These are the raw results from the BenchName benchmark suite, as well as the corresponding model predictions.
Please use the subset dropdown menu to select the necessary data relating to our six benchmarks:
🤗 Library-based code generation
🤗 CI builds repair
🤗 Project-level code completion
🤗 Commit message generation🤗 Bug localization
🤗 Module summarization
