User Tools

Site Tools


cinderella_tomograms

Differences

This shows you the differences between two versions of the page.

Link to this comparison view

Both sides previous revision Previous revision
Next revision
Previous revision
Last revision Both sides next revision
cinderella_tomograms [2019/12/13 15:52]
twagner
cinderella_tomograms [2019/12/13 16:12]
twagner [Training]
Line 28: Line 28:
 After your created your training data, you can start the training :-)  After your created your training data, you can start the training :-) 
  
-You need to specifiy all settings into one config file. To do so, create an empty file with+You need to specify all settings into one config file. To do so, create an empty file using
 <code> <code>
 touch config.json touch config.json
 </code> </code>
  
-Copy the following configuration into the file and adapt it to your needs. The only entries you might want to change are the //input_size//, //good_path//, //bad_path// and //pretrained_weights//.+Copy the following configuration into the new file and adapt it to your needs. The only entries you might want to change are the //input_size//, //good_path//, //bad_path// and //pretrained_weights//.
  
 <code json config.json> <code json config.json>
Line 54: Line 54:
 </code> </code>
 The fields have the following meaning: The fields have the following meaning:
-  * **input_size**: This is the image size to which each central slice is resized to. +  * **input_size**: Each central slice is resized to these dimensions
-  * **batch_size**: How many images are in one mini-batch. If you have memory problemsyou can try to reduce this value. +  * **batch_size**: The number of images in one mini-batch. If you have memory problems you can try to reduce this value. 
-  * **good_path**: Path to folder with good central slices.+  * **good_path**: Path of folder containing good central slices. 
-  * **bad_path**: Path to folder with bad central slices.+  * **bad_path**: Path of folder containing bad central slices.
   * **pretrained_weights**: Path to weights that are used to initialize the network. It can be empty. As Cinderella is using the same network architecture as crYOLO, we are typically using the [[downloads:cryolo_1#general_phosaurusnet_models|general network of crYOLO]] as pretrained weights.   * **pretrained_weights**: Path to weights that are used to initialize the network. It can be empty. As Cinderella is using the same network architecture as crYOLO, we are typically using the [[downloads:cryolo_1#general_phosaurusnet_models|general network of crYOLO]] as pretrained weights.
-  * **saved_weights_name**: Final model filename+  * **saved_weights_name**: Final model filename.
   * **learning_rate**: Learning rate, should not be changed.   * **learning_rate**: Learning rate, should not be changed.
   * **nb_epoch**: Maximum number of epochs to train. However, it will stop earlier (see nb_early_stop).   * **nb_epoch**: Maximum number of epochs to train. However, it will stop earlier (see nb_early_stop).
Line 70: Line 70:
 </code> </code>
  
-This will train a classification network on the GPU with ID=1. After the training finishes, you get a ''my_model.h5'' file. This can then be used to classify subtomograms into good / bad categories.+This will train a classification network on the GPU with ID=1. Once the training finishes, you get a ''my_model.h5'' file. This can then be used to classify subtomograms into good / bad categories.
  
  
cinderella_tomograms.txt · Last modified: 2019/12/13 16:13 by twagner