Egyptian and Nigerian English join Open ASR Leaderboard
The Open ASR Leaderboard now supports Egyptian and Nigerian English.
Stories per day · 30 days
6 RSS
Relatedbemba1bengali1bias-evaluation1catastrophic-forgetting1diarization1egyptian-english1leaderboard1low-resource-languages1model-merging1multilingual1nigerian-english1nyanja1speech-corpus1speech-llm1speech-recognition1youtube1
35 stories, newest first · one card per story
Stories tagged #asr
Speech LLM ASR expansion forgets source languages unless merges are selective
SSU-Wanda+TIES best preserved source languages in single-target expansion; recursive multi-target updates were highly order-sensitive.
Lipi-Ghor: 885-hour diarized Bengali speech corpus from YouTube
Audio from 998 videos across 17 domains is paired with noisy captions plus a 30-hour human gold split; raw audio is not redistributed.
AfriBED evaluates gender and speaker-type bias in Bemba and Nyanja ASR
The set holds about 9 hours of Bemba and 18 hours of Nyanja from 80 and 127 speakers.
TII Releases Falcon-ASR: 1.6B Arabic Speech Model Hits 20.92% WER
TII released Falcon-ASR, a 1.6B Arabic speech model reporting 20.92% average WER; weights are not public.
These execs think voice AI hasn’t reached its ChatGPT moment yet
Voice AI execs say the field still lacks a ChatGPT-style breakthrough despite full-duplex models.
Does Financial Speech Need Continual Learning Regularization? A Study of EWC and mini-KFAC on Earnings Call ASR
Adapting Whisper-small to Earnings22 improves financial ASR, so EWC regularization is often counterproductive.
Speech event detector: Timestamped detection of non-verbal vocalizations and background sounds for rich speech transcription
A 5.8M model timestamps 54 non-speech sounds and merges them into Whisper transcripts.
Introducing Falcon ASR
TII introduced Falcon ASR, a speech-recognition release in the Falcon model family.
Itgan at NADI 2026 shared task: Parameter-Efficient Whisper Adaptation for Robust, Mixed-Dialect and Code-Switched Arabic ASR
LoRA-adapted Whisper systems for NADI 2026 Arabic ASR reach 57.1%, 46.7%, and 14.49% WER on three subtasks.
DirectSpeech2LLM: A Simple End-to-End Framework to Mitigate Prompt Overfitting in Speech-LLMs
DirectSpeech2LLM cuts speech-LLM prompt overfitting, generalizing from ASR-only training to translation and emotion tasks.
HINTT Submission to the 2nd MLC-SLM Challenge: Comparing Cascaded and Unified Approaches to Diarization and ASR
HINTT's cascaded diarization and Qwen3-ASR pipeline beat a unified speech LLM on MLC-SLM Task 1.
Hiding Tool Latency in On-Device Cascaded Voice Agent through Speculative Execution
Speculative tool execution cuts on-device voice-agent median time-to-first-audio from 5.79s to 4.60s.
Cactus-Compute/whistle — new model trending #30 on Hugging Face
Cactus Compute released Whistle, a 16.9 MB on-device speech-to-text model for seven languages without a GPU.
4
Alibaba Qwen Releases Qwen-Audio-3.1-Realtime: A Full-Duplex Voice Model Trained to Think, Act, and Decide When to Speak
Alibaba released API-only Qwen-Audio-3.1-Realtime, a full-duplex voice model for tool-using agents.
FermionResearch/Phonon-2 — new model trending #29 on Hugging Face
Fermion Research released Phonon-2, a 164 MB English speech model averaging 5.21% word error.
3
Sarvam AI Releases Saaras V4: A Speech-to-Text Model for All 22 Indian Languages and Global English
Sarvam AI released Saaras V4, a speech-to-text model for all 22 Indian languages and global English.
Pruned CTC for Memory-Efficient Large-Vocabulary ASR Training
Pruned CTC restricts alignment computation to target tokens and blank, cutting ASR training memory 5.1x for 180K-vocabulary LLMs with minimal accuracy loss.
Muslim: A Deployed Arabic Voice AI Platform for Grounded Islamic Knowledge
Muslim is a production Arabic voice AI for sourced Islamic knowledge, releasing a 6B router and TTS model.
A Training Criterion with Token-Level Tolerance to Transcription Ambiguity for Automatic Speech Recognition
Token-level wildcard paths in CTC cut ASR word error by tolerating localized transcript ambiguity.
Alibaba launches Qwen Audio 3.1 with new models and slashes AI audio prices by up to 95 percent
Alibaba's Qwen team released five Qwen-Audio-3.1 speech models and cut audio prices by up to 95 percent.
Six Layers Less: Encoder Pruning for Whisper with Label-Free Recovery
Six Whisper encoder layers are pruned and partly recovered with label-free distillation.
Edge0/Audio8-ASR-Infinite — new model trending #30 on Hugging Face
Edge0 released Audio8-ASR-Infinite, a bilingual streaming speech model with constant-memory unlimited transcription.
8
SpaceXAI Releases Grok Voice Transcribe 2.0: A Speech-to-Text API Claiming 2x Accuracy Over 1.0 at $0.10 per Hour
SpaceXAI released Grok Voice Transcribe 2.0, a speech-to-text API claiming 2x accuracy over 1.0 at unchanged $0.10-$0.20 per hour pricing.