Research
Tokenizer Transplantation: Mitigating Autoregressive Collapse in Edge-Efficient Bengali ASR
arXiv:2607.09598v2 Announce Type: replace Abstract: Lightweight speech recognition models are critical for edge deployment, yet highly optimized architectures like Moonshine often fail on morphologica
arXiv:2607.09598v2 Announce Type: replace Abstract: Lightweight speech recognition models are critical for edge deployment, yet highly optimized architectures like Moonshine often fail on morphologically rich, non-Latin languages such as Bengali. This study identifies the root cause of this failure as the model's English-centric byte-level tokenizer, which fragments Bengali words into high-fertility byte chains and triggers catastrophic autoregressive collapse during inference. To resolve this, a novel vocabulary transplantation pipeline is proposed to replace the decoder vocabulary with the native-script BanglaBERT WordPiece vocabulary and resize the corresponding token embedding matrix. Experimental results demonstrate a reduction in token fertility from 9.16 to 1.30. By decreasing autoregressive sequence length by 85.8%, decoding instability is entirely mitigated. When evaluated on the 882-hour Lipi-Ghor dataset, the modified architecture achieves a competitive 21.54% Word Error Rate (WER) and a Real-Time Factor (RTF) of 0.0053. Ultimately, this research provides a scalable, reproducible blueprint for cross-script adaptation of compact ASR models without the need for resource-intensive pre-training.
Related
- Diagnostic-Driven Layer-Wise Compensation for Post-Training Quantization of Encoder-Decoder ASR Models
- MELD: Mel-Spectrogram-Based Speech Language Modeling with Discrete Latent Variables
- Improving Speech Recognition of Named Entities in Classroom Speech with LLM Revision and Phonetic-Semantic Context
Source: arXiv cs.CL | 2026-08-03