config command for crYOLO.
Tags
templatejavaimagej2scijava
Citation
Wagner, T., Merino, F., Stabrin, M., Moriya, T., Antoni, C., Apelbaum, A., Hagel, P., Sitsel, O., Raisch, T., Prumbaum, D. and Quentin, D., 2019. SPHIRE-crYOLO is a fast and accurate fully automated particle picker for cryo-EM. Communications biology, 2(1), p.218..https://cryolo.readthedocs.io/en/stable/index.html
Solution written by
Kyle Harrington
License of solution
MIT
Arguments
--train_image_folder
Path to the image folder containing the images to train on. (default value: PARAMETER_VALUE)
--train_annot_folder
Path to folder containing the your annotation files. (default value: PARAMETER_VALUE)
--saved_weights_name
Path for saving final weights. (default value: PARAMETER_VALUE)
--architecture
Backend network architecture. (default value: PARAMETER_VALUE)
--input_size
Input size for the network. (default value: PARAMETER_VALUE)
--norm
Normalization method applied to the images. (default value: PARAMETER_VALUE)
--num_patches
(DEPRECATED) Number of patches if patch mode is used. (default value: PARAMETER_VALUE)
--overlap_patches
(DEPRECATED) Overlap of patches. (default value: PARAMETER_VALUE)
--filtered_output
Output folder for filtered images. (default value: PARAMETER_VALUE)
--filter
Noise filter applied before training/picking. (default value: PARAMETER_VALUE)
--low_pass_cutoff
Low pass filter cutoff frequency. (default value: PARAMETER_VALUE)
--janni_model
Path to JANNI model. (default value: PARAMETER_VALUE)
--janni_overlap
Overlap of patches in pixels for JANNI. (default value: PARAMETER_VALUE)
--janni_batches
Number of batches for JANNI. (default value: PARAMETER_VALUE)
--pretrained_weights
Path to h5 file for initialization for fine-tuning. (default value: PARAMETER_VALUE)
--train_times
How often each image is presented during one epoch. (default value: PARAMETER_VALUE)
--batch_size
Number of images processed in parallel during training. (default value: PARAMETER_VALUE)
--learning_rate
Step size during training. (default value: PARAMETER_VALUE)
--nb_epoch
Maximum number of epochs for training. (default value: PARAMETER_VALUE)
--object_scale
Penalty scaling factor for missing picking particles. (default value: PARAMETER_VALUE)
--no_object_scale
Penalty scaling factor for picking background. (default value: PARAMETER_VALUE)
--coord_scale
Penalty scaling factor for errors in estimating position. (default value: PARAMETER_VALUE)
--class_scale
Penalty scaling factor for class estimation errors. (default value: PARAMETER_VALUE)
--debug
Provide statistics during training if true. (default value: PARAMETER_VALUE)
--valid_image_folder
Path to folder containing validation image files. (default value: PARAMETER_VALUE)
--valid_annot_folder
Path to folder containing validation box files. (default value: PARAMETER_VALUE)
--log_path
Path for log saving. (default value: PARAMETER_VALUE)