A monolingual automatic speech recognition model for
Uzbek
, fine-tuned from
openai/whisper-large-v3
. Part of
BuzzASR
,
a suite of 102 language-specialized ASR models (Findings of EMNLP 2026).
This model uses
simple fine-tuning (Whisper's tokenizer, ASR fine-tuning only)
.
🏆
State-of-the-art (open-source).
On the combined FLEURS + Common Voice test set, this model
achieves the lowest CER of every open system we compare against: Whisper-large-v3, Omnilingual 1B/7B, MMS, Qwen3-ASR, and Cohere Transcribe.
Results (normalized CER / WER, %)
Test set
CER
WER
Whisper-large-v3 (zero-shot) CER
FLEURS
6.6
25.64
28.4
Common Voice 25
2.68
11.84
28.55
Combined
3.86
16.0
28.45
~7.4x CER reduction over Whisper zero-shot on the combined test set.
uzbek huggingface.co is an AI model on huggingface.co that provides uzbek's model effect (), which can be used instantly with this BuzzASR uzbek model. huggingface.co supports a free trial of the uzbek model, and also provides paid use of the uzbek. Support call uzbek model through api, including Node.js, Python, http.
uzbek huggingface.co is an online trial and call api platform, which integrates uzbek's modeling effects, including api services, and provides a free online trial of uzbek, you can try uzbek online for free by clicking the link below.
uzbek is an open source model from GitHub that offers a free installation service, and any user can find uzbek on GitHub to install. At the same time, huggingface.co provides the effect of uzbek install, users can directly use uzbek installed effect in huggingface.co for debugging and trial. It also supports api for free installation.