metadata
tags:
- GGUF
- confucius4
- r2t2
- asr
- streaming
- real-time
- low-latency
- speech-recognition
- multilingual
base_model:
- netease-youdao/Confucius4-R2T2
tasks:
- auto-speech-recognition
pipeline_tag: automatic-speech-recognition
language:
- zh
license: other
license_name: netease-model-use-license-agreement
license_link: >-
https://raw.githubusercontent.com/netease-youdao/Confucius4-R2T2/refs/heads/master/MODEL_LICENSE
Confucius4-R2T2: A Low Latency and High Accuracy Real-Time Speech Recognition Model
Real Real-Time Transcription
Confucius4-R2T2-GGUF
Official GGUF release from NetEase Youdao for netease-youdao/Confucius4-R2T2, a low-latency and high-accuracy true streaming Automatic Speech Recognition (ASR) model that features fine-grained and configurable decoding chunks from 80 ms to 2 s.
Files
| Filename | Quantization | Size | Notes |
|---|---|---|---|
| Confucius4-R2T2-f16.gguf | f16 | 3.2GiB | best quality |
| Confucius4-R2T2-Q8_0.gguf | Q8_0 | 1.7GiB | very good quality |
| Confucius4-R2T2-Q4_K_M.gguf | Q4_K_M | 1.0GiB | fast, lower quality |
| mmproj-Confucius4-R2T2-f16.gguf | mmproj-f16 | 0.6GiB | multi-modal supplement |
| mmproj-Confucius4-R2T2-Q8_0.gguf | mmproj-Q8_0 | 0.3GiB | multi-modal supplement |
The low-bit variants are quantized from the f16 GGUF.