File size: 1,101 Bytes
d0f3698
e3a3acf
 
d0f3698
 
 
 
 
e3a3acf
 
 
 
 
 
d0f3698
 
e3a3acf
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
---
title: speakwrite-azerbaijani
emoji: 🎙️
colorFrom: yellow
colorTo: indigo
sdk: gradio
sdk_version: 6.31.0
app_file: app.py
short_description: Azerbaijani speech recognition with KenLM decoding
python_version: "3.12"
startup_duration_timeout: 30m
license: cc-by-nc-4.0
models:
- LocalDoc/speakwrite-azerbaijani
---

# speakwrite-azerbaijani

Gradio demo of [LocalDoc/speakwrite-azerbaijani](https://huggingface.co/LocalDoc/speakwrite-azerbaijani): Azerbaijani ASR
(FastConformer-CTC + loop refiner + product-key acoustic memory + 6-gram token KenLM).

- Model: **"speakwrite-azerbaijani by LocalDoc"**, CC BY-NC 4.0 — non-commercial use only.
- Inference code: vendored from [balbal](https://github.com/erayyap/balbal) at commit `1be9f88b77c63fbafa7713c6e46e79847f6c390e`
  (Apache-2.0, see `balbal/LICENSE` and `balbal/NOTICE`), with the model repo's `refiner.json`.
- Example audio: [FLEURS](https://huggingface.co/datasets/google/fleurs) Azerbaijani test split
  (via [LocalDoc/fleurs-azerbaijani-asr](https://huggingface.co/datasets/LocalDoc/fleurs-azerbaijani-asr)), CC BY 4.0, Google.