Spaces:
Sleeping
Sleeping
divakaivan
commited on
Commit
•
c6e8353
1
Parent(s):
76fd095
Upload folder using huggingface_hub
Browse files- .gitignore +1 -0
- README.md +2 -9
- app.py +4 -3
- app_old.py +29 -0
- requirements.txt +1 -1
.gitignore
ADDED
@@ -0,0 +1 @@
|
|
|
|
|
1 |
+
.DS_Store
|
README.md
CHANGED
@@ -1,13 +1,6 @@
|
|
1 |
---
|
2 |
-
title:
|
3 |
-
|
4 |
-
colorFrom: indigo
|
5 |
-
colorTo: blue
|
6 |
sdk: gradio
|
7 |
sdk_version: 4.36.1
|
8 |
-
app_file: app.py
|
9 |
-
pinned: false
|
10 |
-
license: apache-2.0
|
11 |
---
|
12 |
-
|
13 |
-
Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference
|
|
|
1 |
---
|
2 |
+
title: glaswegian-whisper
|
3 |
+
app_file: app.py
|
|
|
|
|
4 |
sdk: gradio
|
5 |
sdk_version: 4.36.1
|
|
|
|
|
|
|
6 |
---
|
|
|
|
app.py
CHANGED
@@ -4,11 +4,12 @@ import gradio as gr
|
|
4 |
pipe = pipeline(model="divakaivan/whisper-small-hi_test")
|
5 |
|
6 |
def transcribe(audio):
|
7 |
-
|
|
|
8 |
text = pipe(audio)["text"]
|
9 |
return text
|
10 |
|
11 |
-
|
12 |
fn=transcribe,
|
13 |
inputs=gr.Audio(type="filepath"),
|
14 |
outputs="text",
|
@@ -16,4 +17,4 @@ demo = gr.Interface(
|
|
16 |
description="Realtime demo for Glaswegian speech recognition using a fine-tuned Whisper small model.",
|
17 |
)
|
18 |
|
19 |
-
|
|
|
4 |
pipe = pipeline(model="divakaivan/whisper-small-hi_test")
|
5 |
|
6 |
def transcribe(audio):
|
7 |
+
print(audio)
|
8 |
+
print(type(audio))
|
9 |
text = pipe(audio)["text"]
|
10 |
return text
|
11 |
|
12 |
+
iface = gr.Interface(
|
13 |
fn=transcribe,
|
14 |
inputs=gr.Audio(type="filepath"),
|
15 |
outputs="text",
|
|
|
17 |
description="Realtime demo for Glaswegian speech recognition using a fine-tuned Whisper small model.",
|
18 |
)
|
19 |
|
20 |
+
iface.launch(share=True)
|
app_old.py
ADDED
@@ -0,0 +1,29 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import streamlit as st
|
2 |
+
from st_audiorec import st_audiorec
|
3 |
+
from transformers import AutoProcessor, AutoModelForSpeechSeq2Seq
|
4 |
+
|
5 |
+
processor = AutoProcessor.from_pretrained("openai/whisper-small")
|
6 |
+
model = AutoModelForSpeechSeq2Seq.from_pretrained("openai/whisper-small")
|
7 |
+
|
8 |
+
def transcribe(audio):
|
9 |
+
|
10 |
+
text = processor.batch_decode(model.generate(processor(audio), num_beams=4), skip_special_tokens=True)
|
11 |
+
return text
|
12 |
+
|
13 |
+
wav_audio_data = st_audiorec()
|
14 |
+
|
15 |
+
if wav_audio_data is not None:
|
16 |
+
# st.audio(wav_audio_data, format='audio/wav')
|
17 |
+
st.write("Transcription:")
|
18 |
+
st.write(transcribe(wav_audio_data))
|
19 |
+
|
20 |
+
# Set up the Streamlit app
|
21 |
+
st.title("Glaswegian Transcription with Whisper")
|
22 |
+
api_key = st.sidebar.text_input("Enter your API key")
|
23 |
+
|
24 |
+
# Check if API key is provided
|
25 |
+
if api_key:
|
26 |
+
st.write("API key:", api_key)
|
27 |
+
# Add your code here to use the Whisper model for audio transcription
|
28 |
+
else:
|
29 |
+
st.warning("Please enter your API key in the sidebar.")
|
requirements.txt
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
transformers
|
2 |
torch
|
3 |
-
torchvision
|
|
|
1 |
transformers
|
2 |
torch
|
3 |
+
torchvision
|