Agentic Coding Dataset: Aggregated Programming Tasks for AI Training

An aggregated dataset of coding and instruction-following examples designed to train agentic coding models. It was compiled by author 'ethanker' and last updated on November 30, 2025. The dataset likely contains samples from sources including CodeAlpaca-20k, Evol-CodeAlpaca-v1, Code Review Instruct, and APPS.

Use Cases

• Train models for code generation based on instruction-following tasks from CodeAlpaca-20k.
• Develop models for handling complex, evolved coding instructions similar to those in Evol-CodeAlpaca-v1.
• Fine-tune models for Python code review and revision using examples from Code Review Instruct.
• Benchmark model performance on programming problems from the APPS (Automated Programming Progress Standard) source.

Strengths

• Aggregates multiple established coding datasets, potentially providing diverse task types.
• Specifically designed for training agentic coding models, indicating a focused application.
• Last updated on 2025-11-30, suggesting recent maintenance.

Limitations

• Description metadata is limited; actual data quality, size, and column structure require manual inspection after download.
• Row count, file formats, and license information are unknown, which may limit suitability assessment.
• Data may reflect source-specific biases inherent to the aggregated datasets from huggingface.

Provenance

Source

Aggregated from multiple sources including CodeAlpaca-20k, Evol-CodeAlpaca-v1, Code Review Instruct, and APPS.

Collection Method

Compilation of existing datasets.

Freshness

Last updated 2025-11-30 02:03:17; freshness should be verified.