Whisper-small
PulseAugur coverage of Whisper-small — every cluster mentioning Whisper-small across labs, papers, and developer communities, ranked by signal.
3 day(s) with sentiment data
-
New dataset MGhana-ST targets low-resource Ghanaian languages
Researchers have introduced MGhana-ST, a new speech translation dataset designed for four low-resource Ghanaian languages: Ga, Twi, Ewe, and Fante. The dataset, which includes paired audio and English translations with …
-
Pruning Whisper-small improves ASR accuracy by acting as a regularizer
Researchers have demonstrated that neural network pruning can act as a regularization technique for Automatic Speech Recognition (ASR) systems, rather than just a method for compression. By analyzing the sensitivity of …
-
New research probes acoustic information loss in audio-conditioned LLMs
Researchers have investigated why audio-conditioned language models often fail to utilize crucial acoustic cues like prosody and emotion. Their study, detailed on arXiv, tested various audio encoders including Whisper-T…
-
Whisper model adapted for indigenous Baniwa language speech recognition
Researchers have successfully adapted OpenAI's Whisper model to perform automatic speech recognition for the Baniwa language, an indigenous Arawakan language spoken across Brazil, Colombia, and Venezuela. Using a small …
-
Whisper ASR Models Adapted for Multilingual Medical Use
Researchers have analyzed how multilingual medical adaptation affects the internal representations of Whisper models. The study compared various fine-tuning strategies across different Whisper model sizes, finding that …
-
Apple's SpeechAnalyzer API enhances on-device transcription speed
Apple has introduced its SpeechAnalyzer API, designed for on-device English speech transcription. This new API utilizes the Apple Neural Engine to achieve significant improvements in speed and accuracy, reportedly being…
-
ASR systems evaluated for low-resource African language text corpora
Researchers have evaluated the effectiveness of Automatic Speech Recognition (ASR) systems for creating text corpora for low-resource African languages, specifically Fongbe and Hausa. By fine-tuning the MMS-300M model o…
-
New attack targets ASR by manipulating speech features, not waveforms
Researchers have developed a new adversarial attack method for automatic speech recognition (ASR) systems that operates in the feature space rather than directly on audio waveforms. This approach, termed the Clean-Refer…
-
New method boosts speech recognition for code-switching
Researchers have developed a new contrastive training method to improve automatic speech recognition for code-switching, which involves alternating between languages within a single utterance. This approach identifies c…
-
Quantization study enables smaller, more accurate Whisper-small ASR
A new study published on arXiv evaluates various post-training quantization (PTQ) techniques for the Whisper-small automatic speech recognition model. The research, which tested libraries like PyTorch, Optimum-Quanto, H…