# tf.summary.scalar(name, tensor, collections=None)

### tf.summary.scalar(name, tensor, collections=None)

See the guide: Summary Operations > Generation of Summaries

Outputs a Summary protocol buffer containing a single scalar value.

The generated Summary has a Tensor.proto containing the input Tensor.

#### Args:

• name: A name for the generated node. Will also serve as the series name in TensorBoard.
• tensor: A real numeric Tensor containing a single value.
• collections: Optional list of graph collections keys. The new summary op is added to these collections. Defaults to [GraphKeys.SUMMARIES].

#### Returns:

A scalar Tensor of type string. Which contains a Summary protobuf.

#### Raises:

• ValueError: If tensor has the wrong shape or type.