langchain_experimental.rl_chain.metrics.MetricsTrackerRollingWindow

class langchain_experimental.rl_chain.metrics.MetricsTrackerRollingWindow(window_size: int, step: int)[source]

指标跟踪滚动窗口。

Attributes

score

Methods

__init__(window_size, step)

on_decision()

on_feedback(value)

to_pandas()

Parameters

  • window_size (int) –

  • step (int) –

__init__(window_size: int, step: int)[source]
Parameters

  • window_size (int) –

  • step (int) –

on_decision() → None[source]
Return type

None

on_feedback(value: float) → None[source]
Parameters

value (float) –

Return type

None

to_pandas() → pd.DataFrame[source]
Return type

pd.DataFrame