I'm trying to train a custom model through API and I'm facing many problems along the process.

I can upload a senteces file, attach it to a new project created through the API, but the processing always fails while the same file uploaded through the web interface is processed without problems. The sentences file is publicly accessible.

These are the steps I follow:

First I create a project and with that project id, I try to create the dataset

I can see the uploaded file being processed in speech studio, but always fails. What I'm doing wrong or what I'm missing?