Update repo references after rename to include nict-tib1
Browse files
README.md
CHANGED
|
@@ -14,7 +14,7 @@ metrics:
|
|
| 14 |
- cer
|
| 15 |
- wer
|
| 16 |
model-index:
|
| 17 |
-
- name: tibetan-asr-whisper-small
|
| 18 |
results:
|
| 19 |
- task:
|
| 20 |
type: automatic-speech-recognition
|
|
@@ -38,7 +38,7 @@ Fine-tuned **`openai/whisper-small`** for automatic speech recognition (ASR) on
|
|
| 38 |
|
| 39 |
> J. Moore, S. Li and P. Lauren, "Evaluating Tibetan ASR With Segmented Word Error Rate: Beyond Character-Level Metrics," in *IEEE Access*, vol. 14, pp. 101790-101805, 2026, doi: 10.1109/ACCESS.2026.3709206.
|
| 40 |
|
| 41 |
-
This repo is one of 14 model checkpoints released with the paper, benchmarking 5 ASR architectures (Whisper Tiny/Base/Small, Wav2Vec 2.0 Base, HuBERT Base) under full fine-tuning, LoRA, and QLoRA (8-bit/4-bit) adaptation. See [billingsmoore/tibetan-asr-*](https://huggingface.co/billingsmoore) for the full set.
|
| 42 |
|
| 43 |
## Model description
|
| 44 |
|
|
@@ -75,11 +75,11 @@ This configuration achieved the best score on every metric (CER, SER, and all th
|
|
| 75 |
|
| 76 |
| Model | Repo | CER | SER | BoTok-SWER | BERT-SWER | Gem-SWER |
|
| 77 |
|---|---|---|---|---|---|---|
|
| 78 |
-
| HuBERT Base | [tibetan-asr-hubert-base](https://huggingface.co/billingsmoore/tibetan-asr-hubert-base) | 0.1352 | 0.3690 | 0.4477 | 0.161 | 0.9653 |
|
| 79 |
-
| Wav2Vec 2.0 Base | [tibetan-asr-wav2vec2-base](https://huggingface.co/billingsmoore/tibetan-asr-wav2vec2-base) | 0.0745 | 0.2152 | 0.2747 | 0.097 | 0.7447 |
|
| 80 |
-
| Whisper Tiny | [tibetan-asr-whisper-tiny](https://huggingface.co/billingsmoore/tibetan-asr-whisper-tiny) | 0.1560 | 0.2351 | 0.2975 | 0.118 | 0.6759 |
|
| 81 |
-
| Whisper Base | [tibetan-asr-whisper-base](https://huggingface.co/billingsmoore/tibetan-asr-whisper-base) | 0.1417 | 0.2083 | 0.2600 | 0.105 | 0.6314 |
|
| 82 |
-
| **Whisper Small** | [tibetan-asr-whisper-small](https://huggingface.co/billingsmoore/tibetan-asr-whisper-small) | **0.1185** | **0.1692** | **0.2042** | **0.086** | **0.5337** |
|
| 83 |
|
| 84 |
|
| 85 |
## How to use
|
|
@@ -87,7 +87,7 @@ This configuration achieved the best score on every metric (CER, SER, and all th
|
|
| 87 |
```python
|
| 88 |
from transformers import pipeline
|
| 89 |
|
| 90 |
-
pipe = pipeline("automatic-speech-recognition", model="billingsmoore/tibetan-asr-whisper-small")
|
| 91 |
result = pipe("path/to/audio.wav")
|
| 92 |
print(result["text"])
|
| 93 |
```
|
|
|
|
| 14 |
- cer
|
| 15 |
- wer
|
| 16 |
model-index:
|
| 17 |
+
- name: tibetan-asr-nict-tib1-whisper-small
|
| 18 |
results:
|
| 19 |
- task:
|
| 20 |
type: automatic-speech-recognition
|
|
|
|
| 38 |
|
| 39 |
> J. Moore, S. Li and P. Lauren, "Evaluating Tibetan ASR With Segmented Word Error Rate: Beyond Character-Level Metrics," in *IEEE Access*, vol. 14, pp. 101790-101805, 2026, doi: 10.1109/ACCESS.2026.3709206.
|
| 40 |
|
| 41 |
+
This repo is one of 14 model checkpoints released with the paper, benchmarking 5 ASR architectures (Whisper Tiny/Base/Small, Wav2Vec 2.0 Base, HuBERT Base) under full fine-tuning, LoRA, and QLoRA (8-bit/4-bit) adaptation. See [billingsmoore/tibetan-asr-nict-tib1-*](https://huggingface.co/billingsmoore) for the full set.
|
| 42 |
|
| 43 |
## Model description
|
| 44 |
|
|
|
|
| 75 |
|
| 76 |
| Model | Repo | CER | SER | BoTok-SWER | BERT-SWER | Gem-SWER |
|
| 77 |
|---|---|---|---|---|---|---|
|
| 78 |
+
| HuBERT Base | [tibetan-asr-nict-tib1-hubert-base](https://huggingface.co/billingsmoore/tibetan-asr-nict-tib1-hubert-base) | 0.1352 | 0.3690 | 0.4477 | 0.161 | 0.9653 |
|
| 79 |
+
| Wav2Vec 2.0 Base | [tibetan-asr-nict-tib1-wav2vec2-base](https://huggingface.co/billingsmoore/tibetan-asr-nict-tib1-wav2vec2-base) | 0.0745 | 0.2152 | 0.2747 | 0.097 | 0.7447 |
|
| 80 |
+
| Whisper Tiny | [tibetan-asr-nict-tib1-whisper-tiny](https://huggingface.co/billingsmoore/tibetan-asr-nict-tib1-whisper-tiny) | 0.1560 | 0.2351 | 0.2975 | 0.118 | 0.6759 |
|
| 81 |
+
| Whisper Base | [tibetan-asr-nict-tib1-whisper-base](https://huggingface.co/billingsmoore/tibetan-asr-nict-tib1-whisper-base) | 0.1417 | 0.2083 | 0.2600 | 0.105 | 0.6314 |
|
| 82 |
+
| **Whisper Small** | [tibetan-asr-nict-tib1-whisper-small](https://huggingface.co/billingsmoore/tibetan-asr-nict-tib1-whisper-small) | **0.1185** | **0.1692** | **0.2042** | **0.086** | **0.5337** |
|
| 83 |
|
| 84 |
|
| 85 |
## How to use
|
|
|
|
| 87 |
```python
|
| 88 |
from transformers import pipeline
|
| 89 |
|
| 90 |
+
pipe = pipeline("automatic-speech-recognition", model="billingsmoore/tibetan-asr-nict-tib1-whisper-small")
|
| 91 |
result = pipe("path/to/audio.wav")
|
| 92 |
print(result["text"])
|
| 93 |
```
|