-
Notifications
You must be signed in to change notification settings - Fork 2
/
config_prediction.ini
25 lines (22 loc) · 1.16 KB
/
config_prediction.ini
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
; here the user can define all parameters needed for using models of the binding_in_disorder project for prediction
[workflow]
; overwrite existing files?
overwrite = True
[parameters]
; name or alias of the model used, leave blank to use the project's best model: FNN_all
model_name =
; int between 0 and 4, the id of the validation-fold of the model
; leave blank for using the best performing CV-fold model
fold =
; single float for classification cutoff, between 0.0 and 1.0
; leave blank for using the best cutoff, determined during the project.
cutoff =
; parameter that indicates which entries in the MobiDB annotation file define disordered residues; default: 'curated-disorder-merge'
disorder_param = prediction-disorder-mobidb_lite
[input_files]
; paths to all input files, current location being binding_in_disorder/src folder
dataset_directory = ../dataset/speed_assessment/
; disorder annotation either from MobiDB [.fasta] or predicted by SETH [disorder_pred.csv, ids.txt]
annotations = ../dataset/speed_assessment/mobidb_annotation.fasta
test_set_fasta = ../dataset/speed_assessment/dataset.fasta
test_set_embeddings = ../dataset/speed_assessment/embeddings_file.h5