model.py

import torch
import torch.nn as nn
from operations import *
from torch.autograd import Variable
from utils import drop_path
import utils
from tqdm import tqdm
import numpy as np
import torch.nn.functional as F
from scipy.stats import mode
class Cell(nn.Module):

  def __init__(self, genotype, C_prev_prev, C_prev, C, reduction, reduction_prev):
    super(Cell, self).__init__()
    print(C_prev_prev, C_prev, C)

    if reduction_prev:
      self.preprocess0 = FactorizedReduce(C_prev_prev, C)
    else:
      self.preprocess0 = ReLUConvBN(C_prev_prev, C, 1, 1, 0)
    self.preprocess1 = ReLUConvBN(C_prev, C, 1, 1, 0)
    
    if reduction:
      op_names, indices = zip(*genotype.reduce)
      concat = genotype.reduce_concat
    else:
      op_names, indices = zip(*genotype.normal)
      concat = genotype.normal_concat
    self._compile(C, op_names, indices, concat, reduction)

  def _compile(self, C, op_names, indices, concat, reduction):
    assert len(op_names) == len(indices)
    self._steps = len(op_names) // 2
    self._concat = concat
    self.multiplier = len(concat)

    self._ops = nn.ModuleList()
    for name, index in zip(op_names, indices):
      stride = 2 if reduction and index < 2 else 1
      op = OPS[name](C, stride, True)
      self._ops += [op]
    self._indices = indices

  def forward(self, s0, s1, drop_prob):
    s0 = self.preprocess0(s0)
    s1 = self.preprocess1(s1)

    states = [s0, s1]
    for i in range(self._steps):
      h1 = states[self._indices[2*i]]
      h2 = states[self._indices[2*i+1]]
      op1 = self._ops[2*i]
      op2 = self._ops[2*i+1]
      h1 = op1(h1)
      h2 = op2(h2)
      if self.training and drop_prob > 0.:
        if not isinstance(op1, Identity):
          h1 = drop_path(h1, drop_prob)
        if not isinstance(op2, Identity):
          h2 = drop_path(h2, drop_prob)
      s = h1 + h2
      states += [s]
    return torch.cat([states[i] for i in self._concat], dim=1)


class AuxiliaryHeadCIFAR(nn.Module):

  def __init__(self, C, num_classes):
    """assuming input size 8x8"""
    super(AuxiliaryHeadCIFAR, self).__init__()
    self.features = nn.Sequential(
      nn.ReLU(inplace=True),
      nn.AvgPool2d(5, stride=3, padding=0, count_include_pad=False), # image size = 2 x 2
      nn.Conv2d(C, 128, 1, bias=False),
      nn.BatchNorm2d(128),
      nn.ReLU(inplace=True),
      nn.Conv2d(128, 768, 2, bias=False),
      nn.BatchNorm2d(768),
      nn.ReLU(inplace=True)
    )
    self.classifier = nn.Linear(768, num_classes)

  def forward(self, x):
    x = self.features(x)
    x = self.classifier(x.view(x.size(0),-1))
    return x


class AuxiliaryHeadImageNet(nn.Module):

  def __init__(self, C, num_classes):
    """assuming input size 14x14"""
    super(AuxiliaryHeadImageNet, self).__init__()
    self.features = nn.Sequential(
      nn.ReLU(inplace=True),
      nn.AvgPool2d(5, stride=2, padding=0, count_include_pad=False),
      nn.Conv2d(C, 128, 1, bias=False),
      nn.BatchNorm2d(128),
      nn.ReLU(inplace=True),
      nn.Conv2d(128, 768, 2, bias=False),
      # NOTE: This batchnorm was omitted in my earlier implementation due to a typo.
      # Commenting it out for consistency with the experiments in the paper.
      # nn.BatchNorm2d(768),
      nn.ReLU(inplace=True)
    )
    self.classifier = nn.Linear(768, num_classes)

  def forward(self, x):
    x = self.features(x)
    x = self.classifier(x.view(x.size(0),-1))
    return x


class NetworkCIFAR(nn.Module):

  def __init__(self, C, num_classes, layers, auxiliary, genotype):
    super(NetworkCIFAR, self).__init__()
    self._layers = layers
    self._auxiliary = auxiliary

    stem_multiplier = 3
    C_curr = stem_multiplier*C
    self.stem = nn.Sequential(
      nn.Conv2d(3, C_curr, 3, padding=1, bias=False),
      nn.BatchNorm2d(C_curr)
    )
    
    C_prev_prev, C_prev, C_curr = C_curr, C_curr, C
    self.cells = nn.ModuleList()
    reduction_prev = False
    for i in range(layers):
      if i in [layers//3, 2*layers//3]:
        C_curr *= 2
        reduction = True
      else:
        reduction = False
      cell = Cell(genotype, C_prev_prev, C_prev, C_curr, reduction, reduction_prev)
      reduction_prev = reduction
      self.cells += [cell]
      C_prev_prev, C_prev = C_prev, cell.multiplier*C_curr
      if i == 2*layers//3:
        C_to_auxiliary = C_prev

    if auxiliary:
      self.auxiliary_head = AuxiliaryHeadCIFAR(C_to_auxiliary, num_classes)
    self.global_pooling = nn.AdaptiveAvgPool2d(1)
    self.classifier = nn.Linear(C_prev, num_classes)

  def forward(self, input):
    logits_aux = None
    s0 = s1 = self.stem(input)
    for i, cell in enumerate(self.cells):
      s0, s1 = s1, cell(s0, s1, self.drop_path_prob)
      if i == 2*self._layers//3:
        if self._auxiliary and self.training:
          logits_aux = self.auxiliary_head(s1)
    out = self.global_pooling(s1)
    logits = self.classifier(out.view(out.size(0),-1))
    return logits, logits_aux


class NetworkImageNet(nn.Module):

  def __init__(self, C, num_classes, layers, auxiliary, genotype):
    super(NetworkImageNet, self).__init__()
    self._layers = layers
    self._auxiliary = auxiliary

    self.stem0 = nn.Sequential(
      nn.Conv2d(3, C // 2, kernel_size=3, stride=2, padding=1, bias=False),
      nn.BatchNorm2d(C // 2),
      nn.ReLU(inplace=True),
      nn.Conv2d(C // 2, C, 3, stride=2, padding=1, bias=False),
      nn.BatchNorm2d(C),
    )

    self.stem1 = nn.Sequential(
      nn.ReLU(inplace=True),
      nn.Conv2d(C, C, 3, stride=2, padding=1, bias=False),
      nn.BatchNorm2d(C),
    )

    C_prev_prev, C_prev, C_curr = C, C, C

    self.cells = nn.ModuleList()
    reduction_prev = True
    for i in range(layers):
      if i in [layers // 3, 2 * layers // 3]:
        C_curr *= 2
        reduction = True
      else:
        reduction = False
      cell = Cell(genotype, C_prev_prev, C_prev, C_curr, reduction, reduction_prev)
      reduction_prev = reduction
      self.cells += [cell]
      C_prev_prev, C_prev = C_prev, cell.multiplier * C_curr
      if i == 2 * layers // 3:
        C_to_auxiliary = C_prev

    if auxiliary:
      self.auxiliary_head = AuxiliaryHeadImageNet(C_to_auxiliary, num_classes)
    self.global_pooling = nn.AvgPool2d(7)
    self.classifier = nn.Linear(C_prev, num_classes)

  def forward(self, input):
    logits_aux = None
    s0 = self.stem0(input)
    s1 = self.stem1(s0)
    for i, cell in enumerate(self.cells):
      s0, s1 = s1, cell(s0, s1, self.drop_path_prob)
      if i == 2 * self._layers // 3:
        if self._auxiliary and self.training:
          logits_aux = self.auxiliary_head(s1)
    out = self.global_pooling(s1)
    logits = self.classifier(out.view(out.size(0), -1))
    return logits, logits_aux

class NetworkKMNIST(nn.Module):

  def __init__(self, C, num_classes, layers, genotype):
    super(NetworkKMNIST, self).__init__()
    self._layers = layers

    stem_multiplier = 3
    C_curr = stem_multiplier*C
    self.stem = nn.Sequential(
      nn.Conv2d(1, C_curr, 3, padding=1, bias=False),
      nn.BatchNorm2d(C_curr)
    )
    
    C_prev_prev, C_prev, C_curr = C_curr, C_curr, C
    self.cells = nn.ModuleList()
    reduction_prev = False
    for i in range(layers):
      if i in [layers//3, 2*layers//3]:
        C_curr *= 2
        reduction = True
      else:
        reduction = False
      cell = Cell(genotype, C_prev_prev, C_prev, C_curr, reduction, reduction_prev)
      reduction_prev = reduction
      self.cells += [cell]
      C_prev_prev, C_prev = C_prev, cell.multiplier*C_curr

    self.global_pooling = nn.AdaptiveAvgPool2d(1)
    self.classifier = nn.Linear(C_prev, num_classes)

  def forward(self, input):
    logits_aux = None
    s0 = s1 = self.stem(input)
    for i, cell in enumerate(self.cells):
      s0, s1 = s1, cell(s0, s1, self.drop_path_prob)
    out = self.global_pooling(s1)
    logits = self.classifier(out.view(out.size(0),-1))
    return logits

# class EnsembleModel(nn.Module):
#     def __init__(self, models, out_size):
#         super(EnsembleModel, self).__init__()
#         self.models = list()
#         self.out_size = out_size
#         for model in models:
#           self.models.append(model)
#         self.classifier = nn.Linear(self.out_size, self.out_size)

        
#     def forward(self, x):
        
#         for i, model in enumerate(self.models):
#             out = model(x)
#             if i != 0:
#               out = out + model(x)
#         out = out/(i+1)
#         x = self.classifier(F.relu(out))
#         return x



class EnsembleModel(nn.Module):
    def __init__(self, models, dense_units, out_size):
        super(EnsembleModel, self).__init__()
        self.out_size = out_size
        self.dense_units = dense_units
        self.model_1 = models[0]
        self.model_2 = models[1]
        self.model_3 = models[2]
        for p in self.model_1.parameters():
            p.requires_grad=False
        for p in self.model_2.parameters():
            p.requires_grad=False
        for p in self.model_3.parameters():
            p.requires_grad=False
        # for model in models:
        #     for p in model.parameters():
        #       p.requires_grad=False
        #     self.models.append(model)
        # self.classifiers = []
        # for i in range(len(self.models)):
        #     self.classifiers.append(nn.Linear(self.out_size, dense_units))

        self.out_classifier = nn.Linear(self.out_size*3, self.out_size)

        
    def forward(self, x):
        outputs = list()
        outputs.append(self.model_1(x))
        outputs.append(self.model_2(x))
        outputs.append(self.model_3(x))

        # for i, model in enumerate(self.models):
        #     model_out = model(x)
        #     # class_out = self.classifiers[i].forward(model_out)
        #     outputs.append(model_out)
        out = torch.cat([out for out in outputs], dim=1)
        x = self.out_classifier(F.relu(out))
        return x

class MajorityEnsembleModel(nn.Module):
    def __init__(self, models):
        super(MajorityEnsembleModel, self).__init__()
        self.models = list()
        for model in models:
            for p in model.parameters():
              p.requires_grad=False
            self.models.append(model)
      
    def forward(self, x):
        outputs = list()

        for i, model in enumerate(self.models):
            model_out = model(x)
            model_out = model_out.cpu()
            # class_out = self.classifiers[i].forward(model_out)
            if i == 0:
                outputs = model_out
                outputs = np.array(outputs)
                print("initially:", outputs.shape)
            else:
                np.hstack((outputs, model_out))
        print(outputs.shape)
        # out = torch.cat([out for out in outputs], dim=1)
        x = np.array(mode(outputs, axis=1))
        x = torch.from_numpy(x).float().to('cuda:0')
        print(x.size())
        return x