---
license: apache-2.0
base_model: Qwen/Qwen3-ASR-1.7B
base_model_relation: finetune
language:
- th
pipeline_tag: automatic-speech-recognition
library_name: qwen-asr
datasets:
- google/fleurs
tags:
- automatic-speech-recognition
- asr
- thai
- qwen3-asr
- fleurs
- finetune
metrics:
- cer
- wer
model-index:
- name: Qwen3-ASR-1.7B-th-fleurs
  results:
  - task:
      type: automatic-speech-recognition
      name: Automatic Speech Recognition
    dataset:
      name: FLEURS (Thai)
      type: google/fleurs
      config: th_th
      split: test
    metrics:
    - type: cer
      value: 0.0702
      name: Character Error Rate
    - type: wer
      value: 0.6100
      name: Word Error Rate
---

# Qwen3-ASR-1.7B-th-fleurs

Thai automatic speech recognition model, full fine-tune of [Qwen/Qwen3-ASR-1.7B](https://huggingface.co/Qwen/Qwen3-ASR-1.7B) on the Thai split of [google/fleurs](https://huggingface.co/datasets/google/fleurs).

## Evaluation

FLEURS Thai `test` split (1,021 utterances). Reported via the `evaluate` library — raw model output vs reference, no text normalisation.

| Model | CER (%) | WER (%) |
|---|---:|---:|
| Qwen/Qwen3-ASR-1.7B (base) | 8.32 | 79.56 |
| **This model** | **7.02** | **61.00** |
| Δ relative improvement | **+15.7%** | **+23.3%** |

Lower is better.

## Usage

```python
import torch
from qwen_asr import Qwen3ASRModel

model = Qwen3ASRModel.from_pretrained(
    "PogusTheWhisper/Qwen3-ASR-1.7B-th-fleurs",
    dtype=torch.bfloat16,
    device_map="cuda:0",
    max_inference_batch_size=16,
    max_new_tokens=256,
)
results = model.transcribe(audio="path/to/audio.wav", language="Thai")
print(results[0].text)
```

For maximum throughput, use the vLLM backend:

```python
model = Qwen3ASRModel.LLM(
    model="PogusTheWhisper/Qwen3-ASR-1.7B-th-fleurs",
    gpu_memory_utilization=0.7,
    max_inference_batch_size=128,
    max_new_tokens=4096,
)
results = model.transcribe(audio="path/to/audio.wav", language="Thai")
print(results[0].text)
```

## Training

Full fine-tune (FFT) of the entire 1.7B-parameter base model using the official [QwenLM/Qwen3-ASR `qwen3_asr_sft.py`](https://github.com/QwenLM/Qwen3-ASR/tree/main/finetuning) script.

| Hyperparameter | Value |
|---|---|
| Base model | Qwen/Qwen3-ASR-1.7B |
| Method | Full fine-tune (FFT, not LoRA) |
| Dataset | google/fleurs (`th_th`, 2,602 train / 1,021 test utterances) |
| Label format | `language Thai<asr_text>{{transcript}}` |
| Optimizer | AdamW (HF Trainer defaults) |
| Learning rate | 2e-5 (Official tested at this lr w/ eff bs 32) |
| LR scheduler | cosine (Smoother decay, better final epochs) |
| Warmup ratio | 0.05 (0.1 too aggressive for cosine) |
| weight_decay | 0.01 (Anti-memorization) |
| Effective batch size | 32 (per-device 1 × grad_acc 32) |
| Epochs | 5 |
| Precision | bfloat16 |
| Hardware | 1× NVIDIA RTX 3090 (24 GB) |
| Training time | ~25 minutes |

## Limitations

- Trained on read-speech only — FLEURS is broadcast / audiobook style. Conversational, noisy, and telephony-audio performance is not measured.
- 16 kHz mono audio required (matches the base model).
- Code-switched audio (Thai + English) inherits base-model behaviour; not specifically tuned.
- Small training set (~7 hours). For production use, consider mixing with Common Voice Thai or your own labelled data.

## Citation

```bibtex
@misc{qwen3asr,
  title  = {Qwen3-ASR},
  author = {Qwen Team},
  year   = {2025},
  url    = {https://huggingface.co/Qwen/Qwen3-ASR-1.7B}
}

@inproceedings{conneau2023fleurs,
  title     = {FLEURS: Few-shot Learning Evaluation of Universal Representations of Speech},
  author    = {Conneau, Alexis and Ma, Min and Khanuja, Simran and Zhang, Yu and Axelrod, Vera and Dalmia, Siddharth and Riesa, Jason and Rivera, Clara and Bapna, Ankur},
  booktitle = {SLT},
  year      = {2023}
}
```