X-Git-Url: https://fleuret.org/cgi-bin/gitweb/gitweb.cgi?p=pysvrt.git;a=blobdiff_plain;f=cnn-svrt.py;h=bbce4c92e6426d48d15676372822ff86962066ed;hp=755d1c7bf1afcd74f4d5957e682d5d7bfe74c2e3;hb=6c83bf23d43bdbf2a8cae2df4654b26d46d53046;hpb=1b19eb14d0bd617deacfb7d15aa2e7babad3a5a7 diff --git a/cnn-svrt.py b/cnn-svrt.py index 755d1c7..bbce4c9 100755 --- a/cnn-svrt.py +++ b/cnn-svrt.py @@ -23,6 +23,9 @@ import time import argparse +import math + +from colorama import Fore, Back, Style import torch @@ -42,117 +45,233 @@ parser = argparse.ArgumentParser( formatter_class = argparse.ArgumentDefaultsHelpFormatter ) -parser.add_argument('--nb_train_samples', - type = int, default = 100000, +parser.add_argument('--nb_train_batches', + type = int, default = 1000, help = 'How many samples for train') -parser.add_argument('--nb_test_samples', - type = int, default = 10000, +parser.add_argument('--nb_test_batches', + type = int, default = 100, help = 'How many samples for test') parser.add_argument('--nb_epochs', - type = int, default = 25, + type = int, default = 50, help = 'How many training epochs') +parser.add_argument('--batch_size', + type = int, default = 100, + help = 'Mini-batch size') + +parser.add_argument('--log_file', + type = str, default = 'cnn-svrt.log', + help = 'Log file name') + +parser.add_argument('--compress_vignettes', + action='store_true', default = False, + help = 'Should we use lossless compression of vignette to reduce the memory footprint') + args = parser.parse_args() ###################################################################### -log_file = open('cnn-svrt.log', 'w') +log_file = open(args.log_file, 'w') + +print(Fore.RED + 'Logging into ' + args.log_file + Style.RESET_ALL) def log_string(s): - s = time.ctime() + ' ' + str(problem_number) + ' | ' + s + s = Fore.GREEN + time.ctime() + Style.RESET_ALL + ' ' + s log_file.write(s + '\n') log_file.flush() print(s) ###################################################################### -def generate_set(p, n): - target = torch.LongTensor(n).bernoulli_(0.5) - input = svrt.generate_vignettes(p, target) - input = input.view(input.size(0), 1, input.size(1), input.size(2)).float() - return Variable(input), Variable(target) +class VignetteSet: + def __init__(self, problem_number, nb_batches): + self.batch_size = args.batch_size + self.problem_number = problem_number + self.nb_batches = nb_batches + self.nb_samples = self.nb_batches * self.batch_size + self.targets = [] + self.inputs = [] + + acc = 0.0 + acc_sq = 0.0 + + for k in range(0, self.nb_batches): + target = torch.LongTensor(self.batch_size).bernoulli_(0.5) + input = svrt.generate_vignettes(problem_number, target) + input = input.float().view(input.size(0), 1, input.size(1), input.size(2)) + if torch.cuda.is_available(): + input = input.cuda() + target = target.cuda() + acc += input.float().sum() / input.numel() + acc_sq += input.float().pow(2).sum() / input.numel() + self.targets.append(target) + self.inputs.append(input) + + mean = acc / self.nb_batches + std = math.sqrt(acc_sq / self.nb_batches - mean * mean) + for k in range(0, self.nb_batches): + self.inputs[k].sub_(mean).div_(std) + + def get_batch(self, b): + return self.inputs[b], self.targets[b] + +class CompressedVignetteSet: + def __init__(self, problem_number, nb_batches): + self.batch_size = args.batch_size + self.problem_number = problem_number + self.nb_batches = nb_batches + self.nb_samples = self.nb_batches * self.batch_size + self.targets = [] + self.input_storages = [] + + acc = 0.0 + acc_sq = 0.0 + for k in range(0, self.nb_batches): + target = torch.LongTensor(self.batch_size).bernoulli_(0.5) + input = svrt.generate_vignettes(problem_number, target) + acc += input.float().sum() / input.numel() + acc_sq += input.float().pow(2).sum() / input.numel() + self.targets.append(target) + self.input_storages.append(svrt.compress(input.storage())) + + self.mean = acc / self.nb_batches + self.std = math.sqrt(acc_sq / self.nb_batches - self.mean * self.mean) + + def get_batch(self, b): + input = torch.ByteTensor(svrt.uncompress(self.input_storages[b])).float() + input = input.view(self.batch_size, 1, 128, 128).sub_(self.mean).div_(self.std) + target = self.targets[b] + + if torch.cuda.is_available(): + input = input.cuda() + target = target.cuda() + + return input, target ###################################################################### -# 128x128 --conv(9)-> 120x120 --max(4)-> 30x30 --conv(6)-> 25x25 --max(5)-> 5x5 - -class Net(nn.Module): +# Afroze's ShallowNet + +# map size nb. maps +# ---------------------- +# input 128x128 1 +# -- conv(21x21 x 6) -> 108x108 6 +# -- max(2x2) -> 54x54 6 +# -- conv(19x19 x 16) -> 36x36 16 +# -- max(2x2) -> 18x18 16 +# -- conv(18x18 x 120) -> 1x1 120 +# -- reshape -> 120 1 +# -- full(120x84) -> 84 1 +# -- full(84x2) -> 2 1 + +class AfrozeShallowNet(nn.Module): def __init__(self): - super(Net, self).__init__() - self.conv1 = nn.Conv2d(1, 10, kernel_size=9) - self.conv2 = nn.Conv2d(10, 20, kernel_size=6) - self.fc1 = nn.Linear(500, 100) - self.fc2 = nn.Linear(100, 2) + super(AfrozeShallowNet, self).__init__() + self.conv1 = nn.Conv2d(1, 6, kernel_size=21) + self.conv2 = nn.Conv2d(6, 16, kernel_size=19) + self.conv3 = nn.Conv2d(16, 120, kernel_size=18) + self.fc1 = nn.Linear(120, 84) + self.fc2 = nn.Linear(84, 2) def forward(self, x): - x = fn.relu(fn.max_pool2d(self.conv1(x), kernel_size=4, stride=4)) - x = fn.relu(fn.max_pool2d(self.conv2(x), kernel_size=5, stride=5)) - x = x.view(-1, 500) + x = fn.relu(fn.max_pool2d(self.conv1(x), kernel_size=2)) + x = fn.relu(fn.max_pool2d(self.conv2(x), kernel_size=2)) + x = fn.relu(self.conv3(x)) + x = x.view(-1, 120) x = fn.relu(self.fc1(x)) x = self.fc2(x) return x -def train_model(train_input, train_target): - model, criterion = Net(), nn.CrossEntropyLoss() +def train_model(model, train_set): + batch_size = args.batch_size + criterion = nn.CrossEntropyLoss() if torch.cuda.is_available(): - model.cuda() criterion.cuda() - optimizer, bs = optim.SGD(model.parameters(), lr = 1e-1), 100 + optimizer = optim.SGD(model.parameters(), lr = 1e-2) for k in range(0, args.nb_epochs): acc_loss = 0.0 - for b in range(0, train_input.size(0), bs): - output = model.forward(train_input.narrow(0, b, bs)) - loss = criterion(output, train_target.narrow(0, b, bs)) + for b in range(0, train_set.nb_batches): + input, target = train_set.get_batch(b) + output = model.forward(Variable(input)) + loss = criterion(output, Variable(target)) acc_loss = acc_loss + loss.data[0] model.zero_grad() loss.backward() optimizer.step() - log_string('TRAIN_LOSS {:d} {:f}'.format(k, acc_loss)) + log_string('train_loss {:d} {:f}'.format(k, acc_loss)) return model ###################################################################### -def nb_errors(model, data_input, data_target, bs = 100): +def nb_errors(model, data_set): ne = 0 - - for b in range(0, data_input.size(0), bs): - output = model.forward(data_input.narrow(0, b, bs)) + for b in range(0, data_set.nb_batches): + input, target = data_set.get_batch(b) + output = model.forward(Variable(input)) wta_prediction = output.data.max(1)[1].view(-1) - for i in range(0, bs): - if wta_prediction[i] != data_target.narrow(0, b, bs).data[i]: + for i in range(0, data_set.batch_size): + if wta_prediction[i] != target[i]: ne = ne + 1 return ne ###################################################################### +for arg in vars(args): + log_string('argument ' + str(arg) + ' ' + str(getattr(args, arg))) + for problem_number in range(1, 24): - train_input, train_target = generate_set(problem_number, args.nb_train_samples) - test_input, test_target = generate_set(problem_number, args.nb_test_samples) + if args.compress_vignettes: + train_set = CompressedVignetteSet(problem_number, args.nb_train_batches) + test_set = CompressedVignetteSet(problem_number, args.nb_test_batches) + else: + train_set = VignetteSet(problem_number, args.nb_train_batches) + test_set = VignetteSet(problem_number, args.nb_test_batches) - if torch.cuda.is_available(): - train_input, train_target = train_input.cuda(), train_target.cuda() - test_input, test_target = test_input.cuda(), test_target.cuda() + model = AfrozeShallowNet() - mu, std = train_input.data.mean(), train_input.data.std() - train_input.data.sub_(mu).div_(std) - test_input.data.sub_(mu).div_(std) + if torch.cuda.is_available(): + model.cuda() - model = train_model(train_input, train_target) + nb_parameters = 0 + for p in model.parameters(): + nb_parameters += p.numel() + log_string('nb_parameters {:d}'.format(nb_parameters)) + + model_filename = 'model_' + str(problem_number) + '.param' + + try: + model.load_state_dict(torch.load(model_filename)) + log_string('loaded_model ' + model_filename) + except: + log_string('training_model') + train_model(model, train_set) + torch.save(model.state_dict(), model_filename) + log_string('saved_model ' + model_filename) + + nb_train_errors = nb_errors(model, train_set) + + log_string('train_error {:d} {:.02f}% {:d} {:d}'.format( + problem_number, + 100 * nb_train_errors / train_set.nb_samples, + nb_train_errors, + train_set.nb_samples) + ) - nb_test_errors = nb_errors(model, test_input, test_target) + nb_test_errors = nb_errors(model, test_set) - log_string('TEST_ERROR {:.02f}% ({:d}/{:d})'.format( - 100 * nb_test_errors / test_input.size(0), + log_string('test_error {:d} {:.02f}% {:d} {:d}'.format( + problem_number, + 100 * nb_test_errors / test_set.nb_samples, nb_test_errors, - test_input.size(0)) + test_set.nb_samples) ) ######################################################################