Splitformer: An improved early-exit architecture for automatic speech recognition on edge devices

Documenti analoghi

• Delayed Fusion: Integrating Large Language Models into First-Pass Decoding in End-to-end Speech Recognition
  di: Hori, Takaaki, et al.
  Pubblicazione: (2025)
• Quantifying the effect of speech pathology on automatic and human speaker verification
  di: Halpern, Bence Mark, et al.
  Pubblicazione: (2024)
• Quantization for OpenAI's Whisper Models: A Comparative Analysis
  di: Andreyev, Allison
  Pubblicazione: (2025)
• A Unified Model For Voice and Accent Conversion In Speech and Singing using Self-Supervised Learning and Feature Extraction
  di: Cheripally, Sowmya
  Pubblicazione: (2024)
• Thaka at KSAA-2026 Task 2: Regularized Fine-Tuning for Arabic Speech Diacritization
  di: Alamr, Meshal, et al.
  Pubblicazione: (2026)