Spaces:
Sleeping
Sleeping
| title: JULIA | |
| emoji: 🔥 | |
| colorFrom: blue | |
| colorTo: green | |
| sdk: gradio | |
| sdk_version: 4.28.3 | |
| app_file: app.py | |
| pinned: false | |
| short_description: Voice Chat with JULIA | |
| # JULIA⚡ | |
| A personal assistant inspired by Tony Stark's JARVIS, built using Gradio, edge_tts, Hugging Face Inference API, and streaming STT (Speech-to-Text) with Nemo. | |
| ## Features | |
| - Voice and text input support | |
| - Text-to-Speech response | |
| - Multiple model support via Hugging Face Inference API | |
| - Friendly and concise responses from a virtual assistant named Julia | |
| ## How It Works | |
| - **Speech-to-Text (STT):** Uses the `streaming_stt_nemo` library to transcribe audio inputs. | |
| - **Text Generation:** Uses models from Hugging Face Inference API to generate responses. | |
| - **Text-to-Speech (TTS):** Uses `edge_tts` to convert the generated response into audio. | |
| ## Interface | |
| The Gradio interface includes: | |
| - A dropdown to select the model. | |
| - An audio input for voice commands. | |
| - A text input for typed commands. | |
| - A send button to submit the input. | |
| - A text output to display the assistant's response. | |
| - An audio output to play the assistant's response. | |
| ## License | |
| This project is licensed under the MIT License. See the [LICENSE](LICENSE) file for details. | |
| ## Acknowledgments | |
| Special thanks to the developers of Gradio, Hugging Face, edge_tts, and Nemo for their amazing libraries. | |
| Sure! Here's a detailed description of each model available in the dropdown for your virtual assistant project: | |
| ### Model Descriptions | |
| #### 1. **Mixtral 8x7B** | |
| - **Description:** Mixtral 8x7B is a state-of-the-art large language model developed by Mistralai. It's designed to handle a variety of natural language understanding and generation tasks with high accuracy and coherence. | |
| - **Strengths:** | |
| - Excels in general knowledge and conversational tasks. | |
| - Capable of producing detailed and contextually relevant responses. | |
| - **Use Cases:** Suitable for detailed Q&A, storytelling, and providing comprehensive explanations. | |
| #### 2. **Llama 3 8B** | |
| - **Description:** Llama 3 8B is the latest iteration in the Llama series developed by Meta. It focuses on generating human-like text based on the provided prompts. | |
| - **Strengths:** | |
| - Highly optimized for generating coherent and context-aware text. | |
| - Efficient in understanding and maintaining conversation flow. | |
| - **Use Cases:** Ideal for chatbots, creative writing, and interactive dialogues. | |
| #### 3. **Mistral 7B v0.3** | |
| - **Description:** Mistral 7B v0.3 is a powerful language model developed by Mistralai, designed to perform well in both understanding and generating text across various domains. | |
| - **Strengths:** | |
| - High performance in both technical and casual conversational contexts. | |
| - Robust in handling diverse topics and maintaining context over longer interactions. | |
| - **Use Cases:** Best suited for customer support, technical assistance, and in-depth discussions. | |
| #### 4. **Phi 3 mini** | |
| - **Description:** Phi 3 mini, developed by Microsoft, is a compact yet efficient language model optimized for fast and responsive text generation. | |
| - **Strengths:** | |
| - Lightweight and quick to respond. | |
| - Maintains a good balance between performance and computational efficiency. | |
| - **Use Cases:** Perfect for real-time applications, quick Q&A, and use cases where response time is critical. | |
| ### Available Models | |
| This project supports multiple language models to cater to different needs. Here are the details of each model: | |
| 1. **Mixtral 8x7B** | |
| - **Description:** A state-of-the-art language model by Mistralai, designed for a wide range of natural language tasks. | |
| - **Strengths:** Excels in general knowledge, conversational tasks, and detailed responses. | |
| - **Use Cases:** Q&A, storytelling, detailed explanations. | |
| 2. **Llama 3 8B** | |
| - **Description:** The latest in the Llama series by Meta, focusing on generating human-like text. | |
| - **Strengths:** Highly coherent text generation, maintains conversation flow. | |
| - **Use Cases:** Chatbots, creative writing, interactive dialogues. | |
| 3. **Mistral 7B v0.3** | |
| - **Description:** A powerful model by Mistralai, capable of understanding and generating text across various domains. | |
| - **Strengths:** High performance in technical and casual contexts, robust over longer interactions. | |
| - **Use Cases:** Customer support, technical assistance, in-depth discussions. | |
| 4. **Phi 3 mini** | |
| - **Description:** A compact model by Microsoft, optimized for fast and responsive text generation. | |
| - **Strengths:** Lightweight, quick response times, efficient. | |
| - **Use Cases:** Real-time applications, quick Q&A, scenarios requiring fast responses. | |
| You can select any of these models based on your specific needs from the dropdown menu in the interface. | |