Spaces:
Sleeping
Sleeping
Update README.md
Browse files
README.md
CHANGED
|
@@ -1,42 +1,30 @@
|
|
| 1 |
---
|
| 2 |
-
title:
|
| 3 |
emoji: ποΈ
|
| 4 |
colorFrom: indigo
|
| 5 |
colorTo: purple
|
| 6 |
sdk: gradio
|
| 7 |
-
sdk_version: "
|
| 8 |
app_file: app.py
|
| 9 |
pinned: true
|
| 10 |
-
license: mit
|
| 11 |
---
|
| 12 |
|
| 13 |
-
#
|
| 14 |
|
| 15 |
-
##
|
| 16 |
-
|
| 17 |
-
-
|
| 18 |
-
-
|
| 19 |
-
- Generate a **summary** in either **Urdu or English**
|
| 20 |
|
| 21 |
-
|
| 22 |
-
|
| 23 |
-
|
| 24 |
-
|
| 25 |
-
|
| 26 |
-
|
| 27 |
-
| π§© Summarization | `facebook/bart-large-cnn` (English) <br> `facebook/mbart-large-50-many-to-many-mmt` (Urdu) |
|
| 28 |
-
| π» Interface | Gradio 4.43 |
|
| 29 |
-
| π Deployment | Hugging Face Spaces |
|
| 30 |
-
| ποΈ YouTube Audio | yt-dlp + ffmpeg |
|
| 31 |
-
|
| 32 |
-
---
|
| 33 |
-
|
| 34 |
-
## βοΈ Setup Instructions (for Hugging Face Spaces)
|
| 35 |
|
| 36 |
-
|
| 37 |
-
|
| 38 |
-
|
| 39 |
-
|
| 40 |
|
| 41 |
-
2. Go to **Settings β Variables**
|
| 42 |
-
Add these two environment variables:
|
|
|
|
| 1 |
---
|
| 2 |
+
title: SmartTranscribe β YouTube + Upload (Urdu/English)
|
| 3 |
emoji: ποΈ
|
| 4 |
colorFrom: indigo
|
| 5 |
colorTo: purple
|
| 6 |
sdk: gradio
|
| 7 |
+
sdk_version: "4.44.0"
|
| 8 |
app_file: app.py
|
| 9 |
pinned: true
|
|
|
|
| 10 |
---
|
| 11 |
|
| 12 |
+
# SmartTranscribe β YouTube + Upload (Urdu/English)
|
| 13 |
|
| 14 |
+
## Overview
|
| 15 |
+
- Upload audio **or** paste a YouTube link.
|
| 16 |
+
- App transcribes speech (Groq Whisper `whisper-large-v3`) and summarizes the result in **English** or **Urdu**.
|
| 17 |
+
- Designed to handle long files by automatic chunking.
|
|
|
|
| 18 |
|
| 19 |
+
## Setup (Hugging Face Spaces)
|
| 20 |
+
1. Upload `app.py`, `requirements.txt`, and this `README.md`.
|
| 21 |
+
2. In **Settings β Repository secrets**, add:
|
| 22 |
+
- `GROQ_API_KEY` β your Groq API key
|
| 23 |
+
- `HUGGINGFACE_API_TOKEN` β your Hugging Face token (starts with `hf_`)
|
| 24 |
+
3. Rebuild the Space.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 25 |
|
| 26 |
+
## Notes
|
| 27 |
+
- Large videos: The app downloads YouTube audio into a temp folder, converts to WAV, splits into 5-minute chunks and transcribes each chunk to avoid size limits.
|
| 28 |
+
- Urdu summaries use `facebook/mbart-large-50-many-to-many-mmt` (works reasonably, but if you need more accurate Urdu summarization we can switch to or fine-tune a better Urdu-capable model).
|
| 29 |
+
- Temporary files are cleaned up automatically.
|
| 30 |
|
|
|
|
|
|