5arXiv cs.CL (Computation and Language)·
DONDO: Open w2v-BERT ASR base models for 27 African language varieties
Researchers release DONDO, a family of 21 monolingual and 5 multilingual automatic speech recognition models covering 27 language varieties across Ghana, Sierra Leone, Nigeria, Senegal, Kenya, and Zimbabwe, built on the w2v-BERT 2.0 self-supervised encoder. Models are fine-tuned on read speech from religious texts and achieve average word error rates of 10–13% on multilingual checkpoints, closing most of the gap to monolingual baselines. A lightweight language-conditioning mechanism using one-hot prefix frames allows a single multilingual checkpoint to be steered to a target language at inference. All models are released under Apache-2.0 on Hugging Face via the KhayaAI organisation, covering languages spoken by roughly 100 million first-language speakers.
Related guides (2)
Related events (8)
Qwen2.5-LLM: Alibaba releases open-weight language models from 0.5B to 72B
Alibaba's Qwen team releases the Qwen2.5 series of decoder-only dense language models, open-sourcing seven variants spanning 0.5B to 72B parameters. The release targets production use cases in the 10-30B range and mobile deployments at 3B scale. This represents a significant expansion of the open-weights frontier from a Tier 1 Chinese AI lab.
Qwen releases Qwen3-ASR-1.7B, a multilingual automatic speech recognition model
Qwen has released Qwen3-ASR-1.7B on Hugging Face, a 1.7B parameter automatic speech recognition model supporting multiple languages including Chinese (Mandarin and Cantonese), English, Arabic, German, French, and Spanish. The model uses the Qwen3 architecture and is released in safetensors format. This extends the Qwen3 model family into the speech domain.
Qwen releases Qwen3-ASR-0.6B, a multilingual automatic speech recognition model
Alibaba's Qwen team released Qwen3-ASR-0.6B on Hugging Face, a 0.6B parameter automatic speech recognition model supporting multiple languages including Chinese, English, Cantonese, Arabic, German, French, and Spanish. The model uses the Qwen3 architecture and is available in HuggingFace-native format with safetensors weights. This is a compact ASR model extending the Qwen3 family into speech recognition.
Forgotten Words: Benchmarking NeoBERT for Dementia Detection in Low-Resource Conversational Filipino and English Speech
This paper presents the first NLP-based dementia detection study for Filipino speech, constructing a parallel bilingual dataset of 4,000 DementiaBank-derived transcripts with manual Filipino translations. Five model families are evaluated across monolingual, zero-shot cross-lingual, and bilingual fine-tuning settings. English-trained BERT degrades sharply on Filipino (Macro-F1 = 0.455), but bilingual fine-tuning recovers performance to Macro-F1 = 0.969–0.973 across all transformer models. The key finding is that multilingual clinical NLP performance is driven by linguistic coverage during training rather than model scale or architecture.
BamiBERT: New BERT-based language model sets state of the art for Vietnamese NLP
Qualcomm AI Research introduces BamiBERT, a BERT-based encoder pre-trained from scratch on 129GB of Vietnamese text for 20 epochs, supporting up to 2048-token context without requiring external word segmentation. It outperforms PhoBERT, the previous de facto Vietnamese encoder, achieving best scores on 11 of 15 metrics across 8 Vietnamese benchmarks. The model is released publicly on Hugging Face.
Qwen2.5-VL-32B: Reinforcement-Learning-Optimized Vision-Language Model Released
Alibaba's Qwen team has released Qwen2.5-VL-32B-Instruct, a 32-billion-parameter vision-language model built on the Qwen2.5-VL series and further optimized with reinforcement learning. The model is open-sourced under the Apache 2.0 license and available on Hugging Face and ModelScope. It follows the January 2025 launch of the broader Qwen2.5-VL series, positioning the 32B scale as a balance between capability and deployment practicality.
Dango: A 1.8B LLM trained exclusively on Japanese to study L1-to-L2 language transfer
Researchers introduce Dango, a 1.8B-parameter decoder-only LLM pretrained strictly on Japanese (L1) and fine-tuned on LLM-generated English (L2) learning lessons to simulate second language acquisition. A key contribution is a filtering method to remove L2 contamination from ostensibly monolingual pretraining corpora. Evaluations show Dango produces human-like L2 error patterns, outperforming multilingual and unfiltered baselines. The model, data, and code are released for computational SLA research.
Soofi S 30B-A3B: Sovereign German-English MoE Hybrid Mamba Transformer Released Open-Source
Researchers present Soofi S 30B-A3B, a Mixture-of-Experts hybrid Mamba Transformer foundation model for German and English, activating only 3B of 30B parameters per token with near-constant inference cache for long-context efficiency. Pretrained on ~27 trillion tokens with up-weighted German data, it claims top aggregate scores among fully open models in both languages, outperforming OLMo 3 32B and Apertus 70B, and surpasses all European sovereign baselines tested. The model was built on Deutsche Telekom's sovereign HPC infrastructure in Munich and will be released with weights, checkpoints, hyperparameters, training code, and full data accounting under permissive terms.