Create integration tests for session-based recommendation

[gabrielspmoreira]

Description

Based on the training/eval script created in #806 , create unit/integration tests that ensures the stability of the API when setting the advanced hparams and also allowing to track in our CI accuracy or performance regressions

Additional context
We have implemented similar unit/integration tests for retrieval models

The integration tests will cover four different architectures :

• GPT-2 (CLM) - Item Id feature
• BERT (MLM) - Item Id feature
• XLNet (MLM) - Item Id feature
• XLNET (MLM) - CONCAT - All features

[viswa-nvidia]

@sararb , please provide context on the 4 architecutres in the description