VoiceLessQ
commited on
Commit
•
6c81617
1
Parent(s):
5cfacd2
Update README.md
Browse files
README.md
CHANGED
@@ -1,74 +1,50 @@
|
|
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 |
-
|
28 |
-
|
29 |
-
|
30 |
-
|
31 |
-
|
32 |
-
|
33 |
-
|
34 |
-
|
35 |
-
|
36 |
-
|
37 |
-
|
38 |
-
|
39 |
-
|
40 |
-
|
41 |
-
|
42 |
-
|
43 |
-
|
44 |
-
|
45 |
-
|
46 |
-
|
47 |
-
|
48 |
-
|
49 |
-
|
50 |
-
|
51 |
-
- Experimentation and testing with Kalaallisut language speech-to-text tasks.
|
52 |
-
- It may be helpful for small projects or as a foundation for further fine-tuning with more data.
|
53 |
-
|
54 |
-
### Usage Example
|
55 |
-
|
56 |
-
You can use the model for transcription with the following code:
|
57 |
-
|
58 |
-
```python
|
59 |
-
from transformers import WhisperProcessor, WhisperForConditionalGeneration
|
60 |
-
import torch
|
61 |
-
|
62 |
-
# Load the model and processor
|
63 |
-
processor = WhisperProcessor.from_pretrained("VoiceLessQ/whisper-tiny-kalaallisut")
|
64 |
-
model = WhisperForConditionalGeneration.from_pretrained("VoiceLessQ/whisper-tiny-kalaallisut")
|
65 |
-
|
66 |
-
# Load and process an audio file
|
67 |
-
audio_array = ... # Load your audio file here
|
68 |
-
input_features = processor(audio_array, sampling_rate=16000, return_tensors="pt").input_features
|
69 |
-
|
70 |
-
# Generate transcription
|
71 |
-
generated_ids = model.generate(input_features)
|
72 |
-
transcription = processor.batch_decode(generated_ids, skip_special_tokens=True)[0]
|
73 |
-
|
74 |
-
print(f"Transcription: {transcription}")
|
|
|
1 |
+
# Whisper Tiny Fine-tuned on Kalaallisut
|
2 |
+
|
3 |
+
This model is a fine-tuned version of the [openai/whisper-tiny](https://huggingface.co/openai/whisper-tiny) on a very small dataset of the Kalaallisut (Greenlandic) language. Whisper is a general-purpose speech recognition model trained on a large-scale dataset. However, this fine-tuned version on Kalaallisut is **not fully reliable** due to the limited amount of training data.
|
4 |
+
|
5 |
+
### Model Details
|
6 |
+
|
7 |
+
- **Model**: Whisper Tiny Fine-tuned on Kalaallisut
|
8 |
+
- **Base Model**: [openai/whisper-tiny](https://huggingface.co/openai/whisper-tiny)
|
9 |
+
- **Training Data**: A very small amount of audio-transcription pairs in Kalaallisut
|
10 |
+
- **Purpose**: Speech-to-text for the Kalaallisut language
|
11 |
+
- **License**: [MIT License](https://opensource.org/licenses/MIT)
|
12 |
+
|
13 |
+
### Training Data
|
14 |
+
|
15 |
+
This model was fine-tuned on a small dataset of the Kalaallisut language. The dataset was **not comprehensive** and **does not cover all aspects** of the language. As a result, the model is not reliable for general use cases and may produce incorrect transcriptions.
|
16 |
+
|
17 |
+
### Limitations
|
18 |
+
|
19 |
+
- The model has been trained on a **very small dataset** (only a few hours of audio).
|
20 |
+
- The **transcriptions may not be accurate**, especially for more complex audio inputs.
|
21 |
+
- This model is a **proof of concept** and not intended for production use.
|
22 |
+
|
23 |
+
### Intended Use
|
24 |
+
|
25 |
+
The model can be used for:
|
26 |
+
|
27 |
+
- Experimentation and testing with Kalaallisut language speech-to-text tasks.
|
28 |
+
- It may be helpful for small projects or as a foundation for further fine-tuning with more data.
|
29 |
+
|
30 |
+
### Usage Example
|
31 |
+
|
32 |
+
You can use the model for transcription with the following code:
|
33 |
+
|
34 |
+
```python
|
35 |
+
from transformers import WhisperProcessor, WhisperForConditionalGeneration
|
36 |
+
import torch
|
37 |
+
|
38 |
+
# Load the model and processor
|
39 |
+
processor = WhisperProcessor.from_pretrained("VoiceLessQ/whisper-tiny-kalaallisut")
|
40 |
+
model = WhisperForConditionalGeneration.from_pretrained("VoiceLessQ/whisper-tiny-kalaallisut")
|
41 |
+
|
42 |
+
# Load and process an audio file
|
43 |
+
audio_array = ... # Load your audio file here
|
44 |
+
input_features = processor(audio_array, sampling_rate=16000, return_tensors="pt").input_features
|
45 |
+
|
46 |
+
# Generate transcription
|
47 |
+
generated_ids = model.generate(input_features)
|
48 |
+
transcription = processor.batch_decode(generated_ids, skip_special_tokens=True)[0]
|
49 |
+
|
50 |
+
print(f"Transcription: {transcription}")
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|