APIGen Function Calling: Over 100K Synthetic Examples for LLM Training

Over 100,000 function calling examples were created by merging two existing datasets following the APIGen recipe. The dataset was uploaded by the organization 'argilla' and last updated on October 10, 2024. This version is not ready for fine-tuning and requires running a preparation script to format it for supervised fine-tuning.

Use Cases

• Fine-tuning language models for API interaction based on the described APIGen recipe.
• Training models on synthetic function-calling tasks based on the merged dataset sources.
• Benchmarking model performance on structured output generation for tool use.
• Studying the effects of synthetic data recipes on instruction-following model behavior.

Strengths

• Contains over 100,000 examples, providing a substantial corpus for training.
• Built by merging two established sources (argilla/Synth-APIGen-v0.1 and Salesforce/xlam-function-calling-60k), suggesting a combined effort.

Limitations

• Row count is unknown, which may limit suitability assessment.
• Column-level documentation is absent; field semantics must be inferred after download.
• The dataset is not ready for training and requires a preparation script, adding a step for users.

Provenance

Source

argilla

Collection Method

Merged from argilla/Synth-APIGen-v0.1 and Salesforce/xlam-function-calling-60k datasets.

Freshness

Last updated 2024-10-10 11:37:18; freshness should be verified.