train_val_seg.py

import argparse
import math
from datetime import datetime
import h5py
import numpy as np
import tensorflow as tf
import socket
import importlib
import os
import sys
ROOT_DIR = os.path.dirname(os.path.abspath(__file__))
sys.path.append(os.path.join(ROOT_DIR, 'utils'))
import provider
import tf_util
import part_dataset_all_normal

parser = argparse.ArgumentParser()
parser.add_argument('--gpu', type=int, default=0, help='GPU to use [default: GPU 0]')
parser.add_argument('--model', default='RIConv', help='Model name [default: model]')
parser.add_argument('--log_dir', default='log/part_seg', help='Log dir [default: log]')
parser.add_argument('--num_point', type=int, default=2048, help='Point Number [default: 2048]')
parser.add_argument('--max_epoch', type=int, default=201, help='Epoch to run [default: 201]')
parser.add_argument('--batch_size', type=int, default=16, help='Batch Size during training [default: 32]')
parser.add_argument('--learning_rate', type=float, default=0.001, help='Initial learning rate [default: 0.001]')
parser.add_argument('--momentum', type=float, default=0.9, help='Initial learning rate [default: 0.9]')
parser.add_argument('--optimizer', default='adam', help='adam or momentum [default: adam]')
parser.add_argument('--decay_step', type=int, default=16881*20, help='Decay step for lr decay [default: 200000]')
parser.add_argument('--decay_rate', type=float, default=0.5, help='Decay rate for lr decay [default: 0.7]')
FLAGS = parser.parse_args()

EPOCH_CNT = 0

BATCH_SIZE = FLAGS.batch_size
NUM_POINT = FLAGS.num_point
MAX_EPOCH = FLAGS.max_epoch
BASE_LEARNING_RATE = FLAGS.learning_rate
GPU_INDEX = FLAGS.gpu
MOMENTUM = FLAGS.momentum
OPTIMIZER = FLAGS.optimizer
DECAY_STEP = FLAGS.decay_step
DECAY_RATE = FLAGS.decay_rate

MODEL = importlib.import_module(FLAGS.model) # import network module
MODEL_FILE = os.path.join(ROOT_DIR, 'models', FLAGS.model+'.py')
LOG_DIR = FLAGS.log_dir
if not os.path.exists(LOG_DIR): os.mkdir(LOG_DIR)
LOG_FOUT = open(os.path.join(LOG_DIR, 'log_partseg_rio_z_so3.txt'), 'w')
LOG_FOUT.write(str(FLAGS)+'\n')

BN_INIT_DECAY = 0.5
BN_DECAY_DECAY_RATE = 0.5
BN_DECAY_DECAY_STEP = float(DECAY_STEP)
BN_DECAY_CLIP = 0.99

NUM_CLASSES = 50

# Shapenet official train/test split
DATA_PATH = os.path.join('../data/shapenet_partseg/', 'shapenetcore_partanno_segmentation_benchmark_v0_normal')
TRAIN_DATASET = part_dataset_all_normal.PartNormalDataset(root=DATA_PATH, npoints=NUM_POINT, classification=False, split='trainval', return_cls_label=True)
TEST_DATASET = part_dataset_all_normal.PartNormalDataset(root=DATA_PATH, npoints=NUM_POINT, classification=False, split='test', return_cls_label=True)

riconv_param_name = ('K', 'D', 'P', 'C', 'links')
riconv_params = [dict(zip(riconv_param_name, riconv_param)) for riconv_param in
                [
                 (64, 4, 512, 128, []),
                 (32, 2, 128, 256, []),
                 (16, 1, 32, 512, [])]]

ridconv_param_name = ('K', 'D', 'pts_layer_idx', 'qrs_layer_idx')
ridconv_params = [dict(zip(ridconv_param_name, ridconv_param)) for ridconv_param in
                 [
                  (8,   1, 2, 1),
                  (16,  2, 1, 0),
                  (32,  2, 0, -1)]]
x = 1
fc_param_name = ('C', 'dropout_rate')
fc_params = [dict(zip(fc_param_name, fc_param)) for fc_param in
            [(128 * x, 0),
            (64 * x, 0.5)]]

WITH_LOCAL = True
WITH_MULTI = True

def log_string(out_str):
    LOG_FOUT.write(out_str+'\n')
    LOG_FOUT.flush()
    print(out_str)

def get_learning_rate(batch):
    learning_rate = tf.train.exponential_decay(
                        BASE_LEARNING_RATE,  # Base learning rate.
                        batch * BATCH_SIZE,  # Current index into the dataset.
                        DECAY_STEP,          # Decay step.
                        DECAY_RATE,          # Decay rate.
                        staircase=True)
    learning_rate = tf.maximum(learning_rate, 0.00001) # CLIP THE LEARNING RATE!
    return learning_rate        

def get_bn_decay(batch):
    bn_momentum = tf.train.exponential_decay(
                      BN_INIT_DECAY,
                      batch*BATCH_SIZE,
                      BN_DECAY_DECAY_STEP,
                      BN_DECAY_DECAY_RATE,
                      staircase=True)
    bn_decay = tf.minimum(BN_DECAY_CLIP, 1 - bn_momentum)
    return bn_decay

def train():
    with tf.Graph().as_default():
        with tf.device('/gpu:'+str(GPU_INDEX)):
            pointclouds_pl, labels_pl, _ = MODEL.placeholder_inputs(BATCH_SIZE, NUM_POINT)
            is_training_pl = tf.placeholder(tf.bool, shape=())
            
            # Note the global_step=batch parameter to minimize. 
            # That tells the optimizer to helpfully increment the 'batch' parameter for you every time it trains.
            batch = tf.Variable(0)
            bn_decay = get_bn_decay(batch)
            tf.summary.scalar('bn_decay', bn_decay)

            # Get model and loss 
            pred = MODEL.get_model(pointclouds_pl, is_training_pl, riconv_params, ridconv_params, fc_params, sampling='fps', bn_decay=bn_decay, part_num=NUM_CLASSES)
            loss, per_instance_seg_loss, per_instance_seg_pred_res = MODEL.get_loss(pred, labels_pl)
            

            # Get training operator
            learning_rate = get_learning_rate(batch)
            if OPTIMIZER == 'momentum':
                optimizer = tf.train.MomentumOptimizer(learning_rate, momentum=MOMENTUM)
            elif OPTIMIZER == 'adam':
                optimizer = tf.train.AdamOptimizer(learning_rate)
            # train_op = optimizer.minimize(loss, global_step=batch)
            update_ops = tf.get_collection(tf.GraphKeys.UPDATE_OPS)
            with tf.control_dependencies(update_ops):
                train_op = optimizer.minimize(loss, global_step=batch)
            
            # Add ops to save and restore all the variables.
            saver = tf.train.Saver()
        
        # Create a session
        config = tf.ConfigProto()
        config.gpu_options.allow_growth = True
        config.allow_soft_placement = True
        config.log_device_placement = False
        sess = tf.Session(config=config)

        # Init variables
        init = tf.global_variables_initializer()
        sess.run(init)
        #sess.run(init, {is_training_pl: True})

        ops = {'pointclouds_pl': pointclouds_pl,
               'labels_pl': labels_pl,
               'is_training_pl': is_training_pl,
               'pred': pred,
               'loss': loss,
               'train_op': train_op,
               'step': batch}

        eval_iou_max = 0
        maxIoU_epoch = 0
        for epoch in range(MAX_EPOCH):
            log_string('**** EPOCH %03d ****' % (epoch))
            sys.stdout.flush()
            
            train_one_epoch(sess, ops)
            iou = eval_one_epoch(sess, ops)

            
            # Save the variables to disk.
            if iou > eval_iou_max:
                max_save_path = saver.save(sess, os.path.join(LOG_DIR, 'model_max_%d.ckpt' % epoch))
                maxIoU_epoch = epoch
                eval_iou_max = iou
                log_string("Model saved in file: %s" % max_save_path)
            if epoch == (MAX_EPOCH-1):
                save_path = saver.save(sess, os.path.join(LOG_DIR, "model.ckpt"))
                log_string("Model saved in file: %s" % save_path)
                log_string("Max iou model saved in epoch: %d" % maxIoU_epoch)
                log_string("Max iou is: %f" % eval_iou_max)

def get_batch(dataset, idxs, start_idx, end_idx):
    bsize = end_idx-start_idx
    batch_data = np.zeros((bsize, NUM_POINT, 6))
    batch_label = np.zeros((bsize, NUM_POINT), dtype=np.int32)
    batch_cls_label = np.zeros((bsize,), dtype=np.int32)
    for i in range(bsize):
        ps,normal,seg,cls = dataset[idxs[i+start_idx]]
        batch_data[i,:,0:3] = ps
        batch_data[i,:,3:6] = normal
        batch_label[i,:] = seg
        batch_cls_label[i] = cls
    return batch_data, batch_label, batch_cls_label

def train_one_epoch(sess, ops):
    """ ops: dict mapping from string to tf ops """
    is_training = True
    
    # Shuffle train samples
    train_idxs = np.arange(0, len(TRAIN_DATASET))
    np.random.shuffle(train_idxs)
    num_batches = len(TRAIN_DATASET)/BATCH_SIZE
    
    log_string(str(datetime.now()))

    total_correct = 0
    total_seen = 0
    loss_sum = 0
    for batch_idx in range(int(num_batches)):
        start_idx = batch_idx * BATCH_SIZE
        end_idx = (batch_idx+1) * BATCH_SIZE
        batch_data, batch_label, batch_cls_label = get_batch(TRAIN_DATASET, train_idxs, start_idx, end_idx)
        # Augment batched point clouds by rotation and jittering
        batch_data[:,:,0:3] = provider.jitter_point_cloud(batch_data[:,:,0:3])
        batch_data[:,:,0:3] = provider.rotate_point_cloud(batch_data[:,:,0:3])  # rotation z
        feed_dict = {ops['pointclouds_pl']: batch_data[:,:,0:3],
                     ops['labels_pl']: batch_label,
                     ops['is_training_pl']: is_training,}
        step, _, loss_val, pred_val = sess.run([ops['step'],
            ops['train_op'], ops['loss'], ops['pred']], feed_dict=feed_dict)
        
        pred_val = np.argmax(pred_val, 2)
        correct = np.sum(pred_val == batch_label)
        total_correct += correct
        total_seen += (BATCH_SIZE*NUM_POINT)
        loss_sum += loss_val

        if (batch_idx+1)%100 == 0:
            log_string(' -- %03d / %03d --' % (batch_idx+1, num_batches))
            log_string('mean loss: %f' % (loss_sum / 10))
            log_string('accuracy: %f' % (total_correct / float(total_seen)))
            total_correct = 0
            total_seen = 0
            loss_sum = 0
        

        
def eval_one_epoch(sess, ops):
    """ ops: dict mapping from string to tf ops """
    global EPOCH_CNT
    is_training = False
    test_idxs = np.arange(0, len(TEST_DATASET))
    # Test on all data: last batch might be smaller than BATCH_SIZE
    num_batches = (len(TEST_DATASET)+BATCH_SIZE-1)/BATCH_SIZE

    total_correct = 0
    total_seen = 0
    loss_sum = 0
    total_seen_class = [0 for _ in range(NUM_CLASSES)]
    total_correct_class = [0 for _ in range(NUM_CLASSES)]

    seg_classes = TEST_DATASET.seg_classes
    shape_ious = {cat:[] for cat in seg_classes.keys()}
    seg_label_to_cat = {} # {0:Airplane, 1:Airplane, ...49:Table}
    for cat in seg_classes.keys():
        for label in seg_classes[cat]:
            seg_label_to_cat[label] = cat

    log_string(str(datetime.now()))
    log_string('---- EPOCH %03d EVALUATION ----'%(EPOCH_CNT))
    
    batch_data = np.zeros((BATCH_SIZE, NUM_POINT, 3))
    batch_label = np.zeros((BATCH_SIZE, NUM_POINT)).astype(np.int32)
    batch_cls_label = np.zeros((BATCH_SIZE,)).astype(np.int32)
    for batch_idx in range(int(num_batches)):
        if batch_idx %20==0:
            log_string('%03d/%03d'%(batch_idx, num_batches))
        start_idx = batch_idx * BATCH_SIZE
        end_idx = min(len(TEST_DATASET), (batch_idx+1) * BATCH_SIZE)
        cur_batch_size = end_idx-start_idx
        cur_batch_data, cur_batch_label, cur_batch_cls_label = get_batch(TEST_DATASET, test_idxs, start_idx, end_idx)
        if cur_batch_size == BATCH_SIZE:
            batch_data = cur_batch_data
            batch_label = cur_batch_label
            batch_cls_label = cur_batch_cls_label
        else:
            batch_data[0:cur_batch_size] = cur_batch_data
            batch_label[0:cur_batch_size] = cur_batch_label
            batch_cls_label[0:cur_batch_size] = cur_batch_cls_label

        # ---------------------------------------------------------------------
        batch_data[:,:,0:3] = provider.rotate_point_cloud_so3(batch_data[:,:,0:3]) 
        feed_dict = {ops['pointclouds_pl']: batch_data[:,:,0:3],
                     ops['labels_pl']: batch_label,
                     ops['is_training_pl']: is_training}
        step, loss_val, pred_val = sess.run([ops['step'],
            ops['loss'], ops['pred']], feed_dict=feed_dict)
       
        # ---------------------------------------------------------------------
    
        # Select valid data
        cur_pred_val = pred_val[0:cur_batch_size]
        # Constrain pred to the groundtruth classes (selected by seg_classes[cat])
        cur_pred_val_logits = cur_pred_val
        cur_pred_val = np.zeros((cur_batch_size, NUM_POINT)).astype(np.int32)
        for i in range(cur_batch_size):
            cat = seg_label_to_cat[cur_batch_label[i,0]]
            logits = cur_pred_val_logits[i,:,:]
            cur_pred_val[i,:] = np.argmax(logits[:,seg_classes[cat]], 1) + seg_classes[cat][0]
        correct = np.sum(cur_pred_val == cur_batch_label)
        total_correct += correct
        total_seen += (cur_batch_size*NUM_POINT)
        if cur_batch_size==BATCH_SIZE:
            loss_sum += loss_val
        for l in range(NUM_CLASSES):
            total_seen_class[l] += np.sum(cur_batch_label==l)
            total_correct_class[l] += (np.sum((cur_pred_val==l) & (cur_batch_label==l)))

        for i in range(cur_batch_size):
            segp = cur_pred_val[i,:]
            segl = cur_batch_label[i,:] 
            cat = seg_label_to_cat[segl[0]]
            part_ious = [0.0 for _ in range(len(seg_classes[cat]))]
            for l in seg_classes[cat]:
                if (np.sum(segl==l) == 0) and (np.sum(segp==l) == 0): # part is not present, no prediction as well
                    part_ious[l-seg_classes[cat][0]] = 1.0
                else:
                    part_ious[l-seg_classes[cat][0]] = np.sum((segl==l) & (segp==l)) / float(np.sum((segl==l) | (segp==l)))
            shape_ious[cat].append(np.mean(part_ious))

    all_shape_ious = []
    for cat in shape_ious.keys():
        for iou in shape_ious[cat]:
            all_shape_ious.append(iou)
        shape_ious[cat] = np.mean(shape_ious[cat])
    mean_shape_ious = np.mean(list(shape_ious.values()))
    log_string('eval mean loss: %f' % (loss_sum / float(len(TEST_DATASET)/BATCH_SIZE)))
    log_string('eval accuracy: %f'% (total_correct / float(total_seen)))
    log_string('eval avg class acc: %f' % (np.mean(np.array(total_correct_class)/np.array(total_seen_class,dtype=np.float))))
    for cat in sorted(shape_ious.keys()):
        log_string('eval mIoU of %s:\t %f'%(cat, shape_ious[cat]))
    log_string('eval mean mIoU: %f' % (mean_shape_ious))
    log_string('eval mean mIoU (all shapes): %f' % (np.mean(all_shape_ious)))
         
    EPOCH_CNT += 1
    return np.mean(all_shape_ious)


if __name__ == "__main__":
    log_string('pid: %s'%(str(os.getpid())))
    train()
    LOG_FOUT.close()