pytorch/examples

Query on loss calculation in word language model

Open

#1,082 创建于 2022年10月14日

在 GitHub 查看
 (0 评论) (0 反应) (0 负责人)Python (21,634 star) (9,429 fork)batch import
help wanted

描述

In the main.py of word language model, I find that in the evaluate function the total_loss is getting multiplied by length of data https://github.com/pytorch/examples/blob/ca1bd9167f7216e087532160fc5b98643d53f87e/word_language_model/main.py#L163

However in the train function, total_loss is not getting multiplied by length of data https://github.com/pytorch/examples/blob/ca1bd9167f7216e087532160fc5b98643d53f87e/word_language_model/main.py#L194

Is this proper?

贡献者指南

技术栈
python
领域
machine learningdocumentation
议题类型
documentation
难度面向新贡献者的预计实现难度，1 表示很小改动，5 表示专家级工作。
2
预计时间有经验贡献者完成调查、实现、测试并准备 pull request 的粗略时间范围。
under 1 hour
活动状态议题当前的可参与程度：新鲜、活跃、陈旧、阻塞或等待维护者输入。
stale
清晰度议题是否清楚说明期望改动、验收标准和下一步。
clear
前置要求
basic pythonpytorch familiarity
新手友好度1-100 的估计分数，表示该议题对首次贡献者的友好程度。
80
研究方向
Examine the evaluate and train functions in word language model/main.py lines 163 and 194. The total loss is multiplied by len(data) in evaluate but not in train. Determine if this is intentional to average over the dataset or if it's an inconsistency. If a bug, propose a fix; otherwise add a code comment explaining the difference.