Update README.md
Browse files
README.md
CHANGED
@@ -1,3 +1,52 @@
|
|
1 |
---
|
2 |
license: creativeml-openrail-m
|
|
|
|
|
|
|
|
|
3 |
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
---
|
2 |
license: creativeml-openrail-m
|
3 |
+
language:
|
4 |
+
- en
|
5 |
+
- hi
|
6 |
+
pipeline_tag: automatic-speech-recognition
|
7 |
---
|
8 |
+
---
|
9 |
+
language:
|
10 |
+
- hi
|
11 |
+
license: apache-2.0
|
12 |
+
tags:
|
13 |
+
- whisper-event
|
14 |
+
metrics:
|
15 |
+
- wer
|
16 |
+
model-index:
|
17 |
+
- name: LLM-HINDI-LARGE - Manan Raval
|
18 |
+
results:
|
19 |
+
- task:
|
20 |
+
type: automatic-speech-recognition
|
21 |
+
name: Automatic Speech Recognition
|
22 |
+
dataset:
|
23 |
+
name: google/fleurs
|
24 |
+
type: google/fleurs
|
25 |
+
config: hn_in
|
26 |
+
split: test
|
27 |
+
metrics:
|
28 |
+
- type: wer
|
29 |
+
value: 12.33
|
30 |
+
name: WER
|
31 |
+
|
32 |
+
|
33 |
+
## Usage
|
34 |
+
|
35 |
+
In order to infer a single audio file using this model, the following code snippet can be used:
|
36 |
+
|
37 |
+
```python
|
38 |
+
>>> import torch
|
39 |
+
>>> from transformers import pipeline
|
40 |
+
|
41 |
+
>>> # path to the audio file to be transcribed
|
42 |
+
>>> audio = "/path/to/audio.format"
|
43 |
+
>>> device = "cuda:0" if torch.cuda.is_available() else "cpu"
|
44 |
+
|
45 |
+
>>> transcribe = pipeline(task="automatic-speech-recognition", model="web30india/LLM-Hindi-Large", chunk_length_s=30, device=device)
|
46 |
+
>>> transcribe.model.config.forced_decoder_ids = transcribe.tokenizer.get_decoder_prompt_ids(language="hi", task="transcribe")
|
47 |
+
|
48 |
+
>>> print('Transcription: ', transcribe(audio)["text"])
|
49 |
+
```
|
50 |
+
|
51 |
+
## Acknowledgement
|
52 |
+
This work was done at [Virtual Height IT Services Pvt. Ltd.]
|