Flagging

Description

A Gradio Interface includes a ‘Flag’ button that appears underneath the output. By default, clicking on the Flag button sends the input and output data back to the machine where the gradio demo is running, and saves it to a CSV log file. But this default behavior can be changed. To set what happens when the Flag button is clicked, you pass an instance of a subclass of FlaggingCallback to the flagging_callback parameter in the Interface constructor. You can use one of the FlaggingCallback subclasses that are listed below, or you can create your own, which lets you do whatever you want with the data that is being flagged.

SimpleCSVLogger

gradio.SimpleCSVLogger(···)

Description

A simplified implementation of the FlaggingCallback abstract class provided for illustrative purposes. Each flagged sample (both the input and output data) is logged to a CSV file on the machine running the gradio app.

Example Usage

import gradio as gr
def image_classifier(inp):
    return {'cat': 0.3, 'dog': 0.7}
demo = gr.Interface(fn=image_classifier, inputs="image", outputs="label",
                    flagging_callback=SimpleCSVLogger())

CSVLogger

gradio.CSVLogger(···)

Description

The default implementation of the FlaggingCallback abstract class. Each flagged sample (both the input and output data) is logged to a CSV file with headers on the machine running the gradio app.

Example Usage

import gradio as gr
def image_classifier(inp):
    return {'cat': 0.3, 'dog': 0.7}
demo = gr.Interface(fn=image_classifier, inputs="image", outputs="label",
                    flagging_callback=CSVLogger())

Initialization

simplify_file_data: bool

default = True

Guides

Using Flagging

HuggingFaceDatasetSaver

gradio.HuggingFaceDatasetSaver(hf_token, dataset_name, ···)

Description

A callback that saves each flagged sample (both the input and output data) to a HuggingFace dataset.

Example Usage

import gradio as gr
hf_writer = gr.HuggingFaceDatasetSaver(HF_API_TOKEN, "image-classification-mistakes")
def image_classifier(inp):
    return {'cat': 0.3, 'dog': 0.7}
demo = gr.Interface(fn=image_classifier, inputs="image", outputs="label",
                    allow_flagging="manual", flagging_callback=hf_writer)

Initialization

hf_token: str

The HuggingFace token to use to create (and write the flagged sample to) the HuggingFace dataset (defaults to the registered one).

dataset_name: str

The repo_id of the dataset to save the data to, e.g. "image-classifier-1" or "username/image-classifier-1".

private: bool

default = False

Whether the dataset should be private (defaults to False).

info_filename: str

default = "dataset_info.json"

The name of the file to save the dataset info (defaults to "dataset_infos.json").

separate_dirs: bool

default = False

If True, each flagged item will be saved in a separate directory. This makes the flagging more robust to concurrent editing, but may be less convenient to use.

Guides

Using Flagging

←

Request

Themes

→