Automatic Speech Recognition
Transformers
Safetensors
Chinese
English
qwen3_asr
taiwan-mandarin
traditional-chinese
code-switching
qwen3-asr
speech
Instructions to use JacobLinCool/TEA-ASR-1.1 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use JacobLinCool/TEA-ASR-1.1 with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("automatic-speech-recognition", model="JacobLinCool/TEA-ASR-1.1")# Load model directly from transformers import AutoProcessor, AutoModelForMultimodalLM processor = AutoProcessor.from_pretrained("JacobLinCool/TEA-ASR-1.1") model = AutoModelForMultimodalLM.from_pretrained("JacobLinCool/TEA-ASR-1.1", device_map="auto") - Notebooks
- Google Colab
- Kaggle
card: TEA-ASR-1.1-mini row -> v27 numbers
Browse files
README.md
CHANGED
|
@@ -71,10 +71,10 @@ and two references — **Breeze-ASR-25** (a Taiwan-specialist ASR) and **Whisper
|
|
| 71 |
|
| 72 |
| Benchmark | TEA-ASR-1.1 | TEA-ASR-1.1-mini | Qwen3-ASR-1.7B | Qwen3-ASR-0.6B | Breeze-ASR-25 | Whisper-large-v3 |
|
| 73 |
|---|---|---|---|---|---|---|
|
| 74 |
-
| CommonVoice 19 (zh-TW) | **3.58** | 5.
|
| 75 |
-
| ASCEND (zh-en) | **9.60** | 11.
|
| 76 |
-
| CSZS (zh-en) | **10.94** | 12.
|
| 77 |
-
| NTUML2021 | **6.67** | 7.
|
| 78 |
|
| 79 |
**Generational improvement — TEA-ASR-1.1 vs TEA-ASR-1** (2B flagship, same protocol, lower is better):
|
| 80 |
|
|
|
|
| 71 |
|
| 72 |
| Benchmark | TEA-ASR-1.1 | TEA-ASR-1.1-mini | Qwen3-ASR-1.7B | Qwen3-ASR-0.6B | Breeze-ASR-25 | Whisper-large-v3 |
|
| 73 |
|---|---|---|---|---|---|---|
|
| 74 |
+
| CommonVoice 19 (zh-TW) | **3.58** | 5.12 | 3.90 | 5.79 | 8.03 | 10.17 |
|
| 75 |
+
| ASCEND (zh-en) | **9.60** | 11.20 | 10.57 | 12.54 | 17.53 | 19.61 |
|
| 76 |
+
| CSZS (zh-en) | **10.94** | 12.51 | 11.03 | 16.03 | 12.18 | 23.24 |
|
| 77 |
+
| NTUML2021 | **6.67** | 7.53 | 10.12 | 11.03 | 7.50 | 9.68 |
|
| 78 |
|
| 79 |
**Generational improvement — TEA-ASR-1.1 vs TEA-ASR-1** (2B flagship, same protocol, lower is better):
|
| 80 |
|