Beryex/APIGen-MT-5k-sharegpt

Dataset Card for APIGen-MT-5k-sharegpt

This dataset is the sharegpt format of the original Salesforce/APIGen-MT-5k dataset. It is primarily designed for fine-tuning large language models (LLMs) for function calling and multi-turn conversations.

Dataset Description

The original Salesforce/APIGen-MT-5k dataset contains conversations between users and a language model, focusing on API usage and tool invocation scenarios. We have converted this dataset into the ShareGPT format, which is a common and widely adopted format for training conversational AI models.

The conversion aims to facilitate the training of instruction-tuned LLMs that can handle multi-turn dialogues and execute function calls based on user prompts.

Dataset Structure

The dataset consists of a single train split. Each example in the dataset has the following fields:

• —conversations (list[dict]): A list of dictionary objects representing a multi-turn conversation. Each dictionary has:
• —from (string): Indicates the speaker, typically "human" for user turns or "gpt" for model responses.
• —value (string): The actual content of the message. This can be a natural language query, a function call string (e.g., in JSON format), or a function output.
• —tools (string): This field contains the definitions of available tools in a structured format (e.g., JSON schema) that the model can potentially call. The content and format depend on the original dataset's tool definitions.
• —system (string): This field contains initial system instructions or context provided to the model at the beginning of a conversation.

How to Load

You can easily load this dataset using the Hugging Face datasets library:

from datasets import load_dataset

dataset = load_dataset("Beryex/APIGen-MT-5k-sharegpt")

print(dataset['train'][0])