Is there any ideal parameter values for fasttext train_supervised function?

Question

I working on NLP problem and try to make text classification with word embedding method. I am training my model with fasttext's train_supervised but is there any ideal or best parameter values for this function that you can advise me also I am using Kfold with some values how can I find best K-fold number in this problem ?

My solution is I am using fasttext's autotune function to find best param values for model to train but is there any possible suggestion to give me ? Following image shows my best params in the model. Finally , I am using fasttext's pretrained word vector model for my training.

Answer 1

Let me answer my own question you can look at the default and optimum parameters values by clicking the following link ;

https://fasttext.cc/docs/en/options.html

and also you can use fasttext's libraries autotune function (Automatic hyperparameter optimization) to find best parameters for your special train and validation dataset by clicking the following link ;

https://fasttext.cc/docs/en/autotune.html

and finally this is the pretrained word vectors provided by fasttext library to utilize in your model's training process also making positive progress for model , in the following link's site they are in the Model section ;

https://fasttext.cc/docs/en/crawl-vectors.html