4arXiv cs.CL (Computation and Language)·
FreyaTTS: Compact Turkish-first TTS model with flow-matching DiT architecture
Freya-TTS is a 183.2M-parameter non-autoregressive text-to-speech model optimized for Turkish, using a conditional flow-matching Diffusion Transformer operating in a continuous latent space without phonemizers or discrete speech tokenizers. The model achieves 8.0% WER and 3.0% CER on the Freya-TR-Eval benchmark, outperforming larger open-source TTS systems, with a real-time factor of 0.11 on consumer GPUs and faster-than-realtime performance on laptop CPUs. Weights, training code, and the evaluation benchmark are released under Apache-2.0, making it relevant for edge deployment and low-resource language TTS research.
Related guides (2)
Related events (8)
FlowEdit: Lifelong pronunciation adaptation for flow-matching TTS via associative memory
FlowEdit is a new framework enabling lifelong pronunciation correction in frozen flow-matching text-to-speech systems without retraining model weights. Corrections are stored as token-level perturbations in text embedding space within a Modern Hopfield Network, retrieved at inference via soft attention with fuzzy morphological matching. On a curated benchmark of 312 multilingual proper nouns across 18 language families, the method reduces target-word Phoneme Error Rate by 92.7% relative to the zero-shot baseline, with each correction completing in ~15 seconds on a single GPU.
Mistral Releases Voxtral TTS: 4B-Parameter Multilingual Text-to-Speech Model
Mistral AI has launched Voxtral TTS, its first text-to-speech model, built on a 4B-parameter transformer-based autoregressive flow-matching architecture derived from Ministral 3B. The model supports 9 languages with zero-shot voice adaptation from as little as 3 seconds of reference audio, achieving 70ms latency for typical inputs and a real-time factor of ~9.7x. Human evaluations claim superior naturalness compared to ElevenLabs Flash v2.5 and parity with ElevenLabs v3. The model is available via Mistral Studio and API, targeting enterprise voice agent workflows.
Kyutai Labs releases pocket-tts: CPU-friendly text-to-speech model
Kyutai Labs published pocket-tts, a text-to-speech system designed to run on CPU hardware without requiring a GPU. The repository is trending on GitHub with 5,905 stars and 510 added today, indicating significant community interest. CPU-deployable TTS lowers the barrier for edge and on-device voice applications.
Mistral Releases Voxtral Transcribe 2: State-of-the-Art Speech-to-Text with Sub-200ms Realtime Model
Mistral AI has released Voxtral Transcribe 2, a family of two speech-to-text models: Voxtral Mini Transcribe V2 for batch transcription and Voxtral Realtime for live applications. Voxtral Realtime features a novel streaming architecture with configurable latency down to sub-200ms, a 4B parameter footprint suitable for edge deployment, and is released as open weights under Apache 2.0. Voxtral Mini Transcribe V2 claims state-of-the-art word error rate on FLEURS at $0.003/min, outperforming GPT-4o mini Transcribe, Gemini 2.5 Flash, AssemblyAI, and Deepgram Nova on accuracy benchmarks. Both models support 13 languages with speaker diarization, word-level timestamps, and context biasing.
Thaka Wins KSAA-2026 Arabic Speech Diacritization Task with Regularized Fine-Tuning of CATT-Whisper
The Thaka team describes their winning system for Task 2 of the KSAA-2026 Shared Task on Arabic Speech Dictation with Automatic Diacritization, which requires producing fully diacritized Arabic text from speech audio and undiacritized transcripts. Their approach fine-tunes CATT-Whisper, a multimodal model combining a CATT text encoder with a frozen Whisper speech encoder, under severe data constraints (2,327 training samples, no external data). Key techniques include R-Drop consistency regularization, Optuna-optimized hyperparameters with high weight decay, Focal Loss, and Monte Carlo Dropout inference averaging over 200 stochastic forward passes across four checkpoints. The system achieves 23.26% WER on the primary metric, placing first among all participants.
Nemotron-Labs Audex-30B: Unified audio-text MoE LLM with no text regression
Nvidia's Nemotron Labs introduces Audex-30B-A3B, a 30B-parameter mixture-of-experts audio-text LLM built on the Nemotron-Cascade-2 text backbone. The model handles audio understanding, ASR, translation, TTS, and speech-to-speech generation within a single Transformer decoder by projecting audio into the text embedding space. Training used 157.4B audio tokens and 320.5B text tokens with multi-stage supervised learning, RL, and on-policy distillation. Model checkpoints are publicly released, and the authors report state-of-the-art audio performance with minimal regression on text reasoning and agentic capabilities.
Falcon-Arabic: A Breakthrough in Arabic Language Models
TII UAE has released Falcon-Arabic, a language model specifically designed for Arabic. The announcement highlights it as a significant advancement in Arabic NLP capabilities. As a tier-2 source with minimal body content, specific technical details about model size, training data, or benchmark performance are not available from this item.
Fine-tuning Florence-2 - Microsoft's Cutting-edge Vision Language Models
This Hugging Face blog post provides a technical guide for fine-tuning Microsoft's Florence-2 vision-language models. Florence-2 is a compact yet capable multimodal model supporting tasks like captioning, object detection, and OCR. The post covers practical implementation details for adapting the model to custom datasets using the Hugging Face ecosystem.