mask_rcnn_ros's Introduction

mask_rcnn_ros

A mask_rcnn ros version with realsense d435i image as input.(This repo was cloned and edit from https://github.com/matterport/Mask_RCNN)
An independent ros package for realsense d435i RGB-D info is in here
Some detail infomation need password to open. If you need, please email to [email protected]
If you like this repo, please give it a star, thank you very much!

software environment

ubuntu 16.04
ROS kinetic
python 3.5
tensorflow-gpu 1.9.0
CUDA 9.2.148
CUDNN 7
NVIDIA Driver 396.82
numpy 1.14.5
Keras 2.1.6
scikit-image 0.15.0
IPython 6.4.0
Pillow 6.0.0

GPU

TITAN X (Pascal) / 12GB

Q&A

If you suffer from any problem, you can read or open issues in this repo.

mask_rcnn_ros's People

Contributors

Stargazers

Watchers

mask_rcnn_ros's Issues

TypeError: resize() got an unexpected keyword argument 'clip'

Check your software environment

tensorflow-gpu 1.9.0
numpy 1.14.5
Keras 2.1.6
scikit-image 0.15.0

Could not create cudnn handle: CUDNN_STATUS_INTERNAL_ERROR

If you face this problem

2020-01-16 00:25:55.881381: E tensorflow/stream_executor/cuda/cuda_dnn.cc:329] Could not create cudnn handle: CUDNN_STATUS_INTERNAL_ERROR2020-01-16 00:25:55.901231: E tensorflow/stream_executor/cuda/cuda_dnn.cc:329] Could not create cudnn handle: CUDNN_STATUS_INTERNAL_ERROR2020-01-16 00:25:55.901300: W tensorflow/core/common_runtime/base_collective_executor.cc:216] BaseCollectiveExecutor::StartAbort Unknown: Failed to get convolution algorithm. This is probably because cuDNN failed to initialize, so try looking to see if a warning log message was printed above.

And you already make sure that you have been installed CUDA & Cudnn package correctly, the problems are maybe caused by Tensorflow.
Please refer to this

Solution:

Add this code at the beginning of the training code

for tf 1.xx
import tensorflow as tfconfig = tf.ConfigProto()

config.gpu_options.allow_growth = True

session = tf.InteractiveSession(config=config)

for tf 2.xx
import tensorflow as tf config = tf.compat.v1.ConfigProto()

config.gpu_options.allow_growth = True

session = tf.compat.v1.InteractiveSession(config=config)

Recommend Projects