Skip to content

🧪 ImageDataset

A Dataset for ImageLists as input and ImageLists, Tables or Columns as output.

Parent type: Dataset<ImageList, O>

Parameters:

Name Type Description Default
inputData ImageList The input ImageList -
outputData O The output data -
batchSize Int The batch size used for training 1
shuffle Boolean Whether the data should be shuffled after each epoch of training false

Type parameters:

Name Upper Bound Description Default
O Any? - -

Examples:

pipeline example {
    val image = Image.fromFile("example.png");
    val imageList = ImageList.fromImages([image]);
    val labels = Column("label", ["example"]);
    val dataset = ImageDataset(imageList, labels);
}
Stub code in ImageDataset.sdsstub 

 24
 25
 26
 27
 28
 29
 30
 31
 32
 33
 34
 35
 36
 37
 38
 39
 40
 41
 42
 43
 44
 45
 46
 47
 48
 49
 50
 51
 52
 53
 54
 55
 56
 57
 58
 59
 60
 61
 62
 63
 64
 65
 66
 67
 68
 69
 70
 71
 72
 73
 74
 75
 76
 77
 78
 79
 80
 81
 82
 83
 84
 85
 86
 87
 88
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
@Experimental
class ImageDataset<out O>(
    @PythonName("input_data") inputData: ImageList,
    @PythonName("output_data") outputData: O,
    @PythonName("batch_size") batchSize: Int = 1,
    shuffle: Boolean = false
) sub Dataset<ImageList, O> {
    /**
     * Get the input `ImageSize` of this dataset.
     */
    @PythonName("input_size") attr inputSize: ImageSize
    /**
     * Get the output size of this dataset.
     */
    @PythonName("output_size") attr outputSize: ImageSize

    /**
     * Get the input data of this dataset.
     *
     * @result input the input data of this dataset
     *
     * @example
     * pipeline example {
     *     val image = Image.fromFile("example.png");
     *     val imageList = ImageList.fromImages([image]);
     *     val labels = Column("label", ["example"]);
     *     val dataset = ImageDataset(imageList, labels);
     *     val input = dataset.getInput();
     * }
     */
    @Pure
    @PythonName("get_input")
    fun getInput() -> input: ImageList

    /**
     * Get the output data of this dataset.
     *
     * @result output the output data of this dataset
     *
     * @example
     * pipeline example {
     *     val image = Image.fromFile("example.png");
     *     val imageList = ImageList.fromImages([image]);
     *     val labels = Column("label", ["example"]);
     *     val dataset = ImageDataset(imageList, labels);
     *     val output = dataset.getOutput();
     * }
     */
    @Pure
    @PythonName("get_output")
    fun getOutput() -> output: O

    /**
     * Return a new `ImageDataset` with shuffled data.
     *
     * The original dataset is not modified.
     *
     * @result imageDataset the shuffled `ImageDataset`
     *
     * @example
     * pipeline example {
     *     val image = Image.fromFile("example.png");
     *     val imageList = ImageList.fromImages([image]);
     *     val labels = Column("label", ["example"]);
     *     val dataset = ImageDataset(imageList, labels);
     *     val shuffledDataset = dataset.shuffle();
     * }
     */
    @Pure
    fun shuffle() -> imageDataset: ImageDataset<O>

    /**
     * Create two image datasets by splitting the data of the current dataset.
     *
     * The first dataset contains a percentage of the data specified by `percentage_in_first`, and the second dataset
     * contains the remaining data. By default, the data is shuffled before splitting. You can disable this by setting
     * `shuffle` to False.
     *
     * The original dataset is not modified.
     *
     * @param percentageInFirst The percentage of data to include in the first dataset. Must be between 0 and 1.
     * @param shuffle Whether to shuffle the data before splitting.
     *
     * @result firstDataset The first dataset.
     * @result secondDataset The second dataset.
     */
    @Pure
    fun split(
        @PythonName("percentage_in_first") percentageInFirst: Float,
        shuffle: Boolean = true
    ) -> (firstDataset: ImageDataset<O>, secondDataset: ImageDataset<O>)
}

inputSize

Get the input ImageSize of this dataset.

Type: ImageSize

outputSize

Get the output size of this dataset.

Type: ImageSize

getInput

Get the input data of this dataset.

Results:

Name Type Description
input ImageList the input data of this dataset

Examples:

pipeline example {
    val image = Image.fromFile("example.png");
    val imageList = ImageList.fromImages([image]);
    val labels = Column("label", ["example"]);
    val dataset = ImageDataset(imageList, labels);
    val input = dataset.getInput();
}
Stub code in ImageDataset.sdsstub 

54
55
56
@Pure
@PythonName("get_input")
fun getInput() -> input: ImageList

getOutput

Get the output data of this dataset.

Results:

Name Type Description
output O the output data of this dataset

Examples:

pipeline example {
    val image = Image.fromFile("example.png");
    val imageList = ImageList.fromImages([image]);
    val labels = Column("label", ["example"]);
    val dataset = ImageDataset(imageList, labels);
    val output = dataset.getOutput();
}
Stub code in ImageDataset.sdsstub 

72
73
74
@Pure
@PythonName("get_output")
fun getOutput() -> output: O

shuffle

Return a new ImageDataset with shuffled data.

The original dataset is not modified.

Results:

Name Type Description
imageDataset ImageDataset<O> the shuffled ImageDataset

Examples:

pipeline example {
    val image = Image.fromFile("example.png");
    val imageList = ImageList.fromImages([image]);
    val labels = Column("label", ["example"]);
    val dataset = ImageDataset(imageList, labels);
    val shuffledDataset = dataset.shuffle();
}
Stub code in ImageDataset.sdsstub 

92
93
@Pure
fun shuffle() -> imageDataset: ImageDataset<O>

split

Create two image datasets by splitting the data of the current dataset.

The first dataset contains a percentage of the data specified by percentage_in_first, and the second dataset contains the remaining data. By default, the data is shuffled before splitting. You can disable this by setting shuffle to False.

The original dataset is not modified.

Parameters:

Name Type Description Default
percentageInFirst Float The percentage of data to include in the first dataset. Must be between 0 and 1. -
shuffle Boolean Whether to shuffle the data before splitting. true

Results:

Name Type Description
firstDataset ImageDataset<O> The first dataset.
secondDataset ImageDataset<O> The second dataset.
Stub code in ImageDataset.sdsstub 

110
111
112
113
114
@Pure
fun split(
    @PythonName("percentage_in_first") percentageInFirst: Float,
    shuffle: Boolean = true
) -> (firstDataset: ImageDataset<O>, secondDataset: ImageDataset<O>)