個人代碼閱讀筆記。
# --------------------------------------------------------
# Tensorflow Faster R-CNN
# Licensed under The MIT License [see LICENSE for details]
# Written by Xinlei Chen and Zheqi He
# --------------------------------------------------------
from __future__ import absolute_import
from __future__ import division
from __future__ import print_function
from model.config import cfg
import roi_data_layer.roidb as rdl_roidb
from roi_data_layer.layer import RoIDataLayer
from utils.timer import Timer
try:
import cPickle as pickle
except ImportError:
import pickle
import numpy as np
import os
import sys
import glob
import time
import tensorflow as tf
from tensorflow.python import pywrap_tensorflow
'''
這是訓練的第二層入口，第一層是train_faster_rcnn.sh
'''
class SolverWrapper(object):
"""
A wrapper class for the training process
訓練過程的封裝
"""
#實例初始化
def __init__(self, sess, network, imdb, roidb, valroidb, output_dir, tbdir, pretrained_model=None):
self.net = network
self.imdb = imdb#image database
self.roidb = roidb#roi database
self.valroidb = valroidb#valid roi database
self.output_dir = output_dir#輸出路徑
self.tbdir = tbdir
# Simply put '_val' at the end to save the summaries from the validation set
self.tbvaldir = tbdir + '_val'#？
if not os.path.exists(self.tbvaldir):
os.makedirs(self.tbvaldir)
self.pretrained_model = pretrained_model#pre訓練模型
def snapshot(self, sess, iter):#保存參數snapshot，多久保存一次tf的會話
net = self.net
if not os.path.exists(self.output_dir):
os.makedirs(self.output_dir)
# Store the model snapshot
#TRAIN.SNAPSHOT_PREFIX爲名字前綴‘res101_faster_rcnn'’
filename = cfg.TRAIN.SNAPSHOT_PREFIX + '_iter_{:d}'.format(iter) + '.ckpt'
filename = os.path.join(self.output_dir, filename)
self.saver.save(sess, filename)
print('Wrote snapshot to: {:s}'.format(filename))
# Also store some meta information, random state, etc.
#可以看出來ckpt纔是主要的權重，pkl是保存了一些原始信息和隨機狀態
nfilename = cfg.TRAIN.SNAPSHOT_PREFIX + '_iter_{:d}'.format(iter) + '.pkl'
nfilename = os.path.join(self.output_dir, nfilename)
# current state of numpy random
#當前的numpy.random狀態（隨機種子什麼的
st0 = np.random.get_state()
# current position in the database
#當前載入數據庫
cur = self.data_layer._cur
# current shuffled indexes of the database
#當前打亂的數據索引
perm = self.data_layer._perm
# current position in the validation database
#當前數據在validation database中的位置
cur_val = self.data_layer_val._cur
# current shuffled indexes of the validation database
#當前打亂的驗證數據集索引
perm_val = self.data_layer_val._perm
# Dump the meta info
with open(nfilename, 'wb') as fid:#打開文件
pickle.dump(st0, fid, pickle.HIGHEST_PROTOCOL)
pickle.dump(cur, fid, pickle.HIGHEST_PROTOCOL)
pickle.dump(perm, fid, pickle.HIGHEST_PROTOCOL)
pickle.dump(cur_val, fid, pickle.HIGHEST_PROTOCOL)
pickle.dump(perm_val, fid, pickle.HIGHEST_PROTOCOL)
pickle.dump(iter, fid, pickle.HIGHEST_PROTOCOL)
return filename, nfilename
def from_snapshot(self, sess, sfile, nfile):#載入快照
print('Restoring model snapshots from {:s}'.format(sfile))
self.saver.restore(sess, sfile)
print('Restored.')
# Needs to restore the other hyper-parameters/states for training, (TODO xinlei) I have
# tried my best to find the random states so that it can be recovered exactly
# However the Tensorflow state is currently not available
with open(nfile, 'rb') as fid:
st0 = pickle.load(fid)
cur = pickle.load(fid)
perm = pickle.load(fid)
cur_val = pickle.load(fid)
perm_val = pickle.load(fid)
last_snapshot_iter = pickle.load(fid)
np.random.set_state(st0)
self.data_layer._cur = cur
self.data_layer._perm = perm
self.data_layer_val._cur = cur_val
self.data_layer_val._perm = perm_val
return last_snapshot_iter
def get_variables_in_checkpoint_file(self, file_name):#在checkpoint中獲取變量
try:
reader = pywrap_tensorflow.NewCheckpointReader(file_name)
var_to_shape_map = reader.get_variable_to_shape_map()
return var_to_shape_map
except Exception as e: # pylint: disable=broad-except
print(str(e))
if "corrupted compressed block contents" in str(e):
print("It's likely that your checkpoint file has been compressed "
"with SNAPPY.")
def construct_graph(self, sess):#構建tf的計算圖模型
with sess.graph.as_default():
# Set the random seed for tensorflow
tf.set_random_seed(cfg.RNG_SEED)#設置隨機種子
# Build the main computation graph
layers = self.net.create_architecture('TRAIN', self.imdb.num_classes, tag='default',#創建網絡結構
anchor_scales=cfg.ANCHOR_SCALES,
anchor_ratios=cfg.ANCHOR_RATIOS)
# Define the loss#定義loss
loss = layers['total_loss']
# Set learning rate and momentum
#設置學習率
lr = tf.Variable(cfg.TRAIN.LEARNING_RATE, trainable=False)
self.optimizer = tf.train.MomentumOptimizer(lr, cfg.TRAIN.MOMENTUM)#設置優化器
# Compute the gradients with regard to the loss計算與損失有關的梯度
gvs = self.optimizer.compute_gradients(loss)
# Double the gradient of the bias if set
if cfg.TRAIN.DOUBLE_BIAS:#偏置的學習率加倍，這個參數默認是打開的。這裏就是調用默認設置
final_gvs = []
with tf.variable_scope('Gradient_Mult') as scope:#獲取變量：多梯度
for grad, var in gvs:#對於裏面的每一個梯度和變量
scale = 1.#尺度=1，也就是正常學習率
if cfg.TRAIN.DOUBLE_BIAS and '/biases:' in var.name:
scale *= 2.#尺度=2，雙倍學習率
if not np.allclose(scale, 1.0):
grad = tf.multiply(grad, scale)
final_gvs.append((grad, var))
train_op = self.optimizer.apply_gradients(final_gvs)#其實就是設置if修改了原來的gvs，變成final_gvs
else:
train_op = self.optimizer.apply_gradients(gvs)#直接應用gvs學習率來進行梯度下降
# We will handle the snapshots ourselves
self.saver = tf.train.Saver(max_to_keep=100000)
# Write the train and validation information to tensorboard#信息傳入tensorboard
self.writer = tf.summary.FileWriter(self.tbdir, sess.graph)
self.valwriter = tf.summary.FileWriter(self.tbvaldir)
return lr, train_op#返回學習率以及訓練的計算模型
def find_previous(self):
sfiles = os.path.join(self.output_dir, cfg.TRAIN.SNAPSHOT_PREFIX + '_iter_*.ckpt.meta')
sfiles = glob.glob(sfiles)
sfiles.sort(key=os.path.getmtime)
# Get the snapshot name in TensorFlow
redfiles = []
for stepsize in cfg.TRAIN.STEPSIZE:
redfiles.append(os.path.join(self.output_dir,
cfg.TRAIN.SNAPSHOT_PREFIX + '_iter_{:d}.ckpt.meta'.format(stepsize+1)))
sfiles = [ss.replace('.meta', '') for ss in sfiles if ss not in redfiles]
nfiles = os.path.join(self.output_dir, cfg.TRAIN.SNAPSHOT_PREFIX + '_iter_*.pkl')
nfiles = glob.glob(nfiles)
nfiles.sort(key=os.path.getmtime)
redfiles = [redfile.replace('.ckpt.meta', '.pkl') for redfile in redfiles]
nfiles = [nn for nn in nfiles if nn not in redfiles]
lsf = len(sfiles)
assert len(nfiles) == lsf
return lsf, nfiles, sfiles
def initialize(self, sess):#模型初始化
# Initial file lists are empty
np_paths = []
ss_paths = []
# Fresh train directly from ImageNet weights
print('Loading initial model weights from {:s}'.format(self.pretrained_model))
variables = tf.global_variables()
# Initialize all variables first
sess.run(tf.variables_initializer(variables, name='init'))
var_keep_dic = self.get_variables_in_checkpoint_file(self.pretrained_model)#獲取預訓練模型的參數
# Get the variables to restore, ignoring the variables to fix
#獲取要還原的變量，忽略要固定的變量？
variables_to_restore = self.net.get_variables_to_restore(variables, var_keep_dic)
restorer = tf.train.Saver(variables_to_restore)
restorer.restore(sess, self.pretrained_model)#保存
print('Loaded.')#初始化完畢之後，打印Loaded
# Need to fix the variables before loading, so that the RGB weights are changed to BGR
# For VGG16 it also changes the convolutional weights fc6 and fc7 to
# fully connected weights
#在加載模型之前，需要把變量固定下來，所謂變量就是計算模型的張量節點
self.net.fix_variables(sess, self.pretrained_model)#
print('Fixed.')#fix是在loaded之後進行的。把網絡的形狀固定
last_snapshot_iter = 0#迭代次數置爲0，加載預訓練模型但不使用premodel的迭代次數
rate = cfg.TRAIN.LEARNING_RATE#設置是0.001
stepsizes = list(cfg.TRAIN.STEPSIZE)#步進參數
return rate, last_snapshot_iter, stepsizes, np_paths, ss_paths
def restore(self, sess, sfile, nfile):#sfile和nfile是什麼？
# Get the most recent snapshot and restore
np_paths = [nfile]
ss_paths = [sfile]
# Restore model from snapshots
last_snapshot_iter = self.from_snapshot(sess, sfile, nfile)
# Set the learning rate
rate = cfg.TRAIN.LEARNING_RATE
stepsizes = []
for stepsize in cfg.TRAIN.STEPSIZE:
if last_snapshot_iter > stepsize:
rate *= cfg.TRAIN.GAMMA
else:
stepsizes.append(stepsize)
return rate, last_snapshot_iter, stepsizes, np_paths, ss_paths
def remove_snapshot(self, np_paths, ss_paths):
to_remove = len(np_paths) - cfg.TRAIN.SNAPSHOT_KEPT
for c in range(to_remove):
nfile = np_paths[0]
os.remove(str(nfile))
np_paths.remove(nfile)
to_remove = len(ss_paths) - cfg.TRAIN.SNAPSHOT_KEPT
#刪除緩存，保證緩存最多3個，當然我可以多設置幾個，來看看不同階段的ap
for c in range(to_remove):
sfile = ss_paths[0]
# To make the code compatible to earlier versions of Tensorflow,
# where the naming tradition for checkpoints are different
if os.path.exists(str(sfile)):
os.remove(str(sfile))
else:
os.remove(str(sfile + '.data-00000-of-00001'))
os.remove(str(sfile + '.index'))
sfile_meta = sfile + '.meta'
os.remove(str(sfile_meta))
ss_paths.remove(sfile)
def train_model(self, sess, max_iters):#訓練模型
# Build data layers for both training and validation set
self.data_layer = RoIDataLayer(self.roidb, self.imdb.num_classes) #兩個部分的訓練模型訓練數據
self.data_layer_val = RoIDataLayer(self.valroidb, self.imdb.num_classes, random=True) #val data
# Construct the computation graph
#
lr, train_op = self.construct_graph(sess)#調用前面定義的self.construct_graph實例函數，生成計算圖
# Find previous snapshots if there is any to restore from
#尋找是否有之前保存的訓練快照，有就載入
lsf, nfiles, sfiles = self.find_previous()
# Initialize the variables or restore them from the last snapshot
if lsf == 0:
rate, last_snapshot_iter, stepsizes, np_paths, ss_paths = self.initialize(sess)
else:
rate, last_snapshot_iter, stepsizes, np_paths, ss_paths = self.restore(sess,
str(sfiles[-1]),
str(nfiles[-1]))
timer = Timer()#打開計時器
iter = last_snapshot_iter + 1#迭代次數+1
last_summary_time = time.time()#開始計時
# Make sure the lists are not empty
stepsizes.append(max_iters)#最後加一個迭代次數
stepsizes.reverse()
next_stepsize = stepsizes.pop()#然後再移出來
while iter < max_iters + 1:
# Learning rate
if iter == next_stepsize + 1:
# Add snapshot here before reducing the learning rate#如果到達該保存的次數，就添加快照
self.snapshot(sess, iter)
rate *= cfg.TRAIN.GAMMA
sess.run(tf.assign(lr, rate))
next_stepsize = stepsizes.pop()
timer.tic()#給出時間
# Get training data, one batch at a time
blobs = self.data_layer.forward()#獲取訓練數據
now = time.time()#記錄時間
#如果迭代次數=1（最開始）或者到達了cfg.TRAIN.SUMMARY_INTERVAL次（180），保存一次tf計算states
if iter == 1 or now - last_summary_time > cfg.TRAIN.SUMMARY_INTERVAL:
# Compute the graph with summary
rpn_loss_cls, rpn_loss_box, loss_cls, loss_box, total_loss, summary = \
self.net.train_step_with_summary(sess, blobs, train_op)
self.writer.add_summary(summary, float(iter))
# Also check the summary on the validation set#獲取驗證機上的摘要
blobs_val = self.data_layer_val.forward()
summary_val = self.net.get_summary(sess, blobs_val)
self.valwriter.add_summary(summary_val, float(iter))
last_summary_time = now
else:#如果沒有到達，就直接運行計算模型，不需要保存摘要
# Compute the graph without summary
rpn_loss_cls, rpn_loss_box, loss_cls, loss_box, total_loss = \
self.net.train_step(sess, blobs, train_op)#self.net.train_step計算得到loss
timer.toc()# self.net.train_step再network.py定義，是計算模型
# Display training information
#打印訓練信息
if iter % (cfg.TRAIN.DISPLAY) == 0:#訓練到第幾次，Loss等等。
print('iter: %d / %d, total loss: %.6f\n >>> rpn_loss_cls: %.6f\n '
'>>> rpn_loss_box: %.6f\n >>> loss_cls: %.6f\n >>> loss_box: %.6f\n >>> lr: %f' % \
(iter, max_iters, total_loss, rpn_loss_cls, rpn_loss_box, loss_cls, loss_box, lr.eval()))
print('speed: {:.3f}s / iter'.format(timer.average_time))
# Snapshotting
if iter % cfg.TRAIN.SNAPSHOT_ITERS == 0:#到達了一定次數保存快照
last_snapshot_iter = iter
ss_path, np_path = self.snapshot(sess, iter)
np_paths.append(np_path)
ss_paths.append(ss_path)
# Remove the old snapshots if there are too many
if len(np_paths) > cfg.TRAIN.SNAPSHOT_KEPT:
self.remove_snapshot(np_paths, ss_paths)
iter += 1
if last_snapshot_iter != iter - 1:
self.snapshot(sess, iter - 1)
self.writer.close()
self.valwriter.close()
def get_training_roidb(imdb):#獲取訓練數據roidb
"""Returns a roidb (Region of Interest database) for use in training."""
if cfg.TRAIN.USE_FLIPPED:
print('Appending horizontally-flipped training examples...')
imdb.append_flipped_images()
print('done')
print('Preparing training data...')
rdl_roidb.prepare_roidb(imdb)
print('done')
return imdb.roidb
def filter_roidb(roidb):#刪除沒有用的rois
"""Remove roidb entries that have no usable RoIs."""
def is_valid(entry):
# Valid images have:
# (1) At least one foreground RoI OR
# (2) At least one background RoI
overlaps = entry['max_overlaps']
# find boxes with sufficient overlap
fg_inds = np.where(overlaps >= cfg.TRAIN.FG_THRESH)[0]
# Select background RoIs as those within [BG_THRESH_LO, BG_THRESH_HI)
bg_inds = np.where((overlaps < cfg.TRAIN.BG_THRESH_HI) &
(overlaps >= cfg.TRAIN.BG_THRESH_LO))[0]
# image is only valid if such boxes exist
valid = len(fg_inds) > 0 or len(bg_inds) > 0
return valid#簡單的驗證一下，roi數據是否達標，即至少有一個前景或者一個背景
num = len(roidb)
filtered_roidb = [entry for entry in roidb if is_valid(entry)]
num_after = len(filtered_roidb)
print('Filtered {} roidb entries: {} -> {}'.format(num - num_after,
num, num_after))
return filtered_roidb
def train_net(network, imdb, roidb, valroidb, output_dir, tb_dir,
pretrained_model=None,
max_iters=40000):
"""Train a Faster R-CNN network."""
roidb = filter_roidb(roidb)#訓練網絡
valroidb = filter_roidb(valroidb)
tfconfig = tf.ConfigProto(allow_soft_placement=True)
tfconfig.gpu_options.allow_growth = True
with tf.Session(config=tfconfig) as sess:
#在這裏條用了solverwrapper求解器
sw = SolverWrapper(sess, network, imdb, roidb, valroidb, output_dir, tb_dir,
pretrained_model=pretrained_model)
print('Solving...')#數據都準備好了，模型也張開了。訓練開始
sw.train_model(sess, max_iters)#運行模型訓練指令
print('done solving')#訓練結束