modeling_miewid.py

import torch
import torch.nn as nn
import torch.nn.functional as F

import timm

from transformers import PreTrainedModel

from .heads import ArcMarginProduct, ElasticArcFace, ArcFaceSubCenterDynamic
from .configuration_miewid import MiewIdNetConfig

def weights_init_kaiming(m):
    classname = m.__class__.__name__
    if classname.find('Linear') != -1:
        nn.init.kaiming_normal_(m.weight, a=0, mode='fan_out')
        nn.init.constant_(m.bias, 0.0)
    elif classname.find('Conv') != -1:
        nn.init.kaiming_normal_(m.weight, a=0, mode='fan_in')
        if m.bias is not None:
            nn.init.constant_(m.bias, 0.0)
    elif classname.find('BatchNorm') != -1:
        if m.affine:
            nn.init.constant_(m.weight, 1.0)
            nn.init.constant_(m.bias, 0.0)


def weights_init_classifier(m):
    classname = m.__class__.__name__
    if classname.find('Linear') != -1:
        nn.init.normal_(m.weight, std=0.001)
        if m.bias:
            nn.init.constant_(m.bias, 0.0)

class GeM(nn.Module):
    def __init__(self, p=3, eps=1e-6):
        super(GeM, self).__init__()
        self.p = nn.Parameter(torch.ones(1)*p)
        self.eps = eps

    def forward(self, x):
        return self.gem(x, p=self.p, eps=self.eps)
        
    def gem(self, x, p=3, eps=1e-6):
        return F.avg_pool2d(x.clamp(min=eps).pow(p), (x.size(-2), x.size(-1))).pow(1./p)
        
    def __repr__(self):
        return self.__class__.__name__ + \
                '(' + 'p=' + '{:.4f}'.format(self.p.data.tolist()[0]) + \
                ', ' + 'eps=' + str(self.eps) + ')'

class MiewIdNet(PreTrainedModel):
    config_class = MiewIdNetConfig
    
    def __init__(self, config):
        """
        """
        super(MiewIdNet, self).__init__(config)
        print('Building Model Backbone for {} model'.format(config.model_name))
        print('config.model_name', config.model_name)

        n_classes=config.n_classes
        model_name=config.model_name
        use_fc=False
        fc_dim=512
        dropout=0.0
        loss_module=config.loss_module
        s=30.0
        margin=0.50
        ls_eps=0.0
        theta_zero=0.785
        pretrained=True
        margins=config.k
        k=config.k

        print('model_name', model_name)

        self.backbone = timm.create_model(model_name, pretrained=pretrained, num_classes=0)
        final_in_features = 2152#self.backbone.classifier.in_features

        print('final_in_features', final_in_features)
        
        # self.backbone.classifier = nn.Identity()
        self.backbone.global_pool =  GeM()#nn.Identity()
        
        # self.pooling =  GeM()
        self.bn = nn.BatchNorm1d(final_in_features)
        self.use_fc = use_fc
        if use_fc:
            self.dropout = nn.Dropout(p=dropout)
            self.bn = nn.BatchNorm1d(fc_dim)
            self.bn.bias.requires_grad_(False)
            self.fc = nn.Linear(final_in_features, n_classes, bias = False)            
            self.bn.apply(weights_init_kaiming)
            self.fc.apply(weights_init_classifier)
            final_in_features = fc_dim

        self.loss_module = loss_module
        if loss_module == 'arcface':
            self.final = ElasticArcFace(final_in_features, n_classes,
                                          s=s, m=margin)
        elif loss_module == 'arcface_subcenter_dynamic':
            if margins is None:
                margins = [0.3] * n_classes
            print(final_in_features, n_classes)
            self.final = ArcFaceSubCenterDynamic(
                embedding_dim=final_in_features, 
                output_classes=n_classes, 
                margins=margins,
                s=s,
                k=k )
        # elif loss_module == 'cosface':
        #     self.final = AddMarginProduct(final_in_features, n_classes, s=s, m=margin)
        # elif loss_module == 'adacos':
        #     self.final = AdaCos(final_in_features, n_classes, m=margin, theta_zero=theta_zero)
        else:
            self.final = nn.Linear(final_in_features, n_classes)

    def _init_params(self):
        nn.init.xavier_normal_(self.fc.weight)
        nn.init.constant_(self.fc.bias, 0)
        nn.init.constant_(self.bn.weight, 1)
        nn.init.constant_(self.bn.bias, 0)

    def forward(self, x, label=None):
        feature = self.extract_feat(x)

        return feature
        # if not self.training:
        #     return feature
        # else:
        #     assert label is not None
        # if self.loss_module in ('arcface', 'arcface_subcenter_dynamic'):
        #     logits = self.final(feature, label)
        # else:
        #     logits = self.final(feature)

        # return logits

    def extract_feat(self, x):
        batch_size = x.shape[0]
        x = self.backbone(x).view(batch_size, -1)
        # x = self.pooling(x).view(batch_size, -1)
        x = self.bn(x)
        if self.use_fc:
            x1 = self.dropout(x)
            x1 = self.bn(x1)
            x1 = self.fc(x1)
    
        return x

    def extract_logits(self, x, label=None):
        feature = self.extract_feat(x)
        assert label is not None
        if self.loss_module in ('arcface', 'arcface_subcenter_dynamic'):
            logits = self.final(feature, label)
        else:
            logits = self.final(feature)
        
        return logits