sai-madiraju/memory_recall_pilot

Cognitive Memory Assessment - Ollama

This Hugging Face Docker Space runs:

• —Gradio on port 7860
• —Ollama internally on 127.0.0.1:11434
• —llama3.2:3b by default

The Space starts Ollama, waits for its API, pulls the configured model if needed, verifies inference, and then starts the Gradio application.

Files

• —app.py
• —Dockerfile
• —start.sh
• —requirements.txt
• —README.md

Changing model

Change OLLAMA_MODEL in the Dockerfile, for example:

• —llama3.2:1b for lower memory usage/faster CPU inference
• —llama3.2:3b for better output quality

The model is downloaded at runtime unless it is already available in the Space's local/persistent storage.