# orbit.utils.SummaryManagerInterface

<br />

|---|
| [![](https://www.tensorflow.org/images/GitHub-Mark-32px.png) View source on GitHub](https://github.com/tensorflow/models/blob/v2.20.0/orbit/utils/summary_manager_interface.py#L20-L64) |

A utility interface for managing summary writing.

## Methods

### `flush`

[View source](https://github.com/tensorflow/models/blob/v2.20.0/orbit/utils/summary_manager_interface.py#L23-L26)

    @abc.abstractmethod
    flush()

Flushes the the recorded summaries.

### `summary_writer`

[View source](https://github.com/tensorflow/models/blob/v2.20.0/orbit/utils/summary_manager_interface.py#L28-L31)

    @abc.abstractmethod
    summary_writer(
        relative_path=''
    )

Returns the underlying summary writer for scoped writers.

### `write_summaries`

[View source](https://github.com/tensorflow/models/blob/v2.20.0/orbit/utils/summary_manager_interface.py#L33-L64)

    @abc.abstractmethod
    write_summaries(
        summary_dict
    )

Writes summaries for the given dictionary of values.

The summary_dict can be any nested dict. The SummaryManager should
recursively creates summaries, yielding a hierarchy of summaries which will
then be reflected in the corresponding UIs.

For example, users may evaluate on multiple datasets and return
`summary_dict` as a nested dictionary:

    {
        "dataset1": {
            "loss": loss1,
            "accuracy": accuracy1
        },
        "dataset2": {
            "loss": loss2,
            "accuracy": accuracy2
        },
    }

This will create two set of summaries, "dataset1" and "dataset2". Each
summary dict will contain summaries including both "loss" and "accuracy".

<br />

<br />

<br />

| Args ||
|---|---|
| `summary_dict` | A dictionary of values. If any value in `summary_dict` is itself a dictionary, then the function will create a new summary_dict with name given by the corresponding key. This is performed recursively. Leaf values are then summarized using the parent relative path. |

<br />