vaerdu

> 你好，是要先Compile后才能test是么？我在CUDA安装好后进行Compile操作即运行第一个setup.py时报错 OSError: The nvcc binary could not be located in your $PATH. Either add it to your path, or set $CUDAHOME 我也遇到了同样的问题

> @wuxiaolianggit Have you successful in changing cuda8.0 to cuda9.0? What should I do to make that change? 你好，请问你用了哪个版本的CUDA成功跑通了程序？

您好，用的python3.7 torch1.6 已经调试好了，谢谢大佬的回复。 At 2021-10-14 17:19:45, "yuguangtongxing" ***@***.***> wrote: @vaerdu 请问您是在什么环境下运行的？ Python和PyTorch 版本分别是多少啊 — You are receiving this because you were mentioned. Reply to this email directly, view it on...

好像是import方式不对，要么就直接把要导入的文件添加到环境里，百度可以搜一搜解决办法 在 2021-12-29 22:26:47，"KeqingWu" ***@***.***> 写道： 请问是如何解决的呢，我也遇到了同样的问题 请问您解决了吗？ — Reply to this email directly, view it on GitHub, or unsubscribe. Triage notifications on the go with GitHub Mobile for iOS...

> Could you provide more info about it? e.g. : How do you like to visualize feature map? fmap saved in Tensorboard or Raw image or something more like this....

> @snatch59 I want to apply the cnn-svm-classifier to only extract features and saved to disk as matrix or excel sheet then feed it to another classifier than svm. My...

> 测试发现训练速度瓶颈主要在simOTA，其中的for loop一张一张处理图像无法充分发挥GPU并行计算的优势。有同学(miemie2013)已经实现了并行版simOTA，去掉了for loop，一个batch的数据都可以并行处理，实测训练速度提升了几倍，建议官方采用： https://github.com/miemie2013/miemiedetection/blob/main/mmdet/models/heads/yolox_head_fast.py 直接替换原来的yolo_head.py再简单改一下import就可以使用。 该head.py支持VOC格式的数据集吗？