GLM-5.3 is live. Run Z.ai's latest model on Friendli Model APIs. Try it today ➜
BuzzASR
Available on FriendliAI
Run this model inference on single tenant GPU with unmatched speed and reliability at scale.
Model Details
Model Provider
BuzzASR
Model Tree
openai/whisper-large-v3
Input Modalities
Output Modalities
Supported Functionality
| Test set | CER | WER | Whisper-large-v3 (zero-shot) CER |
|---|---|---|---|
| FLEURS | 14.57 | 43.72 | 21.75 |
| Common Voice 25 | 10.29 | 34.48 | 18.34 |
| Combined | 12.54 | 39.39 | 19.43 |
~1.5x CER reduction over Whisper zero-shot on the combined test set.
python
import torch, torchaudiofrom transformers import WhisperForConditionalGeneration, WhisperProcessor model = WhisperForConditionalGeneration.from_pretrained("BuzzASR/georgian", torch_dtype=torch.float16).to("cuda").eval()proc = WhisperProcessor.from_pretrained("BuzzASR/georgian") wav, sr = torchaudio.load("audio.wav") # 16 kHz monofeats = proc(wav[0], sampling_rate=16000, return_tensors="pt").input_features.to("cuda").half()ids = model.generate(feats, num_beams=1, no_repeat_ngram_size=3, repetition_penalty=1.2)print(proc.batch_decode(ids, skip_special_tokens=True)[0])The language/task prompt is baked into the generation config, so no language= argument is needed.
FLEURS + Common Voice Corpus 25.0 (Mozilla, March 2025; https://commonvoice.mozilla.org/en/datasets), capped per the paper. Text-only data from the Goldfish corpus (Chang et al., 2026).
Monolingual (Georgian only). Evaluated on FLEURS / Common Voice test splits; other domains or dialects may differ.
bibtex
@misc{buzzasr2026, title = {BuzzASR: A Swarm of 100+ Monolingual Speech Recognition Models}, author = {Shivam Singh and Aditya Yadavalli and Catherine Arnett and Alex Warstadt}, year = {2026}, eprint = {2609.09554}, archivePrefix = {arXiv}, primaryClass = {cs.CL}, note = {Findings of the Association for Computational Linguistics: EMNLP 2026}, url = {https://arxiv.org/abs/2609.09554}}import torch, torchaudiofrom transformers import WhisperForConditionalGeneration, WhisperProcessor model = WhisperForConditionalGeneration.from_pretrained("BuzzASR/georgian", torch_dtype=torch.float16).to("cuda").eval()proc = WhisperProcessor.from_pretrained("BuzzASR/georgian") wav, sr = torchaudio.load("audio.wav") # 16 kHz monofeats = proc(wav[0], sampling_rate=16000, return_tensors="pt").input_features.to("cuda").half()ids = model.generate(feats, num_beams=1, no_repeat_ngram_size=3, repetition_penalty=1.2)print(proc.batch_decode(ids, skip_special_tokens=True)[0])@misc{buzzasr2026, title = {BuzzASR: A Swarm of 100+ Monolingual Speech Recognition Models}, author = {Shivam Singh and Aditya Yadavalli and Catherine Arnett and Alex Warstadt}, year = {2026}, eprint = {2609.09554}, archivePrefix = {arXiv}, primaryClass = {cs.CL}, note = {Findings of the Association for Computational Linguistics: EMNLP 2026}, url = {https://arxiv.org/abs/2609.09554}}