[Feature] Support BigCodeBench

[terryyz]

Describe the feature

BigCode (Hugging Face and ServiceNow Research) released a new large-scale benchmark, BigCodeBench, for code generation with diverse function calls and complex instructions, covering 1140 expert-annotated tasks. It has been officially used by DeepSeek and CodeGeeX4. BigCodeBench is considered a better alternative for HumanEval and other function-level code generation benchmarks (see here).

• Preprint
• GitHub
• Leaderboard

Will you implement it?

• [ ] I would like to implement this feature and create a PR!