_actor.get_action_probabilities(group_state) # shape = (batch, group, problem+1) action = action_probs.argmax(dim=2) # shape = (batch, group) action[group_state.finished] = 0 # stay at depot, if you are finished group_state, reward, done = env.step(action) max_reward, _ = reward.max(dim=1) eval_AM.push(-max_reward) # reward was given as negative dist # LOGGING dist_avg = eval_AM.result() eval_result.append(dist_avg) logger.info('--------------------------------------------------------------------------') log_str = ' <<< EVAL after Epoch:{:03d} >>> Avg.dist:{:f}'.format(epoch, dist_avg) logger.info(log_str) logger.info('--------------------------------------------------------------------------') logger.info('eval_result = {}'.format(eval_result)) logger.info('--------------------------------------------------------------------------') logger.info('--------------------------------------------------------------------------') logger.info('--------------------------------------------------------------------------')
3,969
36.45283
94
py
POMO
POMO-master/OLD_ipynb_ver/POMO_CVRP/source/TRAIN_N_EVAL/Train_Grouped_Actors.py
""" The MIT License Copyright (c) 2020 Yeong-Dae Kwon Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions: The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software. THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. """ import numpy as np # For Logging import time # For debugging from IPython.core.debugger import set_trace # Hyper Parameters from HYPER_PARAMS import * from TORCH_OBJECTS import * from source.utilities import Average_Meter from source.cvrp import CVRP_DATA_LOADER__RANDOM, GROUP_ENVIRONMENT ######################################## # TRAIN ######################################## def TRAIN(grouped_actor, epoch, timer_start, logger): grouped_actor.train() dist_AM = Average_Meter() actor_loss_AM = Average_Meter() train_loader = CVRP_DATA_LOADER__RANDOM(num_sample=TRAIN_DATASET_SIZE, num_nodes=PROBLEM_SIZE, batch_size=BATCH_SIZE) logger_start = time.time() episode = 0 for depot_xy, node_xy, node_demand in train_loader: # depot_xy.shape = (batch, 1, 2) # node_xy.shape = (batch, problem, 2) # node_demand.shape = (batch, problem, 1) batch_s = depot_xy.size(0) episode = episode + batch_s # Actor Group Move ############################################### env = GROUP_ENVIRONMENT(depot_xy, node_xy, node_demand) group_s = PROBLEM_SIZE group_state, reward, done = env.reset(group_size=group_s) grouped_actor.reset(group_state) # First Move is given first_action = LongTensor(np.zeros((batch_s, group_s))) # start from node_0-depot group_state, reward, done = env.step(first_action) # Second Move is given second_action = LongTensor(np.arange(group_s)+1)[None, :].expand(batch_s, group_s) group_state, reward, done = env.step(second_action) group_prob_list = Tensor(np.zeros((batch_s, group_s, 0))) while not done: action_probs = grouped_actor.get_action_probabilities(group_state) # shape = (batch, group, problem+1) action = action_probs.reshape(batch_s*group_s, -1).multinomial(1)\ .squeeze(dim=1).reshape(batch_s, group_s) # shape = (batch, group) action[group_state.finished] = 0 # stay at depot, if you are finished group_state, reward, done = env.step(action) batch_idx_mat = torch.arange(batch_s)[:, None].expand(batch_s, group_s) group_idx_mat = torch.arange(group_s)[None, :].expand(batch_s, group_s) chosen_action_prob = action_probs[batch_idx_mat, group_idx_mat, action].reshape(batch_s, group_s) # shape = (batch, group) chosen_action_prob[group_state.finished] = 1 # done episode will gain no more probability group_prob_list = torch.cat((group_prob_list, chosen_action_prob[:, :, None]), dim=2) # shape = (batch, group, x) # LEARNING - Actor ############################################### group_reward = reward # shape = (batch, group) group_log_prob = group_prob_list.log().sum(dim=2) # shape = (batch, group) group_advantage = group_reward - group_reward.mean(dim=1, keepdim=True) group_loss = -group_advantage * group_log_prob # shape = (batch, group) loss = group_loss.mean() grouped_actor.optimizer.zero_grad() loss.backward() grouped_actor.optimizer.step() # RECORDING ############################################### max_reward, _ = group_reward.max(dim=1) dist_AM.push(-max_reward) # reward was given as negative dist actor_loss_AM.push(group_loss.detach()) # LOGGING ############################################### if (time.time()-logger_start > LOG_PERIOD_SEC) or (episode == TRAIN_DATASET_SIZE): timestr = time.strftime("%H:%M:%S", time.gmtime(time.time()-timer_start)) log_str = 'Ep:{:03d}-{:07d}({:5.1f}%) T:{:s} ALoss:{:+5f} Avg.dist:{:5f}' \ .format(epoch, episode, episode/TRAIN_DATASET_SIZE*100, timestr, actor_loss_AM.result(), dist_AM.result()) logger.info(log_str) logger_start = time.time() # LR STEP, after each epoch grouped_actor.lr_stepper.step()
5,310
37.485507
109
py
POMO
POMO-master/NEW_py_ver/CVRP/CVRProblemDef.py
import torch import numpy as np def get_random_problems(batch_size, problem_size): depot_xy = torch.rand(size=(batch_size, 1, 2)) # shape: (batch, 1, 2) node_xy = torch.rand(size=(batch_size, problem_size, 2)) # shape: (batch, problem, 2) if problem_size == 20: demand_scaler = 30 elif problem_size == 50: demand_scaler = 40 elif problem_size == 100: demand_scaler = 50 else: raise NotImplementedError node_demand = torch.randint(1, 10, size=(batch_size, problem_size)) / float(demand_scaler) # shape: (batch, problem) return depot_xy, node_xy, node_demand def augment_xy_data_by_8_fold(xy_data): # xy_data.shape: (batch, N, 2) x = xy_data[:, :, [0]] y = xy_data[:, :, [1]] # x,y shape: (batch, N, 1) dat1 = torch.cat((x, y), dim=2) dat2 = torch.cat((1 - x, y), dim=2) dat3 = torch.cat((x, 1 - y), dim=2) dat4 = torch.cat((1 - x, 1 - y), dim=2) dat5 = torch.cat((y, x), dim=2) dat6 = torch.cat((1 - y, x), dim=2) dat7 = torch.cat((y, 1 - x), dim=2) dat8 = torch.cat((1 - y, 1 - x), dim=2) aug_xy_data = torch.cat((dat1, dat2, dat3, dat4, dat5, dat6, dat7, dat8), dim=0) # shape: (8*batch, N, 2) return aug_xy_data
1,263
25.333333
94
py
POMO
POMO-master/NEW_py_ver/CVRP/POMO/CVRPEnv.py
from dataclasses import dataclass import torch from CVRProblemDef import get_random_problems, augment_xy_data_by_8_fold @dataclass class Reset_State: depot_xy: torch.Tensor = None # shape: (batch, 1, 2) node_xy: torch.Tensor = None # shape: (batch, problem, 2) node_demand: torch.Tensor = None # shape: (batch, problem) @dataclass class Step_State: BATCH_IDX: torch.Tensor = None POMO_IDX: torch.Tensor = None # shape: (batch, pomo) selected_count: int = None load: torch.Tensor = None # shape: (batch, pomo) current_node: torch.Tensor = None # shape: (batch, pomo) ninf_mask: torch.Tensor = None # shape: (batch, pomo, problem+1) finished: torch.Tensor = None # shape: (batch, pomo) class CVRPEnv: def __init__(self, **env_params): # Const @INIT #################################### self.env_params = env_params self.problem_size = env_params['problem_size'] self.pomo_size = env_params['pomo_size'] self.FLAG__use_saved_problems = False self.saved_depot_xy = None self.saved_node_xy = None self.saved_node_demand = None self.saved_index = None # Const @Load_Problem #################################### self.batch_size = None self.BATCH_IDX = None self.POMO_IDX = None # IDX.shape: (batch, pomo) self.depot_node_xy = None # shape: (batch, problem+1, 2) self.depot_node_demand = None # shape: (batch, problem+1) # Dynamic-1 #################################### self.selected_count = None self.current_node = None # shape: (batch, pomo) self.selected_node_list = None # shape: (batch, pomo, 0~) # Dynamic-2 #################################### self.at_the_depot = None # shape: (batch, pomo) self.load = None # shape: (batch, pomo) self.visited_ninf_flag = None # shape: (batch, pomo, problem+1) self.ninf_mask = None # shape: (batch, pomo, problem+1) self.finished = None # shape: (batch, pomo) # states to return #################################### self.reset_state = Reset_State() self.step_state = Step_State() def use_saved_problems(self, filename, device): self.FLAG__use_saved_problems = True loaded_dict = torch.load(filename, map_location=device) self.saved_depot_xy = loaded_dict['depot_xy'] self.saved_node_xy = loaded_dict['node_xy'] self.saved_node_demand = loaded_dict['node_demand'] self.saved_index = 0 def load_problems(self, batch_size, aug_factor=1): self.batch_size = batch_size if not self.FLAG__use_saved_problems: depot_xy, node_xy, node_demand = get_random_problems(batch_size, self.problem_size) else: depot_xy = self.saved_depot_xy[self.saved_index:self.saved_index+batch_size] node_xy = self.saved_node_xy[self.saved_index:self.saved_index+batch_size] node_demand = self.saved_node_demand[self.saved_index:self.saved_index+batch_size] self.saved_index += batch_size if aug_factor > 1: if aug_factor == 8: self.batch_size = self.batch_size * 8 depot_xy = augment_xy_data_by_8_fold(depot_xy) node_xy = augment_xy_data_by_8_fold(node_xy) node_demand = node_demand.repeat(8, 1) else: raise NotImplementedError self.depot_node_xy = torch.cat((depot_xy, node_xy), dim=1) # shape: (batch, problem+1, 2) depot_demand = torch.zeros(size=(self.batch_size, 1)) # shape: (batch, 1) self.depot_node_demand = torch.cat((depot_demand, node_demand), dim=1) # shape: (batch, problem+1) self.BATCH_IDX = torch.arange(self.batch_size)[:, None].expand(self.batch_size, self.pomo_size) self.POMO_IDX = torch.arange(self.pomo_size)[None, :].expand(self.batch_size, self.pomo_size) self.reset_state.depot_xy = depot_xy self.reset_state.node_xy = node_xy self.reset_state.node_demand = node_demand self.step_state.BATCH_IDX = self.BATCH_IDX self.step_state.POMO_IDX = self.POMO_IDX def reset(self): self.selected_count = 0 self.current_node = None # shape: (batch, pomo) self.selected_node_list = torch.zeros((self.batch_size, self.pomo_size, 0), dtype=torch.long) # shape: (batch, pomo, 0~) self.at_the_depot = torch.ones(size=(self.batch_size, self.pomo_size), dtype=torch.bool) # shape: (batch, pomo) self.load = torch.ones(size=(self.batch_size, self.pomo_size)) # shape: (batch, pomo) self.visited_ninf_flag = torch.zeros(size=(self.batch_size, self.pomo_size, self.problem_size+1)) # shape: (batch, pomo, problem+1) self.ninf_mask = torch.zeros(size=(self.batch_size, self.pomo_size, self.problem_size+1)) # shape: (batch, pomo, problem+1) self.finished = torch.zeros(size=(self.batch_size, self.pomo_size), dtype=torch.bool) # shape: (batch, pomo) reward = None done = False return self.reset_state, reward, done def pre_step(self): self.step_state.selected_count = self.selected_count self.step_state.load = self.load self.step_state.current_node = self.current_node self.step_state.ninf_mask = self.ninf_mask self.step_state.finished = self.finished reward = None done = False return self.step_state, reward, done def step(self, selected): # selected.shape: (batch, pomo) # Dynamic-1 #################################### self.selected_count += 1 self.current_node = selected # shape: (batch, pomo) self.selected_node_list = torch.cat((self.selected_node_list, self.current_node[:, :, None]), dim=2) # shape: (batch, pomo, 0~) # Dynamic-2 #################################### self.at_the_depot = (selected == 0) demand_list = self.depot_node_demand[:, None, :].expand(self.batch_size, self.pomo_size, -1) # shape: (batch, pomo, problem+1) gathering_index = selected[:, :, None] # shape: (batch, pomo, 1) selected_demand = demand_list.gather(dim=2, index=gathering_index).squeeze(dim=2) # shape: (batch, pomo) self.load -= selected_demand self.load[self.at_the_depot] = 1 # refill loaded at the depot self.visited_ninf_flag[self.BATCH_IDX, self.POMO_IDX, selected] = float('-inf') # shape: (batch, pomo, problem+1) self.visited_ninf_flag[:, :, 0][~self.at_the_depot] = 0 # depot is considered unvisited, unless you are AT the depot self.ninf_mask = self.visited_ninf_flag.clone() round_error_epsilon = 0.00001 demand_too_large = self.load[:, :, None] + round_error_epsilon < demand_list # shape: (batch, pomo, problem+1) self.ninf_mask[demand_too_large] = float('-inf') # shape: (batch, pomo, problem+1) newly_finished = (self.visited_ninf_flag == float('-inf')).all(dim=2) # shape: (batch, pomo) self.finished = self.finished + newly_finished # shape: (batch, pomo) # do not mask depot for finished episode. self.ninf_mask[:, :, 0][self.finished] = 0 self.step_state.selected_count = self.selected_count self.step_state.load = self.load self.step_state.current_node = self.current_node self.step_state.ninf_mask = self.ninf_mask self.step_state.finished = self.finished # returning values done = self.finished.all() if done: reward = -self._get_travel_distance() # note the minus sign! else: reward = None return self.step_state, reward, done def _get_travel_distance(self): gathering_index = self.selected_node_list[:, :, :, None].expand(-1, -1, -1, 2) # shape: (batch, pomo, selected_list_length, 2) all_xy = self.depot_node_xy[:, None, :, :].expand(-1, self.pomo_size, -1, -1) # shape: (batch, pomo, problem+1, 2) ordered_seq = all_xy.gather(dim=2, index=gathering_index) # shape: (batch, pomo, selected_list_length, 2) rolled_seq = ordered_seq.roll(dims=2, shifts=-1) segment_lengths = ((ordered_seq-rolled_seq)**2).sum(3).sqrt() # shape: (batch, pomo, selected_list_length) travel_distances = segment_lengths.sum(2) # shape: (batch, pomo) return travel_distances
8,799
35.514523
125
py
POMO
POMO-master/NEW_py_ver/CVRP/POMO/CVRPModel.py
import torch import torch.nn as nn import torch.nn.functional as F class CVRPModel(nn.Module): def __init__(self, **model_params): super().__init__() self.model_params = model_params self.encoder = CVRP_Encoder(**model_params) self.decoder = CVRP_Decoder(**model_params) self.encoded_nodes = None # shape: (batch, problem+1, EMBEDDING_DIM) def pre_forward(self, reset_state): depot_xy = reset_state.depot_xy # shape: (batch, 1, 2) node_xy = reset_state.node_xy # shape: (batch, problem, 2) node_demand = reset_state.node_demand # shape: (batch, problem) node_xy_demand = torch.cat((node_xy, node_demand[:, :, None]), dim=2) # shape: (batch, problem, 3) self.encoded_nodes = self.encoder(depot_xy, node_xy_demand) # shape: (batch, problem+1, embedding) self.decoder.set_kv(self.encoded_nodes) def forward(self, state): batch_size = state.BATCH_IDX.size(0) pomo_size = state.BATCH_IDX.size(1) if state.selected_count == 0: # First Move, depot selected = torch.zeros(size=(batch_size, pomo_size), dtype=torch.long) prob = torch.ones(size=(batch_size, pomo_size)) # # Use Averaged encoded nodes for decoder input_1 # encoded_nodes_mean = self.encoded_nodes.mean(dim=1, keepdim=True) # # shape: (batch, 1, embedding) # self.decoder.set_q1(encoded_nodes_mean) # # Use encoded_depot for decoder input_2 # encoded_first_node = self.encoded_nodes[:, [0], :] # # shape: (batch, 1, embedding) # self.decoder.set_q2(encoded_first_node) elif state.selected_count == 1: # Second Move, POMO selected = torch.arange(start=1, end=pomo_size+1)[None, :].expand(batch_size, pomo_size) prob = torch.ones(size=(batch_size, pomo_size)) else: encoded_last_node = _get_encoding(self.encoded_nodes, state.current_node) # shape: (batch, pomo, embedding) probs = self.decoder(encoded_last_node, state.load, ninf_mask=state.ninf_mask) # shape: (batch, pomo, problem+1) if self.training or self.model_params['eval_type'] == 'softmax': while True: # to fix pytorch.multinomial bug on selecting 0 probability elements with torch.no_grad(): selected = probs.reshape(batch_size * pomo_size, -1).multinomial(1) \ .squeeze(dim=1).reshape(batch_size, pomo_size) # shape: (batch, pomo) prob = probs[state.BATCH_IDX, state.POMO_IDX, selected].reshape(batch_size, pomo_size) # shape: (batch, pomo) if (prob != 0).all(): break else: selected = probs.argmax(dim=2) # shape: (batch, pomo) prob = None # value not needed. Can be anything. return selected, prob def _get_encoding(encoded_nodes, node_index_to_pick): # encoded_nodes.shape: (batch, problem, embedding) # node_index_to_pick.shape: (batch, pomo) batch_size = node_index_to_pick.size(0) pomo_size = node_index_to_pick.size(1) embedding_dim = encoded_nodes.size(2) gathering_index = node_index_to_pick[:, :, None].expand(batch_size, pomo_size, embedding_dim) # shape: (batch, pomo, embedding) picked_nodes = encoded_nodes.gather(dim=1, index=gathering_index) # shape: (batch, pomo, embedding) return picked_nodes ######################################## # ENCODER ######################################## class CVRP_Encoder(nn.Module): def __init__(self, **model_params): super().__init__() self.model_params = model_params embedding_dim = self.model_params['embedding_dim'] encoder_layer_num = self.model_params['encoder_layer_num'] self.embedding_depot = nn.Linear(2, embedding_dim) self.embedding_node = nn.Linear(3, embedding_dim) self.layers = nn.ModuleList([EncoderLayer(**model_params) for _ in range(encoder_layer_num)]) def forward(self, depot_xy, node_xy_demand): # depot_xy.shape: (batch, 1, 2) # node_xy_demand.shape: (batch, problem, 3) embedded_depot = self.embedding_depot(depot_xy) # shape: (batch, 1, embedding) embedded_node = self.embedding_node(node_xy_demand) # shape: (batch, problem, embedding) out = torch.cat((embedded_depot, embedded_node), dim=1) # shape: (batch, problem+1, embedding) for layer in self.layers: out = layer(out) return out # shape: (batch, problem+1, embedding) class EncoderLayer(nn.Module): def __init__(self, **model_params): super().__init__() self.model_params = model_params embedding_dim = self.model_params['embedding_dim'] head_num = self.model_params['head_num'] qkv_dim = self.model_params['qkv_dim'] self.Wq = nn.Linear(embedding_dim, head_num * qkv_dim, bias=False) self.Wk = nn.Linear(embedding_dim, head_num * qkv_dim, bias=False) self.Wv = nn.Linear(embedding_dim, head_num * qkv_dim, bias=False) self.multi_head_combine = nn.Linear(head_num * qkv_dim, embedding_dim) self.add_n_normalization_1 = AddAndInstanceNormalization(**model_params) self.feed_forward = FeedForward(**model_params) self.add_n_normalization_2 = AddAndInstanceNormalization(**model_params) def forward(self, input1): # input1.shape: (batch, problem+1, embedding) head_num = self.model_params['head_num'] q = reshape_by_heads(self.Wq(input1), head_num=head_num) k = reshape_by_heads(self.Wk(input1), head_num=head_num) v = reshape_by_heads(self.Wv(input1), head_num=head_num) # qkv shape: (batch, head_num, problem, qkv_dim) out_concat = multi_head_attention(q, k, v) # shape: (batch, problem, head_num*qkv_dim) multi_head_out = self.multi_head_combine(out_concat) # shape: (batch, problem, embedding) out1 = self.add_n_normalization_1(input1, multi_head_out) out2 = self.feed_forward(out1) out3 = self.add_n_normalization_2(out1, out2) return out3 # shape: (batch, problem, embedding) ######################################## # DECODER ######################################## class CVRP_Decoder(nn.Module): def __init__(self, **model_params): super().__init__() self.model_params = model_params embedding_dim = self.model_params['embedding_dim'] head_num = self.model_params['head_num'] qkv_dim = self.model_params['qkv_dim'] # self.Wq_1 = nn.Linear(embedding_dim, head_num * qkv_dim, bias=False) # self.Wq_2 = nn.Linear(embedding_dim, head_num * qkv_dim, bias=False) self.Wq_last = nn.Linear(embedding_dim+1, head_num * qkv_dim, bias=False) self.Wk = nn.Linear(embedding_dim, head_num * qkv_dim, bias=False) self.Wv = nn.Linear(embedding_dim, head_num * qkv_dim, bias=False) self.multi_head_combine = nn.Linear(head_num * qkv_dim, embedding_dim) self.k = None # saved key, for multi-head attention self.v = None # saved value, for multi-head_attention self.single_head_key = None # saved, for single-head attention # self.q1 = None # saved q1, for multi-head attention # self.q2 = None # saved q2, for multi-head attention def set_kv(self, encoded_nodes): # encoded_nodes.shape: (batch, problem+1, embedding) head_num = self.model_params['head_num'] self.k = reshape_by_heads(self.Wk(encoded_nodes), head_num=head_num) self.v = reshape_by_heads(self.Wv(encoded_nodes), head_num=head_num) # shape: (batch, head_num, problem+1, qkv_dim) self.single_head_key = encoded_nodes.transpose(1, 2) # shape: (batch, embedding, problem+1) def set_q1(self, encoded_q1): # encoded_q.shape: (batch, n, embedding) # n can be 1 or pomo head_num = self.model_params['head_num'] self.q1 = reshape_by_heads(self.Wq_1(encoded_q1), head_num=head_num) # shape: (batch, head_num, n, qkv_dim) def set_q2(self, encoded_q2): # encoded_q.shape: (batch, n, embedding) # n can be 1 or pomo head_num = self.model_params['head_num'] self.q2 = reshape_by_heads(self.Wq_2(encoded_q2), head_num=head_num) # shape: (batch, head_num, n, qkv_dim) def forward(self, encoded_last_node, load, ninf_mask): # encoded_last_node.shape: (batch, pomo, embedding) # load.shape: (batch, pomo) # ninf_mask.shape: (batch, pomo, problem) head_num = self.model_params['head_num'] # Multi-Head Attention ####################################################### input_cat = torch.cat((encoded_last_node, load[:, :, None]), dim=2) # shape = (batch, group, EMBEDDING_DIM+1) q_last = reshape_by_heads(self.Wq_last(input_cat), head_num=head_num) # shape: (batch, head_num, pomo, qkv_dim) # q = self.q1 + self.q2 + q_last # # shape: (batch, head_num, pomo, qkv_dim) q = q_last # shape: (batch, head_num, pomo, qkv_dim) out_concat = multi_head_attention(q, self.k, self.v, rank3_ninf_mask=ninf_mask) # shape: (batch, pomo, head_num*qkv_dim) mh_atten_out = self.multi_head_combine(out_concat) # shape: (batch, pomo, embedding) # Single-Head Attention, for probability calculation ####################################################### score = torch.matmul(mh_atten_out, self.single_head_key) # shape: (batch, pomo, problem) sqrt_embedding_dim = self.model_params['sqrt_embedding_dim'] logit_clipping = self.model_params['logit_clipping'] score_scaled = score / sqrt_embedding_dim # shape: (batch, pomo, problem) score_clipped = logit_clipping * torch.tanh(score_scaled) score_masked = score_clipped + ninf_mask probs = F.softmax(score_masked, dim=2) # shape: (batch, pomo, problem) return probs ######################################## # NN SUB CLASS / FUNCTIONS ######################################## def reshape_by_heads(qkv, head_num): # q.shape: (batch, n, head_num*key_dim) : n can be either 1 or PROBLEM_SIZE batch_s = qkv.size(0) n = qkv.size(1) q_reshaped = qkv.reshape(batch_s, n, head_num, -1) # shape: (batch, n, head_num, key_dim) q_transposed = q_reshaped.transpose(1, 2) # shape: (batch, head_num, n, key_dim) return q_transposed def multi_head_attention(q, k, v, rank2_ninf_mask=None, rank3_ninf_mask=None): # q shape: (batch, head_num, n, key_dim) : n can be either 1 or PROBLEM_SIZE # k,v shape: (batch, head_num, problem, key_dim) # rank2_ninf_mask.shape: (batch, problem) # rank3_ninf_mask.shape: (batch, group, problem) batch_s = q.size(0) head_num = q.size(1) n = q.size(2) key_dim = q.size(3) input_s = k.size(2) score = torch.matmul(q, k.transpose(2, 3)) # shape: (batch, head_num, n, problem) score_scaled = score / torch.sqrt(torch.tensor(key_dim, dtype=torch.float)) if rank2_ninf_mask is not None: score_scaled = score_scaled + rank2_ninf_mask[:, None, None, :].expand(batch_s, head_num, n, input_s) if rank3_ninf_mask is not None: score_scaled = score_scaled + rank3_ninf_mask[:, None, :, :].expand(batch_s, head_num, n, input_s) weights = nn.Softmax(dim=3)(score_scaled) # shape: (batch, head_num, n, problem) out = torch.matmul(weights, v) # shape: (batch, head_num, n, key_dim) out_transposed = out.transpose(1, 2) # shape: (batch, n, head_num, key_dim) out_concat = out_transposed.reshape(batch_s, n, head_num * key_dim) # shape: (batch, n, head_num*key_dim) return out_concat class AddAndInstanceNormalization(nn.Module): def __init__(self, **model_params): super().__init__() embedding_dim = model_params['embedding_dim'] self.norm = nn.InstanceNorm1d(embedding_dim, affine=True, track_running_stats=False) def forward(self, input1, input2): # input.shape: (batch, problem, embedding) added = input1 + input2 # shape: (batch, problem, embedding) transposed = added.transpose(1, 2) # shape: (batch, embedding, problem) normalized = self.norm(transposed) # shape: (batch, embedding, problem) back_trans = normalized.transpose(1, 2) # shape: (batch, problem, embedding) return back_trans class AddAndBatchNormalization(nn.Module): def __init__(self, **model_params): super().__init__() embedding_dim = model_params['embedding_dim'] self.norm_by_EMB = nn.BatchNorm1d(embedding_dim, affine=True) # 'Funny' Batch_Norm, as it will normalized by EMB dim def forward(self, input1, input2): # input.shape: (batch, problem, embedding) batch_s = input1.size(0) problem_s = input1.size(1) embedding_dim = input1.size(2) added = input1 + input2 normalized = self.norm_by_EMB(added.reshape(batch_s * problem_s, embedding_dim)) back_trans = normalized.reshape(batch_s, problem_s, embedding_dim) return back_trans class FeedForward(nn.Module): def __init__(self, **model_params): super().__init__() embedding_dim = model_params['embedding_dim'] ff_hidden_dim = model_params['ff_hidden_dim'] self.W1 = nn.Linear(embedding_dim, ff_hidden_dim) self.W2 = nn.Linear(ff_hidden_dim, embedding_dim) def forward(self, input1): # input.shape: (batch, problem, embedding) return self.W2(F.relu(self.W1(input1)))
14,042
36.150794
109
py
POMO
POMO-master/NEW_py_ver/CVRP/POMO/CVRPTrainer.py
import torch from logging import getLogger from CVRPEnv import CVRPEnv as Env from CVRPModel import CVRPModel as Model from torch.optim import Adam as Optimizer from torch.optim.lr_scheduler import MultiStepLR as Scheduler from utils.utils import * class CVRPTrainer: def __init__(self, env_params, model_params, optimizer_params, trainer_params): # save arguments self.env_params = env_params self.model_params = model_params self.optimizer_params = optimizer_params self.trainer_params = trainer_params # result folder, logger self.logger = getLogger(name='trainer') self.result_folder = get_result_folder() self.result_log = LogData() # cuda USE_CUDA = self.trainer_params['use_cuda'] if USE_CUDA: cuda_device_num = self.trainer_params['cuda_device_num'] torch.cuda.set_device(cuda_device_num) device = torch.device('cuda', cuda_device_num) torch.set_default_tensor_type('torch.cuda.FloatTensor') else: device = torch.device('cpu') torch.set_default_tensor_type('torch.FloatTensor') # Main Components self.model = Model(**self.model_params) self.env = Env(**self.env_params) self.optimizer = Optimizer(self.model.parameters(), **self.optimizer_params['optimizer']) self.scheduler = Scheduler(self.optimizer, **self.optimizer_params['scheduler']) # Restore self.start_epoch = 1 model_load = trainer_params['model_load'] if model_load['enable']: checkpoint_fullname = '{path}/checkpoint-{epoch}.pt'.format(**model_load) checkpoint = torch.load(checkpoint_fullname, map_location=device) self.model.load_state_dict(checkpoint['model_state_dict']) self.start_epoch = 1 + model_load['epoch'] self.result_log.set_raw_data(checkpoint['result_log']) self.optimizer.load_state_dict(checkpoint['optimizer_state_dict']) self.scheduler.last_epoch = model_load['epoch']-1 self.logger.info('Saved Model Loaded !!') # utility self.time_estimator = TimeEstimator() def run(self): self.time_estimator.reset(self.start_epoch) for epoch in range(self.start_epoch, self.trainer_params['epochs']+1): self.logger.info('=================================================================') # LR Decay self.scheduler.step() # Train train_score, train_loss = self._train_one_epoch(epoch) self.result_log.append('train_score', epoch, train_score) self.result_log.append('train_loss', epoch, train_loss) ############################ # Logs & Checkpoint ############################ elapsed_time_str, remain_time_str = self.time_estimator.get_est_string(epoch, self.trainer_params['epochs']) self.logger.info("Epoch {:3d}/{:3d}: Time Est.: Elapsed[{}], Remain[{}]".format( epoch, self.trainer_params['epochs'], elapsed_time_str, remain_time_str)) all_done = (epoch == self.trainer_params['epochs']) model_save_interval = self.trainer_params['logging']['model_save_interval'] img_save_interval = self.trainer_params['logging']['img_save_interval'] # Save latest images, every epoch if epoch > 1: self.logger.info("Saving log_image") image_prefix = '{}/latest'.format(self.result_folder) util_save_log_image_with_label(image_prefix, self.trainer_params['logging']['log_image_params_1'], self.result_log, labels=['train_score']) util_save_log_image_with_label(image_prefix, self.trainer_params['logging']['log_image_params_2'], self.result_log, labels=['train_loss']) # Save Model if all_done or (epoch % model_save_interval) == 0: self.logger.info("Saving trained_model") checkpoint_dict = { 'epoch': epoch, 'model_state_dict': self.model.state_dict(), 'optimizer_state_dict': self.optimizer.state_dict(), 'scheduler_state_dict': self.scheduler.state_dict(), 'result_log': self.result_log.get_raw_data() } torch.save(checkpoint_dict, '{}/checkpoint-{}.pt'.format(self.result_folder, epoch)) # Save Image if all_done or (epoch % img_save_interval) == 0: image_prefix = '{}/img/checkpoint-{}'.format(self.result_folder, epoch) util_save_log_image_with_label(image_prefix, self.trainer_params['logging']['log_image_params_1'], self.result_log, labels=['train_score']) util_save_log_image_with_label(image_prefix, self.trainer_params['logging']['log_image_params_2'], self.result_log, labels=['train_loss']) # All-done announcement if all_done: self.logger.info(" *** Training Done *** ") self.logger.info("Now, printing log array...") util_print_log_array(self.logger, self.result_log) def _train_one_epoch(self, epoch): score_AM = AverageMeter() loss_AM = AverageMeter() train_num_episode = self.trainer_params['train_episodes'] episode = 0 loop_cnt = 0 while episode < train_num_episode: remaining = train_num_episode - episode batch_size = min(self.trainer_params['train_batch_size'], remaining) avg_score, avg_loss = self._train_one_batch(batch_size) score_AM.update(avg_score, batch_size) loss_AM.update(avg_loss, batch_size) episode += batch_size # Log First 10 Batch, only at the first epoch if epoch == self.start_epoch: loop_cnt += 1 if loop_cnt <= 10: self.logger.info('Epoch {:3d}: Train {:3d}/{:3d}({:1.1f}%) Score: {:.4f}, Loss: {:.4f}' .format(epoch, episode, train_num_episode, 100. * episode / train_num_episode, score_AM.avg, loss_AM.avg)) # Log Once, for each epoch self.logger.info('Epoch {:3d}: Train ({:3.0f}%) Score: {:.4f}, Loss: {:.4f}' .format(epoch, 100. * episode / train_num_episode, score_AM.avg, loss_AM.avg)) return score_AM.avg, loss_AM.avg def _train_one_batch(self, batch_size): # Prep ############################################### self.model.train() self.env.load_problems(batch_size) reset_state, _, _ = self.env.reset() self.model.pre_forward(reset_state) prob_list = torch.zeros(size=(batch_size, self.env.pomo_size, 0)) # shape: (batch, pomo, 0~problem) # POMO Rollout ############################################### state, reward, done = self.env.pre_step() while not done: selected, prob = self.model(state) # shape: (batch, pomo) state, reward, done = self.env.step(selected) prob_list = torch.cat((prob_list, prob[:, :, None]), dim=2) # Loss ############################################### advantage = reward - reward.float().mean(dim=1, keepdims=True) # shape: (batch, pomo) log_prob = prob_list.log().sum(dim=2) # size = (batch, pomo) loss = -advantage * log_prob # Minus Sign: To Increase REWARD # shape: (batch, pomo) loss_mean = loss.mean() # Score ############################################### max_pomo_reward, _ = reward.max(dim=1) # get best results from pomo score_mean = -max_pomo_reward.float().mean() # negative sign to make positive value # Step & Return ############################################### self.model.zero_grad() loss_mean.backward() self.optimizer.step() return score_mean.item(), loss_mean.item()
8,460
41.094527
120
py
POMO
POMO-master/NEW_py_ver/CVRP/POMO/CVRPTester.py
import torch import os from logging import getLogger from CVRPEnv import CVRPEnv as Env from CVRPModel import CVRPModel as Model from utils.utils import * class CVRPTester: def __init__(self, env_params, model_params, tester_params): # save arguments self.env_params = env_params self.model_params = model_params self.tester_params = tester_params # result folder, logger self.logger = getLogger(name='trainer') self.result_folder = get_result_folder() # cuda USE_CUDA = self.tester_params['use_cuda'] if USE_CUDA: cuda_device_num = self.tester_params['cuda_device_num'] torch.cuda.set_device(cuda_device_num) device = torch.device('cuda', cuda_device_num) torch.set_default_tensor_type('torch.cuda.FloatTensor') else: device = torch.device('cpu') torch.set_default_tensor_type('torch.FloatTensor') self.device = device # ENV and MODEL self.env = Env(**self.env_params) self.model = Model(**self.model_params) # Restore model_load = tester_params['model_load'] checkpoint_fullname = '{path}/checkpoint-{epoch}.pt'.format(**model_load) checkpoint = torch.load(checkpoint_fullname, map_location=device) self.model.load_state_dict(checkpoint['model_state_dict']) # utility self.time_estimator = TimeEstimator() def run(self): self.time_estimator.reset() score_AM = AverageMeter() aug_score_AM = AverageMeter() if self.tester_params['test_data_load']['enable']: self.env.use_saved_problems(self.tester_params['test_data_load']['filename'], self.device) test_num_episode = self.tester_params['test_episodes'] episode = 0 while episode < test_num_episode: remaining = test_num_episode - episode batch_size = min(self.tester_params['test_batch_size'], remaining) score, aug_score = self._test_one_batch(batch_size) score_AM.update(score, batch_size) aug_score_AM.update(aug_score, batch_size) episode += batch_size ############################ # Logs ############################ elapsed_time_str, remain_time_str = self.time_estimator.get_est_string(episode, test_num_episode) self.logger.info("episode {:3d}/{:3d}, Elapsed[{}], Remain[{}], score:{:.3f}, aug_score:{:.3f}".format( episode, test_num_episode, elapsed_time_str, remain_time_str, score, aug_score)) all_done = (episode == test_num_episode) if all_done: self.logger.info(" *** Test Done *** ") self.logger.info(" NO-AUG SCORE: {:.4f} ".format(score_AM.avg)) self.logger.info(" AUGMENTATION SCORE: {:.4f} ".format(aug_score_AM.avg)) def _test_one_batch(self, batch_size): # Augmentation ############################################### if self.tester_params['augmentation_enable']: aug_factor = self.tester_params['aug_factor'] else: aug_factor = 1 # Ready ############################################### self.model.eval() with torch.no_grad(): self.env.load_problems(batch_size, aug_factor) reset_state, _, _ = self.env.reset() self.model.pre_forward(reset_state) # POMO Rollout ############################################### state, reward, done = self.env.pre_step() while not done: selected, _ = self.model(state) # shape: (batch, pomo) state, reward, done = self.env.step(selected) # Return ############################################### aug_reward = reward.reshape(aug_factor, batch_size, self.env.pomo_size) # shape: (augmentation, batch, pomo) max_pomo_reward, _ = aug_reward.max(dim=2) # get best results from pomo # shape: (augmentation, batch) no_aug_score = -max_pomo_reward[0, :].float().mean() # negative sign to make positive value max_aug_pomo_reward, _ = max_pomo_reward.max(dim=0) # get best results from augmentation # shape: (batch,) aug_score = -max_aug_pomo_reward.float().mean() # negative sign to make positive value return no_aug_score.item(), aug_score.item()
4,557
33.793893
115
py
POMO
POMO-master/NEW_py_ver/TSP/TSProblemDef.py
import torch import numpy as np def get_random_problems(batch_size, problem_size): problems = torch.rand(size=(batch_size, problem_size, 2)) # problems.shape: (batch, problem, 2) return problems def augment_xy_data_by_8_fold(problems): # problems.shape: (batch, problem, 2) x = problems[:, :, [0]] y = problems[:, :, [1]] # x,y shape: (batch, problem, 1) dat1 = torch.cat((x, y), dim=2) dat2 = torch.cat((1 - x, y), dim=2) dat3 = torch.cat((x, 1 - y), dim=2) dat4 = torch.cat((1 - x, 1 - y), dim=2) dat5 = torch.cat((y, x), dim=2) dat6 = torch.cat((1 - y, x), dim=2) dat7 = torch.cat((y, 1 - x), dim=2) dat8 = torch.cat((1 - y, 1 - x), dim=2) aug_problems = torch.cat((dat1, dat2, dat3, dat4, dat5, dat6, dat7, dat8), dim=0) # shape: (8*batch, problem, 2) return aug_problems
856
26.645161
85
py
POMO
POMO-master/NEW_py_ver/TSP/POMO/TSPEnv.py
from dataclasses import dataclass import torch from TSProblemDef import get_random_problems, augment_xy_data_by_8_fold @dataclass class Reset_State: problems: torch.Tensor # shape: (batch, problem, 2) @dataclass class Step_State: BATCH_IDX: torch.Tensor POMO_IDX: torch.Tensor # shape: (batch, pomo) current_node: torch.Tensor = None # shape: (batch, pomo) ninf_mask: torch.Tensor = None # shape: (batch, pomo, node) class TSPEnv: def __init__(self, **env_params): # Const @INIT #################################### self.env_params = env_params self.problem_size = env_params['problem_size'] self.pomo_size = env_params['pomo_size'] # Const @Load_Problem #################################### self.batch_size = None self.BATCH_IDX = None self.POMO_IDX = None # IDX.shape: (batch, pomo) self.problems = None # shape: (batch, node, node) # Dynamic #################################### self.selected_count = None self.current_node = None # shape: (batch, pomo) self.selected_node_list = None # shape: (batch, pomo, 0~problem) def load_problems(self, batch_size, aug_factor=1): self.batch_size = batch_size self.problems = get_random_problems(batch_size, self.problem_size) # problems.shape: (batch, problem, 2) if aug_factor > 1: if aug_factor == 8: self.batch_size = self.batch_size * 8 self.problems = augment_xy_data_by_8_fold(self.problems) # shape: (8*batch, problem, 2) else: raise NotImplementedError self.BATCH_IDX = torch.arange(self.batch_size)[:, None].expand(self.batch_size, self.pomo_size) self.POMO_IDX = torch.arange(self.pomo_size)[None, :].expand(self.batch_size, self.pomo_size) def reset(self): self.selected_count = 0 self.current_node = None # shape: (batch, pomo) self.selected_node_list = torch.zeros((self.batch_size, self.pomo_size, 0), dtype=torch.long) # shape: (batch, pomo, 0~problem) # CREATE STEP STATE self.step_state = Step_State(BATCH_IDX=self.BATCH_IDX, POMO_IDX=self.POMO_IDX) self.step_state.ninf_mask = torch.zeros((self.batch_size, self.pomo_size, self.problem_size)) # shape: (batch, pomo, problem) reward = None done = False return Reset_State(self.problems), reward, done def pre_step(self): reward = None done = False return self.step_state, reward, done def step(self, selected): # selected.shape: (batch, pomo) self.selected_count += 1 self.current_node = selected # shape: (batch, pomo) self.selected_node_list = torch.cat((self.selected_node_list, self.current_node[:, :, None]), dim=2) # shape: (batch, pomo, 0~problem) # UPDATE STEP STATE self.step_state.current_node = self.current_node # shape: (batch, pomo) self.step_state.ninf_mask[self.BATCH_IDX, self.POMO_IDX, self.current_node] = float('-inf') # shape: (batch, pomo, node) # returning values done = (self.selected_count == self.problem_size) if done: reward = -self._get_travel_distance() # note the minus sign! else: reward = None return self.step_state, reward, done def _get_travel_distance(self): gathering_index = self.selected_node_list.unsqueeze(3).expand(self.batch_size, -1, self.problem_size, 2) # shape: (batch, pomo, problem, 2) seq_expanded = self.problems[:, None, :, :].expand(self.batch_size, self.pomo_size, self.problem_size, 2) ordered_seq = seq_expanded.gather(dim=2, index=gathering_index) # shape: (batch, pomo, problem, 2) rolled_seq = ordered_seq.roll(dims=2, shifts=-1) segment_lengths = ((ordered_seq-rolled_seq)**2).sum(3).sqrt() # shape: (batch, pomo, problem) travel_distances = segment_lengths.sum(2) # shape: (batch, pomo) return travel_distances
4,228
32.039063
113
py
POMO
POMO-master/NEW_py_ver/TSP/POMO/TSPModel.py
import torch import torch.nn as nn import torch.nn.functional as F class TSPModel(nn.Module): def __init__(self, **model_params): super().__init__() self.model_params = model_params self.encoder = TSP_Encoder(**model_params) self.decoder = TSP_Decoder(**model_params) self.encoded_nodes = None # shape: (batch, problem, EMBEDDING_DIM) def pre_forward(self, reset_state): self.encoded_nodes = self.encoder(reset_state.problems) # shape: (batch, problem, EMBEDDING_DIM) self.decoder.set_kv(self.encoded_nodes) def forward(self, state): batch_size = state.BATCH_IDX.size(0) pomo_size = state.BATCH_IDX.size(1) if state.current_node is None: selected = torch.arange(pomo_size)[None, :].expand(batch_size, pomo_size) prob = torch.ones(size=(batch_size, pomo_size)) encoded_first_node = _get_encoding(self.encoded_nodes, selected) # shape: (batch, pomo, embedding) self.decoder.set_q1(encoded_first_node) else: encoded_last_node = _get_encoding(self.encoded_nodes, state.current_node) # shape: (batch, pomo, embedding) probs = self.decoder(encoded_last_node, ninf_mask=state.ninf_mask) # shape: (batch, pomo, problem) if self.training or self.model_params['eval_type'] == 'softmax': while True: selected = probs.reshape(batch_size * pomo_size, -1).multinomial(1) \ .squeeze(dim=1).reshape(batch_size, pomo_size) # shape: (batch, pomo) prob = probs[state.BATCH_IDX, state.POMO_IDX, selected] \ .reshape(batch_size, pomo_size) # shape: (batch, pomo) if (prob != 0).all(): break else: selected = probs.argmax(dim=2) # shape: (batch, pomo) prob = None return selected, prob def _get_encoding(encoded_nodes, node_index_to_pick): # encoded_nodes.shape: (batch, problem, embedding) # node_index_to_pick.shape: (batch, pomo) batch_size = node_index_to_pick.size(0) pomo_size = node_index_to_pick.size(1) embedding_dim = encoded_nodes.size(2) gathering_index = node_index_to_pick[:, :, None].expand(batch_size, pomo_size, embedding_dim) # shape: (batch, pomo, embedding) picked_nodes = encoded_nodes.gather(dim=1, index=gathering_index) # shape: (batch, pomo, embedding) return picked_nodes ######################################## # ENCODER ######################################## class TSP_Encoder(nn.Module): def __init__(self, **model_params): super().__init__() self.model_params = model_params embedding_dim = self.model_params['embedding_dim'] encoder_layer_num = self.model_params['encoder_layer_num'] self.embedding = nn.Linear(2, embedding_dim) self.layers = nn.ModuleList([EncoderLayer(**model_params) for _ in range(encoder_layer_num)]) def forward(self, data): # data.shape: (batch, problem, 2) embedded_input = self.embedding(data) # shape: (batch, problem, embedding) out = embedded_input for layer in self.layers: out = layer(out) return out class EncoderLayer(nn.Module): def __init__(self, **model_params): super().__init__() self.model_params = model_params embedding_dim = self.model_params['embedding_dim'] head_num = self.model_params['head_num'] qkv_dim = self.model_params['qkv_dim'] self.Wq = nn.Linear(embedding_dim, head_num * qkv_dim, bias=False) self.Wk = nn.Linear(embedding_dim, head_num * qkv_dim, bias=False) self.Wv = nn.Linear(embedding_dim, head_num * qkv_dim, bias=False) self.multi_head_combine = nn.Linear(head_num * qkv_dim, embedding_dim) self.addAndNormalization1 = Add_And_Normalization_Module(**model_params) self.feedForward = Feed_Forward_Module(**model_params) self.addAndNormalization2 = Add_And_Normalization_Module(**model_params) def forward(self, input1): # input.shape: (batch, problem, EMBEDDING_DIM) head_num = self.model_params['head_num'] q = reshape_by_heads(self.Wq(input1), head_num=head_num) k = reshape_by_heads(self.Wk(input1), head_num=head_num) v = reshape_by_heads(self.Wv(input1), head_num=head_num) # q shape: (batch, HEAD_NUM, problem, KEY_DIM) out_concat = multi_head_attention(q, k, v) # shape: (batch, problem, HEAD_NUM*KEY_DIM) multi_head_out = self.multi_head_combine(out_concat) # shape: (batch, problem, EMBEDDING_DIM) out1 = self.addAndNormalization1(input1, multi_head_out) out2 = self.feedForward(out1) out3 = self.addAndNormalization2(out1, out2) return out3 # shape: (batch, problem, EMBEDDING_DIM) ######################################## # DECODER ######################################## class TSP_Decoder(nn.Module): def __init__(self, **model_params): super().__init__() self.model_params = model_params embedding_dim = self.model_params['embedding_dim'] head_num = self.model_params['head_num'] qkv_dim = self.model_params['qkv_dim'] self.Wq_first = nn.Linear(embedding_dim, head_num * qkv_dim, bias=False) self.Wq_last = nn.Linear(embedding_dim, head_num * qkv_dim, bias=False) self.Wk = nn.Linear(embedding_dim, head_num * qkv_dim, bias=False) self.Wv = nn.Linear(embedding_dim, head_num * qkv_dim, bias=False) self.multi_head_combine = nn.Linear(head_num * qkv_dim, embedding_dim) self.k = None # saved key, for multi-head attention self.v = None # saved value, for multi-head_attention self.single_head_key = None # saved, for single-head attention self.q_first = None # saved q1, for multi-head attention def set_kv(self, encoded_nodes): # encoded_nodes.shape: (batch, problem, embedding) head_num = self.model_params['head_num'] self.k = reshape_by_heads(self.Wk(encoded_nodes), head_num=head_num) self.v = reshape_by_heads(self.Wv(encoded_nodes), head_num=head_num) # shape: (batch, head_num, pomo, qkv_dim) self.single_head_key = encoded_nodes.transpose(1, 2) # shape: (batch, embedding, problem) def set_q1(self, encoded_q1): # encoded_q.shape: (batch, n, embedding) # n can be 1 or pomo head_num = self.model_params['head_num'] self.q_first = reshape_by_heads(self.Wq_first(encoded_q1), head_num=head_num) # shape: (batch, head_num, n, qkv_dim) def forward(self, encoded_last_node, ninf_mask): # encoded_last_node.shape: (batch, pomo, embedding) # ninf_mask.shape: (batch, pomo, problem) head_num = self.model_params['head_num'] # Multi-Head Attention ####################################################### q_last = reshape_by_heads(self.Wq_last(encoded_last_node), head_num=head_num) # shape: (batch, head_num, pomo, qkv_dim) q = self.q_first + q_last # shape: (batch, head_num, pomo, qkv_dim) out_concat = multi_head_attention(q, self.k, self.v, rank3_ninf_mask=ninf_mask) # shape: (batch, pomo, head_num*qkv_dim) mh_atten_out = self.multi_head_combine(out_concat) # shape: (batch, pomo, embedding) # Single-Head Attention, for probability calculation ####################################################### score = torch.matmul(mh_atten_out, self.single_head_key) # shape: (batch, pomo, problem) sqrt_embedding_dim = self.model_params['sqrt_embedding_dim'] logit_clipping = self.model_params['logit_clipping'] score_scaled = score / sqrt_embedding_dim # shape: (batch, pomo, problem) score_clipped = logit_clipping * torch.tanh(score_scaled) score_masked = score_clipped + ninf_mask probs = F.softmax(score_masked, dim=2) # shape: (batch, pomo, problem) return probs ######################################## # NN SUB CLASS / FUNCTIONS ######################################## def reshape_by_heads(qkv, head_num): # q.shape: (batch, n, head_num*key_dim) : n can be either 1 or PROBLEM_SIZE batch_s = qkv.size(0) n = qkv.size(1) q_reshaped = qkv.reshape(batch_s, n, head_num, -1) # shape: (batch, n, head_num, key_dim) q_transposed = q_reshaped.transpose(1, 2) # shape: (batch, head_num, n, key_dim) return q_transposed def multi_head_attention(q, k, v, rank2_ninf_mask=None, rank3_ninf_mask=None): # q shape: (batch, head_num, n, key_dim) : n can be either 1 or PROBLEM_SIZE # k,v shape: (batch, head_num, problem, key_dim) # rank2_ninf_mask.shape: (batch, problem) # rank3_ninf_mask.shape: (batch, group, problem) batch_s = q.size(0) head_num = q.size(1) n = q.size(2) key_dim = q.size(3) input_s = k.size(2) score = torch.matmul(q, k.transpose(2, 3)) # shape: (batch, head_num, n, problem) score_scaled = score / torch.sqrt(torch.tensor(key_dim, dtype=torch.float)) if rank2_ninf_mask is not None: score_scaled = score_scaled + rank2_ninf_mask[:, None, None, :].expand(batch_s, head_num, n, input_s) if rank3_ninf_mask is not None: score_scaled = score_scaled + rank3_ninf_mask[:, None, :, :].expand(batch_s, head_num, n, input_s) weights = nn.Softmax(dim=3)(score_scaled) # shape: (batch, head_num, n, problem) out = torch.matmul(weights, v) # shape: (batch, head_num, n, key_dim) out_transposed = out.transpose(1, 2) # shape: (batch, n, head_num, key_dim) out_concat = out_transposed.reshape(batch_s, n, head_num * key_dim) # shape: (batch, n, head_num*key_dim) return out_concat class Add_And_Normalization_Module(nn.Module): def __init__(self, **model_params): super().__init__() embedding_dim = model_params['embedding_dim'] self.norm = nn.InstanceNorm1d(embedding_dim, affine=True, track_running_stats=False) def forward(self, input1, input2): # input.shape: (batch, problem, embedding) added = input1 + input2 # shape: (batch, problem, embedding) transposed = added.transpose(1, 2) # shape: (batch, embedding, problem) normalized = self.norm(transposed) # shape: (batch, embedding, problem) back_trans = normalized.transpose(1, 2) # shape: (batch, problem, embedding) return back_trans class Feed_Forward_Module(nn.Module): def __init__(self, **model_params): super().__init__() embedding_dim = model_params['embedding_dim'] ff_hidden_dim = model_params['ff_hidden_dim'] self.W1 = nn.Linear(embedding_dim, ff_hidden_dim) self.W2 = nn.Linear(ff_hidden_dim, embedding_dim) def forward(self, input1): # input.shape: (batch, problem, embedding) return self.W2(F.relu(self.W1(input1)))
11,293
34.074534
109
py
POMO
POMO-master/NEW_py_ver/TSP/POMO/TSPTester.py
import torch import os from logging import getLogger from TSPEnv import TSPEnv as Env from TSPModel import TSPModel as Model from utils.utils import * class TSPTester: def __init__(self, env_params, model_params, tester_params): # save arguments self.env_params = env_params self.model_params = model_params self.tester_params = tester_params # result folder, logger self.logger = getLogger(name='trainer') self.result_folder = get_result_folder() # cuda USE_CUDA = self.tester_params['use_cuda'] if USE_CUDA: cuda_device_num = self.tester_params['cuda_device_num'] torch.cuda.set_device(cuda_device_num) device = torch.device('cuda', cuda_device_num) torch.set_default_tensor_type('torch.cuda.FloatTensor') else: device = torch.device('cpu') torch.set_default_tensor_type('torch.FloatTensor') self.device = device # ENV and MODEL self.env = Env(**self.env_params) self.model = Model(**self.model_params) # Restore model_load = tester_params['model_load'] checkpoint_fullname = '{path}/checkpoint-{epoch}.pt'.format(**model_load) checkpoint = torch.load(checkpoint_fullname, map_location=device) self.model.load_state_dict(checkpoint['model_state_dict']) # utility self.time_estimator = TimeEstimator() def run(self): self.time_estimator.reset() score_AM = AverageMeter() aug_score_AM = AverageMeter() test_num_episode = self.tester_params['test_episodes'] episode = 0 while episode < test_num_episode: remaining = test_num_episode - episode batch_size = min(self.tester_params['test_batch_size'], remaining) score, aug_score = self._test_one_batch(batch_size) score_AM.update(score, batch_size) aug_score_AM.update(aug_score, batch_size) episode += batch_size ############################ # Logs ############################ elapsed_time_str, remain_time_str = self.time_estimator.get_est_string(episode, test_num_episode) self.logger.info("episode {:3d}/{:3d}, Elapsed[{}], Remain[{}], score:{:.3f}, aug_score:{:.3f}".format( episode, test_num_episode, elapsed_time_str, remain_time_str, score, aug_score)) all_done = (episode == test_num_episode) if all_done: self.logger.info(" *** Test Done *** ") self.logger.info(" NO-AUG SCORE: {:.4f} ".format(score_AM.avg)) self.logger.info(" AUGMENTATION SCORE: {:.4f} ".format(aug_score_AM.avg)) def _test_one_batch(self, batch_size): # Augmentation ############################################### if self.tester_params['augmentation_enable']: aug_factor = self.tester_params['aug_factor'] else: aug_factor = 1 # Ready ############################################### self.model.eval() with torch.no_grad(): self.env.load_problems(batch_size, aug_factor) reset_state, _, _ = self.env.reset() self.model.pre_forward(reset_state) # POMO Rollout ############################################### state, reward, done = self.env.pre_step() while not done: selected, _ = self.model(state) # shape: (batch, pomo) state, reward, done = self.env.step(selected) # Return ############################################### aug_reward = reward.reshape(aug_factor, batch_size, self.env.pomo_size) # shape: (augmentation, batch, pomo) max_pomo_reward, _ = aug_reward.max(dim=2) # get best results from pomo # shape: (augmentation, batch) no_aug_score = -max_pomo_reward[0, :].float().mean() # negative sign to make positive value max_aug_pomo_reward, _ = max_pomo_reward.max(dim=0) # get best results from augmentation # shape: (batch,) aug_score = -max_aug_pomo_reward.float().mean() # negative sign to make positive value return no_aug_score.item(), aug_score.item()
4,389
33.296875
115
py
POMO
POMO-master/NEW_py_ver/TSP/POMO/TSPTrainer.py
import torch from logging import getLogger from TSPEnv import TSPEnv as Env from TSPModel import TSPModel as Model from torch.optim import Adam as Optimizer from torch.optim.lr_scheduler import MultiStepLR as Scheduler from utils.utils import * class TSPTrainer: def __init__(self, env_params, model_params, optimizer_params, trainer_params): # save arguments self.env_params = env_params self.model_params = model_params self.optimizer_params = optimizer_params self.trainer_params = trainer_params # result folder, logger self.logger = getLogger(name='trainer') self.result_folder = get_result_folder() self.result_log = LogData() # cuda USE_CUDA = self.trainer_params['use_cuda'] if USE_CUDA: cuda_device_num = self.trainer_params['cuda_device_num'] torch.cuda.set_device(cuda_device_num) device = torch.device('cuda', cuda_device_num) torch.set_default_tensor_type('torch.cuda.FloatTensor') else: device = torch.device('cpu') torch.set_default_tensor_type('torch.FloatTensor') # Main Components self.model = Model(**self.model_params) self.env = Env(**self.env_params) self.optimizer = Optimizer(self.model.parameters(), **self.optimizer_params['optimizer']) self.scheduler = Scheduler(self.optimizer, **self.optimizer_params['scheduler']) # Restore self.start_epoch = 1 model_load = trainer_params['model_load'] if model_load['enable']: checkpoint_fullname = '{path}/checkpoint-{epoch}.pt'.format(**model_load) checkpoint = torch.load(checkpoint_fullname, map_location=device) self.model.load_state_dict(checkpoint['model_state_dict']) self.start_epoch = 1 + model_load['epoch'] self.result_log.set_raw_data(checkpoint['result_log']) self.optimizer.load_state_dict(checkpoint['optimizer_state_dict']) self.scheduler.last_epoch = model_load['epoch']-1 self.logger.info('Saved Model Loaded !!') # utility self.time_estimator = TimeEstimator() def run(self): self.time_estimator.reset(self.start_epoch) for epoch in range(self.start_epoch, self.trainer_params['epochs']+1): self.logger.info('=================================================================') # LR Decay self.scheduler.step() # Train train_score, train_loss = self._train_one_epoch(epoch) self.result_log.append('train_score', epoch, train_score) self.result_log.append('train_loss', epoch, train_loss) ############################ # Logs & Checkpoint ############################ elapsed_time_str, remain_time_str = self.time_estimator.get_est_string(epoch, self.trainer_params['epochs']) self.logger.info("Epoch {:3d}/{:3d}: Time Est.: Elapsed[{}], Remain[{}]".format( epoch, self.trainer_params['epochs'], elapsed_time_str, remain_time_str)) all_done = (epoch == self.trainer_params['epochs']) model_save_interval = self.trainer_params['logging']['model_save_interval'] img_save_interval = self.trainer_params['logging']['img_save_interval'] if epoch > 1: # save latest images, every epoch self.logger.info("Saving log_image") image_prefix = '{}/latest'.format(self.result_folder) util_save_log_image_with_label(image_prefix, self.trainer_params['logging']['log_image_params_1'], self.result_log, labels=['train_score']) util_save_log_image_with_label(image_prefix, self.trainer_params['logging']['log_image_params_2'], self.result_log, labels=['train_loss']) if all_done or (epoch % model_save_interval) == 0: self.logger.info("Saving trained_model") checkpoint_dict = { 'epoch': epoch, 'model_state_dict': self.model.state_dict(), 'optimizer_state_dict': self.optimizer.state_dict(), 'scheduler_state_dict': self.scheduler.state_dict(), 'result_log': self.result_log.get_raw_data() } torch.save(checkpoint_dict, '{}/checkpoint-{}.pt'.format(self.result_folder, epoch)) if all_done or (epoch % img_save_interval) == 0: image_prefix = '{}/img/checkpoint-{}'.format(self.result_folder, epoch) util_save_log_image_with_label(image_prefix, self.trainer_params['logging']['log_image_params_1'], self.result_log, labels=['train_score']) util_save_log_image_with_label(image_prefix, self.trainer_params['logging']['log_image_params_2'], self.result_log, labels=['train_loss']) if all_done: self.logger.info(" *** Training Done *** ") self.logger.info("Now, printing log array...") util_print_log_array(self.logger, self.result_log) def _train_one_epoch(self, epoch): score_AM = AverageMeter() loss_AM = AverageMeter() train_num_episode = self.trainer_params['train_episodes'] episode = 0 loop_cnt = 0 while episode < train_num_episode: remaining = train_num_episode - episode batch_size = min(self.trainer_params['train_batch_size'], remaining) avg_score, avg_loss = self._train_one_batch(batch_size) score_AM.update(avg_score, batch_size) loss_AM.update(avg_loss, batch_size) episode += batch_size # Log First 10 Batch, only at the first epoch if epoch == self.start_epoch: loop_cnt += 1 if loop_cnt <= 10: self.logger.info('Epoch {:3d}: Train {:3d}/{:3d}({:1.1f}%) Score: {:.4f}, Loss: {:.4f}' .format(epoch, episode, train_num_episode, 100. * episode / train_num_episode, score_AM.avg, loss_AM.avg)) # Log Once, for each epoch self.logger.info('Epoch {:3d}: Train ({:3.0f}%) Score: {:.4f}, Loss: {:.4f}' .format(epoch, 100. * episode / train_num_episode, score_AM.avg, loss_AM.avg)) return score_AM.avg, loss_AM.avg def _train_one_batch(self, batch_size): # Prep ############################################### self.model.train() self.env.load_problems(batch_size) reset_state, _, _ = self.env.reset() self.model.pre_forward(reset_state) prob_list = torch.zeros(size=(batch_size, self.env.pomo_size, 0)) # shape: (batch, pomo, 0~problem) # POMO Rollout ############################################### state, reward, done = self.env.pre_step() while not done: selected, prob = self.model(state) # shape: (batch, pomo) state, reward, done = self.env.step(selected) prob_list = torch.cat((prob_list, prob[:, :, None]), dim=2) # Loss ############################################### advantage = reward - reward.float().mean(dim=1, keepdims=True) # shape: (batch, pomo) log_prob = prob_list.log().sum(dim=2) # size = (batch, pomo) loss = -advantage * log_prob # Minus Sign: To Increase REWARD # shape: (batch, pomo) loss_mean = loss.mean() # Score ############################################### max_pomo_reward, _ = reward.max(dim=1) # get best results from pomo score_mean = -max_pomo_reward.float().mean() # negative sign to make positive value # Step & Return ############################################### self.model.zero_grad() loss_mean.backward() self.optimizer.step() return score_mean.item(), loss_mean.item()
8,357
41.642857
120
py
exoplanet-atlas
exoplanet-atlas-main/docs/conf.py
#!/usr/bin/env python3 # -*- coding: utf-8 -*- # # exoplanet-atlas documentation build configuration file, created by # sphinx-quickstart on Sun Nov 17 17:31:38 2019. # # This file is execfile()d with the current directory set to its # containing dir. # # Note that not all possible configuration values are present in this # autogenerated file. # # All configuration values have a default; values that are commented out # serve to show the default. # 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 # sys.path.insert(0, os.path.abspath('.')) # -- 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 = [ "sphinx.ext.autodoc", "sphinx.ext.viewcode", "sphinx.ext.githubpages", "sphinx.ext.mathjax", "nbsphinx", ] html_sidebars = { "**": ["globaltoc.html", "relations.html", "sourcelink.html", "searchbox.html"] } # Add any paths that contain templates here, relative to this directory. templates_path = ["_templates"] # The suffix(es) of source filenames. # You can specify multiple suffix as a list of string: # # source_suffix = ['.rst', '.md'] source_suffix = ".rst" # The master toctree document. master_doc = "index" # General information about the project. project = "exoplanet-atlas" copyright = "2019, Zach Berta-Thompson" author = "Zach Berta-Thompson" # The version info for the project you're documenting, acts as replacement for # |version| and |release|, also used in various other places throughout the # built documents. # # The short X.Y version. # import exoatlas version = "0.2.6" # exoatlas.__version__ # The full version, including alpha/beta/rc tags. release = "0.2.6" # exoatlas.__version__ # 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 patterns also effect to html_static_path and html_extra_path exclude_patterns = [] # The name of the Pygments (syntax highlighting) style to use. pygments_style = "sphinx" # If true, `todo` and `todoList` produce output, else they produce nothing. todo_include_todos = False # -- 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 = "alabaster" # 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_theme_options = {} # 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"] # -- Options for HTMLHelp output ------------------------------------------ # Output file base name for HTML help builder. htmlhelp_basename = "exoplanet-atlasdoc" # -- 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, "exoplanet-atlas.tex", "exoplanet-atlas Documentation", "Zach Berta-Thompson", "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, "exoplanet-atlas", "exoplanet-atlas 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, "exoplanet-atlas", "exoplanet-atlas Documentation", author, "exoplanet-atlas", "One line description of project.", "Miscellaneous", ), ] # Example configuration for intersphinx: refer to the Python standard library. intersphinx_mapping = {"https://docs.python.org/": None}
5,305
29.147727
83
py
animeGAN
animeGAN-master/main.py
from __future__ import print_function import os import time import random import argparse import torch import torch.nn as nn import torch.nn.parallel import torch.backends.cudnn as cudnn import torch.optim as optim import torch.utils.data import torchvision.datasets as dset import torchvision.transforms as transforms import torchvision.utils as vutils from torch.autograd import Variable ### load project files import models from models import weights_init parser = argparse.ArgumentParser() parser.add_argument('--dataRoot', required=True, help='path to dataset') parser.add_argument('--workers', type=int, default=2, help='number of data loading workers') parser.add_argument('--batchSize', type=int, default=64, help='input batch size') parser.add_argument('--imageSize', type=int, default=64, help='the height / width of the input image to network') parser.add_argument('--nz', type=int, default=100, help='size of the latent z vector') parser.add_argument('--ngf', type=int, default=64) parser.add_argument('--ndf', type=int, default=64) parser.add_argument('--niter', type=int, default=25, help='number of epochs to train for') parser.add_argument('--lr', type=float, default=0.0002, help='learning rate, default=0.0002') parser.add_argument('--beta1', type=float, default=0.5, help='beta1 for adam. default=0.5') parser.add_argument('--cuda' , action='store_true', help='enables cuda') parser.add_argument('--ngpu' , type=int, default=1, help='number of GPUs to use') parser.add_argument('--netG', default='', help="path to netG (to continue training)") parser.add_argument('--netD', default='', help="path to netD (to continue training)") parser.add_argument('--outDir', default='.', help='folder to output images and model checkpoints') parser.add_argument('--model', type=int, default=1, help='1 for dcgan, 2 for illustrationGAN-like-GAN') parser.add_argument('--d_labelSmooth', type=float, default=0, help='for D, use soft label "1-labelSmooth" for real samples') parser.add_argument('--n_extra_layers_d', type=int, default=0, help='number of extra conv layers in D') parser.add_argument('--n_extra_layers_g', type=int, default=1, help='number of extra conv layers in G') parser.add_argument('--binary', action='store_true', help='z from bernoulli distribution, with prob=0.5') # simply prefer this way # arg_list = [ # '--dataRoot', '/home/jielei/data/danbooru-faces', # '--workers', '12', # '--batchSize', '128', # '--imageSize', '64', # '--nz', '100', # '--ngf', '64', # '--ndf', '64', # '--niter', '80', # '--lr', '0.0002', # '--beta1', '0.5', # '--cuda', # '--ngpu', '1', # '--netG', '', # '--netD', '', # '--outDir', './results', # '--model', '1', # '--d_labelSmooth', '0.1', # 0.25 from imporved-GAN paper # '--n_extra_layers_d', '0', # '--n_extra_layers_g', '1', # in the sense that generator should be more powerful # ] args = parser.parse_args() # opt = parser.parse_args(arg_list) print(opt) try: os.makedirs(opt.outDir) except OSError: pass opt.manualSeed = random.randint(1,10000) # fix seed, a scalar random.seed(opt.manualSeed) torch.manual_seed(opt.manualSeed) cudnn.benchmark = True if torch.cuda.is_available() and not opt.cuda: print("WARNING: You have a CUDA device, so you should probably run with --cuda") nc = 3 ngpu = opt.ngpu nz = opt.nz ngf = opt.ngf ndf = opt.ndf n_extra_d = opt.n_extra_layers_d n_extra_g = opt.n_extra_layers_g dataset = dset.ImageFolder( root=opt.dataRoot, transform=transforms.Compose([ transforms.Scale(opt.imageSize), # transforms.CenterCrop(opt.imageSize), transforms.ToTensor(), transforms.Normalize((0.5,0.5,0.5), (0.5,0.5,0.5)), # bring images to (-1,1) ]) ) dataloader = torch.utils.data.DataLoader(dataset, batch_size=opt.batchSize, shuffle=True, num_workers=opt.workers) # load models if opt.model == 1: netG = models._netG_1(ngpu, nz, nc, ngf, n_extra_g) netD = models._netD_1(ngpu, nz, nc, ndf, n_extra_d) elif opt.model == 2: netG = models._netG_2(ngpu, nz, nc, ngf) netD = models._netD_2(ngpu, nz, nc, ndf) netG.apply(weights_init) if opt.netG != '': netG.load_state_dict(torch.load(opt.netG)) print(netG) netD.apply(weights_init) if opt.netD != '': netD.load_state_dict(torch.load(opt.netD)) print(netD) criterion = nn.BCELoss() criterion_MSE = nn.MSELoss() input = torch.FloatTensor(opt.batchSize, 3, opt.imageSize, opt.imageSize) noise = torch.FloatTensor(opt.batchSize, nz, 1, 1) if opt.binary: bernoulli_prob = torch.FloatTensor(opt.batchSize, nz, 1, 1).fill_(0.5) fixed_noise = torch.bernoulli(bernoulli_prob) else: fixed_noise = torch.FloatTensor(opt.batchSize, nz, 1, 1).normal_(0, 1) label = torch.FloatTensor(opt.batchSize) real_label = 1 fake_label = 0 if opt.cuda: netD.cuda() netG.cuda() criterion.cuda() criterion_MSE.cuda() input, label = input.cuda(), label.cuda() noise, fixed_noise = noise.cuda(), fixed_noise.cuda() input = Variable(input) label = Variable(label) noise = Variable(noise) fixed_noise = Variable(fixed_noise) # setup optimizer optimizerD = optim.Adam(netD.parameters(), lr = opt.lr, betas = (opt.beta1, 0.999)) optimizerG = optim.Adam(netG.parameters(), lr = opt.lr, betas = (opt.beta1, 0.999)) for epoch in range(opt.niter): for i, data in enumerate(dataloader, 0): start_iter = time.time() ############################ # (1) Update D network: maximize log(D(x)) + log(1 - D(G(z))) ########################### # train with real netD.zero_grad() real_cpu, _ = data batch_size = real_cpu.size(0) input.data.resize_(real_cpu.size()).copy_(real_cpu) label.data.resize_(batch_size).fill_(real_label - opt.d_labelSmooth) # use smooth label for discriminator output = netD(input) errD_real = criterion(output, label) errD_real.backward() D_x = output.data.mean() # train with fake noise.data.resize_(batch_size, nz, 1, 1) if opt.binary: bernoulli_prob.resize_(noise.data.size()) noise.data.copy_(2*(torch.bernoulli(bernoulli_prob)-0.5)) else: noise.data.normal_(0, 1) fake,z_prediction = netG(noise) label.data.fill_(fake_label) output = netD(fake.detach()) # add ".detach()" to avoid backprop through G errD_fake = criterion(output, label) errD_fake.backward() # gradients for fake/real will be accumulated D_G_z1 = output.data.mean() errD = errD_real + errD_fake optimizerD.step() # .step() can be called once the gradients are computed ############################ # (2) Update G network: maximize log(D(G(z))) ########################### netG.zero_grad() label.data.fill_(real_label) # fake labels are real for generator cost output = netD(fake) errG = criterion(output, label) errG.backward(retain_variables=True) # True if backward through the graph for the second time if opt.model == 2: # with z predictor errG_z = criterion_MSE(z_prediction, noise) errG_z.backward() D_G_z2 = output.data.mean() optimizerG.step() end_iter = time.time() print('[%d/%d][%d/%d] Loss_D: %.4f Loss_G: %.4f D(x): %.4f D(G(z)): %.4f / %.4f Elapsed %.2f s' % (epoch, opt.niter, i, len(dataloader), errD.data[0], errG.data[0], D_x, D_G_z1, D_G_z2, end_iter-start_iter)) if i % 100 == 0: # the first 64 samples from the mini-batch are saved. vutils.save_image(real_cpu[0:64,:,:,:], '%s/real_samples.png' % opt.outDir, nrow=8) fake,_ = netG(fixed_noise) vutils.save_image(fake.data[0:64,:,:,:], '%s/fake_samples_epoch_%03d.png' % (opt.outDir, epoch), nrow=8) if epoch % 1 == 0: # do checkpointing torch.save(netG.state_dict(), '%s/netG_epoch_%d.pth' % (opt.outDir, epoch)) torch.save(netD.state_dict(), '%s/netD_epoch_%d.pth' % (opt.outDir, epoch))
8,259
37.418605
124
py
animeGAN
animeGAN-master/models.py
import torch import torch.nn as nn import torch.nn.parallel def weights_init(m): classname = m.__class__.__name__ if classname.find('Conv') != -1: m.weight.data.normal_(0.0, 0.02) elif classname.find('BatchNorm') != -1: m.weight.data.normal_(1.0, 0.02) m.bias.data.fill_(0) # DCGAN model, fully convolutional architecture class _netG_1(nn.Module): def __init__(self, ngpu, nz, nc , ngf, n_extra_layers_g): super(_netG_1, self).__init__() self.ngpu = ngpu #self.nz = nz #self.nc = nc #self.ngf = ngf main = nn.Sequential( # input is Z, going into a convolution # state size. nz x 1 x 1 nn.ConvTranspose2d( nz, ngf * 8, 4, 1, 0, bias=False), nn.BatchNorm2d(ngf * 8), nn.LeakyReLU(0.2, inplace=True), # state size. (ngf*8) x 4 x 4 nn.ConvTranspose2d(ngf * 8, ngf * 4, 4, 2, 1, bias=False), nn.BatchNorm2d(ngf * 4), nn.LeakyReLU(0.2, inplace=True), # state size. (ngf*4) x 8 x 8 nn.ConvTranspose2d(ngf * 4, ngf * 2, 4, 2, 1, bias=False), nn.BatchNorm2d(ngf * 2), nn.LeakyReLU(0.2, inplace=True), # state size. (ngf*2) x 16 x 16 nn.ConvTranspose2d(ngf * 2, ngf, 4, 2, 1, bias=False), nn.BatchNorm2d(ngf), nn.LeakyReLU(0.2, inplace=True), # state size. (ngf) x 32 x 32 ) # Extra layers for t in range(n_extra_layers_g): main.add_module('extra-layers-{0}.{1}.conv'.format(t, ngf), nn.Conv2d(ngf, ngf, 3, 1, 1, bias=False)) main.add_module('extra-layers-{0}.{1}.batchnorm'.format(t, ngf), nn.BatchNorm2d(ngf)) main.add_module('extra-layers-{0}.{1}.relu'.format(t, ngf), nn.LeakyReLU(0.2, inplace=True)) main.add_module('final_layer.deconv', nn.ConvTranspose2d(ngf, nc, 4, 2, 1, bias=False)) # 5,3,1 for 96x96 main.add_module('final_layer.tanh', nn.Tanh()) # state size. (nc) x 96 x 96 self.main = main def forward(self, input): gpu_ids = None if isinstance(input.data, torch.cuda.FloatTensor) and self.ngpu > 1: gpu_ids = range(self.ngpu) return nn.parallel.data_parallel(self.main, input, gpu_ids), 0 class _netD_1(nn.Module): def __init__(self, ngpu, nz, nc, ndf, n_extra_layers_d): super(_netD_1, self).__init__() self.ngpu = ngpu main = nn.Sequential( # input is (nc) x 96 x 96 nn.Conv2d(nc, ndf, 4, 2, 1, bias=False), # 5,3,1 for 96x96 nn.LeakyReLU(0.2, inplace=True), # state size. (ndf) x 32 x 32 nn.Conv2d(ndf, ndf * 2, 4, 2, 1, bias=False), nn.BatchNorm2d(ndf * 2), nn.LeakyReLU(0.2, inplace=True), # state size. (ndf*2) x 16 x 16 nn.Conv2d(ndf * 2, ndf * 4, 4, 2, 1, bias=False), nn.BatchNorm2d(ndf * 4), nn.LeakyReLU(0.2, inplace=True), # state size. (ndf*4) x 8 x 8 nn.Conv2d(ndf * 4, ndf * 8, 4, 2, 1, bias=False), nn.BatchNorm2d(ndf * 8), nn.LeakyReLU(0.2, inplace=True), # state size. (ndf*8) x 4 x 4 ) # Extra layers for t in range(n_extra_layers_d): main.add_module('extra-layers-{0}.{1}.conv'.format(t, ndf * 8), nn.Conv2d(ndf * 8, ndf * 8, 3, 1, 1, bias=False)) main.add_module('extra-layers-{0}.{1}.batchnorm'.format(t, ndf * 8), nn.BatchNorm2d(ndf * 8)) main.add_module('extra-layers-{0}.{1}.relu'.format(t, ndf * 8), nn.LeakyReLU(0.2, inplace=True)) main.add_module('final_layers.conv', nn.Conv2d(ndf * 8, 1, 4, 1, 0, bias=False)) main.add_module('final_layers.sigmoid', nn.Sigmoid()) # state size. 1 x 1 x 1 self.main = main def forward(self, input): gpu_ids = None if isinstance(input.data, torch.cuda.FloatTensor) and self.ngpu > 1: gpu_ids = range(self.ngpu) output = nn.parallel.data_parallel(self.main, input, gpu_ids) return output.view(-1, 1) class _netD_2(nn.Module): def __init__(self, ngpu, nz, nc , ndf): super(_netD_2, self).__init__() self.ngpu = ngpu self.convs = nn.Sequential( # input is (nc) x 96 x 96 nn.Conv2d(nc, ndf, 4, 2, 1, bias=False), nn.LeakyReLU(0.2, inplace=True), # state size. (ndf) x 32 x 32 nn.Conv2d(ndf, ndf * 2, 4, 2, 1, bias=False), nn.BatchNorm2d(ndf * 2), nn.LeakyReLU(0.2, inplace=True), # state size. (ndf*2) x 16 x 16 nn.Conv2d(ndf * 2, ndf * 4, 4, 2, 1, bias=False), nn.BatchNorm2d(ndf * 4), nn.LeakyReLU(0.2, inplace=True), # state size. (ndf*4) x 8 x 8 nn.Conv2d(ndf * 4, ndf * 8, 4, 2, 1, bias=False), nn.BatchNorm2d(ndf * 8), nn.LeakyReLU(0.2, inplace=True), # state size. (ndf*8) x 4 x 4 nn.Conv2d(ndf * 8, 1024, 4, 1, 0, bias=False), nn.LeakyReLU(inplace=True), nn.Dropout(0.5), # state size. 1024 x 1 x 1 ) self.fcs = nn.Sequential( nn.Linear(1024, 1024), nn.LeakyReLU(inplace=True), nn.Dropout(0.5), nn.Linear(1024, 1), nn.Sigmoid() ) def forward(self, input): gpu_ids = None if isinstance(input.data, torch.cuda.FloatTensor) and self.ngpu > 1: gpu_ids = range(self.ngpu) output = nn.parallel.data_parallel(self.convs, input, gpu_ids) output = self.fcs(output.view(-1,1024)) return output.view(-1, 1) # with z decoder and fc layers class _netG_2(nn.Module): def __init__(self, ngpu, nz, nc , ngf): super(_netG_2, self).__init__() self.ngpu = ngpu self.nz = nz self.fcs = nn.Sequential( # input is Z, going into a convolution # state size. nz x 1 x 1 nn.Linear(nz, 1024), nn.ReLU(inplace=True), nn.Dropout(0.5), nn.Linear(1024, 1024), nn.ReLU(inplace=True), nn.Dropout(0.5), ) self.decode_fcs = nn.Sequential( nn.Linear(1024, 1024), nn.ReLU(inplace=True), nn.Dropout(0.5), nn.Linear(1024, nz), ) self.convs = nn.Sequential( # 1024x1x1 nn.ConvTranspose2d(1024, ngf * 8, 4, 1, 0, bias=False), nn.BatchNorm2d(ngf * 8), nn.ReLU(inplace=True), # state size. (ngf*8) x 4 x 4 nn.ConvTranspose2d(ngf * 8, ngf * 4, 4, 2, 1, bias=False), nn.BatchNorm2d(ngf * 4), nn.ReLU(inplace=True), # state size. (ngf*4) x 8 x 8 nn.ConvTranspose2d(ngf * 4, ngf * 2, 4, 2, 1, bias=False), nn.BatchNorm2d(ngf * 2), nn.ReLU(inplace=True), # state size. (ngf*2) x 16 x 16 nn.ConvTranspose2d(ngf * 2, ngf, 4, 2, 1, bias=False), nn.BatchNorm2d(ngf), nn.ReLU(inplace=True), # state size. (ngf) x 32 x 32 nn.ConvTranspose2d( ngf, nc, 4, 2, 1, bias=False), nn.Tanh() # state size. (nc) x 96 x 96 ) def forward(self, input): input = self.fcs(input.view(-1,self.nz)) gpu_ids = None if isinstance(input.data, torch.cuda.FloatTensor) and self.ngpu > 1: gpu_ids = range(self.ngpu) z_prediction = self.decode_fcs(input) input = input.view(-1,1024,1,1) output = nn.parallel.data_parallel(self.convs, input, gpu_ids) return output, z_prediction # DCGAN model with fc layers class _netG_3(nn.Module): def __init__(self, ngpu, nz, nc , ngf): super(_netG_3, self).__init__() self.ngpu = ngpu self.fcs = nn.Sequential( # input is Z, going into a convolution # state size. nz x 1 x 1 nn.Linear(nz, 1024), nn.ReLU(inplace=True), nn.Dropout(0.5), nn.Linear(1024, 1024), nn.ReLU(inplace=True), nn.Dropout(0.5), ) self.convs = nn.Sequential( # 1024x1x1 nn.ConvTranspose2d(1024, ngf * 8, 4, 1, 0, bias=False), nn.BatchNorm2d(ngf * 8), nn.ReLU(inplace=True), # state size. (ngf*8) x 4 x 4 nn.ConvTranspose2d(ngf * 8, ngf * 4, 4, 2, 1, bias=False), nn.BatchNorm2d(ngf * 4), nn.ReLU(inplace=True), # state size. (ngf*4) x 8 x 8 nn.ConvTranspose2d(ngf * 4, ngf * 2, 4, 2, 1, bias=False), nn.BatchNorm2d(ngf * 2), nn.ReLU(inplace=True), # state size. (ngf*2) x 16 x 16 nn.ConvTranspose2d(ngf * 2, ngf, 4, 2, 1, bias=False), nn.BatchNorm2d(ngf), nn.ReLU(inplace=True), # state size. (ngf) x 32 x 32 nn.ConvTranspose2d( ngf, nc, 4, 2, 1, bias=False), nn.Tanh() # state size. (nc) x 96 x 96 ) def forward(self, input): input = self.fcs(input.view(-1,nz)) gpu_ids = None if isinstance(input.data, torch.cuda.FloatTensor) and self.ngpu > 1: gpu_ids = range(self.ngpu) input = input.view(-1,1024,1,1) return nn.parallel.data_parallel(self.convs, input, gpu_ids)
9,855
37.20155
89
py
bnp
bnp-master/bayesian_optimization/run_bo.py
import os import argparse from attrdict import AttrDict import numpy as np import os.path as osp import yaml import torch from data.gp import * import bayeso import bayeso.gp as bayesogp from bayeso import covariance from bayeso import acquisition from utils.paths import results_path from utils.misc import load_module def get_str_file(path_, str_kernel, str_model, noise, seed=None): if noise is not None: str_all = 'bo_{}_{}_{}'.format(str_kernel, 'noisy', str_model) else: str_all = 'bo_{}_{}'.format(str_kernel, str_model) if seed is not None: str_all += '_' + str(seed) + '.npy' else: str_all += '.npy' return osp.join(path_, str_all) def main(): parser = argparse.ArgumentParser() parser.add_argument('--mode', choices=['oracle', 'bo'], default='bo') parser.add_argument('--expid', type=str, default='run1') parser.add_argument('--gpu', type=str, default='0') parser.add_argument('--model', type=str, default='cnp') parser.add_argument('--bo_num_samples', type=int, default=200) parser.add_argument('--bo_num_init', type=int, default=1) parser.add_argument('--bo_kernel', type=str, default='periodic') parser.add_argument('--t_noise', type=float, default=None) args = parser.parse_args() os.environ['CUDA_VISIBLE_DEVICES'] = args.gpu model_cls = getattr(load_module(f'models/{args.model}.py'), args.model.upper()) with open(f'configs/gp/{args.model}.yaml', 'r') as f: config = yaml.safe_load(f) model = model_cls(**config).cuda() args.root = osp.join(results_path, 'gp', args.model, args.expid) if args.mode == 'oracle': oracle(args, model) elif args.mode == 'bo': bo(args, model) def oracle(args, model): seed = 42 num_all = 100 num_iter = 50 num_init = args.bo_num_init str_cov = 'se' list_dict = [] if args.bo_kernel == 'rbf': kernel = RBFKernel() elif args.bo_kernel == 'matern': kernel = Matern52Kernel() elif args.bo_kernel == 'periodic': kernel = PeriodicKernel() else: raise ValueError(f'Invalid kernel {args.bo_kernel}') for ind_seed in range(1, num_all + 1): seed_ = seed * ind_seed if seed_ is not None: torch.manual_seed(seed_) torch.cuda.manual_seed(seed_) if os.path.exists(get_str_file('./results', args.bo_kernel, 'oracle', args.t_noise, seed=ind_seed)): dict_exp = np.load(get_str_file('./results', args.bo_kernel, 'oracle', args.t_noise, seed=ind_seed), allow_pickle=True) dict_exp = dict_exp[()] list_dict.append(dict_exp) print(dict_exp) print(dict_exp['global']) print(np.array2string(dict_exp['minima'], separator=',')) print(np.array2string(dict_exp['regrets'], separator=',')) continue sampler = GPPriorSampler(kernel, t_noise=args.t_noise) xp = torch.linspace(-2, 2, 1000).cuda() xp_ = xp.unsqueeze(0).unsqueeze(2) yp = sampler.sample(xp_) min_yp = yp.min() print(min_yp.cpu().numpy()) model.eval() batch = AttrDict() indices_permuted = torch.randperm(yp.shape[1]) batch.x = xp_[:, indices_permuted[:2*num_init], :] batch.y = yp[:, indices_permuted[:2*num_init], :] batch.xc = xp_[:, indices_permuted[:num_init], :] batch.yc = yp[:, indices_permuted[:num_init], :] batch.xt = xp_[:, indices_permuted[num_init:2*num_init], :] batch.yt = yp[:, indices_permuted[num_init:2*num_init], :] X_train = batch.xc.squeeze(0).cpu().numpy() Y_train = batch.yc.squeeze(0).cpu().numpy() X_test = xp_.squeeze(0).cpu().numpy() list_min = [] list_min.append(batch.yc.min().cpu().numpy()) for ind_iter in range(0, num_iter): print('ind_seed {} seed {} iter {}'.format(ind_seed, seed_, ind_iter + 1)) cov_X_X, inv_cov_X_X, hyps = bayesogp.get_optimized_kernel(X_train, Y_train, None, str_cov, is_fixed_noise=False, debug=False) prior_mu_train = bayesogp.get_prior_mu(None, X_train) prior_mu_test = bayesogp.get_prior_mu(None, X_test) cov_X_Xs = covariance.cov_main(str_cov, X_train, X_test, hyps, False) cov_Xs_Xs = covariance.cov_main(str_cov, X_test, X_test, hyps, True) cov_Xs_Xs = (cov_Xs_Xs + cov_Xs_Xs.T) / 2.0 mu_ = np.dot(np.dot(cov_X_Xs.T, inv_cov_X_X), Y_train - prior_mu_train) + prior_mu_test Sigma_ = cov_Xs_Xs - np.dot(np.dot(cov_X_Xs.T, inv_cov_X_X), cov_X_Xs) sigma_ = np.expand_dims(np.sqrt(np.maximum(np.diag(Sigma_), 0.0)), axis=1) acq_vals = -1.0 * acquisition.ei(np.ravel(mu_), np.ravel(sigma_), Y_train) ind_ = np.argmin(acq_vals) x_new = xp[ind_, None, None, None] y_new = yp[:, ind_, None, :] batch.x = torch.cat([batch.x, x_new], axis=1) batch.y = torch.cat([batch.y, y_new], axis=1) batch.xc = torch.cat([batch.xc, x_new], axis=1) batch.yc = torch.cat([batch.yc, y_new], axis=1) X_train = batch.xc.squeeze(0).cpu().numpy() Y_train = batch.yc.squeeze(0).cpu().numpy() min_cur = batch.yc.min() list_min.append(min_cur.cpu().numpy()) print(min_yp.cpu().numpy()) print(np.array2string(np.array(list_min), separator=',')) print(np.array2string(np.array(list_min) - min_yp.cpu().numpy(), separator=',')) dict_exp = { 'seed': seed_, 'str_cov': str_cov, 'global': min_yp.cpu().numpy(), 'minima': np.array(list_min), 'regrets': np.array(list_min) - min_yp.cpu().numpy(), 'xc': X_train, 'yc': Y_train, 'model': 'oracle', } np.save(get_str_file('./results', args.bo_kernel, 'oracle', args.t_noise, seed=ind_seed), dict_exp) list_dict.append(dict_exp) np.save(get_str_file('./figures/results', args.bo_kernel, 'oracle', args.t_noise), list_dict) def bo(args, model): if args.mode == 'bo': ckpt = torch.load(os.path.join(args.root, 'ckpt.tar')) model.load_state_dict(ckpt.model) if args.bo_kernel == 'rbf': kernel = RBFKernel() elif args.bo_kernel == 'matern': kernel = Matern52Kernel() elif args.bo_kernel == 'periodic': kernel = PeriodicKernel() else: raise ValueError(f'Invalid kernel {args.bo_kernel}') seed = 42 str_cov = 'se' num_all = 100 num_iter = 50 num_init = args.bo_num_init list_dict = [] for ind_seed in range(1, num_all + 1): seed_ = seed * ind_seed if seed_ is not None: torch.manual_seed(seed_) torch.cuda.manual_seed(seed_) obj_prior = GPPriorSampler(kernel, t_noise=args.t_noise) xp = torch.linspace(-2, 2, 1000).cuda() xp_ = xp.unsqueeze(0).unsqueeze(2) yp = obj_prior.sample(xp_) min_yp = yp.min() print(min_yp.cpu().numpy()) model.eval() batch = AttrDict() indices_permuted = torch.randperm(yp.shape[1]) batch.x = xp_[:, indices_permuted[:2*num_init], :] batch.y = yp[:, indices_permuted[:2*num_init], :] batch.xc = xp_[:, indices_permuted[:num_init], :] batch.yc = yp[:, indices_permuted[:num_init], :] batch.xt = xp_[:, indices_permuted[num_init:2*num_init], :] batch.yt = yp[:, indices_permuted[num_init:2*num_init], :] X_train = batch.xc.squeeze(0).cpu().numpy() Y_train = batch.yc.squeeze(0).cpu().numpy() X_test = xp_.squeeze(0).cpu().numpy() list_min = [] list_min.append(batch.yc.min().cpu().numpy()) for ind_iter in range(0, num_iter): print('ind_seed {} seed {} iter {}'.format(ind_seed, seed_, ind_iter + 1)) with torch.no_grad(): outs = model(batch, num_samples=args.bo_num_samples) print('ctx_ll {:.4f} tar ll {:.4f}'.format( outs.ctx_ll.item(), outs.tar_ll.item())) py = model.predict(batch.xc, batch.yc, xp[None,:,None].repeat(1, 1, 1), num_samples=args.bo_num_samples) mu, sigma = py.mean.squeeze(0), py.scale.squeeze(0) if mu.dim() == 4: print(mu.shape, sigma.shape) var = sigma.pow(2).mean(0) + mu.pow(2).mean(0) - mu.mean(0).pow(2) sigma = var.sqrt().squeeze(0) mu = mu.mean(0).squeeze(0) mu_ = mu.cpu().numpy() sigma_ = sigma.cpu().numpy() acq_vals = -1.0 * acquisition.ei(np.ravel(mu_), np.ravel(sigma_), Y_train) # acq_vals = [] # for ind_mu in range(0, mu.shape[0]): # acq_vals_ = -1.0 * acquisition.ei(np.ravel(mu[ind_mu].cpu().numpy()), np.ravel(sigma[ind_mu].cpu().numpy()), Y_train) # acq_vals.append(acq_vals_) # acq_vals = np.mean(acq_vals, axis=0) else: mu_ = mu.cpu().numpy() sigma_ = sigma.cpu().numpy() acq_vals = -1.0 * acquisition.ei(np.ravel(mu_), np.ravel(sigma_), Y_train) # var = sigma.pow(2).mean(0) + mu.pow(2).mean(0) - mu.mean(0).pow(2) # sigma = var.sqrt().squeeze(0) # mu = mu.mean(0).squeeze(0) ind_ = np.argmin(acq_vals) x_new = xp[ind_, None, None, None] y_new = yp[:, ind_, None, :] batch.x = torch.cat([batch.x, x_new], axis=1) batch.y = torch.cat([batch.y, y_new], axis=1) batch.xc = torch.cat([batch.xc, x_new], axis=1) batch.yc = torch.cat([batch.yc, y_new], axis=1) X_train = batch.xc.squeeze(0).cpu().numpy() Y_train = batch.yc.squeeze(0).cpu().numpy() min_cur = batch.yc.min() list_min.append(min_cur.cpu().numpy()) print(min_yp.cpu().numpy()) print(np.array2string(np.array(list_min), separator=',')) print(np.array2string(np.array(list_min) - min_yp.cpu().numpy(), separator=',')) dict_exp = { 'seed': seed_, 'global': min_yp.cpu().numpy(), 'minima': np.array(list_min), 'regrets': np.array(list_min) - min_yp.cpu().numpy(), 'xc': X_train, 'yc': Y_train, 'model': args.model, 'cov': str_cov, } list_dict.append(dict_exp) np.save(get_str_file('./figures/results', args.bo_kernel, args.model, args.t_noise), list_dict) if __name__ == '__main__': main()
10,873
32.875389
138
py
bnp
bnp-master/bayesian_optimization/models/anp.py
import torch import torch.nn as nn from torch.distributions import kl_divergence from attrdict import AttrDict from utils.misc import stack, logmeanexp from utils.sampling import sample_subset from models.modules import CrossAttnEncoder, PoolingEncoder, Decoder class ANP(nn.Module): def __init__(self, dim_x=1, dim_y=1, dim_hid=128, dim_lat=128, enc_v_depth=4, enc_qk_depth=2, enc_pre_depth=4, enc_post_depth=2, dec_depth=3): super().__init__() self.denc = CrossAttnEncoder( dim_x=dim_x, dim_y=dim_y, dim_hid=dim_hid, v_depth=enc_v_depth, qk_depth=enc_qk_depth) self.lenc = PoolingEncoder( dim_x=dim_x, dim_y=dim_y, dim_hid=dim_hid, dim_lat=dim_lat, self_attn=True, pre_depth=enc_pre_depth, post_depth=enc_post_depth) self.dec = Decoder( dim_x=dim_x, dim_y=dim_y, dim_enc=dim_hid+dim_lat, dim_hid=dim_hid, depth=dec_depth) def predict(self, xc, yc, xt, z=None, num_samples=None): theta = stack(self.denc(xc, yc, xt), num_samples) if z is None: pz = self.lenc(xc, yc) z = pz.rsample() if num_samples is None \ else pz.rsample([num_samples]) z = stack(z, xt.shape[-2], -2) encoded = torch.cat([theta, z], -1) return self.dec(encoded, stack(xt, num_samples)) def forward(self, batch, num_samples=None, reduce_ll=True): outs = AttrDict() if self.training: pz = self.lenc(batch.xc, batch.yc) qz = self.lenc(batch.x, batch.y) z = qz.rsample() if num_samples is None else \ qz.rsample([num_samples]) py = self.predict(batch.xc, batch.yc, batch.x, z=z, num_samples=num_samples) if num_samples > 1: # K * B * N recon = py.log_prob(stack(batch.y, num_samples)).sum(-1) # K * B log_qz = qz.log_prob(z).sum(-1) log_pz = pz.log_prob(z).sum(-1) # K * B log_w = recon.sum(-1) + log_pz - log_qz outs.loss = -logmeanexp(log_w).mean() / batch.x.shape[-2] else: outs.recon = py.log_prob(batch.y).sum(-1).mean() outs.kld = kl_divergence(qz, pz).sum(-1).mean() outs.loss = -outs.recon + outs.kld / batch.x.shape[-2] else: py = self.predict(batch.xc, batch.yc, batch.x, num_samples=num_samples) if num_samples is None: ll = py.log_prob(batch.y).sum(-1) else: y = torch.stack([batch.y]*num_samples) if reduce_ll: ll = logmeanexp(py.log_prob(y).sum(-1)) else: ll = py.log_prob(y).sum(-1) num_ctx = batch.xc.shape[-2] if reduce_ll: outs.ctx_ll = ll[...,:num_ctx].mean() outs.tar_ll = ll[...,num_ctx:].mean() else: outs.ctx_ll = ll[...,:num_ctx] outs.tar_ll = ll[...,num_ctx:] return outs
3,447
32.153846
83
py
bnp
bnp-master/bayesian_optimization/models/cnp.py
import torch import torch.nn as nn from attrdict import AttrDict from models.modules import PoolingEncoder, Decoder class CNP(nn.Module): def __init__(self, dim_x=1, dim_y=1, dim_hid=128, enc_pre_depth=4, enc_post_depth=2, dec_depth=3): super().__init__() self.enc1 = PoolingEncoder( dim_x=dim_x, dim_y=dim_y, dim_hid=dim_hid, pre_depth=enc_pre_depth, post_depth=enc_post_depth) self.enc2 = PoolingEncoder( dim_x=dim_x, dim_y=dim_y, dim_hid=dim_hid, pre_depth=enc_pre_depth, post_depth=enc_post_depth) self.dec = Decoder( dim_x=dim_x, dim_y=dim_y, dim_enc=2*dim_hid, dim_hid=dim_hid, depth=dec_depth) def predict(self, xc, yc, xt, num_samples=None): encoded = torch.cat([self.enc1(xc, yc), self.enc2(xc, yc)], -1) encoded = torch.stack([encoded]*xt.shape[-2], -2) return self.dec(encoded, xt) def forward(self, batch, num_samples=None, reduce_ll=True): outs = AttrDict() py = self.predict(batch.xc, batch.yc, batch.x) ll = py.log_prob(batch.y).sum(-1) if self.training: outs.loss = -ll.mean() else: num_ctx = batch.xc.shape[-2] if reduce_ll: outs.ctx_ll = ll[...,:num_ctx].mean() outs.tar_ll = ll[...,num_ctx:].mean() else: outs.ctx_ll = ll[...,:num_ctx] outs.tar_ll = ll[...,num_ctx:] return outs
1,748
27.672131
71
py
bnp
bnp-master/bayesian_optimization/models/modules.py
import torch import torch.nn as nn import torch.nn.functional as F from torch.distributions import Normal from models.attention import MultiHeadAttn, SelfAttn __all__ = ['PoolingEncoder', 'CrossAttnEncoder', 'Decoder'] def build_mlp(dim_in, dim_hid, dim_out, depth): modules = [nn.Linear(dim_in, dim_hid), nn.ReLU(True)] for _ in range(depth-2): modules.append(nn.Linear(dim_hid, dim_hid)) modules.append(nn.ReLU(True)) modules.append(nn.Linear(dim_hid, dim_out)) return nn.Sequential(*modules) class PoolingEncoder(nn.Module): def __init__(self, dim_x=1, dim_y=1, dim_hid=128, dim_lat=None, self_attn=False, pre_depth=4, post_depth=2): super().__init__() self.use_lat = dim_lat is not None self.net_pre = build_mlp(dim_x+dim_y, dim_hid, dim_hid, pre_depth) \ if not self_attn else \ nn.Sequential( build_mlp(dim_x+dim_y, dim_hid, dim_hid, pre_depth-2), nn.ReLU(True), SelfAttn(dim_hid, dim_hid)) self.net_post = build_mlp(dim_hid, dim_hid, 2*dim_lat if self.use_lat else dim_hid, post_depth) def forward(self, xc, yc, mask=None): out = self.net_pre(torch.cat([xc, yc], -1)) if mask is None: out = out.mean(-2) else: mask = mask.to(xc.device) out = (out * mask.unsqueeze(-1)).sum(-2) / \ (mask.sum(-1, keepdim=True).detach() + 1e-5) if self.use_lat: mu, sigma = self.net_post(out).chunk(2, -1) sigma = 0.1 + 0.9 * torch.sigmoid(sigma) return Normal(mu, sigma) else: return self.net_post(out) class CrossAttnEncoder(nn.Module): def __init__(self, dim_x=1, dim_y=1, dim_hid=128, dim_lat=None, self_attn=True, v_depth=4, qk_depth=2): super().__init__() self.use_lat = dim_lat is not None if not self_attn: self.net_v = build_mlp(dim_x+dim_y, dim_hid, dim_hid, v_depth) else: self.net_v = build_mlp(dim_x+dim_y, dim_hid, dim_hid, v_depth-2) self.self_attn = SelfAttn(dim_hid, dim_hid) self.net_qk = build_mlp(dim_x, dim_hid, dim_hid, qk_depth) self.attn = MultiHeadAttn(dim_hid, dim_hid, dim_hid, 2*dim_lat if self.use_lat else dim_hid) def forward(self, xc, yc, xt, mask=None): q, k = self.net_qk(xt), self.net_qk(xc) v = self.net_v(torch.cat([xc, yc], -1)) if hasattr(self, 'self_attn'): v = self.self_attn(v, mask=mask) out = self.attn(q, k, v, mask=mask) if self.use_lat: mu, sigma = out.chunk(2, -1) sigma = 0.1 + 0.9 * torch.sigmoid(sigma) return Normal(mu, sigma) else: return out class Decoder(nn.Module): def __init__(self, dim_x=1, dim_y=1, dim_enc=128, dim_hid=128, depth=3): super().__init__() self.fc = nn.Linear(dim_x+dim_enc, dim_hid) self.dim_hid = dim_hid modules = [nn.ReLU(True)] for _ in range(depth-2): modules.append(nn.Linear(dim_hid, dim_hid)) modules.append(nn.ReLU(True)) modules.append(nn.Linear(dim_hid, 2*dim_y)) self.mlp = nn.Sequential(*modules) def add_ctx(self, dim_ctx): self.dim_ctx = dim_ctx self.fc_ctx = nn.Linear(dim_ctx, self.dim_hid, bias=False) def forward(self, encoded, x, ctx=None): packed = torch.cat([encoded, x], -1) hid = self.fc(packed) if ctx is not None: hid = hid + self.fc_ctx(ctx) out = self.mlp(hid) mu, sigma = out.chunk(2, -1) sigma = 0.1 + 0.9 * F.softplus(sigma) return Normal(mu, sigma)
3,867
33.535714
78
py
bnp
bnp-master/bayesian_optimization/models/banp.py
import torch import torch.nn as nn from attrdict import AttrDict from models.canp import CANP from utils.misc import stack, logmeanexp from utils.sampling import sample_with_replacement as SWR, sample_subset class BANP(CANP): def __init__(self, *args, **kwargs): super().__init__(*args, **kwargs) self.dec.add_ctx(2*kwargs['dim_hid']) def encode(self, xc, yc, xt, mask=None): theta1 = self.enc1(xc, yc, xt) theta2 = self.enc2(xc, yc) encoded = torch.cat([theta1, torch.stack([theta2]*xt.shape[-2], -2)], -1) return encoded def predict(self, xc, yc, xt, num_samples=None, return_base=False): with torch.no_grad(): bxc, byc = SWR(xc, yc, num_samples=num_samples) sxc, syc = stack(xc, num_samples), stack(yc, num_samples) encoded = self.encode(bxc, byc, sxc) py_res = self.dec(encoded, sxc) mu, sigma = py_res.mean, py_res.scale res = SWR((syc - mu)/sigma).detach() res = (res - res.mean(-2, keepdim=True)) bxc = sxc byc = mu + sigma * res encoded_base = self.encode(xc, yc, xt) sxt = stack(xt, num_samples) encoded_bs = self.encode(bxc, byc, sxt) py = self.dec(stack(encoded_base, num_samples), sxt, ctx=encoded_bs) if self.training or return_base: py_base = self.dec(encoded_base, xt) return py_base, py else: return py def forward(self, batch, num_samples=None, reduce_ll=True): outs = AttrDict() def compute_ll(py, y): ll = py.log_prob(y).sum(-1) if ll.dim() == 3 and reduce_ll: ll = logmeanexp(ll) return ll if self.training: py_base, py = self.predict(batch.xc, batch.yc, batch.x, num_samples=num_samples) outs.ll_base = compute_ll(py_base, batch.y).mean() outs.ll = compute_ll(py, batch.y).mean() outs.loss = -outs.ll_base - outs.ll else: py = self.predict(batch.xc, batch.yc, batch.x, num_samples=num_samples) ll = compute_ll(py, batch.y) num_ctx = batch.xc.shape[-2] if reduce_ll: outs.ctx_ll = ll[...,:num_ctx].mean() outs.tar_ll = ll[...,num_ctx:].mean() else: outs.ctx_ll = ll[...,:num_ctx] outs.tar_ll = ll[...,num_ctx:] return outs
2,555
31.35443
72
py
bnp
bnp-master/bayesian_optimization/models/canp.py
import torch import torch.nn as nn from attrdict import AttrDict from models.modules import CrossAttnEncoder, Decoder, PoolingEncoder class CANP(nn.Module): def __init__(self, dim_x=1, dim_y=1, dim_hid=128, enc_v_depth=4, enc_qk_depth=2, enc_pre_depth=4, enc_post_depth=2, dec_depth=3): super().__init__() self.enc1 = CrossAttnEncoder( dim_x=dim_x, dim_y=dim_y, dim_hid=dim_hid, v_depth=enc_v_depth, qk_depth=enc_qk_depth) self.enc2 = PoolingEncoder( dim_x=dim_x, dim_y=dim_y, dim_hid=dim_hid, self_attn=True, pre_depth=enc_pre_depth, post_depth=enc_post_depth) self.dec = Decoder( dim_x=dim_x, dim_y=dim_y, dim_enc=2*dim_hid, dim_hid=dim_hid, depth=dec_depth) def predict(self, xc, yc, xt, num_samples=None): theta1 = self.enc1(xc, yc, xt) theta2 = self.enc2(xc, yc) encoded = torch.cat([theta1, torch.stack([theta2]*xt.shape[-2], -2)], -1) return self.dec(encoded, xt) def forward(self, batch, num_samples=None, reduce_ll=True): outs = AttrDict() py = self.predict(batch.xc, batch.yc, batch.x) ll = py.log_prob(batch.y).sum(-1) if self.training: outs.loss = -ll.mean() else: num_ctx = batch.xc.shape[-2] if reduce_ll: outs.ctx_ll = ll[...,:num_ctx].mean() outs.tar_ll = ll[...,num_ctx:].mean() else: outs.ctx_ll = ll[...,:num_ctx] outs.tar_ll = ll[...,num_ctx:] return outs
1,886
27.590909
68
py
bnp
bnp-master/bayesian_optimization/models/bnp.py
import torch import torch.nn as nn from attrdict import AttrDict from models.cnp import CNP from utils.misc import stack, logmeanexp from utils.sampling import sample_with_replacement as SWR, sample_subset class BNP(CNP): def __init__(self, *args, **kwargs): super().__init__(*args, **kwargs) self.dec.add_ctx(2*kwargs['dim_hid']) def encode(self, xc, yc, xt, mask=None): encoded = torch.cat([ self.enc1(xc, yc, mask=mask), self.enc2(xc, yc, mask=mask)], -1) return stack(encoded, xt.shape[-2], -2) def predict(self, xc, yc, xt, num_samples=None, return_base=False): with torch.no_grad(): bxc, byc = SWR(xc, yc, num_samples=num_samples) sxc, syc = stack(xc, num_samples), stack(yc, num_samples) encoded = self.encode(bxc, byc, sxc) py_res = self.dec(encoded, sxc) mu, sigma = py_res.mean, py_res.scale res = SWR((syc - mu)/sigma).detach() res = (res - res.mean(-2, keepdim=True)) bxc = sxc byc = mu + sigma * res encoded_base = self.encode(xc, yc, xt) sxt = stack(xt, num_samples) encoded_bs = self.encode(bxc, byc, sxt) py = self.dec(stack(encoded_base, num_samples), sxt, ctx=encoded_bs) if self.training or return_base: py_base = self.dec(encoded_base, xt) return py_base, py else: return py def forward(self, batch, num_samples=None, reduce_ll=True): outs = AttrDict() def compute_ll(py, y): ll = py.log_prob(y).sum(-1) if ll.dim() == 3 and reduce_ll: ll = logmeanexp(ll) return ll if self.training: py_base, py = self.predict(batch.xc, batch.yc, batch.x, num_samples=num_samples) outs.ll_base = compute_ll(py_base, batch.y).mean() outs.ll = compute_ll(py, batch.y).mean() outs.loss = -outs.ll_base - outs.ll else: py = self.predict(batch.xc, batch.yc, batch.x, num_samples=num_samples) ll = compute_ll(py, batch.y) num_ctx = batch.xc.shape[-2] if reduce_ll: outs.ctx_ll = ll[...,:num_ctx].mean() outs.tar_ll = ll[...,num_ctx:].mean() else: outs.ctx_ll = ll[...,:num_ctx] outs.tar_ll = ll[...,num_ctx:] return outs
2,527
31.410256
72
py
bnp
bnp-master/bayesian_optimization/models/np.py
import torch import torch.nn as nn from torch.distributions import kl_divergence from attrdict import AttrDict from utils.misc import stack, logmeanexp from utils.sampling import sample_subset from models.modules import PoolingEncoder, Decoder class NP(nn.Module): def __init__(self, dim_x=1, dim_y=1, dim_hid=128, dim_lat=128, enc_pre_depth=4, enc_post_depth=2, dec_depth=3): super().__init__() self.denc = PoolingEncoder( dim_x=dim_x, dim_y=dim_y, dim_hid=dim_hid, pre_depth=enc_pre_depth, post_depth=enc_post_depth) self.lenc = PoolingEncoder( dim_x=dim_x, dim_y=dim_y, dim_hid=dim_hid, dim_lat=dim_lat, pre_depth=enc_pre_depth, post_depth=enc_post_depth) self.dec = Decoder( dim_x=dim_x, dim_y=dim_y, dim_enc=dim_hid+dim_lat, dim_hid=dim_hid, depth=dec_depth) def predict(self, xc, yc, xt, z=None, num_samples=None): theta = stack(self.denc(xc, yc), num_samples) if z is None: pz = self.lenc(xc, yc) z = pz.rsample() if num_samples is None \ else pz.rsample([num_samples]) encoded = torch.cat([theta, z], -1) encoded = stack(encoded, xt.shape[-2], -2) return self.dec(encoded, stack(xt, num_samples)) def forward(self, batch, num_samples=None, reduce_ll=True): outs = AttrDict() if self.training: pz = self.lenc(batch.xc, batch.yc) qz = self.lenc(batch.x, batch.y) z = qz.rsample() if num_samples is None else \ qz.rsample([num_samples]) py = self.predict(batch.xc, batch.yc, batch.x, z=z, num_samples=num_samples) if num_samples > 1: # K * B * N recon = py.log_prob(stack(batch.y, num_samples)).sum(-1) # K * B log_qz = qz.log_prob(z).sum(-1) log_pz = pz.log_prob(z).sum(-1) # K * B log_w = recon.sum(-1) + log_pz - log_qz outs.loss = -logmeanexp(log_w).mean() / batch.x.shape[-2] else: outs.recon = py.log_prob(batch.y).sum(-1).mean() outs.kld = kl_divergence(qz, pz).sum(-1).mean() outs.loss = -outs.recon + outs.kld / batch.x.shape[-2] else: py = self.predict(batch.xc, batch.yc, batch.x, num_samples=num_samples) if num_samples is None: ll = py.log_prob(batch.y).sum(-1) else: y = torch.stack([batch.y]*num_samples) if reduce_ll: ll = logmeanexp(py.log_prob(y).sum(-1)) else: ll = py.log_prob(y).sum(-1) num_ctx = batch.xc.shape[-2] if reduce_ll: outs.ctx_ll = ll[...,:num_ctx].mean() outs.tar_ll = ll[...,num_ctx:].mean() else: outs.ctx_ll = ll[...,:num_ctx] outs.tar_ll = ll[...,num_ctx:] return outs
3,352
33.214286
83
py
bnp
bnp-master/bayesian_optimization/models/attention.py
import torch import torch.nn as nn import torch.nn.functional as F import math class MultiHeadAttn(nn.Module): def __init__(self, dim_q, dim_k, dim_v, dim_out, num_heads=8): super().__init__() self.num_heads = num_heads self.dim_out = dim_out self.fc_q = nn.Linear(dim_q, dim_out, bias=False) self.fc_k = nn.Linear(dim_k, dim_out, bias=False) self.fc_v = nn.Linear(dim_v, dim_out, bias=False) self.fc_out = nn.Linear(dim_out, dim_out) self.ln1 = nn.LayerNorm(dim_out) self.ln2 = nn.LayerNorm(dim_out) def scatter(self, x): return torch.cat(x.chunk(self.num_heads, -1), -3) def gather(self, x): return torch.cat(x.chunk(self.num_heads, -3), -1) def attend(self, q, k, v, mask=None): q_, k_, v_ = [self.scatter(x) for x in [q, k, v]] A_logits = q_ @ k_.transpose(-2, -1) / math.sqrt(self.dim_out) if mask is not None: mask = mask.bool().to(q.device) mask = torch.stack([mask]*q.shape[-2], -2) mask = torch.cat([mask]*self.num_heads, -3) A = torch.softmax(A_logits.masked_fill(mask, -float('inf')), -1) A = A.masked_fill(torch.isnan(A), 0.0) else: A = torch.softmax(A_logits, -1) return self.gather(A @ v_) def forward(self, q, k, v, mask=None): q, k, v = self.fc_q(q), self.fc_k(k), self.fc_v(v) out = self.ln1(q + self.attend(q, k, v, mask=mask)) out = self.ln2(out + F.relu(self.fc_out(out))) return out class SelfAttn(MultiHeadAttn): def __init__(self, dim_in, dim_out, num_heads=8): super().__init__(dim_in, dim_in, dim_in, dim_out, num_heads) def forward(self, x, mask=None): return super().forward(x, x, x, mask=mask)
1,805
35.857143
76
py
bnp
bnp-master/bayesian_optimization/utils/misc.py
import os from importlib.machinery import SourceFileLoader import math import torch def gen_load_func(parser, func): def load(args, cmdline): sub_args, cmdline = parser.parse_known_args(cmdline) for k, v in sub_args.__dict__.items(): args.__dict__[k] = v return func(**sub_args.__dict__), cmdline return load def load_module(filename): module_name = os.path.splitext(os.path.basename(filename))[0] return SourceFileLoader(module_name, filename).load_module() def logmeanexp(x, dim=0): return x.logsumexp(dim) - math.log(x.shape[dim]) def stack(x, num_samples=None, dim=0): return x if num_samples is None \ else torch.stack([x]*num_samples, dim=dim)
726
29.291667
65
py
bnp
bnp-master/bayesian_optimization/utils/log.py
import torch import time import logging from collections import OrderedDict def get_logger(filename, mode='a'): logging.basicConfig(level=logging.INFO, format='%(message)s') logger = logging.getLogger() logger.addHandler(logging.FileHandler(filename, mode=mode)) return logger class RunningAverage(object): def __init__(self, *keys): self.sum = OrderedDict() self.cnt = OrderedDict() self.clock = time.time() for key in keys: self.sum[key] = 0 self.cnt[key] = 0 def update(self, key, val): if isinstance(val, torch.Tensor): val = val.item() if self.sum.get(key, None) is None: self.sum[key] = val self.cnt[key] = 1 else: self.sum[key] = self.sum[key] + val self.cnt[key] += 1 def reset(self): for key in self.sum.keys(): self.sum[key] = 0 self.cnt[key] = 0 self.clock = time.time() def clear(self): self.sum = OrderedDict() self.cnt = OrderedDict() self.clock = time.time() def keys(self): return self.sum.keys() def get(self, key): assert(self.sum.get(key, None) is not None) return self.sum[key] / self.cnt[key] def info(self, show_et=True): line = '' for key in self.sum.keys(): val = self.sum[key] / self.cnt[key] if type(val) == float: line += f'{key} {val:.4f} ' else: line += f'{key} {val} '.format(key, val) if show_et: line += f'({time.time()-self.clock:.3f} secs)' return line
1,679
27
65
py
bnp
bnp-master/bayesian_optimization/utils/sampling.py
import torch def gather(items, idxs): K = idxs.shape[0] idxs = idxs.to(items[0].device) gathered = [] for item in items: gathered.append(torch.gather( torch.stack([item]*K), -2, torch.stack([idxs]*item.shape[-1], -1)).squeeze(0)) return gathered[0] if len(gathered) == 1 else gathered def sample_subset(*items, r_N=None, num_samples=None): r_N = r_N or torch.rand(1).item() K = num_samples or 1 N = items[0].shape[-2] Ns = max(1, int(r_N * N)) batch_shape = items[0].shape[:-2] idxs = torch.rand((K,)+batch_shape+(Ns,)).argsort(-1) return gather(items, idxs[...,:Ns]), gather(items, idxs[...,Ns:]) def sample_with_replacement(*items, num_samples=None, r_N=1.0, N_s=None): K = num_samples or 1 N = items[0].shape[-2] N_s = N_s or max(1, int(r_N * N)) batch_shape = items[0].shape[:-2] idxs = torch.randint(N, size=(K,)+batch_shape+(N_s,)) return gather(items, idxs) def sample_mask(B, N, num_samples=None, min_num=3, prob=0.5): min_num = min(min_num, N) K = num_samples or 1 fixed = torch.ones(K, B, min_num) if N - min_num > 0: rand = torch.bernoulli(prob*torch.ones(K, B, N-min_num)) mask = torch.cat([fixed, rand], -1) return mask.squeeze(0) else: return fixed.squeeze(0)
1,334
32.375
73
py
bnp
bnp-master/bayesian_optimization/data/gp.py
import torch import torch.nn as nn import torch.nn.functional as F from torch.distributions import MultivariateNormal, StudentT from attrdict import AttrDict import math __all__ = ['GPPriorSampler', 'GPSampler', 'RBFKernel', 'PeriodicKernel', 'Matern52Kernel'] class GPPriorSampler(object): def __init__(self, kernel, t_noise=None): self.kernel = kernel self.t_noise = t_noise def sample(self, bx, device='cuda:0'): # bx: 1 * num_points * 1 # 1 * num_points * num_points cov = self.kernel(bx) mean = torch.zeros(1, bx.shape[1], device=device) mean = mean.cuda() by = MultivariateNormal(mean, cov).rsample().unsqueeze(-1) if self.t_noise is not None: by += self.t_noise * StudentT(2.1).rsample(by.shape).to(device) return by class GPSampler(object): def __init__(self, kernel, t_noise=None): self.kernel = kernel self.t_noise = t_noise def sample(self, batch_size=16, num_ctx=None, max_num_points=50, x_range=(-2, 2), device='cpu'): batch = AttrDict() num_ctx = num_ctx or torch.randint(low=3, high=max_num_points-3, size=[1]).item() num_tar = torch.randint(low=3, high=max_num_points-num_ctx, size=[1]).item() num_points = num_ctx + num_tar batch.x = x_range[0] + (x_range[1] - x_range[0]) \ * torch.rand([batch_size, num_points, 1], device=device) batch.xc = batch.x[:,:num_ctx] batch.xt = batch.x[:,num_ctx:] # batch_size * num_points * num_points cov = self.kernel(batch.x) mean = torch.zeros(batch_size, num_points, device=device) batch.y = MultivariateNormal(mean, cov).rsample().unsqueeze(-1) batch.yc = batch.y[:,:num_ctx] batch.yt = batch.y[:,num_ctx:] if self.t_noise is not None: batch.y += self.t_noise * StudentT(2.1).rsample(batch.y.shape).to(device) return batch class RBFKernel(object): def __init__(self, sigma_eps=2e-2, max_length=0.6, max_scale=1.0): self.sigma_eps = sigma_eps self.max_length = max_length self.max_scale = max_scale # x: batch_size * num_points * dim def __call__(self, x): length = 0.1 + (self.max_length-0.1) \ * torch.rand([x.shape[0], 1, 1, 1], device=x.device) scale = 0.1 + (self.max_scale-0.1) \ * torch.rand([x.shape[0], 1, 1], device=x.device) # batch_size * num_points * num_points * dim dist = (x.unsqueeze(-2) - x.unsqueeze(-3))/length # batch_size * num_points * num_points cov = scale.pow(2) * torch.exp(-0.5 * dist.pow(2).sum(-1)) \ + self.sigma_eps**2 * torch.eye(x.shape[-2]).to(x.device) return cov class Matern52Kernel(object): def __init__(self, sigma_eps=2e-2, max_length=0.6, max_scale=1.0): self.sigma_eps = sigma_eps self.max_length = max_length self.max_scale = max_scale # x: batch_size * num_points * dim def __call__(self, x): length = 0.1 + (self.max_length-0.1) \ * torch.rand([x.shape[0], 1, 1, 1], device=x.device) scale = 0.1 + (self.max_scale-0.1) \ * torch.rand([x.shape[0], 1, 1], device=x.device) # batch_size * num_points * num_points dist = torch.norm((x.unsqueeze(-2) - x.unsqueeze(-3))/length, dim=-1) cov = scale.pow(2)*(1 + math.sqrt(5.0)*dist + 5.0*dist.pow(2)/3.0) \ * torch.exp(-math.sqrt(5.0) * dist) \ + self.sigma_eps**2 * torch.eye(x.shape[-2]).to(x.device) return cov class PeriodicKernel(object): def __init__(self, sigma_eps=2e-2, max_length=0.6, max_scale=1.0): #self.p = p self.sigma_eps = sigma_eps self.max_length = max_length self.max_scale = max_scale # x: batch_size * num_points * dim def __call__(self, x): p = 0.1 + 0.4*torch.rand([x.shape[0], 1, 1], device=x.device) length = 0.1 + (self.max_length-0.1) \ * torch.rand([x.shape[0], 1, 1], device=x.device) scale = 0.1 + (self.max_scale-0.1) \ * torch.rand([x.shape[0], 1, 1], device=x.device) dist = x.unsqueeze(-2) - x.unsqueeze(-3) cov = scale.pow(2) * torch.exp(\ - 2*(torch.sin(math.pi*dist.abs().sum(-1)/p)/length).pow(2)) \ + self.sigma_eps**2 * torch.eye(x.shape[-2]).to(x.device) return cov
4,576
34.207692
90
py
bnp
bnp-master/regression/gp.py
import os import os.path as osp import argparse import yaml import torch import torch.nn as nn import math import time import matplotlib.pyplot as plt from attrdict import AttrDict from tqdm import tqdm from copy import deepcopy from data.gp import * from utils.misc import load_module, logmeanexp from utils.paths import results_path, evalsets_path from utils.log import get_logger, RunningAverage def main(): parser = argparse.ArgumentParser() parser.add_argument('--mode', choices=['train', 'eval', 'plot', 'ensemble'], default='train') parser.add_argument('--expid', type=str, default='trial') parser.add_argument('--resume', action='store_true', default=False) parser.add_argument('--gpu', type=str, default='0') parser.add_argument('--max_num_points', type=int, default=50) parser.add_argument('--model', type=str, default='cnp') parser.add_argument('--train_batch_size', type=int, default=100) parser.add_argument('--train_num_samples', type=int, default=4) parser.add_argument('--lr', type=float, default=5e-4) parser.add_argument('--num_steps', type=int, default=100000) parser.add_argument('--print_freq', type=int, default=200) parser.add_argument('--eval_freq', type=int, default=5000) parser.add_argument('--save_freq', type=int, default=1000) parser.add_argument('--eval_seed', type=int, default=42) parser.add_argument('--eval_num_batches', type=int, default=3000) parser.add_argument('--eval_batch_size', type=int, default=16) parser.add_argument('--eval_num_samples', type=int, default=50) parser.add_argument('--eval_logfile', type=str, default=None) parser.add_argument('--plot_seed', type=int, default=None) parser.add_argument('--plot_batch_size', type=int, default=16) parser.add_argument('--plot_num_samples', type=int, default=30) parser.add_argument('--plot_num_ctx', type=int, default=None) # OOD settings parser.add_argument('--eval_kernel', type=str, default='rbf') parser.add_argument('--t_noise', type=float, default=None) args = parser.parse_args() os.environ['CUDA_VISIBLE_DEVICES'] = args.gpu model_cls = getattr(load_module(f'models/{args.model}.py'), args.model.upper()) with open(f'configs/gp/{args.model}.yaml', 'r') as f: config = yaml.safe_load(f) model = model_cls(**config).cuda() args.root = osp.join(results_path, 'gp', args.model, args.expid) if args.mode == 'train': train(args, model) elif args.mode == 'eval': eval(args, model) elif args.mode == 'plot': plot(args, model) elif args.mode == 'ensemble': ensemble(args, model) def train(args, model): if not osp.isdir(args.root): os.makedirs(args.root) with open(osp.join(args.root, 'args.yaml'), 'w') as f: yaml.dump(args.__dict__, f) sampler = GPSampler(RBFKernel()) optimizer = torch.optim.Adam(model.parameters(), lr=args.lr) scheduler = torch.optim.lr_scheduler.CosineAnnealingLR( optimizer, T_max=args.num_steps) if args.resume: ckpt = torch.load(os.path.join(args.root, 'ckpt.tar')) model.load_state_dict(ckpt.model) optimizer.load_state_dict(ckpt.optimizer) scheduler.load_state_dict(ckpt.scheduler) logfilename = ckpt.logfilename start_step = ckpt.step else: logfilename = os.path.join(args.root, f'train_{time.strftime("%Y%m%d-%H%M")}.log') start_step = 1 logger = get_logger(logfilename) ravg = RunningAverage() if not args.resume: logger.info('Total number of parameters: {}\n'.format( sum(p.numel() for p in model.parameters()))) for step in range(start_step, args.num_steps+1): model.train() optimizer.zero_grad() batch = sampler.sample( batch_size=args.train_batch_size, max_num_points=args.max_num_points, device='cuda') outs = model(batch, num_samples=args.train_num_samples) outs.loss.backward() optimizer.step() scheduler.step() for key, val in outs.items(): ravg.update(key, val) if step % args.print_freq == 0: line = f'{args.model}:{args.expid} step {step} ' line += f'lr {optimizer.param_groups[0]["lr"]:.3e} ' line += ravg.info() logger.info(line) if step % args.eval_freq == 0: line = eval(args, model) logger.info(line + '\n') ravg.reset() if step % args.save_freq == 0 or step == args.num_steps: ckpt = AttrDict() ckpt.model = model.state_dict() ckpt.optimizer = optimizer.state_dict() ckpt.scheduler = scheduler.state_dict() ckpt.logfilename = logfilename ckpt.step = step + 1 torch.save(ckpt, os.path.join(args.root, 'ckpt.tar')) args.mode = 'eval' eval(args, model) def gen_evalset(args): if args.eval_kernel == 'rbf': kernel = RBFKernel() elif args.eval_kernel == 'matern': kernel = Matern52Kernel() elif args.eval_kernel == 'periodic': kernel = PeriodicKernel() else: raise ValueError(f'Invalid kernel {args.eval_kernel}') torch.manual_seed(args.eval_seed) torch.cuda.manual_seed(args.eval_seed) sampler = GPSampler(kernel, t_noise=args.t_noise) batches = [] for i in tqdm(range(args.eval_num_batches)): batches.append(sampler.sample( batch_size=args.eval_batch_size, max_num_points=args.max_num_points)) torch.manual_seed(time.time()) torch.cuda.manual_seed(time.time()) path = osp.join(evalsets_path, 'gp') if not osp.isdir(path): os.makedirs(path) filename = f'{args.eval_kernel}' if args.t_noise is not None: filename += f'_{args.t_noise}' filename += '.tar' torch.save(batches, osp.join(path, filename)) def eval(args, model): if args.mode == 'eval': ckpt = torch.load(os.path.join(args.root, 'ckpt.tar')) model.load_state_dict(ckpt.model) if args.eval_logfile is None: eval_logfile = f'eval_{args.eval_kernel}' if args.t_noise is not None: eval_logfile += f'_tn_{args.t_noise}' eval_logfile += '.log' else: eval_logfile = args.eval_logfile filename = os.path.join(args.root, eval_logfile) logger = get_logger(filename, mode='w') else: logger = None if args.eval_kernel == 'rbf': kernel = RBFKernel() elif args.eval_kernel == 'matern': kernel = Matern52Kernel() elif args.eval_kernel == 'periodic': kernel = PeriodicKernel() else: raise ValueError(f'Invalid kernel {args.eval_kernel}') path = osp.join(evalsets_path, 'gp') filename = f'{args.eval_kernel}' if args.t_noise is not None: filename += f'_{args.t_noise}' filename += '.tar' if not osp.isfile(osp.join(path, filename)): print('generating evaluation sets...') gen_evalset(args) eval_batches = torch.load(osp.join(path, filename)) torch.manual_seed(args.eval_seed) torch.cuda.manual_seed(args.eval_seed) ravg = RunningAverage() model.eval() with torch.no_grad(): for batch in tqdm(eval_batches): for key, val in batch.items(): batch[key] = val.cuda() outs = model(batch, num_samples=args.eval_num_samples) for key, val in outs.items(): ravg.update(key, val) torch.manual_seed(time.time()) torch.cuda.manual_seed(time.time()) line = f'{args.model}:{args.expid} {args.eval_kernel} ' if args.t_noise is not None: line += f'tn {args.t_noise} ' line += ravg.info() if logger is not None: logger.info(line) return line def plot(args, model): ckpt = torch.load(os.path.join(args.root, 'ckpt.tar')) model.load_state_dict(ckpt.model) def tnp(x): return x.squeeze().cpu().data.numpy() if args.plot_seed is not None: torch.manual_seed(args.plot_seed) torch.cuda.manual_seed(args.plot_seed) kernel = RBFKernel() if args.pp is None else PeriodicKernel(p=args.pp) sampler = GPSampler(kernel, t_noise=args.t_noise) xp = torch.linspace(-2, 2, 200).cuda() batch = sampler.sample( batch_size=args.plot_batch_size, max_num_points=args.max_num_points, num_ctx=args.plot_num_ctx, device='cuda') model.eval() with torch.no_grad(): outs = model(batch, num_samples=args.eval_num_samples) print(f'ctx_ll {outs.ctx_ll.item():.4f}, tar_ll {outs.tar_ll.item():.4f}') py = model.predict(batch.xc, batch.yc, xp[None,:,None].repeat(args.plot_batch_size, 1, 1), num_samples=args.plot_num_samples) mu, sigma = py.mean.squeeze(0), py.scale.squeeze(0) if args.plot_batch_size > 1: nrows = max(args.plot_batch_size//4, 1) ncols = min(4, args.plot_batch_size) fig, axes = plt.subplots(nrows, ncols, figsize=(5*ncols, 5*nrows)) axes = axes.flatten() else: fig = plt.figure(figsize=(5, 5)) axes = [plt.gca()] # multi sample if mu.dim() == 4: #var = sigma.pow(2).mean(0) + mu.pow(2).mean(0) - mu.mean(0).pow(2) #sigma = var.sqrt() #mu = mu.mean(0) for i, ax in enumerate(axes): #ax.plot(tnp(xp), tnp(mu[i]), color='steelblue', alpha=0.5) #ax.fill_between(tnp(xp), tnp(mu[i]-sigma[i]), tnp(mu[i]+sigma[i]), # color='skyblue', alpha=0.2, linewidth=0.0) for s in range(mu.shape[0]): ax.plot(tnp(xp), tnp(mu[s][i]), color='steelblue', alpha=max(0.5/args.plot_num_samples, 0.1)) ax.fill_between(tnp(xp), tnp(mu[s][i])-tnp(sigma[s][i]), tnp(mu[s][i])+tnp(sigma[s][i]), color='skyblue', alpha=max(0.2/args.plot_num_samples, 0.02), linewidth=0.0) ax.scatter(tnp(batch.xc[i]), tnp(batch.yc[i]), color='k', label='context', zorder=mu.shape[0]+1) ax.scatter(tnp(batch.xt[i]), tnp(batch.yt[i]), color='orchid', label='target', zorder=mu.shape[0]+1) ax.legend() else: for i, ax in enumerate(axes): ax.plot(tnp(xp), tnp(mu[i]), color='steelblue', alpha=0.5) ax.fill_between(tnp(xp), tnp(mu[i]-sigma[i]), tnp(mu[i]+sigma[i]), color='skyblue', alpha=0.2, linewidth=0.0) ax.scatter(tnp(batch.xc[i]), tnp(batch.yc[i]), color='k', label='context') ax.scatter(tnp(batch.xt[i]), tnp(batch.yt[i]), color='orchid', label='target') ax.legend() plt.tight_layout() plt.show() def ensemble(args, model): num_runs = 5 models = [] for i in range(num_runs): model_ = deepcopy(model) ckpt = torch.load(osp.join(results_path, 'gp', args.model, f'run{i+1}', 'ckpt.tar')) model_.load_state_dict(ckpt['model']) model_.cuda() model_.eval() models.append(model_) path = osp.join(evalsets_path, 'gp') filename = f'{args.eval_kernel}' if args.t_noise is not None: filename += f'_{args.t_noise}' filename += '.tar' if not osp.isfile(osp.join(path, filename)): print('generating evaluation sets...') gen_evalset(args) eval_batches = torch.load(osp.join(path, filename)) torch.manual_seed(args.eval_seed) torch.cuda.manual_seed(args.eval_seed) ravg = RunningAverage() with torch.no_grad(): for batch in tqdm(eval_batches): for key, val in batch.items(): batch[key] = val.cuda() ctx_ll = [] tar_ll = [] for model in models: outs = model(batch, num_samples=args.eval_num_samples, reduce_ll=False) ctx_ll.append(outs.ctx_ll) tar_ll.append(outs.tar_ll) if ctx_ll[0].dim() == 2: ctx_ll = torch.stack(ctx_ll) tar_ll = torch.stack(tar_ll) else: ctx_ll = torch.cat(ctx_ll) tar_ll = torch.cat(tar_ll) ctx_ll = logmeanexp(ctx_ll).mean() tar_ll = logmeanexp(tar_ll).mean() ravg.update('ctx_ll', ctx_ll) ravg.update('tar_ll', tar_ll) filename = f'ensemble_{args.eval_kernel}' if args.t_noise is not None: filename += f'_{args.t_noise}' filename += '.log' logger = get_logger(osp.join(results_path, 'gp', args.model, filename), mode='w') logger.info(ravg.info()) if __name__ == '__main__': main()
13,075
32.875648
92
py
bnp
bnp-master/regression/emnist.py
import os import os.path as osp import argparse import yaml import torch import torch.nn as nn import math import time import matplotlib.pyplot as plt from attrdict import AttrDict from tqdm import tqdm from copy import deepcopy from data.image import img_to_task, task_to_img from data.emnist import EMNIST from utils.misc import load_module, logmeanexp from utils.paths import results_path, evalsets_path from utils.log import get_logger, RunningAverage def main(): parser = argparse.ArgumentParser() parser.add_argument('--mode', choices=['train', 'eval', 'plot', 'ensemble'], default='train') parser.add_argument('--expid', type=str, default='trial') parser.add_argument('--resume', action='store_true', default=False) parser.add_argument('--gpu', type=str, default='0') parser.add_argument('--max_num_points', type=int, default=200) parser.add_argument('--class_range', type=int, nargs='*', default=[0,10]) parser.add_argument('--model', type=str, default='cnp') parser.add_argument('--train_batch_size', type=int, default=100) parser.add_argument('--train_num_samples', type=int, default=4) parser.add_argument('--lr', type=float, default=5e-4) parser.add_argument('--num_epochs', type=int, default=200) parser.add_argument('--eval_freq', type=int, default=10) parser.add_argument('--save_freq', type=int, default=10) parser.add_argument('--eval_seed', type=int, default=42) parser.add_argument('--eval_batch_size', type=int, default=16) parser.add_argument('--eval_num_samples', type=int, default=50) parser.add_argument('--eval_logfile', type=str, default=None) parser.add_argument('--plot_seed', type=int, default=None) parser.add_argument('--plot_batch_size', type=int, default=16) parser.add_argument('--plot_num_samples', type=int, default=30) parser.add_argument('--plot_num_ctx', type=int, default=100) # OOD settings parser.add_argument('--t_noise', type=float, default=None) args = parser.parse_args() os.environ['CUDA_VISIBLE_DEVICES'] = args.gpu model_cls = getattr(load_module(f'models/{args.model}.py'), args.model.upper()) with open(f'configs/emnist/{args.model}.yaml', 'r') as f: config = yaml.safe_load(f) model = model_cls(**config).cuda() args.root = osp.join(results_path, 'emnist', args.model, args.expid) if args.mode == 'train': train(args, model) elif args.mode == 'eval': eval(args, model) elif args.mode == 'plot': plot(args, model) elif args.mode == 'ensemble': ensemble(args, model) def train(args, model): if not osp.isdir(args.root): os.makedirs(args.root) with open(osp.join(args.root, 'args.yaml'), 'w') as f: yaml.dump(args.__dict__, f) train_ds = EMNIST(train=True, class_range=args.class_range) eval_ds = EMNIST(train=False, class_range=args.class_range) train_loader = torch.utils.data.DataLoader(train_ds, batch_size=args.train_batch_size, shuffle=True, num_workers=4) optimizer = torch.optim.Adam(model.parameters(), lr=args.lr) scheduler = torch.optim.lr_scheduler.CosineAnnealingLR( optimizer, T_max=len(train_loader)*args.num_epochs) if args.resume: ckpt = torch.load(osp.join(args.root, 'ckpt.tar')) model.load_state_dict(ckpt.model) optimizer.load_state_dict(ckpt.optimizer) scheduler.load_state_dict(ckpt.scheduler) logfilename = ckpt.logfilename start_epoch = ckpt.epoch else: logfilename = osp.join(args.root, 'train_{}.log'.format( time.strftime('%Y%m%d-%H%M'))) start_epoch = 1 logger = get_logger(logfilename) ravg = RunningAverage() if not args.resume: logger.info('Total number of parameters: {}\n'.format( sum(p.numel() for p in model.parameters()))) for epoch in range(start_epoch, args.num_epochs+1): model.train() for (x, _) in tqdm(train_loader): batch = img_to_task(x, max_num_points=args.max_num_points, device='cuda') optimizer.zero_grad() outs = model(batch, num_samples=args.train_num_samples) outs.loss.backward() optimizer.step() scheduler.step() for key, val in outs.items(): ravg.update(key, val) line = f'{args.model}:{args.expid} epoch {epoch} ' line += f'lr {optimizer.param_groups[0]["lr"]:.3e} ' line += ravg.info() logger.info(line) if epoch % args.eval_freq == 0: logger.info(eval(args, model) + '\n') ravg.reset() if epoch % args.save_freq == 0 or epoch == args.num_epochs: ckpt = AttrDict() ckpt.model = model.state_dict() ckpt.optimizer = optimizer.state_dict() ckpt.scheduler = scheduler.state_dict() ckpt.logfilename = logfilename ckpt.epoch = epoch + 1 torch.save(ckpt, osp.join(args.root, 'ckpt.tar')) args.mode = 'eval' eval(args, model) def gen_evalset(args): torch.manual_seed(args.eval_seed) torch.cuda.manual_seed(args.eval_seed) eval_ds = EMNIST(train=False, class_range=args.class_range) eval_loader = torch.utils.data.DataLoader(eval_ds, batch_size=args.eval_batch_size, shuffle=False, num_workers=4) batches = [] for x, _ in tqdm(eval_loader): batches.append(img_to_task(x, t_noise=args.t_noise, max_num_points=args.max_num_points)) torch.manual_seed(time.time()) torch.cuda.manual_seed(time.time()) path = osp.join(evalsets_path, 'emnist') if not osp.isdir(path): os.makedirs(path) c1, c2 = args.class_range filename = f'{c1}-{c2}' if args.t_noise is not None: filename += f'_{args.t_noise}' filename += '.tar' torch.save(batches, osp.join(path, filename)) def eval(args, model): if args.mode == 'eval': ckpt = torch.load(osp.join(args.root, 'ckpt.tar')) model.load_state_dict(ckpt.model) if args.eval_logfile is None: c1, c2 = args.class_range eval_logfile = f'eval_{c1}-{c2}' if args.t_noise is not None: eval_logfile += f'_{args.t_noise}' eval_logfile += '.log' else: eval_logfile = args.eval_logfile filename = osp.join(args.root, eval_logfile) logger = get_logger(filename, mode='w') else: logger = None path = osp.join(evalsets_path, 'emnist') c1, c2 = args.class_range filename = f'{c1}-{c2}' if args.t_noise is not None: filename += f'_{args.t_noise}' filename += '.tar' if not osp.isfile(osp.join(path, filename)): print('generating evaluation sets...') gen_evalset(args) eval_batches = torch.load(osp.join(path, filename)) torch.manual_seed(args.eval_seed) torch.cuda.manual_seed(args.eval_seed) ravg = RunningAverage() model.eval() with torch.no_grad(): for batch in tqdm(eval_batches): for key, val in batch.items(): batch[key] = val.cuda() outs = model(batch, num_samples=args.eval_num_samples) for key, val in outs.items(): ravg.update(key, val) torch.manual_seed(time.time()) torch.cuda.manual_seed(time.time()) c1, c2 = args.class_range line = f'{args.model}:{args.expid} {c1}-{c2} ' if args.t_noise is not None: line += f'tn {args.t_noise} ' line += ravg.info() if logger is not None: logger.info(line) return line def ensemble(args, model): num_runs = 5 models = [] for i in range(num_runs): model_ = deepcopy(model) ckpt = torch.load(osp.join(results_path, 'emnist', args.model, f'run{i+1}', 'ckpt.tar')) model_.load_state_dict(ckpt['model']) model_.cuda() model_.eval() models.append(model_) path = osp.join(evalsets_path, 'emnist') c1, c2 = args.class_range filename = f'{c1}-{c2}' if args.t_noise is not None: filename += f'_{args.t_noise}' filename += '.tar' if not osp.isfile(osp.join(path, filename)): print('generating evaluation sets...') gen_evalset(args) eval_batches = torch.load(osp.join(path, filename)) ravg = RunningAverage() with torch.no_grad(): for batch in tqdm(eval_batches): for key, val in batch.items(): batch[key] = val.cuda() ctx_ll = [] tar_ll = [] for model in models: outs = model(batch, num_samples=args.eval_num_samples, reduce_ll=False) ctx_ll.append(outs.ctx_ll) tar_ll.append(outs.tar_ll) if ctx_ll[0].dim() == 2: ctx_ll = torch.stack(ctx_ll) tar_ll = torch.stack(tar_ll) else: ctx_ll = torch.cat(ctx_ll) tar_ll = torch.cat(tar_ll) ctx_ll = logmeanexp(ctx_ll).mean() tar_ll = logmeanexp(tar_ll).mean() ravg.update('ctx_ll', ctx_ll) ravg.update('tar_ll', tar_ll) torch.manual_seed(time.time()) torch.cuda.manual_seed(time.time()) filename = f'ensemble_{c1}-{c2}' if args.t_noise is not None: filename += f'_{args.t_noise}' filename += '.log' logger = get_logger(osp.join(results_path, 'emnist', args.model, filename), mode='w') logger.info(ravg.info()) if __name__ == '__main__': main()
9,732
31.335548
96
py
bnp
bnp-master/regression/lotka_volterra.py
import os import os.path as osp import argparse import yaml import torch import torch.nn as nn import math import time import matplotlib.pyplot as plt from attrdict import AttrDict from tqdm import tqdm from copy import deepcopy from utils.misc import load_module, logmeanexp from utils.paths import results_path, datasets_path, evalsets_path from utils.log import get_logger, RunningAverage from data.lotka_volterra import load_hare_lynx def standardize(batch): with torch.no_grad(): mu, sigma = batch.xc.mean(-2, keepdim=True), batch.xc.std(-2, keepdim=True) sigma[sigma==0] = 1.0 batch.x = (batch.x - mu) / (sigma + 1e-5) batch.xc = (batch.xc - mu) / (sigma + 1e-5) batch.xt = (batch.xt - mu) / (sigma + 1e-5) mu, sigma = batch.yc.mean(-2, keepdim=True), batch.yc.std(-2, keepdim=True) batch.y = (batch.y - mu) / (sigma + 1e-5) batch.yc = (batch.yc - mu) / (sigma + 1e-5) batch.yt = (batch.yt - mu) / (sigma + 1e-5) return batch def main(): parser = argparse.ArgumentParser() parser.add_argument('--mode', choices=['train', 'eval', 'plot', 'ensemble'], default='train') parser.add_argument('--expid', type=str, default='trial') parser.add_argument('--resume', action='store_true', default=False) parser.add_argument('--gpu', type=str, default='0') parser.add_argument('--max_num_points', type=int, default=50) parser.add_argument('--model', type=str, default='cnp') parser.add_argument('--train_batch_size', type=int, default=100) parser.add_argument('--train_num_samples', type=int, default=4) parser.add_argument('--lr', type=float, default=5e-4) parser.add_argument('--print_freq', type=int, default=200) parser.add_argument('--eval_freq', type=int, default=5000) parser.add_argument('--save_freq', type=int, default=1000) parser.add_argument('--eval_seed', type=int, default=42) parser.add_argument('--hare_lynx', action='store_true') parser.add_argument('--eval_num_samples', type=int, default=50) parser.add_argument('--eval_logfile', type=str, default=None) parser.add_argument('--plot_seed', type=int, default=None) parser.add_argument('--plot_batch_size', type=int, default=16) parser.add_argument('--plot_num_samples', type=int, default=30) args = parser.parse_args() os.environ['CUDA_VISIBLE_DEVICES'] = args.gpu model_cls = getattr(load_module(f'models/{args.model}.py'), args.model.upper()) with open(f'configs/lotka_volterra/{args.model}.yaml', 'r') as f: config = yaml.safe_load(f) model = model_cls(**config).cuda() args.root = osp.join(results_path, 'lotka_volterra', args.model, args.expid) if args.mode == 'train': train(args, model) elif args.mode == 'eval': eval(args, model) elif args.mode == 'plot': plot(args, model) elif args.mode == 'ensemble': ensemble(args, model) def train(args, model): if not osp.isdir(args.root): os.makedirs(args.root) with open(osp.join(args.root, 'args.yaml'), 'w') as f: yaml.dump(args.__dict__, f) train_data = torch.load(osp.join(datasets_path, 'lotka_volterra', 'train.tar')) eval_data = torch.load(osp.join(datasets_path, 'lotka_volterra', 'eval.tar')) num_steps = len(train_data) optimizer = torch.optim.Adam(model.parameters(), lr=args.lr) scheduler = torch.optim.lr_scheduler.CosineAnnealingLR( optimizer, T_max=num_steps) if args.resume: ckpt = torch.load(osp.join(args.root, 'ckpt.tar')) model.load_state_dict(ckpt.model) optimizer.load_state_dict(ckpt.optimizer) scheduler.load_state_dict(ckpt.scheduler) logfilename = ckpt.logfilename start_step = ckpt.step else: logfilename = osp.join(args.root, f'train_{time.strftime("%Y%m%d-%H%M")}.log') start_step = 1 logger = get_logger(logfilename) ravg = RunningAverage() if not args.resume: logger.info('Total number of parameters: {}\n'.format( sum(p.numel() for p in model.parameters()))) for step in range(start_step, num_steps+1): model.train() optimizer.zero_grad() batch = standardize(train_data[step-1]) for key, val in batch.items(): batch[key] = val.cuda() outs = model(batch, num_samples=args.train_num_samples) outs.loss.backward() optimizer.step() scheduler.step() for key, val in outs.items(): ravg.update(key, val) if step % args.print_freq == 0: line = f'{args.model}:{args.expid} step {step} ' line += f'lr {optimizer.param_groups[0]["lr"]:.3e} ' line += ravg.info() logger.info(line) if step % args.eval_freq == 0: line = eval(args, model, eval_data=eval_data) logger.info(line + '\n') ravg.reset() if step % args.save_freq == 0 or step == num_steps: ckpt = AttrDict() ckpt.model = model.state_dict() ckpt.optimizer = optimizer.state_dict() ckpt.scheduler = scheduler.state_dict() ckpt.logfilename = logfilename ckpt.step = step + 1 torch.save(ckpt, osp.join(args.root, 'ckpt.tar')) args.mode = 'eval' eval(args, model, eval_data=eval_data) def eval(args, model, eval_data=None): if args.mode == 'eval': ckpt = torch.load(osp.join(args.root, 'ckpt.tar')) model.load_state_dict(ckpt.model) if args.eval_logfile is None: if args.hare_lynx: eval_logfile = 'hare_lynx.log' else: eval_logfile = 'eval.log' else: eval_logfile = args.eval_logfile filename = osp.join(args.root, eval_logfile) logger = get_logger(filename, mode='w') else: logger = None torch.manual_seed(args.eval_seed) torch.cuda.manual_seed(args.eval_seed) if eval_data is None: if args.hare_lynx: eval_data = load_hare_lynx(1000, 16) else: eval_data = torch.load(osp.join(datasets_path, 'lotka_volterra', 'eval.tar')) ravg = RunningAverage() model.eval() with torch.no_grad(): for batch in tqdm(eval_data): batch = standardize(batch) for key, val in batch.items(): batch[key] = val.cuda() outs = model(batch, num_samples=args.eval_num_samples) for key, val in outs.items(): ravg.update(key, val) torch.manual_seed(time.time()) torch.cuda.manual_seed(time.time()) line = f'{args.model}:{args.expid} ' line += ravg.info() if logger is not None: logger.info(line) return line def ensemble(args, model): num_runs = 5 models = [] for i in range(num_runs): model_ = deepcopy(model) ckpt = torch.load(osp.join(results_path, 'lotka_volterra', args.model, f'run{i+1}', 'ckpt.tar')) model_.load_state_dict(ckpt['model']) model_.cuda() model_.eval() models.append(model_) torch.manual_seed(args.eval_seed) torch.cuda.manual_seed(args.eval_seed) if args.hare_lynx: eval_data = load_hare_lynx(1000, 16) else: eval_data = torch.load(osp.join(datasets_path, 'lotka_volterra', 'eval.tar')) ravg = RunningAverage() with torch.no_grad(): for batch in tqdm(eval_data): batch = standardize(batch) for key, val in batch.items(): batch[key] = val.cuda() ctx_ll = [] tar_ll = [] for model_ in models: outs = model_(batch, num_samples=args.eval_num_samples, reduce_ll=False) ctx_ll.append(outs.ctx_ll) tar_ll.append(outs.tar_ll) if ctx_ll[0].dim() == 2: ctx_ll = torch.stack(ctx_ll) tar_ll = torch.stack(tar_ll) else: ctx_ll = torch.cat(ctx_ll) tar_ll = torch.cat(tar_ll) ctx_ll = logmeanexp(ctx_ll).mean() tar_ll = logmeanexp(tar_ll).mean() ravg.update('ctx_ll', ctx_ll) ravg.update('tar_ll', tar_ll) torch.manual_seed(time.time()) torch.cuda.manual_seed(time.time()) filename = 'ensemble' if args.hare_lynx: filename += '_hare_lynx' filename += '.log' logger = get_logger(osp.join(results_path, 'lotka_volterra', args.model, filename), mode='w') logger.info(ravg.info()) def plot(args, model): ckpt = torch.load(osp.join(args.root, 'ckpt.tar')) model.load_state_dict(ckpt.model) def tnp(x): return x.squeeze().cpu().data.numpy() if args.hare_lynx: eval_data = load_hare_lynx(1000, 16) else: eval_data = torch.load(osp.join(datasets_path, 'lotka_volterra', 'eval.tar')) bid = torch.randint(len(eval_data), [1]).item() batch = standardize(eval_data[bid]) for k, v in batch.items(): batch[k] = v.cuda() model.eval() outs = model(batch, num_samples=args.eval_num_samples) print(outs.tar_ll) fig, axes = plt.subplots(4, 4, figsize=(20, 20)) xp = [] for b in range(batch.x.shape[0]): bx = batch.x[b] xp.append(torch.linspace(bx.min()-0.1, bx.max()+0.1, 200)) xp = torch.stack(xp).unsqueeze(-1).cuda() model.eval() with torch.no_grad(): py = model.predict(batch.xc, batch.yc, xp, num_samples=args.plot_num_samples) mu, sigma = py.mean, py.scale if mu.dim() > 3: bmu = mu.mean(0) bvar = sigma.pow(2).mean(0) + mu.pow(2).mean(0) - mu.mean(0).pow(2) bsigma = bvar.sqrt() else: bmu = mu bsigma = sigma for i, ax in enumerate(axes.flatten()): ax.plot(tnp(xp[i]), tnp(bmu[i]), alpha=0.5) upper = tnp(bmu[i][:,0] + bsigma[i][:,0]) lower = tnp(bmu[i][:,0] - bsigma[i][:,0]) ax.fill_between(tnp(xp[i]), lower, upper, alpha=0.2, linewidth=0.0, label='predator') upper = tnp(bmu[i][:,1] + bsigma[i][:,1]) lower = tnp(bmu[i][:,1] - bsigma[i][:,1]) ax.fill_between(tnp(xp[i]), lower, upper, alpha=0.2, linewidth=0.0, label='prey') ax.scatter(tnp(batch.xc[i]), tnp(batch.yc[i][:,0]), color='k', marker='*') ax.scatter(tnp(batch.xc[i]), tnp(batch.yc[i][:,1]), color='k', marker='*') ax.scatter(tnp(batch.xt[i]), tnp(batch.yt[i][:,0]), color='orchid', marker='x') ax.scatter(tnp(batch.xt[i]), tnp(batch.yt[i][:,1]), color='orchid', marker='x') plt.tight_layout() plt.show() if __name__ == '__main__': main()
10,852
32.291411
104
py
bnp
bnp-master/regression/celeba.py
import os import os.path as osp import argparse import yaml import torch import torch.nn as nn import math import time import matplotlib.pyplot as plt from attrdict import AttrDict from tqdm import tqdm from copy import deepcopy from data.image import img_to_task, task_to_img from data.celeba import CelebA from utils.misc import load_module, logmeanexp from utils.paths import results_path, evalsets_path from utils.log import get_logger, RunningAverage def main(): parser = argparse.ArgumentParser() parser.add_argument('--mode', choices=['train', 'eval', 'plot', 'ensemble'], default='train') parser.add_argument('--expid', type=str, default='trial') parser.add_argument('--resume', action='store_true', default=False) parser.add_argument('--gpu', type=str, default='0') parser.add_argument('--max_num_points', type=int, default=200) parser.add_argument('--model', type=str, default='cnp') parser.add_argument('--train_batch_size', type=int, default=100) parser.add_argument('--train_num_samples', type=int, default=4) parser.add_argument('--lr', type=float, default=5e-4) parser.add_argument('--num_epochs', type=int, default=200) parser.add_argument('--eval_freq', type=int, default=10) parser.add_argument('--save_freq', type=int, default=10) parser.add_argument('--eval_seed', type=int, default=42) parser.add_argument('--eval_batch_size', type=int, default=16) parser.add_argument('--eval_num_samples', type=int, default=50) parser.add_argument('--eval_logfile', type=str, default=None) parser.add_argument('--plot_seed', type=int, default=None) parser.add_argument('--plot_batch_size', type=int, default=16) parser.add_argument('--plot_num_samples', type=int, default=30) parser.add_argument('--plot_num_ctx', type=int, default=100) # OOD settings parser.add_argument('--t_noise', type=float, default=None) args = parser.parse_args() os.environ['CUDA_VISIBLE_DEVICES'] = args.gpu model_cls = getattr(load_module(f'models/{args.model}.py'), args.model.upper()) with open(f'configs/celeba/{args.model}.yaml', 'r') as f: config = yaml.safe_load(f) model = model_cls(**config).cuda() args.root = osp.join(results_path, 'celeba', args.model, args.expid) if args.mode == 'train': train(args, model) elif args.mode == 'eval': eval(args, model) elif args.mode == 'plot': plot(args, model) elif args.mode == 'ensemble': ensemble(args, model) def train(args, model): if not osp.isdir(args.root): os.makedirs(args.root) with open(osp.join(args.root, 'args.yaml'), 'w') as f: yaml.dump(args.__dict__, f) train_ds = CelebA(train=True) eval_ds = CelebA(train=False) train_loader = torch.utils.data.DataLoader(train_ds, batch_size=args.train_batch_size, shuffle=True, num_workers=4) #eval_loader = torch.utils.data.DataLoader(eval_ds, # batch_size=args.eval_batch_size, # shuffle=False, num_workers=4) optimizer = torch.optim.Adam(model.parameters(), lr=args.lr) scheduler = torch.optim.lr_scheduler.CosineAnnealingLR( optimizer, T_max=len(train_loader)*args.num_epochs) if args.resume: ckpt = torch.load(osp.join(args.root, 'ckpt.tar')) model.load_state_dict(ckpt.model) optimizer.load_state_dict(ckpt.optimizer) scheduler.load_state_dict(ckpt.scheduler) logfilename = ckpt.logfilename start_epoch = ckpt.epoch else: logfilename = osp.join(args.root, 'train_{}.log'.format( time.strftime('%Y%m%d-%H%M'))) start_epoch = 1 logger = get_logger(logfilename) ravg = RunningAverage() if not args.resume: logger.info('Total number of parameters: {}\n'.format( sum(p.numel() for p in model.parameters()))) for epoch in range(start_epoch, args.num_epochs+1): model.train() for (x, _) in tqdm(train_loader): batch = img_to_task(x, max_num_points=args.max_num_points, device='cuda') optimizer.zero_grad() outs = model(batch, num_samples=args.train_num_samples) outs.loss.backward() optimizer.step() scheduler.step() for key, val in outs.items(): ravg.update(key, val) line = f'{args.model}:{args.expid} epoch {epoch} ' line += f'lr {optimizer.param_groups[0]["lr"]:.3e} ' line += ravg.info() logger.info(line) if epoch % args.eval_freq == 0: logger.info(eval(args, model) + '\n') ravg.reset() if epoch % args.save_freq == 0 or epoch == args.num_epochs: ckpt = AttrDict() ckpt.model = model.state_dict() ckpt.optimizer = optimizer.state_dict() ckpt.scheduler = scheduler.state_dict() ckpt.logfilename = logfilename ckpt.epoch = epoch + 1 torch.save(ckpt, osp.join(args.root, 'ckpt.tar')) args.mode = 'eval' eval(args, model) def gen_evalset(args): torch.manual_seed(args.eval_seed) torch.cuda.manual_seed(args.eval_seed) eval_ds = CelebA(train=False) eval_loader = torch.utils.data.DataLoader(eval_ds, batch_size=args.eval_batch_size, shuffle=False, num_workers=4) batches = [] for x, _ in tqdm(eval_loader): batches.append(img_to_task(x, t_noise=args.t_noise, max_num_points=args.max_num_points)) torch.manual_seed(time.time()) torch.cuda.manual_seed(time.time()) path = osp.join(evalsets_path, 'celeba') if not osp.isdir(path): os.makedirs(path) filename = 'no_noise.tar' if args.t_noise is None else \ f'{args.t_noise}.tar' torch.save(batches, osp.join(path, filename)) def eval(args, model): if args.mode == 'eval': ckpt = torch.load(osp.join(args.root, 'ckpt.tar')) model.load_state_dict(ckpt.model) if args.eval_logfile is None: eval_logfile = f'eval' if args.t_noise is not None: eval_logfile += f'_{args.t_noise}' eval_logfile += '.log' else: eval_logfile = args.eval_logfile filename = osp.join(args.root, eval_logfile) logger = get_logger(filename, mode='w') else: logger = None path = osp.join(evalsets_path, 'celeba') if not osp.isdir(path): os.makedirs(path) filename = 'no_noise.tar' if args.t_noise is None else \ f'{args.t_noise}.tar' if not osp.isfile(osp.join(path, filename)): print('generating evaluation sets...') gen_evalset(args) eval_batches = torch.load(osp.join(path, filename)) torch.manual_seed(args.eval_seed) torch.cuda.manual_seed(args.eval_seed) ravg = RunningAverage() model.eval() with torch.no_grad(): for batch in tqdm(eval_batches): for key, val in batch.items(): batch[key] = val.cuda() outs = model(batch, num_samples=args.eval_num_samples) for key, val in outs.items(): ravg.update(key, val) torch.manual_seed(time.time()) torch.cuda.manual_seed(time.time()) line = f'{args.model}:{args.expid} ' if args.t_noise is not None: line += f'tn {args.t_noise} ' line += ravg.info() if logger is not None: logger.info(line) return line def ensemble(args, model): num_runs = 5 models = [] for i in range(num_runs): model_ = deepcopy(model) ckpt = torch.load(osp.join(results_path, 'celeba', args.model, f'run{i+1}', 'ckpt.tar')) model_.load_state_dict(ckpt['model']) model_.cuda() model_.eval() models.append(model_) path = osp.join(evalsets_path, 'celeba') if not osp.isdir(path): os.makedirs(path) filename = 'no_noise.tar' if args.t_noise is None else \ f'{args.t_noise}.tar' if not osp.isfile(osp.join(path, filename)): print('generating evaluation sets...') gen_evalset(args) eval_batches = torch.load(osp.join(path, filename)) ravg = RunningAverage() with torch.no_grad(): for batch in tqdm(eval_batches): for key, val in batch.items(): batch[key] = val.cuda() ctx_ll = [] tar_ll = [] for model in models: outs = model(batch, num_samples=args.eval_num_samples, reduce_ll=False) ctx_ll.append(outs.ctx_ll) tar_ll.append(outs.tar_ll) if ctx_ll[0].dim() == 2: ctx_ll = torch.stack(ctx_ll) tar_ll = torch.stack(tar_ll) else: ctx_ll = torch.cat(ctx_ll) tar_ll = torch.cat(tar_ll) ctx_ll = logmeanexp(ctx_ll).mean() tar_ll = logmeanexp(tar_ll).mean() ravg.update('ctx_ll', ctx_ll) ravg.update('tar_ll', tar_ll) torch.manual_seed(time.time()) torch.cuda.manual_seed(time.time()) filename = f'ensemble' if args.t_noise is not None: filename += f'_{args.t_noise}' filename += '.log' logger = get_logger(osp.join(results_path, 'celeba', args.model, filename), mode='w') logger.info(ravg.info()) if __name__ == '__main__': main()
9,536
31.328814
96
py
bnp
bnp-master/regression/models/anp.py
import torch import torch.nn as nn from torch.distributions import kl_divergence from attrdict import AttrDict from utils.misc import stack, logmeanexp from utils.sampling import sample_subset from models.modules import CrossAttnEncoder, PoolingEncoder, Decoder class ANP(nn.Module): def __init__(self, dim_x=1, dim_y=1, dim_hid=128, dim_lat=128, enc_v_depth=4, enc_qk_depth=2, enc_pre_depth=4, enc_post_depth=2, dec_depth=3): super().__init__() self.denc = CrossAttnEncoder( dim_x=dim_x, dim_y=dim_y, dim_hid=dim_hid, v_depth=enc_v_depth, qk_depth=enc_qk_depth) self.lenc = PoolingEncoder( dim_x=dim_x, dim_y=dim_y, dim_hid=dim_hid, dim_lat=dim_lat, self_attn=True, pre_depth=enc_pre_depth, post_depth=enc_post_depth) self.dec = Decoder( dim_x=dim_x, dim_y=dim_y, dim_enc=dim_hid+dim_lat, dim_hid=dim_hid, depth=dec_depth) def predict(self, xc, yc, xt, z=None, num_samples=None): theta = stack(self.denc(xc, yc, xt), num_samples) if z is None: pz = self.lenc(xc, yc) z = pz.rsample() if num_samples is None \ else pz.rsample([num_samples]) z = stack(z, xt.shape[-2], -2) encoded = torch.cat([theta, z], -1) return self.dec(encoded, stack(xt, num_samples)) def forward(self, batch, num_samples=None, reduce_ll=True): outs = AttrDict() if self.training: pz = self.lenc(batch.xc, batch.yc) qz = self.lenc(batch.x, batch.y) z = qz.rsample() if num_samples is None else \ qz.rsample([num_samples]) py = self.predict(batch.xc, batch.yc, batch.x, z=z, num_samples=num_samples) if num_samples > 1: # K * B * N recon = py.log_prob(stack(batch.y, num_samples)).sum(-1) # K * B log_qz = qz.log_prob(z).sum(-1) log_pz = pz.log_prob(z).sum(-1) # K * B log_w = recon.sum(-1) + log_pz - log_qz outs.loss = -logmeanexp(log_w).mean() / batch.x.shape[-2] else: outs.recon = py.log_prob(batch.y).sum(-1).mean() outs.kld = kl_divergence(qz, pz).sum(-1).mean() outs.loss = -outs.recon + outs.kld / batch.x.shape[-2] else: py = self.predict(batch.xc, batch.yc, batch.x, num_samples=num_samples) if num_samples is None: ll = py.log_prob(batch.y).sum(-1) else: y = torch.stack([batch.y]*num_samples) if reduce_ll: ll = logmeanexp(py.log_prob(y).sum(-1)) else: ll = py.log_prob(y).sum(-1) num_ctx = batch.xc.shape[-2] if reduce_ll: outs.ctx_ll = ll[...,:num_ctx].mean() outs.tar_ll = ll[...,num_ctx:].mean() else: outs.ctx_ll = ll[...,:num_ctx] outs.tar_ll = ll[...,num_ctx:] return outs
3,447
32.153846
83
py
bnp
bnp-master/regression/models/cnp.py
import torch import torch.nn as nn from attrdict import AttrDict from models.modules import PoolingEncoder, Decoder class CNP(nn.Module): def __init__(self, dim_x=1, dim_y=1, dim_hid=128, enc_pre_depth=4, enc_post_depth=2, dec_depth=3): super().__init__() self.enc1 = PoolingEncoder( dim_x=dim_x, dim_y=dim_y, dim_hid=dim_hid, pre_depth=enc_pre_depth, post_depth=enc_post_depth) self.enc2 = PoolingEncoder( dim_x=dim_x, dim_y=dim_y, dim_hid=dim_hid, pre_depth=enc_pre_depth, post_depth=enc_post_depth) self.dec = Decoder( dim_x=dim_x, dim_y=dim_y, dim_enc=2*dim_hid, dim_hid=dim_hid, depth=dec_depth) def predict(self, xc, yc, xt, num_samples=None): encoded = torch.cat([self.enc1(xc, yc), self.enc2(xc, yc)], -1) encoded = torch.stack([encoded]*xt.shape[-2], -2) return self.dec(encoded, xt) def forward(self, batch, num_samples=None, reduce_ll=True): outs = AttrDict() py = self.predict(batch.xc, batch.yc, batch.x) ll = py.log_prob(batch.y).sum(-1) if self.training: outs.loss = -ll.mean() else: num_ctx = batch.xc.shape[-2] if reduce_ll: outs.ctx_ll = ll[...,:num_ctx].mean() outs.tar_ll = ll[...,num_ctx:].mean() else: outs.ctx_ll = ll[...,:num_ctx] outs.tar_ll = ll[...,num_ctx:] return outs
1,748
27.672131
71
py
bnp
bnp-master/regression/models/modules.py
import torch import torch.nn as nn import torch.nn.functional as F from torch.distributions import Normal from models.attention import MultiHeadAttn, SelfAttn __all__ = ['PoolingEncoder', 'CrossAttnEncoder', 'Decoder'] def build_mlp(dim_in, dim_hid, dim_out, depth): modules = [nn.Linear(dim_in, dim_hid), nn.ReLU(True)] for _ in range(depth-2): modules.append(nn.Linear(dim_hid, dim_hid)) modules.append(nn.ReLU(True)) modules.append(nn.Linear(dim_hid, dim_out)) return nn.Sequential(*modules) class PoolingEncoder(nn.Module): def __init__(self, dim_x=1, dim_y=1, dim_hid=128, dim_lat=None, self_attn=False, pre_depth=4, post_depth=2): super().__init__() self.use_lat = dim_lat is not None self.net_pre = build_mlp(dim_x+dim_y, dim_hid, dim_hid, pre_depth) \ if not self_attn else \ nn.Sequential( build_mlp(dim_x+dim_y, dim_hid, dim_hid, pre_depth-2), nn.ReLU(True), SelfAttn(dim_hid, dim_hid)) self.net_post = build_mlp(dim_hid, dim_hid, 2*dim_lat if self.use_lat else dim_hid, post_depth) def forward(self, xc, yc, mask=None): out = self.net_pre(torch.cat([xc, yc], -1)) if mask is None: out = out.mean(-2) else: mask = mask.to(xc.device) out = (out * mask.unsqueeze(-1)).sum(-2) / \ (mask.sum(-1, keepdim=True).detach() + 1e-5) if self.use_lat: mu, sigma = self.net_post(out).chunk(2, -1) sigma = 0.1 + 0.9 * torch.sigmoid(sigma) return Normal(mu, sigma) else: return self.net_post(out) class CrossAttnEncoder(nn.Module): def __init__(self, dim_x=1, dim_y=1, dim_hid=128, dim_lat=None, self_attn=True, v_depth=4, qk_depth=2): super().__init__() self.use_lat = dim_lat is not None if not self_attn: self.net_v = build_mlp(dim_x+dim_y, dim_hid, dim_hid, v_depth) else: self.net_v = build_mlp(dim_x+dim_y, dim_hid, dim_hid, v_depth-2) self.self_attn = SelfAttn(dim_hid, dim_hid) self.net_qk = build_mlp(dim_x, dim_hid, dim_hid, qk_depth) self.attn = MultiHeadAttn(dim_hid, dim_hid, dim_hid, 2*dim_lat if self.use_lat else dim_hid) def forward(self, xc, yc, xt, mask=None): q, k = self.net_qk(xt), self.net_qk(xc) v = self.net_v(torch.cat([xc, yc], -1)) if hasattr(self, 'self_attn'): v = self.self_attn(v, mask=mask) out = self.attn(q, k, v, mask=mask) if self.use_lat: mu, sigma = out.chunk(2, -1) sigma = 0.1 + 0.9 * torch.sigmoid(sigma) return Normal(mu, sigma) else: return out class Decoder(nn.Module): def __init__(self, dim_x=1, dim_y=1, dim_enc=128, dim_hid=128, depth=3): super().__init__() self.fc = nn.Linear(dim_x+dim_enc, dim_hid) self.dim_hid = dim_hid modules = [nn.ReLU(True)] for _ in range(depth-2): modules.append(nn.Linear(dim_hid, dim_hid)) modules.append(nn.ReLU(True)) modules.append(nn.Linear(dim_hid, 2*dim_y)) self.mlp = nn.Sequential(*modules) def add_ctx(self, dim_ctx): self.dim_ctx = dim_ctx self.fc_ctx = nn.Linear(dim_ctx, self.dim_hid, bias=False) def forward(self, encoded, x, ctx=None): packed = torch.cat([encoded, x], -1) hid = self.fc(packed) if ctx is not None: hid = hid + self.fc_ctx(ctx) out = self.mlp(hid) mu, sigma = out.chunk(2, -1) sigma = 0.1 + 0.9 * F.softplus(sigma) return Normal(mu, sigma)
3,867
33.535714
78
py
bnp
bnp-master/regression/models/banp.py
import torch import torch.nn as nn from attrdict import AttrDict from models.canp import CANP from utils.misc import stack, logmeanexp from utils.sampling import sample_with_replacement as SWR, sample_subset class BANP(CANP): def __init__(self, *args, **kwargs): super().__init__(*args, **kwargs) self.dec.add_ctx(2*kwargs['dim_hid']) def encode(self, xc, yc, xt, mask=None): theta1 = self.enc1(xc, yc, xt) theta2 = self.enc2(xc, yc) encoded = torch.cat([theta1, torch.stack([theta2]*xt.shape[-2], -2)], -1) return encoded def predict(self, xc, yc, xt, num_samples=None, return_base=False): with torch.no_grad(): bxc, byc = SWR(xc, yc, num_samples=num_samples) sxc, syc = stack(xc, num_samples), stack(yc, num_samples) encoded = self.encode(bxc, byc, sxc) py_res = self.dec(encoded, sxc) mu, sigma = py_res.mean, py_res.scale res = SWR((syc - mu)/sigma).detach() res = (res - res.mean(-2, keepdim=True)) bxc = sxc byc = mu + sigma * res encoded_base = self.encode(xc, yc, xt) sxt = stack(xt, num_samples) encoded_bs = self.encode(bxc, byc, sxt) py = self.dec(stack(encoded_base, num_samples), sxt, ctx=encoded_bs) if self.training or return_base: py_base = self.dec(encoded_base, xt) return py_base, py else: return py def forward(self, batch, num_samples=None, reduce_ll=True): outs = AttrDict() def compute_ll(py, y): ll = py.log_prob(y).sum(-1) if ll.dim() == 3 and reduce_ll: ll = logmeanexp(ll) return ll if self.training: py_base, py = self.predict(batch.xc, batch.yc, batch.x, num_samples=num_samples) outs.ll_base = compute_ll(py_base, batch.y).mean() outs.ll = compute_ll(py, batch.y).mean() outs.loss = -outs.ll_base - outs.ll else: py = self.predict(batch.xc, batch.yc, batch.x, num_samples=num_samples) ll = compute_ll(py, batch.y) num_ctx = batch.xc.shape[-2] if reduce_ll: outs.ctx_ll = ll[...,:num_ctx].mean() outs.tar_ll = ll[...,num_ctx:].mean() else: outs.ctx_ll = ll[...,:num_ctx] outs.tar_ll = ll[...,num_ctx:] return outs
2,555
31.35443
72
py
bnp
bnp-master/regression/models/canp.py
import torch import torch.nn as nn from attrdict import AttrDict from models.modules import CrossAttnEncoder, Decoder, PoolingEncoder class CANP(nn.Module): def __init__(self, dim_x=1, dim_y=1, dim_hid=128, enc_v_depth=4, enc_qk_depth=2, enc_pre_depth=4, enc_post_depth=2, dec_depth=3): super().__init__() self.enc1 = CrossAttnEncoder( dim_x=dim_x, dim_y=dim_y, dim_hid=dim_hid, v_depth=enc_v_depth, qk_depth=enc_qk_depth) self.enc2 = PoolingEncoder( dim_x=dim_x, dim_y=dim_y, dim_hid=dim_hid, self_attn=True, pre_depth=enc_pre_depth, post_depth=enc_post_depth) self.dec = Decoder( dim_x=dim_x, dim_y=dim_y, dim_enc=2*dim_hid, dim_hid=dim_hid, depth=dec_depth) def predict(self, xc, yc, xt, num_samples=None): theta1 = self.enc1(xc, yc, xt) theta2 = self.enc2(xc, yc) encoded = torch.cat([theta1, torch.stack([theta2]*xt.shape[-2], -2)], -1) return self.dec(encoded, xt) def forward(self, batch, num_samples=None, reduce_ll=True): outs = AttrDict() py = self.predict(batch.xc, batch.yc, batch.x) ll = py.log_prob(batch.y).sum(-1) if self.training: outs.loss = -ll.mean() else: num_ctx = batch.xc.shape[-2] if reduce_ll: outs.ctx_ll = ll[...,:num_ctx].mean() outs.tar_ll = ll[...,num_ctx:].mean() else: outs.ctx_ll = ll[...,:num_ctx] outs.tar_ll = ll[...,num_ctx:] return outs
1,886
27.590909
68
py
bnp
bnp-master/regression/models/bnp.py
import torch import torch.nn as nn from attrdict import AttrDict from models.cnp import CNP from utils.misc import stack, logmeanexp from utils.sampling import sample_with_replacement as SWR, sample_subset class BNP(CNP): def __init__(self, *args, **kwargs): super().__init__(*args, **kwargs) self.dec.add_ctx(2*kwargs['dim_hid']) def encode(self, xc, yc, xt, mask=None): encoded = torch.cat([ self.enc1(xc, yc, mask=mask), self.enc2(xc, yc, mask=mask)], -1) return stack(encoded, xt.shape[-2], -2) def predict(self, xc, yc, xt, num_samples=None, return_base=False): with torch.no_grad(): bxc, byc = SWR(xc, yc, num_samples=num_samples) sxc, syc = stack(xc, num_samples), stack(yc, num_samples) encoded = self.encode(bxc, byc, sxc) py_res = self.dec(encoded, sxc) mu, sigma = py_res.mean, py_res.scale res = SWR((syc - mu)/sigma).detach() res = (res - res.mean(-2, keepdim=True)) bxc = sxc byc = mu + sigma * res encoded_base = self.encode(xc, yc, xt) sxt = stack(xt, num_samples) encoded_bs = self.encode(bxc, byc, sxt) py = self.dec(stack(encoded_base, num_samples), sxt, ctx=encoded_bs) if self.training or return_base: py_base = self.dec(encoded_base, xt) return py_base, py else: return py def forward(self, batch, num_samples=None, reduce_ll=True): outs = AttrDict() def compute_ll(py, y): ll = py.log_prob(y).sum(-1) if ll.dim() == 3 and reduce_ll: ll = logmeanexp(ll) return ll if self.training: py_base, py = self.predict(batch.xc, batch.yc, batch.x, num_samples=num_samples) outs.ll_base = compute_ll(py_base, batch.y).mean() outs.ll = compute_ll(py, batch.y).mean() outs.loss = -outs.ll_base - outs.ll else: py = self.predict(batch.xc, batch.yc, batch.x, num_samples=num_samples) ll = compute_ll(py, batch.y) num_ctx = batch.xc.shape[-2] if reduce_ll: outs.ctx_ll = ll[...,:num_ctx].mean() outs.tar_ll = ll[...,num_ctx:].mean() else: outs.ctx_ll = ll[...,:num_ctx] outs.tar_ll = ll[...,num_ctx:] return outs
2,527
31.410256
72
py
bnp
bnp-master/regression/models/np.py
import torch import torch.nn as nn from torch.distributions import kl_divergence from attrdict import AttrDict from utils.misc import stack, logmeanexp from utils.sampling import sample_subset from models.modules import PoolingEncoder, Decoder class NP(nn.Module): def __init__(self, dim_x=1, dim_y=1, dim_hid=128, dim_lat=128, enc_pre_depth=4, enc_post_depth=2, dec_depth=3): super().__init__() self.denc = PoolingEncoder( dim_x=dim_x, dim_y=dim_y, dim_hid=dim_hid, pre_depth=enc_pre_depth, post_depth=enc_post_depth) self.lenc = PoolingEncoder( dim_x=dim_x, dim_y=dim_y, dim_hid=dim_hid, dim_lat=dim_lat, pre_depth=enc_pre_depth, post_depth=enc_post_depth) self.dec = Decoder( dim_x=dim_x, dim_y=dim_y, dim_enc=dim_hid+dim_lat, dim_hid=dim_hid, depth=dec_depth) def predict(self, xc, yc, xt, z=None, num_samples=None): theta = stack(self.denc(xc, yc), num_samples) if z is None: pz = self.lenc(xc, yc) z = pz.rsample() if num_samples is None \ else pz.rsample([num_samples]) encoded = torch.cat([theta, z], -1) encoded = stack(encoded, xt.shape[-2], -2) return self.dec(encoded, stack(xt, num_samples)) def forward(self, batch, num_samples=None, reduce_ll=True): outs = AttrDict() if self.training: pz = self.lenc(batch.xc, batch.yc) qz = self.lenc(batch.x, batch.y) z = qz.rsample() if num_samples is None else \ qz.rsample([num_samples]) py = self.predict(batch.xc, batch.yc, batch.x, z=z, num_samples=num_samples) if num_samples > 1: # K * B * N recon = py.log_prob(stack(batch.y, num_samples)).sum(-1) # K * B log_qz = qz.log_prob(z).sum(-1) log_pz = pz.log_prob(z).sum(-1) # K * B log_w = recon.sum(-1) + log_pz - log_qz outs.loss = -logmeanexp(log_w).mean() / batch.x.shape[-2] else: outs.recon = py.log_prob(batch.y).sum(-1).mean() outs.kld = kl_divergence(qz, pz).sum(-1).mean() outs.loss = -outs.recon + outs.kld / batch.x.shape[-2] else: py = self.predict(batch.xc, batch.yc, batch.x, num_samples=num_samples) if num_samples is None: ll = py.log_prob(batch.y).sum(-1) else: y = torch.stack([batch.y]*num_samples) if reduce_ll: ll = logmeanexp(py.log_prob(y).sum(-1)) else: ll = py.log_prob(y).sum(-1) num_ctx = batch.xc.shape[-2] if reduce_ll: outs.ctx_ll = ll[...,:num_ctx].mean() outs.tar_ll = ll[...,num_ctx:].mean() else: outs.ctx_ll = ll[...,:num_ctx] outs.tar_ll = ll[...,num_ctx:] return outs
3,352
33.214286
83
py