pytorch/examples

add support for mock data input to imagenet example

Open

#194 建立於 2017年8月7日

在 GitHub 查看
 (1 留言) (0 反應) (0 負責人)Python (21,634 star) (9,429 fork)batch import
distributedgood first issuetriaged

描述

in order to test the performance conveniently, especially distributed training, it would be great to add an option to allow mock data input, which means directly random some data in program as input.

貢獻者指南

技術棧
python
領域
machine learningtestingperformance
議題類型
feature
難度面向新貢獻者的預計實作難度，1 表示很小改動，5 表示專家級工作。
3
預計時間有經驗貢獻者完成調查、實作、測試並準備 pull request 的粗略時間範圍。
half day
活動狀態議題目前的可參與程度：新鮮、活躍、陳舊、阻塞或等待維護者輸入。
stale
清晰度議題是否清楚說明預期改動、驗收標準和下一步。
clear
前置要求
Basic PythonBasic PyTorchUnderstand imagenet example
新手友善度1-100 的估計分數，表示該議題對首次貢獻者的友善程度。
70
研究方向
Examine the imagenet example code, likely in `main.py`, to locate the data loading section. Add an option (e.g., ` mock data`) that generates random tensors as input instead of loading from disk. Ensure the mock data is consistent across distributed processes by using the same random seed. Review the single comment on the issue (if any) for additional context or requirements. Test the changes with a dummy run to confirm functionality.