3D convolutional neural network for video classification
270
stars
107
forks
source link
tf version updated to 1.9, same error (for 3D input) still occurs: Executor failed to create kernel. Invalid argument: Value for attr 'data_format' of "NDHWC" is not in the list of allowed values: "NHWC", "NCHW" #16
Have I written custom code:
Not very sure what custom code means here, I am trying to replicate this work: https://github.com/harvitronix/five-video-classification-methods. This error occurs only when I try model "Conv3d" or "c3d" both of which call keras.layers.Conv3D().
OS Platform and Distribution:
Ubuntu--16-0 with AWS DLAMI
TensorFlow installed from:
tensorflow 1.9.0 mkl_py36h6d6ce78_1
GPU model and memory:
No GPU, 4G
I am using AWS EC2 c5.large
Have I written custom code: Not very sure what custom code means here, I am trying to replicate this work: https://github.com/harvitronix/five-video-classification-methods. This error occurs only when I try model "Conv3d" or "c3d" both of which call keras.layers.Conv3D(). OS Platform and Distribution: Ubuntu--16-0 with AWS DLAMI TensorFlow installed from: tensorflow 1.9.0 mkl_py36h6d6ce78_1 GPU model and memory: No GPU, 4G I am using AWS EC2 c5.large
_Originally posted by @sailormoon2016 in https://github.com/kcct-fujimotolab/3DCNN/issue_comments#issuecomment-431714786_