nvidia/NVIDIA-Nemotron-3.5-Transcribe-30B-A4B
NVIDIA-Nemotron-3.5-Transcribe-30B-A4B !Nemotron 3.5 Transcribe overview: one speech LLM transcribes a recording in the style selected by the instruction(hero.png) Nemotron 3.5 Transcribe is an open-weight, offline speech and audio model for transcr...
Thành tựu nổi bật
Những danh hiệu mới nhất của dự án
#1 Repository Của Ngày
#1 Repository Của Ngày
#1 Repository Của Ngày
🎙️ NVIDIA Nemotron 3.5 Transcribe: 40-Language Speech LLM
Transcribe up to one hour of audio in 40 languages with speaker diarization, timestamps, and instruction following.
NVIDIA introduces Nemotron 3.5 Transcribe, an open-weight speech-augmented language model for offline transcription, translation, and rich audio understanding. The model accepts recordings up to one hour and outputs text in 40 languages, supporting speaker attribution, timestamp prediction, language identification, and audio question answering. It built on a hybrid Mamba-2–Transformer mixture-of-experts architecture that activates roughly 4.3 billion parameters while maintaining a 65,536-token context window.
✨ Highlights
- 40-language transcription: Supports transcription across 40 languages.
- Long-form audio: Handles recordings up to one hour with a 65,536-token context.
- Efficient mixture-of-experts: Activates ~4.3B of 34B parameters for fast inference.
- Rich transcription features: Includes speaker diarization, timestamps, language identification, and code-switching.
- Instruction following: Accepts text instructions to guide output style.
- Multiple precision formats: Available in BF16, FP8, and NVFP4 for varied deployment.
💡 Why it matters
Nemotron 3.5 Transcribe stands out by combining a hybrid Mamba-2–Transformer backbone with a parallel expert encoder, enabling both high accuracy and low latency. Its offline capability means no internet connection is required, making it suitable for privacy-sensitive or edge deployments. The model's support for speaker diarization, timestamps, and language identification adds value for meeting transcription, media captioning, and multilingual applications.
#SpeechRecognition #NVIDIA #Nemotron #TiniX
Similar Projects
Discover other similar projects and AI models.
Whisper Whisper is a state-of-the-art model for automatic speech recognition (ASR) and speech translation, proposed in the paper Robust Speech Recognition via Large-Scale Weak Supervision(https://huggingface.co/papers/2212.04356) by Alec Radford et...
Whisper Whisper is a state-of-the-art model for automatic speech recognition (ASR) and speech translation, proposed in the paper Robust Speech Recognition via Large-Scale Weak Supervision(https://huggingface.co/papers/2212.04356) by Alec Radford et...
Whisper Whisper is a pre-trained model for automatic speech recognition (ASR) and speech translation. Trained on 680k hours of labelled data, Whisper models demonstrate a strong ability to generalise to many datasets and domains without the need for...
Whisper Whisper is a pre-trained model for automatic speech recognition (ASR) and speech translation. Trained on 680k hours of labelled data, Whisper models demonstrate a strong ability to generalise to many datasets and domains without the need for...
Whisper Whisper is a pre-trained model for automatic speech recognition (ASR) and speech translation. Trained on 680k hours of labelled data, Whisper models demonstrate a strong ability to generalise to many datasets and domains without the need for...
sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2 This is a sentence-transformers(https://www.SBERT.net) model: It maps sentences & paragraphs to a 384 dimensional dense vector space and can be used for tasks like clustering or semantic se...
Whisper small model for CTranslate2 This repository contains the conversion of openai/whisper-small(https://huggingface.co/openai/whisper-small) to the CTranslate2(https://github.com/OpenNMT/CTranslate2) model format. This model can be used in CTran...
Multilingual-E5-large Multilingual E5 Text Embeddings: A Technical Report(https://arxiv.org/pdf/2402.05672). Liang Wang, Nan Yang, Xiaolong Huang, Linjun Yang, Rangan Majumder, Furu Wei, arXiv 2024 This model has 24 layers and the embedding size is...
XLM-RoBERTa (base-sized model) XLM-RoBERTa model pre-trained on 2.5TB of filtered CommonCrawl data containing 100 languages. It was introduced in the paper Unsupervised Cross-lingual Representation Learning at Scale(https://arxiv.org/abs/1911.02116)...