Although the recall difference with XLM-RoBERTa was marginally non-significant, better precision and F1-score confirmed the model’s robustness.
← all excerpts
Ensemble transformer with post-hoc explanations for depression emotion and severity detection.
1
—
—