Code Dataset Generator for Falcon 40B

Upload Source Files

Drag & drop your source files here

Supported: Python, C/C++, Rust, Go, JavaScript, Java, PHP, Ruby, TypeScript

Output Format

Language Detection

Extract functions/methods Extract classes Include comments/docstrings Clean code (remove noise)

Processing files... {{ processedFiles }} / {{ selectedFiles.length }}

{{ previewData.length }} items

Instruction

Input Code

{{ previewData[selectedSampleIndex].input }}

Output

Your processed dataset will appear here

Automatically detects and processes code in Python, C/C++, Rust, Go, JavaScript and more.

Automatically generates meaningful instruction-output pairs from your source code.

Tokenization and formatting specifically optimized for Falcon 40B model fine-tuning.