Skip to content

ForwardBackwardResult

Source code in docs_src/capo_sagemakertrainingsessionruntime/types/forward_backward_result.py 
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
class ForwardBackwardResult(TypedDict, closed=True):
    loss_fn_output_type: NotRequired["str"]
    """<p>The data type of the loss function output tensors.</p>"""
    loss_fn_outputs: NotRequired[
        "capo_sagemakertrainingsessionruntime.types.tensor_data_map_list.TensorDataMapList"
    ]
    """<p>The loss function output tensors, keyed by name.</p>"""
    metrics: NotRequired[
        "capo_sagemakertrainingsessionruntime.types.metrics_map.MetricsMap"
    ]
    """<p>The training metrics emitted by the engine, such as loss values.</p>"""
    token_usage: NotRequired[
        "capo_sagemakertrainingsessionruntime.types.training_token_usage.TrainingTokenUsage"
    ]
    """<p>The token usage statistics for metering.</p>"""

loss_fn_output_type instance-attribute 

loss_fn_output_type: NotRequired[str]

The data type of the loss function output tensors.

loss_fn_outputs instance-attribute 

loss_fn_outputs: NotRequired[TensorDataMapList]

The loss function output tensors, keyed by name.

metrics instance-attribute 

metrics: NotRequired[MetricsMap]

The training metrics emitted by the engine, such as loss values.

token_usage instance-attribute 

token_usage: NotRequired[TrainingTokenUsage]

The token usage statistics for metering.