<<<<<<< HEAD
Content Creation with LLM
A project that revolutionizes content creation using Large Language Models (LLMs). This offline system leverages GPT-Neo for text generation, providing flexibility, contextual awareness, and high-quality outputs.
Features:
Text Generation: Generates high-quality, context-aware content. Grammar Feedback: Uses language-tool-python for grammar and style checks. Fully Offline: Runs without internet access, including the GPT-Neo model (~5GB). Customizable Settings: Adjust parameters like max length and temperature for output control. Local Execution: Access the application via your browser on http://127.0.0.1:5000.
System Requirements:
Operating System: Windows 10/11 Disk Space: ~6GB (includes GPT-Neo model and Java runtime) RAM: 8GB or more recommended Dependencies: Python 3.9+ Java Runtime Environment (JRE)
Installation:
Step 1: Extract Files Unzip the provided archive (ContentCreationLLM.zip) into a folder. Ensure the following folders and files are present:
ContentCreationLLM/ ├── ContentCreationLLM.exe ├── models/ │ └── gpt-neo-1.3B/ ├── java/ │ └── bin/ ├── app/ │ ├── templates/ │ └── static/
Step 2: Run the Application Double-click ContentCreationLLM.exe. Open your browser and navigate to: arduino Copy Edit http://127.0.0.1:5000
Step 3: Generate Content Enter a prompt in the provided input box. Adjust Max Length and Temperature as needed. Click Generate to create content.
How It Works:
Text Generation: Uses the GPT-Neo model loaded from the local models/ folder. Grammar Feedback: Integrated language-tool-python checks for grammar and style errors. Java Runtime is used for the LanguageTool backend. Frontend: Flask serves a local web app accessible through your browser.
Configuration:
Modify Model Parameters: To customize model behavior, edit the app/models.py file before creating the .exe: python Copy Edit result = text_generator( prompt, max_length=300, temperature=0.7, top_k=50, top_p=0.9, )
Known Issues:
Startup Issues: Ensure no other service is using port 5000. Java Not Found: Ensure the java/ folder is included and configured correctly.
Credits: Transformers Library: Hugging Face LanguageTool: Open source grammar checker Flask: Python web framework
origin/main