Lightweight CNN-transformer hybrid network for English speech recognition

Key Points

• The aim is to develop a lightweight network combining CNN and transformer architectures for improved speech recognition.
• Developed a hybrid neural network using CNN and transformer components.
• Evaluated the model's performance on English speech recognition tasks.
• Compared with existing speech recognition systems to assess accuracy and efficiency.
• Achieved higher accuracy rates than traditional models in speech recognition tasks.
• Reduced computational requirements, indicating a more efficient processing capability.