faster rcnn源码解读（三）train_faster_rcnn_alt

转载自：faster rcnn源码解读（三）train_faster_rcnn_alt_opt.py - 野孩子的专栏 - 博客频道 - CSDN.NET

http://blog.csdn.net/u010668907/article/details/51945320

faster用python版本的https://github.com/rbgirshick/py-faster-rcnn

train_faster_rcnn_alt_opt.py源码在https://github.com/rbgirshick/py-faster-rcnn/blob/master/tools/train_faster_rcnn_alt_opt.py

faster rcnn训练的开始是：faster_rcnn_alt_opt.sh。下面命令是训练的，还有它的参数说明。

1.调用最初脚本的说明

cd $FRCN_ROOT

# ./experiments/scripts/faster_rcnn_alt_opt.sh GPU NET DATASET [options args to {train,test}_net.py]

# GPU_ID is the GPU you want to train on

# NET in {ZF, VGG_CNN_M_1024, VGG16} is the network arch to use

# DATASET is only pascal_voc for now

train_faster_rcnn_alt_opt.py的源码：

[python] view plaincopy print?

#!/usr/bin/env python
# --------------------------------------------------------
# Faster R-CNN
# Copyright (c) 2015 Microsoft
# Licensed under The MIT License [see LICENSE for details]
# Written by Ross Girshick
# --------------------------------------------------------
"""Train a Faster R-CNN network using alternating optimization.
This tool implements the alternating optimization algorithm described in our
NIPS 2015 paper ("Faster R-CNN: Towards Real-time Object Detection with Region
Proposal Networks." Shaoqing Ren, Kaiming He, Ross Girshick, Jian Sun.)
"""
import _init_paths
from fast_rcnn.train import get_training_roidb, train_net
from fast_rcnn.config import cfg, cfg_from_file, cfg_from_list, get_output_dir
from datasets.factory import get_imdb
from rpn.generate import imdb_proposals
import argparse
import pprint
import numpy as np
import sys, os
import multiprocessing as mp
import cPickle
import shutil
def parse_args():
"""
Parse input arguments
"""
parser = argparse.ArgumentParser(description='Train a Faster R-CNN network')
parser.add_argument('--gpu', dest='gpu_id',
help='GPU device id to use [0]',
default=0, type=int)
parser.add_argument('--net_name', dest='net_name',
help='network name (e.g., "ZF")',
default=None, type=str)
parser.add_argument('--weights', dest='pretrained_model',
help='initialize with pretrained model weights',
default=None, type=str)
parser.add_argument('--cfg', dest='cfg_file',
help='optional config file',
default=None, type=str)
parser.add_argument('--imdb', dest='imdb_name',
help='dataset to train on',
default='voc_2007_trainval', type=str)
parser.add_argument('--set', dest='set_cfgs',
help='set config keys', default=None,
nargs=argparse.REMAINDER)
if len(sys.argv) == 1:
parser.print_help()
sys.exit(1)
args = parser.parse_args()
return args
def get_roidb(imdb_name, rpn_file=None):
imdb = get_imdb(imdb_name)
print 'Loaded dataset `{:s}` for training'.format(imdb.name)
imdb.set_proposal_method(cfg.TRAIN.PROPOSAL_METHOD)
print 'Set proposal method: {:s}'.format(cfg.TRAIN.PROPOSAL_METHOD)
if rpn_file is not None:
imdb.config['rpn_file'] = rpn_file
roidb = get_training_roidb(imdb)
return roidb, imdb
def get_solvers(net_name):
# Faster R-CNN Alternating Optimization
n = 'faster_rcnn_alt_opt'
# Solver for each training stage
solvers = [[net_name, n, 'stage1_rpn_solver60k80k.pt'],
[net_name, n, 'stage1_fast_rcnn_solver30k40k.pt'],
[net_name, n, 'stage2_rpn_solver60k80k.pt'],
[net_name, n, 'stage2_fast_rcnn_solver30k40k.pt']]
solvers = [os.path.join(cfg.MODELS_DIR, *s) for s in solvers]
# Iterations for each training stage
max_iters = [80000, 40000, 80000, 40000]
# max_iters = [100, 100, 100, 100]
# Test prototxt for the RPN
rpn_test_prototxt = os.path.join(
cfg.MODELS_DIR, net_name, n, 'rpn_test.pt')
return solvers, max_iters, rpn_test_prototxt
# ------------------------------------------------------------------------------
# Pycaffe doesn't reliably free GPU memory when instantiated nets are discarded
# (e.g. "del net" in Python code). To work around this issue, each training
# stage is executed in a separate process using multiprocessing.Process.
# ------------------------------------------------------------------------------
def _init_caffe(cfg):
"""Initialize pycaffe in a training process.
"""
import caffe
# fix the random seeds (numpy and caffe) for reproducibility
np.random.seed(cfg.RNG_SEED)
caffe.set_random_seed(cfg.RNG_SEED)
# set up caffe
caffe.set_mode_gpu()
caffe.set_device(cfg.GPU_ID)
def train_rpn(queue=None, imdb_name=None, init_model=None, solver=None,
max_iters=None, cfg=None):
"""Train a Region Proposal Network in a separate training process.
"""
# Not using any proposals, just ground-truth boxes
cfg.TRAIN.HAS_RPN = True
cfg.TRAIN.BBOX_REG = False # applies only to Fast R-CNN bbox regression
cfg.TRAIN.PROPOSAL_METHOD = 'gt'
cfg.TRAIN.IMS_PER_BATCH = 1
print 'Init model: {}'.format(init_model)
print('Using config:')
pprint.pprint(cfg)
import caffe
_init_caffe(cfg)
roidb, imdb = get_roidb(imdb_name)
print 'roidb len: {}'.format(len(roidb))
output_dir = get_output_dir(imdb)
print 'Output will be saved to `{:s}`'.format(output_dir)
model_paths = train_net(solver, roidb, output_dir,
pretrained_model=init_model,
max_iters=max_iters)
# Cleanup all but the final model
for i in model_paths[:-1]:
os.remove(i)
rpn_model_path = model_paths[-1]
# Send final model path through the multiprocessing queue
queue.put({'model_path': rpn_model_path})
def rpn_generate(queue=None, imdb_name=None, rpn_model_path=None, cfg=None,
rpn_test_prototxt=None):
"""Use a trained RPN to generate proposals.
"""
cfg.TEST.RPN_PRE_NMS_TOP_N = -1 # no pre NMS filtering
cfg.TEST.RPN_POST_NMS_TOP_N = 2000 # limit top boxes after NMS
print 'RPN model: {}'.format(rpn_model_path)
print('Using config:')
pprint.pprint(cfg)
import caffe
_init_caffe(cfg)
# NOTE: the matlab implementation computes proposals on flipped images, too.
# We compute them on the image once and then flip the already computed
# proposals. This might cause a minor loss in mAP (less proposal jittering).
imdb = get_imdb(imdb_name)
print 'Loaded dataset `{:s}` for proposal generation'.format(imdb.name)
# Load RPN and configure output directory
rpn_net = caffe.Net(rpn_test_prototxt, rpn_model_path, caffe.TEST)
output_dir = get_output_dir(imdb)
print 'Output will be saved to `{:s}`'.format(output_dir)
# Generate proposals on the imdb
rpn_proposals = imdb_proposals(rpn_net, imdb)
# Write proposals to disk and send the proposal file path through the
# multiprocessing queue
rpn_net_name = os.path.splitext(os.path.basename(rpn_model_path))[0]
rpn_proposals_path = os.path.join(
output_dir, rpn_net_name + '_proposals.pkl')
with open(rpn_proposals_path, 'wb') as f:
cPickle.dump(rpn_proposals, f, cPickle.HIGHEST_PROTOCOL)
print 'Wrote RPN proposals to {}'.format(rpn_proposals_path)
queue.put({'proposal_path': rpn_proposals_path})
def train_fast_rcnn(queue=None, imdb_name=None, init_model=None, solver=None,
max_iters=None, cfg=None, rpn_file=None):
"""Train a Fast R-CNN using proposals generated by an RPN.
"""
cfg.TRAIN.HAS_RPN = False # not generating prosals on-the-fly
cfg.TRAIN.PROPOSAL_METHOD = 'rpn' # use pre-computed RPN proposals instead
cfg.TRAIN.IMS_PER_BATCH = 2
print 'Init model: {}'.format(init_model)
print 'RPN proposals: {}'.format(rpn_file)
print('Using config:')
pprint.pprint(cfg)
import caffe
_init_caffe(cfg)
roidb, imdb = get_roidb(imdb_name, rpn_file=rpn_file)
output_dir = get_output_dir(imdb)
print 'Output will be saved to `{:s}`'.format(output_dir)
# Train Fast R-CNN
model_paths = train_net(solver, roidb, output_dir,
pretrained_model=init_model,
max_iters=max_iters)
# Cleanup all but the final model
for i in model_paths[:-1]:
os.remove(i)
fast_rcnn_model_path = model_paths[-1]
# Send Fast R-CNN model path over the multiprocessing queue
queue.put({'model_path': fast_rcnn_model_path})
if __name__ == '__main__':
args = parse_args()
print('Called with args:')
print(args)
if args.cfg_file is not None:
cfg_from_file(args.cfg_file)
if args.set_cfgs is not None:
cfg_from_list(args.set_cfgs)
cfg.GPU_ID = args.gpu_id
# --------------------------------------------------------------------------
# Pycaffe doesn't reliably free GPU memory when instantiated nets are
# discarded (e.g. "del net" in Python code). To work around this issue, each
# training stage is executed in a separate process using
# multiprocessing.Process.
# --------------------------------------------------------------------------
# queue for communicated results between processes
mp_queue = mp.Queue()
# solves, iters, etc. for each training stage
solvers, max_iters, rpn_test_prototxt = get_solvers(args.net_name)
print '~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~'
print 'Stage 1 RPN, init from ImageNet model'
print '~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~'
cfg.TRAIN.SNAPSHOT_INFIX = 'stage1'
mp_kwargs = dict(
queue=mp_queue,
imdb_name=args.imdb_name,
init_model=args.pretrained_model,
solver=solvers[0],
max_iters=max_iters[0],
cfg=cfg)
p = mp.Process(target=train_rpn, kwargs=mp_kwargs)
p.start()
rpn_stage1_out = mp_queue.get()
p.join()
print '~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~'
print 'Stage 1 RPN, generate proposals'
print '~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~'
mp_kwargs = dict(
queue=mp_queue,
imdb_name=args.imdb_name,
rpn_model_path=str(rpn_stage1_out['model_path']),
cfg=cfg,
rpn_test_prototxt=rpn_test_prototxt)
p = mp.Process(target=rpn_generate, kwargs=mp_kwargs)
p.start()
rpn_stage1_out['proposal_path'] = mp_queue.get()['proposal_path']
p.join()
print '~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~'
print 'Stage 1 Fast R-CNN using RPN proposals, init from ImageNet model'
print '~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~'
cfg.TRAIN.SNAPSHOT_INFIX = 'stage1'
mp_kwargs = dict(
queue=mp_queue,
imdb_name=args.imdb_name,
init_model=args.pretrained_model,
solver=solvers[1],
max_iters=max_iters[1],
cfg=cfg,
rpn_file=rpn_stage1_out['proposal_path'])
p = mp.Process(target=train_fast_rcnn, kwargs=mp_kwargs)
p.start()
fast_rcnn_stage1_out = mp_queue.get()
p.join()
print '~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~'
print 'Stage 2 RPN, init from stage 1 Fast R-CNN model'
print '~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~'
cfg.TRAIN.SNAPSHOT_INFIX = 'stage2'
mp_kwargs = dict(
queue=mp_queue,
imdb_name=args.imdb_name,
init_model=str(fast_rcnn_stage1_out['model_path']),
solver=solvers[2],
max_iters=max_iters[2],
cfg=cfg)
p = mp.Process(target=train_rpn, kwargs=mp_kwargs)
p.start()
rpn_stage2_out = mp_queue.get()
p.join()
print '~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~'
print 'Stage 2 RPN, generate proposals'
print '~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~'
mp_kwargs = dict(
queue=mp_queue,
imdb_name=args.imdb_name,
rpn_model_path=str(rpn_stage2_out['model_path']),
cfg=cfg,
rpn_test_prototxt=rpn_test_prototxt)
p = mp.Process(target=rpn_generate, kwargs=mp_kwargs)
p.start()
rpn_stage2_out['proposal_path'] = mp_queue.get()['proposal_path']
p.join()
print '~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~'
print 'Stage 2 Fast R-CNN, init from stage 2 RPN R-CNN model'
print '~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~'
cfg.TRAIN.SNAPSHOT_INFIX = 'stage2'
mp_kwargs = dict(
queue=mp_queue,
imdb_name=args.imdb_name,
init_model=str(rpn_stage2_out['model_path']),
solver=solvers[3],
max_iters=max_iters[3],
cfg=cfg,
rpn_file=rpn_stage2_out['proposal_path'])
p = mp.Process(target=train_fast_rcnn, kwargs=mp_kwargs)
p.start()
fast_rcnn_stage2_out = mp_queue.get()
p.join()
# Create final model (just a copy of the last stage)
final_path = os.path.join(
os.path.dirname(fast_rcnn_stage2_out['model_path']),
args.net_name + '_faster_rcnn_final.caffemodel')
print 'cp {} -> {}'.format(
fast_rcnn_stage2_out['model_path'], final_path)
shutil.copy(fast_rcnn_stage2_out['model_path'], final_path)
print 'Final model: {}'.format(final_path)

2. train_faster_rcnn_alt_opt.py的部分参数说明

net_name: {ZF, VGG_CNN_M_1024, VGG16}

pretrained_model: data/imagenet_models/${net_name}.v2.caffemodel

cfg_file: experiments/cfgs/faster_rcnn_alt_opt.yml

imdb_name: "voc_2007_trainval" or "voc_2007_test"

cfg.TRAIN.HAS_RPN = True表示用xml提供的propoal

cfg是配置文件，它的默认值放在上面的cfg_file里，其他还可以自己写配置文件之后与默认配置文件融合。

2.1 net_name是用get_solvers()找到网络。还要用到cfg的参数MODELS_DIR，

例子是join（MODELS_DIR， net_name， 'faster_rcnn_alt_opt', 'stage1_rpn_solver60k80k.pt'）

2.2 imdb_name在factory中被拆成‘2007’（year）和‘trainval’/‘test’（split）到类pascal_voc中产生相应的imdb

2.3 整个step的大致流程：

(ImageNet model)->stage1_rpn_train->rpn_test

|(proposal_path)

(ImageNetmodel)->stage1_fast_rcnn_train-> stage2_rpn_train-> rpn_test-> stage2_fast_rcnn_train

2.4 数据imdb和roidb

roidb原本是imdb的一个属性，但imdb其实是为了计算roidb存在的，他所有的其他属性和方法都是为了计算roidb

faster rcnn源码解读（三）train_faster_rcnn_alt_opt.py相关推荐

faster rcnn源码解读（四）之数据类型imdb.py和pascal_voc.py（主要是imdb和roidb数据类型的解说）
转载自:faster rcnn源码解读(四)之数据类型imdb.py和pascal_voc.py(主要是imdb和roidb数据类型的解说) - 野孩子的专栏 - 博客频道 - CSDN.NET ht ...
faster rcnn源码解读（六）之minibatch
转载自:faster rcnn源码解读(六)之minibatch - 野孩子的专栏 - 博客频道 - CSDN.NET http://blog.csdn.net/u010668907/article/ ...
faster rcnn源码解读（五）之layer（网络里的input-data）
转载自:faster rcnn源码解读(五)之layer(网络里的input-data) - 野孩子的专栏 - 博客频道 - CSDN.NET http://blog.csdn.net/u010668 ...
faster rcnn源码解读总结
转载自:faster rcnn源码解读总结 - 野孩子的专栏 - 博客频道 - CSDN.NET http://blog.csdn.net/u010668907/article/details/519 ...
还不懂目标检测嘛？一起来看看Faster R-CNN源码解读
【Faster R-CNN论文精度系列】从Faster R-CNN源码中，我们“学习”到了什么？
[Faster R-CNN论文精度系列] (如下为建议阅读顺序) 1[Faster R-CNN论文精度系列]从Faster R-CNN源码中,我们"学习"到了什么? 2[Faste ...
Mask RCNN源码解读
Mask RCNN源码解读前言数据集数据载入模型搭建模型输入模型输出 resnet101 RPN网络 ProposalLayer DetectionTargetLayer fpn_clas ...
faster rcnn源码理解（二）之AnchorTargetLayer（网络中的rpn_data）
转载自:faster rcnn源码理解(二)之AnchorTargetLayer(网络中的rpn_data) - 野孩子的专栏 - 博客频道 - CSDN.NET http://blog.csdn.n ...
Faster R-CNN源码中ＲＰＮ的解析（自用）
参考博客(一定要看前面两个) 一文看懂Faster R-CNN 详细的Faster R-CNN源码解析之RPN源码解析关于RPN一些我的想法 rpn的中心思想就是在了anchors了,如何产生anc ...

faster rcnn源码解读（三）train_faster_rcnn_alt_opt.py

faster rcnn源码解读（三）train_faster_rcnn_alt_opt.py相关推荐

最新文章

热门文章