aidotse/LeakPro

Loss in training loops

Aberta

#239 aberto em 24 de fev. de 2025

 (1 comentário) (0 reação) (0 responsável)Python (26 forks)auto 404
good first issue

Métricas do repositório

Stars
 (23 estrelas)
Métricas de merge de PR
 (Métricas PR pendentes)

Description

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

Guia do colaborador