scSemiProfiler.utils.loss_curve

scSemiProfiler.utils.loss_curve(name, reprepid=None, tgtpid=None, stage=1, save=None)[source]

Visualize the training loss curves

Parameters

  • name (str) – Project name

  • reprepid (typing.Optional[int]) – Representative sanmple ID

  • tgtpid (typing.Optional[int]) – target sample IDs

  • stage (int) – The training stage to visualize, 1: pretrain1; 2: pretrain2; 3: inference

  • save – Path within the ‘figures’ folder to save the plot

Return type

None

Returns

None

Example

>>> name = 'project name'
>>> loss_curve(name, reprepid='BGCV09_CV0279',tgtpid=None,stage=1) # or loss_curve(name, sids, reprepid=6,tgtpid=None,stage=1)