File size: 2,298 Bytes
df453a5 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 |
---
title: Video Transcript Chatbot
emoji: 🎥
colorFrom: yellow
colorTo: indigo
sdk: gradio
python_version: '3.10'
---
# Video Transcript Chatbot
A beginner-friendly Gradio app that turns any YouTube video into a conversational chatbot using LangChain and Hugging Face Inference API.
---
## Features
- **Dynamic Video Input**: Paste a full YouTube URL or raw video ID.
- **Embedding Model Selection**: Pick any HF embedding model (default: `sentence-transformers/all-MiniLM-L6-v2`).
- **LLM Model Selection**: Choose any HF text-generation model (default: `meta-llama/Llama-3.1-8B-Instruct`).
- **Secure Token Entry**: You must enter your own HF API token at runtime—no hard-coded defaults.
- **Conversational Memory**: Multi-turn chat history is preserved.
- **Retrieval-Augmented Generation**: Uses FAISS + transcript context to ground answers.
---
## Prerequisites
- **Python 3.8+**
- **Hugging Face API Token** with Inference access:
https://huggingface.co/settings/tokens
- **Git** (for cloning the repo)
---
## Installation
1. **Clone the repo**
```bash
git clone https://github.com/<your-username>/yt-rag-chatbot.git
cd yt-rag-chatbot
2. **(Optional) Create a virtual environment**
```bash
python -m venv venv
source venv/bin/activate # macOS/Linux
venv\Scripts\activate # Windows
3. **Install dependencies**
```bash
python -m venv venv
source venv/bin/activate # macOS/Linux
venv\Scripts\activate # Windows
**Usage**
1. **Start the app:**
```bash
python app.py
2. **Open** your browser at the local URL (e.g. http://127.0.0.1:7860)
3. **Use the UI:**
- **YouTube Video URL or ID:** Paste your link/ID.
- **Embedding Model:** Leave default or enter another HF embedding model.
- **LLM Model:** Enter your desired HF LLM repo.
- **Your HF API Token:** Paste your token (input hidden).
- Click **Initialize Chat** to load and index the transcript.
- Ask questions in the chat window to interact with the video content.
**Customization**
- **Default Models:** Edit the default values for embedding_model_input and llm_model_input in app.py.
- **Retrieval Size:** Change the k value in the retriever configuration:
```python
retriever = vector_store.as_retriever(search_kwargs={'k': 4}) |