{"slug": "pruned-ctc-for-memory-efficient-large-vocabulary-asr-training", "title": "Pruned CTC for Memory-Efficient Large-Vocabulary ASR Training", "summary": "Researchers present pruned CTC, a method that avoids materializing frame-by-vocabulary activations in memory to make large-vocabulary ASR training feasible. Conventional CTC implementations materialize these activations, which the work states makes CTC training with native LLM vocabularies prohibitively memory-intensive. The approach targets memory-efficient training of connectionist temporal classification models that support both offline and streaming speech recognition under utterance-level supervision.", "body_md": "Connectionist temporal classification (CTC) naturally supports offline and streaming speech recognition with utterance-level supervision, but conventional implementations materialize frame-by-vocabulary activations in memory, making CTC training with native LLM vocabularies prohibitively memory-inte", "url": "https://wpnews.pro/news/pruned-ctc-for-memory-efficient-large-vocabulary-asr-training", "canonical_source": "https://aiflash.com/news/128586/", "published_at": "2026-09-29 15:01:10+00:00", "updated_at": "2026-09-29 15:22:07.481357+00:00", "lang": "en", "topics": ["artificial-intelligence", "machine-learning", "natural-language-processing", "ai-research"], "entities": [], "also_reported_by": [], "alternates": {"html": "https://wpnews.pro/news/pruned-ctc-for-memory-efficient-large-vocabulary-asr-training", "markdown": "https://wpnews.pro/news/pruned-ctc-for-memory-efficient-large-vocabulary-asr-training.md", "text": "https://wpnews.pro/news/pruned-ctc-for-memory-efficient-large-vocabulary-asr-training.txt", "jsonld": "https://wpnews.pro/news/pruned-ctc-for-memory-efficient-large-vocabulary-asr-training.jsonld"}}