robustdg/algorithms/csd.py

178 строки
7.3 KiB
Python

import sys
import numpy as np
import argparse
import copy
import random
import json
import torch
from torch.autograd import grad
from torch import nn, optim
from torch.nn import functional as F
from torchvision import datasets, transforms
from torchvision.utils import save_image
from torch.autograd import Variable
import torch.utils.data as data_utils
from .algo import BaseAlgo
from utils.helper import l1_dist, l2_dist, embedding_dist, cosine_similarity
class CSD(BaseAlgo):
def __init__(self, args, train_dataset, val_dataset, test_dataset, base_res_dir, post_string, cuda):
super().__init__(args, train_dataset, val_dataset, test_dataset, base_res_dir, post_string, cuda)
# H_dim as per the feature layer dimension of ResNet-18
H_dim= self.args.rep_dim
self.K, m, self.num_classes = 1, H_dim, self.args.out_classes
num_domains = self.total_domains
self.sms = torch.nn.Parameter(torch.normal(0, 1e-1, size=[self.K+1, m, self.num_classes], dtype=torch.float, device='cuda:0'), requires_grad=True)
self.sm_biases = torch.nn.Parameter(torch.normal(0, 1e-1, size=[self.K+1, self.num_classes], dtype=torch.float, device='cuda:0'), requires_grad=True)
self.embs = torch.nn.Parameter(torch.normal(mean=0., std=1e-1, size=[num_domains, self.K], dtype=torch.float, device='cuda:0'), requires_grad=True)
self.cs_wt = torch.nn.Parameter(torch.normal(mean=.1, std=1e-4, size=[], dtype=torch.float, device='cuda:0'), requires_grad=True)
self.opt= optim.SGD([
{'params': filter(lambda p: p.requires_grad, self.phi.parameters()) },
{'params': self.sms },
{'params': self.sm_biases },
{'params': self.embs },
{'params': self.cs_wt }
], lr= self.args.lr, weight_decay= 5e-4, momentum= 0.9, nesterov=True )
self.criterion = torch.nn.CrossEntropyLoss()
def forward(self, x, y, di, eval_case=0):
x = self.phi(x)
w_c, b_c = self.sms[0, :, :], self.sm_biases[0, :]
logits_common = torch.matmul(x, w_c) + b_c
if eval_case:
return logits_common
domains= di
c_wts = torch.matmul(domains, self.embs)
# B x K
batch_size = x.shape[0]
c_wts = torch.cat((torch.ones((batch_size, 1), dtype=torch.float).to(self.cuda)*self.cs_wt, c_wts), 1)
c_wts = torch.tanh(c_wts).to(self.cuda)
w_d, b_d = torch.einsum("bk,krl->brl", c_wts, self.sms), torch.einsum("bk,kl->bl", c_wts, self.sm_biases)
logits_specialized = torch.einsum("brl,br->bl", w_d, x) + b_d
specific_loss = self.criterion(logits_specialized, y)
class_loss = self.criterion(logits_common, y)
sms = self.sms
diag_tensor = torch.stack([torch.eye(self.K+1).to(self.cuda) for _ in range(self.num_classes)], dim=0)
cps = torch.stack([torch.matmul(sms[:, :, _], torch.transpose(sms[:, :, _], 0, 1)) for _ in range(self.num_classes)], dim=0)
orth_loss = torch.mean((1-diag_tensor)*(cps - diag_tensor)**2)
loss = class_loss + specific_loss + orth_loss
return loss, logits_common
def epoch_callback(self, nepoch, final=False):
if nepoch % 100 == 0:
print (self.embs, torch.norm(self.sms[0]), torch.norm(self.sms[1]))
def train(self):
self.max_epoch=-1
self.max_val_acc=0.0
for epoch in range(self.args.epochs):
if epoch ==0 or (epoch % self.args.match_interrupt == 0 and self.args.match_flag):
data_match_tensor, label_match_tensor= self.get_match_function(epoch)
penalty_csd=0
train_acc= 0.0
train_size=0
perm = torch.randperm(data_match_tensor.size(0))
data_match_tensor_split= torch.split(data_match_tensor[perm], self.args.batch_size, dim=0)
label_match_tensor_split= torch.split(label_match_tensor[perm], self.args.batch_size, dim=0)
print('Split Matched Data: ', len(data_match_tensor_split), data_match_tensor_split[0].shape, len(label_match_tensor_split))
#Batch iteration over single epoch
for batch_idx, (x_e, y_e ,d_e, idx_e) in enumerate(self.train_dataset):
# print('Batch Idx: ', batch_idx)
self.opt.zero_grad()
loss_e= torch.tensor(0.0).to(self.cuda)
x_e= x_e.to(self.cuda)
y_e= torch.argmax(y_e, dim=1).to(self.cuda)
#Forward Pass
csd_loss, out= self.forward(x_e, y_e, d_e.to(self.cuda), eval_case=0)
loss_e+= csd_loss
penalty_csd += float(loss_e)
#Backprorp
loss_e.backward(retain_graph=False)
self.opt.step()
del csd_loss
del loss_e
torch.cuda.empty_cache()
train_acc+= torch.sum(torch.argmax(out, dim=1) == y_e ).item()
train_size+= y_e.shape[0]
print('Train Loss Basic : ', penalty_csd )
print('Train Acc Env : ', 100*train_acc/train_size )
print('Done Training for epoch: ', epoch)
#Train Dataset Accuracy
self.train_acc.append( 100*train_acc/train_size )
#Val Dataset Accuracy
self.val_acc.append( self.get_test_accuracy('val') )
#Test Dataset Accuracy
self.final_acc.append( self.get_test_accuracy('test') )
#Save the model if current best epoch as per validation loss
if self.val_acc[-1] > self.max_val_acc:
self.max_val_acc=self.val_acc[-1]
self.max_epoch= epoch
self.save_model()
print('Current Best Epoch: ', self.max_epoch, ' with Test Accuracy: ', self.final_acc[self.max_epoch])
def get_test_accuracy(self, case):
#Test Env Code
test_acc= 0.0
test_size=0
if case == 'val':
dataset= self.val_dataset
elif case == 'test':
dataset= self.test_dataset
for batch_idx, (x_e, y_e ,d_e, idx_e) in enumerate(dataset):
with torch.no_grad():
x_e= x_e.to(self.cuda)
y_e= torch.argmax(y_e, dim=1).to(self.cuda)
#Forward Pass
out= self.forward(x_e, y_e, d_e.to(self.cuda), eval_case=1)
test_acc+= torch.sum( torch.argmax(out, dim=1) == y_e ).item()
test_size+= y_e.shape[0]
print(' Accuracy: ', case, 100*test_acc/test_size )
return 100*test_acc/test_size
def save_model(self):
# Store the weights of the model
torch.save(self.phi.state_dict(), self.base_res_dir + '/Model_' + self.post_string + '.pth')
# Store the parameters
torch.save(self.sms, self.base_res_dir + '/Sms_' + self.post_string + ".pt")
torch.save(self.sm_biases, self.base_res_dir + '/SmBiases_' + self.post_string + ".pt")