arXiv Computation and Language

Antalia-2 Mini: A 7.6-million-parameter Turkish text-to-speech model trained entirely on synthetic speech

Antalia-2 Mini is a single‑voice Turkish text‑to‑speech model with 7.6 million parameters that reads characters and outputs 48 kHz audio at 41× real time on an Apple M5 laptop CPU. It was trained solely on synthetic speech from the larger VoxCPM2 teacher, and fine‑tuning its vocoder on the model’s own mel spectrograms reduced Whisper’s word error rate from 3.89 % to 1.75 %. On the 495‑sentence Freya‑TR‑Eval benchmark, version 1.0.1 achieves a 1.09 % WER and 0.22 % character error rate, comparable to the EMA Lightning model in full‑band UTMOS scores.

arXiv Computation and Language
6d ago

Alania-2: Turkish text-to-speech with instant and fine-tuned voice cloning

Alania-2 is a Turkish text‑to‑speech system featuring 488 M parameters and the VoxCPM architecture, trained from scratch on 18,195 hours of Turkish speech—90 % of which is synthetic. It offers two built‑in voices, voice design from text descriptions, and two voice‑cloning modes, achieving a 1.44 % word error rate on the Freya‑TR‑Eval benchmark and competitive intelligibility with far fewer parameters. Practical insights include improved cloning accuracy with reference audio conditioning, optimal 20‑30 s reference length, and rapid low‑rank adapter training for speaker similarity, along with inference optimizations that reduce first‑audio latency to 32 ms.

By Sezgin Saygili, Emre Kaplaner, Oncel Ozgul, Fikri San Koktas
arXiv Computation and Language
Nov 4

ufakzeka-kulak: Building and Evaluating a 31.6M-Parameter Turkish Speech Recognizer from Scratch

ufakzeka-kulak is an open Turkish speech‑to‑text model with 31.6 M parameters, built using a FastConformer encoder, a token‑and‑duration transducer (TDT), and a CTC head. It was trained from scratch on 4,977 hours of data—281 hours of human transcription and 4,696 hours of machine‑generated subtitles that were verified by a second reading—under licenses permitting commercial use. The model outputs cased, punctuated Turkish text with numbers written as words, achieving a normalised WER of 0.0921 on FLEURS Turkish and 0.0916 on Common Voice 27.0 Turkish, outperforming Whisper base and small models, and it is released under Apache‑2.0.

By Sait Furkan Teke (ufak AI)
arXiv Computation and Language
Oct 12

O\'ido: Open-Vocabulary Speech Recognition on a $5 Microcontroller without a Neural Accelerator

O'ido is an inference engine that performs open‑vocabulary English speech recognition entirely on an ESP32‑S3 microcontroller, running a 13‑million‑parameter Conformer‑CTC model with int8 quantization and a 1.3‑million‑parameter language model. On LibriSpeech it achieves 3.70%/8.23% WER on test‑clean/test‑other, only 0.12 points below full precision, and outperforms the vendor’s recognizer and several laptop‑based tiny models. The firmware processes audio in about twice the duration of the input on a single core, with a streaming variant that delivers partial transcripts while speaking, though it is not yet real‑time. whyItMatters":"This work demonstrates that sophisticated, open‑vocabulary speech recognition can run on inexpensive, low‑power microcontrollers without dedicated neural accelerators, expanding the reach of voice interfaces to devices that previously could only support keyword spotting or fixed command sets."

By Daniel Varela S\'anchez, Guillermo Mart\'in-Coello Ju\'arez
arXiv Computation and Language
Sep 24

Mizar: A 159M-Parameter Audio-Language Model for Audio Understanding

Mizar is a 159.3‑million‑parameter audio‑language model designed for devices with limited memory and computation. It couples a compact CED‑Small audio encoder with SmolLM2‑135M via a frequency‑merging mapper and is trained in three stages—audio‑language alignment, audio‑dependent fine‑tuning, and post‑training—to improve performance on audio‑question tasks. Across five random seeds, Mizar outperforms all other sub‑200M‑parameter ALMs on MMAU, MMAR, and ADQA‑clean, achieving mean accuracies of 52.92%, 42.42%, and 36.02% respectively, while enabling local inference on a single CPU with an average latency of 1.09 seconds for MMAU questions.

By Kaiyang Li, Shaobo Han, Yue Tian, Shihao Ji
arXiv Computation and Language
Oct 22

How Far Can the Gemma 4 Speech Encoder Go Without Changing Its Weights? Saudi and Gulf Arabic Speech Recognition with a Frozen and an Adapter-Augmented Encoder

The paper evaluates a 305 M‑parameter audio encoder shared by Gemma 4 and EmbeddingGemma 2 for Saudi and Gulf Arabic speech recognition without altering its weights. Two approaches are compared: a frozen system that trains only a 22.6 M‑parameter CTC reader, and an adapted system that adds 3.1 M low‑rank adapters, a 41.5 M‑parameter hybrid CTC/attention reader, and joint decoding with a sub‑word language model. The adapted system matches or surpasses a 0.6 B‑parameter fine‑tuned recogniser on several datasets, while the frozen system remains competitive in character error rate and runs three‑quarters faster on CPU with greedy decoding.

By Amr Abdelkhalik
arXiv Computation and Language
6d ago

Earlybird: Predictive Streaming Speech Recognition in 29M Parameters

Earlybird is a 29‑million‑parameter streaming speech recognizer that predicts words before the speaker finishes them, achieving median correct predictions 90 ms early on clean audiobook speech. It learns from audio cut off at random points, including mid‑word, and outputs confidence scores for every word, including the one still being spoken. Across six diverse test sets—audible books, TED talks, public speech, earnings calls, and synthetic speech—Earlybird consistently outperforms existing streaming models, correctly predicting 69 % of words before they finish on audiobooks, compared to 3–25 % for other models.

By Sahil Mahendrakar
arXiv AI
Sep 15

Building a Production Greek-English Speech Recognizer

The paper details a multi‑month engineering effort to build Sophea, a production Greek‑English automatic speech recognition system. It evaluates the system against nine production gates—including word error rate, language identification, and hallucinations—across twenty‑three training iterations and two model architectures, finding that no single training‑data composition satisfies all gates. The authors describe a six‑stage data pipeline that dramatically reduces discarded Greek audio, isolate a hallucination defect to a specific training package, and demonstrate that a three‑model ROVER ensemble achieves full gate coverage and a 29 % relative WER improvement. A separate learned arbiter model is released on the Open ASR Leaderboard with 4.26 % average WER on English tests and 25.88 % on live Greek noisy traffic.

By Christos Petrocheilos, Cleopatra Papadopoulou, Chris Porikis, Ioakeim Perros, Ayoub Kirouane, Themistoklis Nikolis