sess, val_loss, val_accuracy, val_labels,
val_probabilities)
val_loss_arr.append(loss_values)
val_acc_arr.append(accuracy_values)
logging.info('Loss on validation batch %s is : %s' %
(i, loss_values))
logging.info('AUC on validaton batch %s is : %s' %
(i, auc))
# for label_idx in range(len(auc)):
# auc_arr[label_idx] += auc[label_idx]
logging.info(
'Mean loss on this validation epoch is: %s' %
(float(sum(val_loss_arr)) / max(len(val_loss_arr), 1)))
logging.info(
'Mean accuracy on this validation epoch is: %s' %
(float(sum(val_acc_arr)) / max(len(val_acc_arr), 1)))
# mean_auc = [auc / val_num_batches_per_epoch for auc in auc_arr]
# logging.info('Mean auc on this validation epoch is: %s' % mean_auc)
# Log the summaries every 10 step.
if step % 10 == 0:
summaries = sess.run(my_summary_ops)
sv.summary_computed(sess, summaries)
#Once all the training has been done, save the log files and checkpoint model
logging.info('Finished training! Saving model to disk now.')
sv.saver.save(sess, sv.save_path, global_step=sv.global_step)
if __name__ == '__main__':
mlog.initlog(FLAGS.log_dir)
run()
'shuffle': True, # shuffle dataset every epoch or not
'normalization': 'divide_255',
}
def get_train_params_by_name(name):
if name in ['C10', 'C10+', 'C100', 'C100+']:
return train_params_cifar
if name == 'SVHN':
return train_params_svhn
if name == 'chexnet':
return train_params_chexnet
if __name__ == '__main__':
initlog()
logging.basicConfig(
format='%(asctime)s(%(relativeCreated)d) - %(levelname)s %(filename)s(%(lineno)d) :: %(message)s',
level=logging.DEBUG)
parser = argparse.ArgumentParser()
parser.add_argument(
'--train', action='store_true',
help='Train the model')
parser.add_argument(
'--test', action='store_true',
help='Test model for required dataset if pretrained model exists.'
'If provided together with `--train` flag testing will be'
'performed right after training.')
parser.add_argument(
'--model_type', '-m', type=str, choices=['DenseNet', 'DenseNet-BC'],
default='DenseNet',
sys.stdout.write('\n')
sys.stdout.flush()
def main():
LAYERS = 3
pkl_fname = "data/preprocess/stage1_train_set_rgb.pkl"
images, masks = get_dataset(pkl_fname)
logging.info("read train set: %s, %s", images.shape, masks.shape)
logging.info("image:[%s, %s], mask:[%s, %s]", np.max(images), np.min(images), np.max(masks), np.min(masks))
# pred_size, offset = unet_size(256, LAYERS)
# logging.info("pred_size: %d, offset: %d", pred_size, offset)
# images = padding_array(images, offset, default_val=0.0)
# masks = padding_array(masks, offset, default_val=False)
# args.data_dir = args.data_dir.strip()
# if len(args.data_dir) >= 0:
# fnames = [os.path.join(args.data_dir, x) for x in fnames]
train_ratio = 0.9
n_train = int(len(images)*train_ratio)
logging.info("train_ratio: %s, n_train: %s, n_val: %s", train_ratio, n_train, len(images)-n_train)
convert_dataset(images[:n_train], masks[:n_train], "data/tfrecords/256x256/train", 4)
convert_dataset(images[n_train:], masks[n_train:], "data/tfrecords/256x256/val", 2)
if __name__ == "__main__":
initlog("log")
main()
print 'logits: \n', logits_value
print 'Probabilities: \n', probabilities_value
print 'predictions: \n', predictions_value
print 'Labels:\n:', labels_value
#Log the summaries every 10 step.
if step % 10 == 0:
loss, _ = train_step(sess, train_op, sv.global_step)
mean_loss_arr.append(loss)
summaries = sess.run(my_summary_op)
sv.summary_computed(sess, summaries)
#If not, simply run the training step
else:
loss, _ = train_step(sess, train_op, sv.global_step)
mean_loss_arr.append(loss)
#We log the final training loss and accuracy
logging.info('Final Loss: %s', loss)
logging.info('Final Accuracy: %s', sess.run(accuracy))
#Once all the training has been done, save the log files and checkpoint model
logging.info('Finished training! Saving model to disk now.')
# saver.save(sess, "./flowers_model.ckpt")
sv.saver.save(sess, sv.save_path, global_step=sv.global_step)
if __name__ == '__main__':
mlog.initlog('./log/pne_log')
run()