|
---
|
|
language: vie
|
|
datasets:
|
|
- legacy-datasets/common_voice
|
|
- vlsp2020_vinai_100h
|
|
- AILAB-VNUHCM/vivos
|
|
- doof-ferb/vlsp2020_vinai_100h
|
|
- doof-ferb/fpt_fosd
|
|
- doof-ferb/infore1_25hours
|
|
- linhtran92/viet_bud500
|
|
- doof-ferb/LSVSC
|
|
- doof-ferb/vais1000
|
|
- doof-ferb/VietMed_labeled
|
|
- NhutP/VSV-1100
|
|
- doof-ferb/Speech-MASSIVE_vie
|
|
- doof-ferb/BibleMMS_vie
|
|
- capleaf/viVoice
|
|
metrics:
|
|
- wer
|
|
pipeline_tag: automatic-speech-recognition
|
|
tags:
|
|
- transcription
|
|
- audio
|
|
- speech
|
|
- chunkformer
|
|
- asr
|
|
- automatic-speech-recognition
|
|
- long-form transcription
|
|
license: cc-by-nc-4.0
|
|
model-index:
|
|
- name: ChunkFormer Large Vietnamese
|
|
results:
|
|
- task:
|
|
name: Speech Recognition
|
|
type: automatic-speech-recognition
|
|
dataset:
|
|
name: common-voice-vietnamese
|
|
type: common_voice
|
|
args: vi
|
|
metrics:
|
|
- name: Test WER
|
|
type: wer
|
|
value: x
|
|
- task:
|
|
name: Speech Recognition
|
|
type: automatic-speech-recognition
|
|
dataset:
|
|
name: VIVOS
|
|
type: vivos
|
|
args: vi
|
|
metrics:
|
|
- name: Test WER
|
|
type: wer
|
|
value: x
|
|
- task:
|
|
name: Speech Recognition
|
|
type: automatic-speech-recognition
|
|
dataset:
|
|
name: VLSP - Task 1
|
|
type: vlsp
|
|
args: vi
|
|
metrics:
|
|
- name: Test WER
|
|
type: wer
|
|
value: x
|
|
---
|
|
|
|
# **ChunkFormer-Large-Vie: Large-Scale Pretrained ChunkFormer for Vietnamese Automatic Speech Recognition**
|
|
[](https://creativecommons.org/licenses/by-nc/4.0/)
|
|
[](https://github.com/khanld/chunkformer)
|
|
[](https://your-paper-link)
|
|
|
|
---
|
|
## Table of contents
|
|
1. [Model Description](#description)
|
|
2. [Documentation and Implementation](#implementation)
|
|
3. [Benchmark Results](#benchmark)
|
|
4. [Usage](#usage)
|
|
6. [Citation](#citation)
|
|
7. [Contact](#contact)
|
|
|
|
---
|
|
<a name = "description" ></a>
|
|
## Model Description
|
|
**ChunkFormer-Large-Vie** is a large-scale Vietnamese Automatic Speech Recognition (ASR) model based on the **ChunkFormer** architecture, introduced at **ICASSP 2025**. The model has been fine-tuned on approximately **3000 hours** of public Vietnamese speech data sourced from diverse datasets. A list of datasets can be found [**HERE**](dataset.tsv).
|
|
|
|
**!!! Please note that only the \[train-subset\] was used for tuning the model.**
|
|
|
|
---
|
|
<a name = "implementation" ></a>
|
|
## Documentation and Implementation
|
|
The [Documentation](#) and [Implementation](#) of ChunkFormer are publicly available.
|
|
|
|
---
|
|
<a name = "benchmark" ></a>
|
|
## Benchmark Results
|
|
We evaluate the models using **Word Error Rate (WER)**. To ensure consistency and fairness in comparison, we manually apply **Text Normalization**, including the handling of numbers, uppercase letters, and punctuation.
|
|
|
|
1. **Public Models**:
|
|
| STT | Model | #Params | Vivos | Common Voice | VLSP - Task 1 | Avg. |
|
|
|-----|------------------------------------------------------------------------|---------|-------|--------------|---------------|------|
|
|
| 1 | **ChunkFormer** | 110M | 4.18 | 6.72 | 13.91 | **8.27** |
|
|
| 2 | [vinai/PhoWhisper-large](https://huggingface.co/vinai/PhoWhisper-large) | 1.55B | 4.67 | 8.14 | 13.75 | 8.85 |
|
|
| 3 | [nguyenvulebinh/wav2vec2-base-vietnamese-250h](https://huggingface.co/nguyenvulebinh/wav2vec2-base-vietnamese-250h) | 95M | 10.77 | 18.34 | 13.33 | 14.15 |
|
|
| 4 | [openai/whisper-large-v3](https://huggingface.co/openai/whisper-large-v3) | 1.55B | 8.81 | 15.45 | 20.41 | 14.89 |
|
|
| 5 | [khanhld/wav2vec2-base-vietnamese-160h](https://huggingface.co/khanhld/wav2vec2-base-vietnamese-160h) | 95M | 15.05 | 10.78 | 31.62 | 19.16 |
|
|
| 6 | [homebrewltd/Ichigo-whisper-v0.1](https://huggingface.co/homebrewltd/Ichigo-whisper-v0.1) | 22M | 13.46 | 23.52 | 21.64 | 19.54 |
|
|
|
|
2. **Private Models (API)**:
|
|
| STT | Model | VLSP - Task 1 |
|
|
|-----|--------|---------------|
|
|
| 1 | **ChunkFormer** | **13.9** |
|
|
| 2 | Viettel | 14.5 |
|
|
| 3 | Google | 19.5 |
|
|
| 4 | FPT | 28.8 |
|
|
|
|
---
|
|
<a name = "usage" ></a>
|
|
## Quick Usage
|
|
To use the ChunkFormer model for Vietnamese Automatic Speech Recognition, follow these steps:
|
|
|
|
1. **Download the ChunkFormer Repository**
|
|
```bash
|
|
git clone https://github.com/khanld/chunkformer.git
|
|
cd chunkformer
|
|
pip install -r requirements.txt
|
|
```
|
|
2. **Download the Model Checkpoint from Hugging Face**
|
|
```bash
|
|
git lfs install
|
|
git clone https://huggingface.co/khanhld/chunkformer-large-vie
|
|
```
|
|
This will download the model checkpoint to the checkpoints folder inside your chunkformer directory.
|
|
|
|
3. **Run the model**
|
|
```bash
|
|
python decode.py \
|
|
--model_checkpoint path/to/local/chunkformer-large-vie \
|
|
--max_duration 14400 \ #in second, default is 1800
|
|
--long_form_audio path/to/audio.wav \
|
|
--chunk_size 64 \
|
|
--left_context_size 128 \
|
|
--right_context_size 128
|
|
```
|
|
|
|
---
|
|
<a name = "citation" ></a>
|
|
## Citation
|
|
If you use this work in your research, please cite:
|
|
|
|
```bibtex
|
|
@inproceedings{chunkformer,
|
|
title={ChunkFormer: Masked Chunking Conformer For Long-Form Speech Transcription},
|
|
author={Khanh Le, Tuan Vu Ho, Dung Tran and Duc Thanh Chau},
|
|
booktitle={ICASSP},
|
|
year={2025}
|
|
}
|
|
```
|
|
|
|
---
|
|
<a name = "contact"></a>
|
|
## Contact
|
|
- [email protected]
|
|
- [](https://github.com/khanld)
|
|
- [](https://www.linkedin.com/in/khanhld257/)
|
|
|
|
|
|
|
|
|
|
|
