Skip to content
Permalink
Branch: master
Find file Copy path
Find file Copy path
Fetching contributors…
Cannot retrieve contributors at this time
267 lines (249 sloc) 12.7 KB
import torch
import torch.nn as nn
import torch.nn.functional as F
from torch.nn.parameter import Parameter
from torch.optim import Adam
from torch.autograd import Variable
import numpy as np
from sklearn.cluster import KMeans
from sklearn.metrics.cluster import normalized_mutual_info_score as nmi_score
from sklearn.metrics import adjusted_rand_score as ari_score
from sklearn.decomposition import PCA
from data.omniglotloader import omniglot_alphabet_func, omniglot_evaluation_alphabets_mapping
from utils.util import cluster_acc, Identity, AverageMeter, seed_torch, str2bool
from utils import ramps
from models.vgg import VGG
from modules.module import feat2prob, target_distribution
from tqdm import tqdm
import warnings
import os
warnings.filterwarnings("ignore", category=UserWarning)
def init_prob_kmeans(model, eval_loader, args):
torch.manual_seed(1)
model = model.to(device)
# cluster parameter initiate
model.eval()
targets = np.zeros(len(eval_loader.dataset))
feats = np.zeros((len(eval_loader.dataset), 1024))
for _, (x, _, label, idx) in enumerate(eval_loader):
x = x.to(device)
_, feat = model(x)
feat = feat.view(x.size(0), -1)
idx = idx.data.cpu().numpy()
feats[idx, :] = feat.data.cpu().numpy()
targets[idx] = label.data.cpu().numpy()
# evaluate clustering performance
pca = PCA(n_components=args.n_clusters)
feats = pca.fit_transform(feats)
kmeans = KMeans(n_clusters=args.n_clusters, n_init=20)
y_pred = kmeans.fit_predict(feats)
acc, nmi, ari = cluster_acc(targets, y_pred), nmi_score(targets, y_pred), ari_score(targets, y_pred)
print('Init acc {:.4f}, nmi {:.4f}, ari {:.4f}'.format(acc, nmi, ari))
probs = feat2prob(torch.from_numpy(feats), torch.from_numpy(kmeans.cluster_centers_))
return kmeans.cluster_centers_, probs
def warmup_train(model, alphabetStr, train_loader, eval_loader, args):
optimizer = Adam(model.parameters(), lr=args.warmup_lr)
for epoch in range(args.warmup_epochs):
loss_record = AverageMeter()
model.train()
for batch_idx, (x, g_x, _, idx) in enumerate(train_loader):
_, feat = model(x.to(device))
prob = feat2prob(feat, model.center)
loss = F.kl_div(prob.log(), args.p_targets[idx].float().to(device))
optimizer.zero_grad()
loss.backward()
optimizer.step()
loss_record.update(loss.item(), x.size(0))
print('Warmup Train Epoch: {} Avg Loss: {:.4f}'.format(epoch, loss_record.avg))
test(model, eval_loader, args)
def Baseline_train(model, alphabetStr, train_loader, eval_loader, args):
optimizer = Adam(model.parameters(), lr=args.lr)
for epoch in range(args.epochs):
loss_record = AverageMeter()
model.train()
for batch_idx, (x, g_x, _, idx) in enumerate(train_loader):
_, feat = model(x.to(device))
prob = feat2prob(feat, model.center)
loss = F.kl_div(prob.log(), args.p_targets[idx].float().to(device))
optimizer.zero_grad()
loss.backward()
optimizer.step()
loss_record.update(loss.item(), x.size(0))
print('Train Epoch: {} Avg Loss: {:.4f}'.format(epoch, loss_record.avg))
_, _, _, probs = test(model, eval_loader, args)
if epoch % args.update_interval==0:
args.p_targets= target_distribution(probs)
torch.save(model.state_dict(), args.model_dir)
print("model saved to {}.".format(args.model_dir))
def PI_train(model, alphabetStr, train_loader, eval_loader, args):
optimizer = Adam(model.parameters(), lr=args.lr)
for epoch in range(args.epochs):
loss_record = AverageMeter()
model.train()
w = args.rampup_coefficient * ramps.sigmoid_rampup(epoch, args.rampup_length)
for batch_idx, (x, g_x, _, idx) in enumerate(train_loader):
_, feat = model(x.to(device))
_, feat_g = model(g_x.to(device))
prob = feat2prob(feat, model.center)
prob_g = feat2prob(feat_g, model.center)
loss = F.kl_div(prob.log(), args.p_targets[idx].float().to(device))
mse_loss = F.mse_loss(prob, prob_g)
loss=loss + w*mse_loss
optimizer.zero_grad()
loss.backward()
optimizer.step()
loss_record.update(loss.item(), x.size(0))
print('Train Epoch: {} Avg Loss: {:.4f}'.format(epoch, loss_record.avg))
_, _, _, probs = test(model, eval_loader, args)
if epoch % args.update_interval==0:
args.p_targets= target_distribution(probs)
torch.save(model.state_dict(), args.model_dir)
print("model saved to {}.".format(args.model_dir))
def TE_train(model, alphabetStr, train_loader, eval_loader, args):
optimizer = Adam(model.parameters(), lr=args.lr)
w = 0
alpha = 0.6
ntrain = len(train_loader.dataset)
Z = torch.zeros(ntrain, args.n_clusters).float().to(device) # intermediate values
z_ema = torch.zeros(ntrain, args.n_clusters).float().to(device) # temporal outputs
z_epoch = torch.zeros(ntrain, args.n_clusters).float().to(device) # current outputs
for epoch in range(args.epochs):
loss_record = AverageMeter()
model.train()
w = args.rampup_coefficient * ramps.sigmoid_rampup(epoch, args.rampup_length)
for batch_idx, (x, _, _, idx) in enumerate(train_loader):
_, feat = model(x.to(device))
prob = feat2prob(feat, model.center)
z_epoch[idx, :] = prob
prob_bar = Variable(z_ema[idx, :], requires_grad=False)
loss = F.kl_div(prob.log(), args.p_targets[idx].float().to(device))
mse_loss = F.mse_loss(prob, prob_bar)
loss=loss+w*mse_loss
optimizer.zero_grad()
loss.backward()
optimizer.step()
loss_record.update(loss.item(), x.size(0))
Z = alpha * Z + (1. - alpha) * z_epoch
z_ema = Z * (1. / (1. - alpha ** (epoch + 1)))
print('Train Epoch: {} Avg Loss: {:.4f}'.format(epoch, loss_record.avg))
_, _, _, probs = test(model, eval_loader, args)
if epoch % args.update_interval==0:
args.p_targets = target_distribution(probs)
torch.save(model.state_dict(), args.model_dir)
print("model saved to {}.".format(args.model_dir))
def TEP_train(model, alphabetStr, train_loader, eval_loader, args):
optimizer = Adam(model.parameters(), lr=args.lr)
w = 0
alpha = 0.6
ntrain = len(train_loader.dataset)
Z = torch.zeros(ntrain, args.n_clusters).float().to(device) # intermediate values
z_ema = torch.zeros(ntrain, args.n_clusters).float().to(device) # temporal outputs
z_epoch = torch.zeros(ntrain, args.n_clusters).float().to(device) # current outputs
for epoch in range(args.epochs):
loss_record = AverageMeter()
model.train()
for batch_idx, (x, g_x, _, idx) in enumerate(train_loader):
_, feat = model(x.to(device))
prob = feat2prob(feat, model.center)
loss = F.kl_div(prob.log(), args.p_targets[idx].float().to(device))
optimizer.zero_grad()
loss.backward()
optimizer.step()
loss_record.update(loss.item(), x.size(0))
print('Train Epoch: {} Avg Loss: {:.4f}'.format(epoch, loss_record.avg))
_, _, _, probs = test(model, eval_loader, args)
z_epoch = probs.float().to(device)
Z = alpha * Z + (1. - alpha) * z_epoch
z_bars = Z * (1. / (1. - alpha ** (epoch + 1)))
if epoch % args.update_interval==0:
args.p_targets = target_distribution(z_bars).float().to(device)
torch.save(model.state_dict(), args.model_dir)
print("model saved to {}.".format(args.model_dir))
def test(model, eval_loader, args):
model.eval()
targets = np.zeros(len(eval_loader.dataset))
y_pred = np.zeros(len(eval_loader.dataset))
probs= np.zeros((len(eval_loader.dataset), args.n_clusters))
for _, (x, _, label, idx) in enumerate(eval_loader):
x = x.to(device)
_, feat = model(x)
prob = feat2prob(feat, model.center)
# prob = F.softmax(logit, dim=1)
idx = idx.data.cpu().numpy()
y_pred[idx] = prob.data.cpu().detach().numpy().argmax(1)
targets[idx] = label.data.cpu().numpy()
probs[idx, :] = prob.cpu().detach().numpy()
# evaluate clustering performance
y_pred = y_pred.astype(np.int64)
acc, nmi, ari = cluster_acc(targets, y_pred), nmi_score(targets, y_pred), ari_score(targets, y_pred)
print('Test acc {:.4f}, nmi {:.4f}, ari {:.4f}'.format(acc, nmi, ari))
probs = torch.from_numpy(probs)
return acc, nmi, ari, probs
if __name__ == "__main__":
import argparse
parser = argparse.ArgumentParser(
description='cluster',
formatter_class=argparse.ArgumentDefaultsHelpFormatter)
parser.add_argument('--lr', type=float, default=0.001)
parser.add_argument('--warmup_lr', type=float, default=0.001)
parser.add_argument('--warmup_epochs', default=10, type=int)
parser.add_argument('--epochs', default=90, type=int)
parser.add_argument('--batch_size', default=100, type=int)
parser.add_argument('--update_interval', default=1, type=int)
parser.add_argument('--seed', default=1, type=int)
parser.add_argument('--save_txt', default=False, type=str2bool, help='save txt or not', metavar='BOOL')
parser.add_argument('--rampup_length', default=5, type=int)
parser.add_argument('--rampup_coefficient', default=100.0, type=float)
parser.add_argument('--n_clusters', default=10, type=int)
parser.add_argument('--num_workers', default=2, type=int)
parser.add_argument('--pretrain_dir', type=str, default='./data/experiments/pretrained/vgg6_omniglot_proto.pth')
parser.add_argument('--dataset_root', type=str, default='./data/datasets')
parser.add_argument('--exp_root', type=str, default='./data/experiments/')
parser.add_argument('--subfolder_name', type=str, default='run')
parser.add_argument('--save_txt_name', type=str, default='result.txt')
parser.add_argument('--DTC', type=str, default='PI')
args = parser.parse_args()
args.cuda = torch.cuda.is_available()
print("use cuda: {}".format(args.cuda))
device = torch.device("cuda" if args.cuda else "cpu")
seed_torch(args.seed)
model = VGG(n_layer='4+2', in_channels=1).to(device)
model.load_state_dict(torch.load(args.pretrain_dir), strict=False)
model.last = Identity()
init_feat_extractor = model
acc = {}
nmi = {}
ari = {}
for _, alphabetStr in omniglot_evaluation_alphabets_mapping.items():
runner_name = os.path.basename(__file__).split(".")[0]
model_dir= args.exp_root + '{}/{}/{}'.format(runner_name, args.DTC, args.subfolder_name)
if not os.path.exists(model_dir):
os.makedirs(model_dir)
args.model_dir = model_dir+'/'+'vgg6_{}.pth'.format(alphabetStr)
args.save_txt_path= args.exp_root + '{}/{}/{}'.format(runner_name, args.DTC, args.save_txt_name)
train_Dloader, eval_Dloader = omniglot_alphabet_func(alphabet=alphabetStr, background=False, root=args.dataset_root)(batch_size=args.batch_size, num_workers=args.num_workers)
args.n_clusters = train_Dloader.num_classes
model = VGG(n_layer='4+2', out_dim=args.n_clusters, in_channels=1).to(device)
model.load_state_dict(torch.load(args.pretrain_dir), strict=False)
model.center= Parameter(torch.Tensor(args.n_clusters, args.n_clusters))
init_centers, init_probs = init_prob_kmeans(init_feat_extractor, eval_Dloader, args)
args.p_targets = target_distribution(init_probs)
model.center.data = torch.tensor(init_centers).float().to(device)
warmup_train(model, alphabetStr, train_Dloader, eval_Dloader, args)
if args.DTC == 'Baseline':
Baseline_train(model, alphabetStr, train_Dloader, eval_Dloader, args)
elif args.DTC == 'PI':
PI_train(model, alphabetStr, train_Dloader, eval_Dloader, args)
elif args.DTC == 'TE':
TE_train(model, alphabetStr, train_Dloader, eval_Dloader, args)
elif args.DTC == 'TEP':
TEP_train(model, alphabetStr, train_Dloader, eval_Dloader, args)
acc[alphabetStr], nmi[alphabetStr], ari[alphabetStr], _ = test(model, eval_Dloader, args)
print('ACC for all alphabets:',acc)
print('NMI for all alphabets:',nmi)
print('ARI for all alphabets:',ari)
avg_acc, avg_nmi, avg_ari = sum(acc.values())/float(len(acc)), sum(nmi.values())/float(len(nmi)), sum(ari.values())/float(len(ari))
print('avg ACC {:.4f}, NMI {:.4f} ARI {:.4f}'.format(avg_acc, avg_nmi, avg_ari))
if args.save_txt:
with open(args.save_txt_path, 'a') as f:
f.write("{:.4f}, {:.4f}, {:.4f}\n".format(avg_acc, avg_nmi, avg_ari))
You can’t perform that action at this time.