šŸ§¬ LLM Hill Climber

Efficient program evolution using StarCoder2-3B on HuggingFace Spaces

šŸš€ HuggingFace Spaces Deployment

Model: bigcode/starcoder2-3b (specialized coding AI) ā€¢ Strategy: Single mutation per generation ā€¢ Auth: Add your HF token above or get one free at huggingface.co/settings/tokens

šŸŽÆ Problem Configuration

šŸ¤– AI Model Selection

Specialized coding models by BigCode and Meta. Excellent at JavaScript generation.

Get a free token at huggingface.co/settings/tokens

Select a problem to begin

Choose a programming challenge for the AI to evolve solutions for.

āš” API Efficiency

Single Mutation Strategy: One API call per generation for maximum efficiency. Estimated total: 12 API calls.

Ready to evolve JavaScript programs
0 / 0 generations

šŸ“Š Fitness Evolution

šŸ”¬ Evolution Progress

Ready to start evolution process... šŸ§¬ Hill Climber Algorithm: ā€¢ Generate initial solution with StarCoder2-3B ā€¢ Each generation: create one mutation ā€¢ Keep mutation if fitness improves ā€¢ Stop when perfect solution found or max generations reached

šŸ’» Current Best Solution

No solution generated yet