arXiv - CSCL: "End-to-end spoken language understanding using jo…" - Qoto Mastodon

arXiv - CSCL @arxiv_cscl@qoto.org

End-to-end spoken language understanding using joint CTC loss and self-supervised, pretrained acoustic encoders. (arXiv:2305.02937v2 [cs.CL] UPDATED)

http://arxiv.org/abs/2305.02937 #arXiv #NLProc

Jun 05, 2023, 03:20 · · arxiv-cscl · · ·

Sign in to participate in the conversation