Upload README.md
Browse files
README.md
ADDED
|
@@ -0,0 +1,63 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
title: Audio/Video Translation Toolkit
|
| 3 |
+
emoji: 🚀
|
| 4 |
+
colorFrom: indigo
|
| 5 |
+
colorTo: purple
|
| 6 |
+
sdk: gradio
|
| 7 |
+
python_version: 3.10.0
|
| 8 |
+
app_file: app.py
|
| 9 |
+
tags:
|
| 10 |
+
- translation
|
| 11 |
+
- audio
|
| 12 |
+
- video
|
| 13 |
+
- speech-synthesis
|
| 14 |
+
- voice-cloning
|
| 15 |
+
- gradio
|
| 16 |
+
models:
|
| 17 |
+
- openai/whisper-large-v3
|
| 18 |
+
- JaesungHuh/voice-gender-classifier
|
| 19 |
+
- ai4bharat/IndicF5
|
| 20 |
+
preload_from_hub:
|
| 21 |
+
- openai/whisper-large-v3
|
| 22 |
+
- JaesungHuh/voice-gender-classifier
|
| 23 |
+
- ai4bharat/IndicF5
|
| 24 |
+
---
|
| 25 |
+
|
| 26 |
+
# 🚀 Audio/Video Translation Toolkit
|
| 27 |
+
|
| 28 |
+
This application provides a complete pipeline for translating the audio of a video or audio file from English to various Indian languages. It handles everything from vocal separation and transcription to translation, speech synthesis, and voice cloning.
|
| 29 |
+
|
| 30 |
+
---
|
| 31 |
+
|
| 32 |
+
## ## Key Features 🛠️
|
| 33 |
+
|
| 34 |
+
* **🎬 Full Video Translation:** Upload a video, and the app will extract the audio, translate it, and merge it back into the original video.
|
| 35 |
+
* **🎵 Full Audio Translation:** Translate standalone audio files.
|
| 36 |
+
* **🗣️ Vocal Separation:** Isolate vocals from background music before processing.
|
| 37 |
+
* **✍️ Transcription & Pace Detection:** Uses Whisper to transcribe the audio and determine the original speaker's pace.
|
| 38 |
+
* **🌐 Multi-Lingual Translation:** Translate text to Tamil, Telugu, or Hindi using either local models or the Sarvam API.
|
| 39 |
+
* **🔊 Speech Synthesis:** Generate new speech in the target language using models from `ai4bharat`.
|
| 40 |
+
* **🧬 Voice Cloning:** Clone the voice from the original speaker onto the newly synthesized audio for a more natural result.
|
| 41 |
+
|
| 42 |
+
---
|
| 43 |
+
|
| 44 |
+
## ## How to Use the Main Pipeline
|
| 45 |
+
|
| 46 |
+
1. Navigate to the **Translate Video** or **Translate Audio** tab.
|
| 47 |
+
2. Upload your file.
|
| 48 |
+
3. Select the **Target Language**.
|
| 49 |
+
4. Choose the **Translation Via** method (`local` or `api`).
|
| 50 |
+
5. Click the **Translate** button and wait for the process to complete.
|
| 51 |
+
|
| 52 |
+
---
|
| 53 |
+
|
| 54 |
+
## ## ‼️ Important Setup Note for Duplication
|
| 55 |
+
|
| 56 |
+
This Space relies on several local modules and data files that are not installed via `pip`. If you are duplicating this Space, you **must** manually upload the following directories to the root of your repository for the application to function correctly:
|
| 57 |
+
|
| 58 |
+
* `gender/` (Contains the gender prediction model code)
|
| 59 |
+
* `openvoice/` (Contains the voice cloning API and extractor code)
|
| 60 |
+
* `checkpoints_v2/` (Contains the pre-trained model checkpoints for voice cloning)
|
| 61 |
+
* `reference/` (Contains the reference audio and text files for speech synthesis)
|
| 62 |
+
|
| 63 |
+
Without these directories, the application will fail with `ImportError` or `FileNotFoundError`.
|