Sci-Fi Image Recognition AI

I’m building a small end-to-end AI tool that first recognises what it sees in an image and then spins those findings into original science-fiction prose or poetry.

Core needs
The heart of the project is a solid image-recognition model—think TensorFlow or PyTorch—capable of tagging objects, settings and moods with good accuracy. Once the scene is understood, a text-generation layer (GPT-style or similar) should take those tags and weave them into creative, space-age narratives or concise poems. The writing must sound natural, imaginative and unmistakably sci-fi.

What you’ll deliver
• Trained image-recognition model with its labelled dataset or clear notes on the dataset used
• Fine-tuned creative-writing module focused on science-fiction tone and vocabulary
• Lightweight demo app (web or desktop) that lets me upload an image and instantly read the generated story/poem
• Read-me explaining setup, retraining steps and how to expand the genre list in future

Acceptance check
If I can drop in ten unseen images—starfields, robots, alien landscapes—and receive coherent, vivid sci-fi text that references what’s actually in each picture, the job is done.