Demos using real datasets

[kevinnkansah]

Could someone please make demos on how to use the Encoders correctly on an actual dataset like (emotions dataset) for example because I am struggling to use SequenceEncoder. Noob in NLP btw

[koaning]

I am not behind a laptop now, but could you elaborate on the sequence encoder? I dont recall a component with that name in this repository.

[koaning]

It looks ok when I glance at this code, do you receive an error message?

[kevinnkansah]

No there are no syntax or runtime errors but i think some sort of logic error. I get a model score of 0, but i restarted my notebook and it was fine.

[koaning]

Could you share the confusion matrix?

[kevinnkansah]

I believe it is like that but I accidently made the target variable the training variable, so this is my fault

This is it after I restarted my notebook, also I am sampling only 250 data entries instead of 416809 in the original dataset but it takes forever even just to do 500 data entries , my score is also 0.34 btw

Link to dataset:

[koaning]

embeddings can be slow yeah but thats something i cant fix directly because the underlying models carry heavy compute workloads.

[kevinnkansah]

Here is one but with 10000 data entries