Is there any way to track the dataset name, from which we trained our NER model using prodigy, Actually I am having lots of model folder from different datasets, I want to check the dataset of a particular model. If there is a way , please help me out.
Prodigy doesn't save out the dataset with the model, but you could use your own system to track this – for instance, use the dataset name as the name of the output directory, or run
db-out before/after you train. If you wanted this to be more elegant, you could also write your own shell script that reads the dataset name from an environment variable and then triggers the training, saves out the data and logs the command etc.