README.md

metadata

title: OralCoachZeroGPU
emoji: ⚡
colorFrom: red
colorTo: purple
sdk: gradio
sdk_version: 4.31.4
app_file: app.py
pinned: false
license: mit

Oral Coach 🎤✨ powered by ZeroGPU ⚡ϞϞ(๑⚈ ․̫ ⚈๑)∩ ⚡ and Mixtral 🎭🎓 The Oral Coach is an AI-powered conversational coach designed to guide students through their oral responses. It is built using Gradio, HuggingFace's InferenceClient, and edge_tts for text-to-speech conversion. The Oral Coach is participating in the Hugging Face ZeroGPU initiative. Project Objectives

Enhance students' critical thinking when structuring their responses to oral prompts. Provide personalized feedback to students. Improve students' structured response techniques for oral communication.

Features

Student information input (class and index number) User acceptance policy checkbox Question selection Thinking frame selection Feedback level selection Audio recording and transcription AI-generated personalized feedback Text-to-speech feedback playback Teacher's dashboard (not shown in the provided code)

Getting Started The Oral Coach app was piloted in 2024 by over 1000+ students in 5 S4 Cluster Schools. This repository contains the source code for the Oral Coach app. You can run the app locally or deploy it to Render.com. To get started with the Oral Coach, simply clone the Hugging Space repository as the app is hosted on Hugging Face: Copy codegit clone https://huggingface.co/spaces/your-username/oral-coach Usage

Access the application through the provided URL. Enter the student information, agree to the user acceptance policy, and click "Submit Info". Choose a question, thinking frame, and feedback level. Record the oral response using the microphone. Click "Submit Oral Response" to generate personalized feedback. Review the feedback in the chatbot and listen to the audio playback.

Code Structure

app.py: The main application file containing the Gradio interface and prediction logic. thinkingframes.py: Contains the question prompts, thinking frames, and feedback generation prompts. styles.py: Defines the theme and styling for the Gradio interface. utils.py: Utility functions for displaying images and collecting student information. database_functions.py: Functions for interacting with the database (not shown in the provided code). tab_teachers_dashboard.py: Functions for creating the teacher's dashboard tab (not shown in the provided code). config.py: Configuration variables (not shown in the provided code).

ZeroGPU Integration The Oral Coach leverages Hugging Face's ZeroGPU initiative to enable GPU acceleration for inference. The @spaces.GPU(duration=120) decorator is used to allocate GPU resources for the model function, which performs the inference using the "mistralai/Mixtral-8x7B-Instruct-v0.1" model. The Oral Coach utilizes the Mixtral model, a powerful language model specifically designed for educational applications. Mixtral provides the underlying intelligence for generating personalized feedback and guiding students through their oral responses. Contributing Contributions are welcome! If you find any issues or have suggestions for improvements, please create an issue or submit a pull request. License This project is licensed under the MIT License. Acknowledgements

Hugging Face for providing the InferenceClient and ZeroGPU initiative. Gradio for the user interface library. edge_tts for text-to-speech conversion.

Feel free to customize and expand upon this README template based on your specific Oral Coach application and requirements.