Function Calling Chatml: Synthetic Conversations for AI Assistant Training

A dataset converted from glaiveai/Glaive-function-calling-v2 into the ChatML message format. The data likely contains structured conversations where an AI assistant is prompted to use specific functions, such as creating a contact. Created by Locutusque and last updated on Hugging Face on July 16, 2024.

Use Cases

• Fine-tuning language models to follow system prompts based on the described ChatML format.
• Training models to generate structured function calls based on natural language user requests.
• Benchmarking model performance on tasks requiring interaction with external APIs or tools.
• Studying conversation structure and system instruction adherence in synthetic dialogue data.

Strengths

• Data is structured in the standardized ChatML message format, which is commonly used for training conversational models.
• The dataset is derived from an established source (Glaive-function-calling-v2), suggesting a specific focus on function-calling tasks.

Limitations

• Column-level documentation and sample data are unavailable, making field semantics and data structure difficult to assess prior to download.
• Row count and total size are unknown, which limits suitability assessment for large-scale training.
• The description is brief; data quality, source details, and conversion methodology require manual inspection.

Provenance

Source

Converted from glaiveai/Glaive-function-calling-v2.

Collection Method

Conversion to ChatML format; original data generation method is not specified.

Freshness

Last updated 2024-07-16 11:50:45.