6b Streaming 1120ms Fp16 is an English-language automatic speech recognition (ASR) model available on Hugging Face. The model is described as a streaming ASR solution, with a focus on processing input in buffered or unified streaming modes. It is provided in the ONNX format and supports FP16 (half-precision floating point), which reduces download size and VRAM requirements compared to full-precision models. The FP16 version was converted using onnxruntime’s transformers.float16 converter, with the claim that transcripts produced are identical to those generated by the FP32 version when run end-to-end on DirectML. The model is associated with the NeMo and Parakeet frameworks and is designed for use with libraries, inference providers, notebooks, and local applications. The evidence mentions compatibility with Google Colab and Kaggle for notebook-based usage. It is released under the CC-BY-4.0 license. There is no explicit information regarding target users, pricing, or additional features beyond those related to streaming ASR and FP16 optimization.
In the Other AI space, 6b Streaming 1120ms Fp16 takes a focused approach. It enables developers to perform efficient, streaming English speech-to-text transcription locally. 6b Streaming 1120ms Fp16 is an open-source project aimed at machine learning engineers. The project is open source (MIT). It runs on the web and the command line, and it can be self-hosted.
It is developed by Masterx, and the product first shipped in 2026. The project is developed in the open on GitHub with 27 stars and 244 commits in the last 90 days. Across PulseGate's embedding index, 6b Streaming 1120ms Fp16 has few near neighbours, marking it as relatively distinct. Among its 5 catalogued features are Streaming ASR, FP16 weights, and ONNX compatible.
Latest indexed changes and source events
Masterx/sherpa-onnx-nemo-parakeet-unified-en-0.6b-streaming-1120ms-fp16 discovered by the PulseGate indexer
