Houdinivexbench: VEX Code Generation Tasks from Houdini 21.0.596

86 benchmark tasks for VEX (Vector Expression Language) code generation and understanding, extracted from Houdini 21.0.596. The dataset, created by kelvincai, was last updated on February 11, 2026. It includes tasks for code completion, documentation-to-code, and code explanation.

Use Cases

• Benchmarking code generation models based on the 61 code completion tasks.
• Training models for documentation-to-code translation based on the 12 doc_to_code tasks.
• Evaluating models for code explanation and understanding based on the 13 code_explanation tasks.
• Developing specialized AI assistants for Houdini VEX scripting based on the extracted VEX language examples.

Strengths

• Contains 86 specific benchmark tasks with defined splits for different evaluation types.
• Source data is explicitly extracted from a specific software version, Houdini 21.0.596.
• Dataset version 1.0.0 includes an official VEX compilation fix, suggesting attention to correctness.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• Data may reflect the specific features and biases of the Houdini 21.0.596 software version.

Provenance

Source

Extracted from Houdini 21.0.596 software.

Collection Method

Likely extracted and processed from the Houdini software environment.

Time Range

Based on Houdini version 21.0.596.

Freshness

Last updated 2026-02-11 01:48:02.

Geography

null