harsh2ai
commited on
Commit
Β·
d87796a
1
Parent(s):
508b24f
Reorder UI and update Hindi focus
Browse files
app.py
CHANGED
|
@@ -133,15 +133,70 @@ def create_interface():
|
|
| 133 |
gr.Markdown("""
|
| 134 |
<div class="main-header">
|
| 135 |
<h1>ποΈ Ringg STT V0</h1>
|
| 136 |
-
<p>
|
| 137 |
</div>
|
| 138 |
""")
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 139 |
|
|
|
|
|
|
|
|
|
|
|
|
|
| 140 |
# Performance Comparison Table
|
| 141 |
gr.Markdown("""
|
| 142 |
## Performance Benchmarks
|
| 143 |
|
| 144 |
-
Our model achieves **state-of-the-art performance** on
|
| 145 |
""")
|
| 146 |
|
| 147 |
with gr.Row():
|
|
@@ -163,71 +218,8 @@ def create_interface():
|
|
| 163 |
gr.Markdown("""
|
| 164 |
**Ringg STT V0** ranks **2nd** among top models, outperforming OpenAI Whisper Large-v3 and other leading solutions.
|
| 165 |
|
| 166 |
-
Lower WER (Word Error Rate) indicates better accuracy. Our model
|
| 167 |
-
""")
|
| 168 |
-
|
| 169 |
-
gr.Markdown("""
|
| 170 |
-
### β¨ Features
|
| 171 |
-
- π **Bilingual Support**: Transcribe English and Hindi speech
|
| 172 |
-
- π― **High Accuracy**: Competitive with leading ASR models
|
| 173 |
-
- π **File Upload**: Support for various audio formats (WAV, MP3, FLAC, etc.)
|
| 174 |
-
- β‘ **Fast Processing**: Optimized for quick transcription
|
| 175 |
-
- π **Private Infrastructure**: Secure and controlled deployment
|
| 176 |
-
""")
|
| 177 |
-
|
| 178 |
-
gr.Markdown("""
|
| 179 |
-
### π Links
|
| 180 |
-
- **Organization**: [RinggAI on Hugging Face](https://huggingface.co/RinggAI)
|
| 181 |
-
- **TTS Space**: [Ringg TTS V0](https://huggingface.co/spaces/RinggAI/Ringg-TTS-v0.0)
|
| 182 |
-
|
| 183 |
-
### π Acknowledgements
|
| 184 |
-
- Special thanks to [@jeremylee12](https://huggingface.co/jeremylee12) for their contributions
|
| 185 |
""")
|
| 186 |
-
|
| 187 |
-
# API Status indicator
|
| 188 |
-
with gr.Row():
|
| 189 |
-
with gr.Column(scale=4):
|
| 190 |
-
api_status = gr.Textbox(
|
| 191 |
-
label="π API Status",
|
| 192 |
-
value=health_status["message"],
|
| 193 |
-
interactive=False
|
| 194 |
-
)
|
| 195 |
-
with gr.Column(scale=1):
|
| 196 |
-
check_btn = gr.Button("π Check Status", size="sm")
|
| 197 |
-
check_btn.click(check_api_status, outputs=api_status)
|
| 198 |
-
|
| 199 |
-
with gr.Tab("π File Upload"):
|
| 200 |
-
gr.Markdown("### Upload Audio File")
|
| 201 |
-
gr.Markdown("Upload an audio file for transcription (supports WAV, MP3, FLAC, M4A, etc.)")
|
| 202 |
-
|
| 203 |
-
audio_input = gr.Audio(
|
| 204 |
-
label="π Upload Audio File",
|
| 205 |
-
type="filepath",
|
| 206 |
-
sources=["upload"]
|
| 207 |
-
)
|
| 208 |
-
|
| 209 |
-
transcribe_btn = gr.Button("π Transcribe", variant="primary", size="lg")
|
| 210 |
-
|
| 211 |
-
file_output = gr.Textbox(
|
| 212 |
-
label="Transcription Result",
|
| 213 |
-
lines=8,
|
| 214 |
-
interactive=False,
|
| 215 |
-
placeholder="Upload a file and click Transcribe..."
|
| 216 |
-
)
|
| 217 |
-
|
| 218 |
-
transcribe_btn.click(
|
| 219 |
-
transcribe_audio,
|
| 220 |
-
inputs=audio_input,
|
| 221 |
-
outputs=file_output
|
| 222 |
-
)
|
| 223 |
-
|
| 224 |
-
gr.Markdown("""
|
| 225 |
-
### π‘ Tips for Best Results
|
| 226 |
-
- Use clear audio with minimal background noise
|
| 227 |
-
- Speak naturally at a moderate pace
|
| 228 |
-
- For file upload, ensure audio quality is good (16kHz or higher recommended)
|
| 229 |
-
- Model handles code-switching between English and Hindi
|
| 230 |
-
""")
|
| 231 |
|
| 232 |
return demo
|
| 233 |
|
|
|
|
| 133 |
gr.Markdown("""
|
| 134 |
<div class="main-header">
|
| 135 |
<h1>ποΈ Ringg STT V0</h1>
|
| 136 |
+
<p>High-Accuracy Hindi Speech-to-Text</p>
|
| 137 |
</div>
|
| 138 |
""")
|
| 139 |
+
|
| 140 |
+
# API Status indicator
|
| 141 |
+
with gr.Row():
|
| 142 |
+
with gr.Column(scale=4):
|
| 143 |
+
api_status = gr.Textbox(
|
| 144 |
+
label="π API Status",
|
| 145 |
+
value=health_status["message"],
|
| 146 |
+
interactive=False
|
| 147 |
+
)
|
| 148 |
+
with gr.Column(scale=1):
|
| 149 |
+
check_btn = gr.Button("π Check Status", size="sm")
|
| 150 |
+
check_btn.click(check_api_status, outputs=api_status)
|
| 151 |
+
|
| 152 |
+
gr.Markdown("""
|
| 153 |
+
### π File Upload
|
| 154 |
+
Upload an audio file for transcription (supports WAV, MP3, FLAC, M4A, etc.)
|
| 155 |
+
""")
|
| 156 |
+
|
| 157 |
+
audio_input = gr.Audio(
|
| 158 |
+
label="π Upload Audio File",
|
| 159 |
+
type="filepath",
|
| 160 |
+
sources=["upload"]
|
| 161 |
+
)
|
| 162 |
+
|
| 163 |
+
transcribe_btn = gr.Button("π Transcribe", variant="primary", size="lg")
|
| 164 |
+
|
| 165 |
+
file_output = gr.Textbox(
|
| 166 |
+
label="Transcription Result",
|
| 167 |
+
lines=8,
|
| 168 |
+
interactive=False,
|
| 169 |
+
placeholder="Upload a file and click Transcribe..."
|
| 170 |
+
)
|
| 171 |
+
|
| 172 |
+
transcribe_btn.click(
|
| 173 |
+
transcribe_audio,
|
| 174 |
+
inputs=audio_input,
|
| 175 |
+
outputs=file_output
|
| 176 |
+
)
|
| 177 |
+
|
| 178 |
+
gr.Markdown("""
|
| 179 |
+
### β¨ Features
|
| 180 |
+
- π **Hindi Support**: Accurate transcription for Hindi audio
|
| 181 |
+
- π― **High Accuracy**: Competitive with leading ASR models
|
| 182 |
+
- π **File Upload**: Support for various audio formats (WAV, MP3, FLAC, etc.)
|
| 183 |
+
- β‘ **Fast Processing**: Optimized for quick transcription
|
| 184 |
+
""")
|
| 185 |
+
|
| 186 |
+
gr.Markdown("""
|
| 187 |
+
### π Links
|
| 188 |
+
- **Organization**: [RinggAI on Hugging Face](https://huggingface.co/RinggAI)
|
| 189 |
+
- **TTS Space**: [Ringg TTS V0](https://huggingface.co/spaces/RinggAI/Ringg-TTS-v0.0)
|
| 190 |
|
| 191 |
+
### π Acknowledgements
|
| 192 |
+
- Special thanks to [@jeremylee12](https://huggingface.co/jeremylee12) for their contributions
|
| 193 |
+
""")
|
| 194 |
+
|
| 195 |
# Performance Comparison Table
|
| 196 |
gr.Markdown("""
|
| 197 |
## Performance Benchmarks
|
| 198 |
|
| 199 |
+
Our model achieves **state-of-the-art performance** on Hindi speech recognition benchmarks:
|
| 200 |
""")
|
| 201 |
|
| 202 |
with gr.Row():
|
|
|
|
| 218 |
gr.Markdown("""
|
| 219 |
**Ringg STT V0** ranks **2nd** among top models, outperforming OpenAI Whisper Large-v3 and other leading solutions.
|
| 220 |
|
| 221 |
+
Lower WER (Word Error Rate) indicates better accuracy. Our model delivers competitive performance for Hindi transcription tasks.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 222 |
""")
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 223 |
|
| 224 |
return demo
|
| 225 |
|