Semantic training spunet folder

operators.ml3d.semantic_training_spunet_folder(client, data_in_folder='/data_in_folder', out_model_parameters_path='trained_model/model_1', class_names='1,2,3,4,5,6,7,8', feature_names='red,green,blue', point_names='X,Y,Z', label_name='classification', resolution=0.05, max_epochs=500, learning_rate=0.01, batch_size=10, final_div_factor=100, div_factor=10, weight_decay=0.005, channels='32,64,128,256,256,128,96,96', layers='2,3,4,6,2,2,2,2', worker_instance_type='P2', manager_instance_type='small', extension_data_in_path='.laz', skip_existing_files=False)
Spunet Training

semantic_training_spunet_folder(client,
data_in_folder=’/data_in_folder’,
out_model_parameters_path=’trained_model/model_1’,
class_names=’1,2,3,4,5,6,7,8’,
feature_names=’red,green,blue’,
point_names=’X,Y,Z’,
label_name=’classification’,
resolution=0.05,
max_epochs=500,
learning_rate=0.01,
batch_size=10,
final_div_factor=100,
div_factor=10,
weight_decay=0.005,
channels=’32,64,128,256,256,128,96,96’,
layers=’2,3,4,6,2,2,2,2’,
worker_instance_type=’P2’,
manager_instance_type=”small”,
extension_data_in_folder=”./data/files/”,
skip_existing_files = False )
Parameters:

  • out_model_parameters_path – path to model

  • class_names – comma separated list of class names. Class 0 is always given and is used to denote unlabeled points.

  • feature_names – comma separated list of features that are provided

  • point_names – comma separated list of point identifiers in (las/laz)

  • label_name – label name for (las/laz)

  • resolution – resolution of the subsampled point cloud

  • max_epochs – maximum number of epochs

  • learning_rate – learning rate

  • batch_size – batch size

  • final_div_factor – final div factor for learning rate

  • div_factor – div factor for learning rate

  • weight_decay – weight decay

  • channels – comma separated list of channels

  • layers – comma separated list of layers

  • data_in_folder – folder to folder that contains the training data

  • worker_instance_type – cloud instance type of worker nodes

  • manager_instance_type – cloud instance type of manager node

  • extension_data_in_folder – File extension of files in folder for data_in_folder

  • skip_existing_files – skip files that already exist in the output folder