一、前言:
学无止境,在学习Mask RCNN的理论及使用方法时,有以下文章及github开源代码为我提供了很大帮助,特此附上链接,表示感谢!希望对您也有参考借鉴的意义!
二、学习笔记 / 翻译分析
1. 论文链接:
2. 论文翻译:
3. 论文详解:
(1) mask-rcnn解读
4. 整理笔记:
三、源码
1.(推荐)https://github.com/matterport/Mask_RCNN
四、运行出现问题及解决办法
1. 大佬文章,感激不尽!
Fix multiple GPUs fails in training Mask_RCNN
2. 选择多GPU并行运算时,部分报错内容如下:
No node-device colocations were active during op 'tower_1/mask_rcnn/anchors/Variable/anchors/Variable/read_tower_1/mask_rcnn/anchors/Variable_0' creation.
Device assignments active during op 'tower_1/mask_rcnn/anchors/Variable/anchors/Variable/read_tower_1/mask_rcnn/anchors/Variable_0' creation:
with tf.device(/gpu:1): <M:\new\mrcnn\parallel_model.py:70>
No node-device colocations were active during op 'anchors/Variable' creation.
No device assignments were active during op 'anchors/Variable' creation.
解决办法:
在train.py中,添加‘config.allow_soft_placement=True’:
# 设置占用的GPU显存
if ('tensorflow' == K.backend()):
import tensorflow as tf
from keras.backend.tensorflow_backend import set_session
config = tf.ConfigProto()
# config.gpu_options.allow_growth = False
config.allow_soft_placement=True # 新加
config.gpu_options.per_process_gpu_memory_fraction = 0.5
set_session(tf.Session(config=config))