Giter Club home page Giter Club logo

Comments (3)

HuanzhiMao avatar HuanzhiMao commented on June 12, 2024 1

Hi @MayankAgarwal ,
Thanks for the feedback. We went over the index you provided. Most of them are already fixed in the April 1 release. There are a few that are still incorrect. We have addressed those in #309.

from gorilla.

HuanzhiMao avatar HuanzhiMao commented on June 12, 2024

Hi Mayank,

Thank you very much for putting up such a detailed issue report. We really appreciate your effort.

This is indeed an oversight on our end. The human_eval_answer is outdated and we are not using it in the evaluation process at all. (Apologies, we should mention this clearly). We only use the possible_answer as the ground truth. Thus, some human_eval_answer are wrong and misaligned with possible_answer, very sorry about the confusion caused. We have removed the human_eval_answer in our April 1 release.

from gorilla.

MayankAgarwal avatar MayankAgarwal commented on June 12, 2024

Hi @HuanzhiMao, thank you for your reply. I realized with yesterday's PR that the human_eval_answer fields were removed :)

Please note that some of the problems highlighted by the previous procedure still hold. For example, please refer to the simple function dataset index 25. The API specification mentions only the height parameter as required, while the possible_answer file marks the initial_velocity as required as well. Therefore, if the model follows the API specification and does not generate the initial_velocity parameter, it would still be incorrectly marked as failed. I checked and this is still the case with the latest version of the dataset.

{
    "question": "Calculate the final velocity of an object falling from a 150 meter building, assuming initial velocity is zero.",
    "function": [
        {
            "name": "calculate_final_velocity",
            "description": "Calculate the final velocity of a free falling object given the height it's dropped from, the initial velocity and acceleration due to gravity. Ignore air resistance.",
            "parameters": {
                "type": "object",
                "properties": {
                    "height": {
                        "type": "number",
                        "description": "The height the object is dropped from, in meters."
                    },
                    "initial_velocity": {
                        "type": "number",
                        "description": "The initial velocity of the object in m/s. Default is zero."
                    },
                    "gravity": {
                        "type": "number",
                        "description": "Acceleration due to gravity. Default value is approximately 9.81 m/s^2, earth's gravity."
                    }
                },
                "required": [
                    "height"
                ]
            }
        }
    ],
    "ground_truth": "calculate_final_velocity(height=150)",
    "result": "calculate_final_velocity(height=150)",
    "index": 25,
    "failure_categorization": [
        "MISSING_REQUIRED_PARAM"
    ],
    "potential_answer": {
        "calculate_final_velocity": {
            "height": [
                150
            ],
            "initial_velocity": [
                0
            ],
            "gravity": [
                9.81,
                ""
            ]
        }
    }
}

What I would like to do is to go through the indices mentioned in the original bug report, identify which problems still hold with the latest release, and reopen this bug with the updated indices, but please let me what would you prefer.

from gorilla.

Related Issues (20)

Recommend Projects

  • React photo React

    A declarative, efficient, and flexible JavaScript library for building user interfaces.

  • Vue.js photo Vue.js

    🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.

  • Typescript photo Typescript

    TypeScript is a superset of JavaScript that compiles to clean JavaScript output.

  • TensorFlow photo TensorFlow

    An Open Source Machine Learning Framework for Everyone

  • Django photo Django

    The Web framework for perfectionists with deadlines.

  • D3 photo D3

    Bring data to life with SVG, Canvas and HTML. 📊📈🎉

Recommend Topics

  • javascript

    JavaScript (JS) is a lightweight interpreted programming language with first-class functions.

  • web

    Some thing interesting about web. New door for the world.

  • server

    A server is a program made to process requests and deliver data to clients.

  • Machine learning

    Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.

  • Game

    Some thing interesting about game, make everyone happy.

Recommend Org

  • Facebook photo Facebook

    We are working to build community through open source technology. NB: members must have two-factor auth.

  • Microsoft photo Microsoft

    Open source projects and samples from Microsoft.

  • Google photo Google

    Google ❤️ Open Source for everyone.

  • D3 photo D3

    Data-Driven Documents codes.