biobot500 / neuralgym Goto Github PK

python app for training spaCy models

Option 1: No installation required. Download and unpack(extract) NGym.7z
Launch ng.exe
Option 2: Download NGym folder with python files. Run ng.pyw (You will need pyqt5, spaCy and matpotlib to be installed on your PC). In this mode you can use spacy.prefer_gpu() option.

1. Create an output directory where the trained model will be saved.
   
2. Select a training data file. Training data should be in spaCy data format. You can use TagEditor to create your training data. See axample of training data in file train_data.txt. Make sure your training data is utf-8 encoded.
   
3. Select a model to train (it can be any spaCy model compatible with spaCy 2.2+) or create a blank model. Labels in the training data should match labels in the original model otherwise start from blank model.
   
4. Check on Parameter averaging so the model to be saved with parameter averaging after training is done.
5. Press Start. You can disrupt training process at any time by clicking stop or stop and save.