aidotse/LeakPro

Loss in training loops

开放

#239 创建于 2025年2月24日

 (1 条评论) (0 个反应) (0 位负责人)Python (26 个派生)auto 404
good first issue

仓库指标

星标
 (23 个星标)
PR 合并指标
 (PR 指标待抓取)

描述

Issue

Problem Description

The last batch of each epoch is usually smaller than the others. One approach is to sum the losses and divide by the total number of samples.

Currently, loss calculations in handlers sum the average loss of each batch and divide by the length of the dataloader.

Expected Behavior

What Needs to be Done

How Can It Be Tested or Reproduced

贡献者指南