Dataset format for FineTuning

Before pushing your dataset, you need to make sure that it is in the right format. The format that we will be using is called the Alpaca format. The dataset should be a json file containing the following columns:

• instruction : the instruction to execute

• input : the input of the instruction

• output : the output of the instruction

Here is an example of a dataset in the right format:

[
    {
        "instruction": "Assess the company's waste management practices.",
        "input": "How does your company manage hazardous and non-hazardous waste?",
        "output": "We segregate hazardous and non-hazardous waste at the source. Hazardous waste is handled by a licensed contractor, and we regularly audit their practices for compliance. Non-hazardous waste is recycled where possible."
    },
    {
        "instruction": "Evaluate the company's water usage efficiency.",
        "input": "What measures has your company implemented to reduce water consumption?",
        "output": "We have installed low-flow fixtures and automated sensors in our facilities. We also use a water recycling system in our manufacturing process to reduce overall water consumption."
    },
    {
        "instruction": "Check the company's energy efficiency initiatives.",
        "input": "What steps has your company taken to improve energy efficiency?",
        "output": "We've upgraded to energy-efficient lighting, HVAC systems, and machinery. Additionally, we conduct regular energy audits to identify further improvement areas."
    },
    {
        "instruction": "Review the company's air emissions control.",
        "input": "How does your company manage and monitor air emissions?",
        "output": "We have installed emission control systems like scrubbers and electrostatic precipitators. We conduct regular emissions testing and maintain records as per regulatory requirements."
    }
]