search

yangyanli / PointCNN

PointCNN: Convolution On X-Transformed Points (NeurIPS 2018)
https://arxiv.org/abs/1801.07791
Other
1.36k stars 365 forks source link

The Problem about the PointCNN #219

Closed kevin-steiner closed 4 years ago

kevin-steiner commented 4 years ago

2ce0093a597a45490849e281fb23805

When I try to run the code with GPU, tensorflow=1.14, it comes the error above. How can I solve it?

danbull1 commented 4 years ago

Hi Kevin I stumbled across exactly the same error when attempting to run this code on tf1.14. I fixed this by updating the following line of code:

iterator = tf.data.Iterator.from_string_handle(handle, dataset_train.output_types,dataset_train.output_shapes) at line 147 of train_vas_cls.py i.e. adding an explicit shape to the iterator.

kevin-steiner commented 4 years ago

Hi Kevin I stumbled across exactly the same error when attempting to run this code on tf1.14. I fixed this by updating the following line of code:

iterator = tf.data.Iterator.from_string_handle(handle, dataset_train.output_types,dataset_train.output_shapes) at line 147 of train_vas_cls.py i.e. adding an explicit shape to the iterator.

Thank you very much, it works well. BTW, do you have any idea of how to get the feature vector of the input data.

danbull1 commented 4 years ago

Look - Im just in the process of stepping through the code to try and understand what is going on. Points are loaded in this function: Line122 data_utils: load_cls loads pts into array e.g. [0:6] = [0.098043844, -0.16375712, -0.88815236, 0.0, 0.0, 1.0]