ImageSets folder

[MatteoT9890]

https://drive.google.com/drive/folders/1S5L-YmIiFMAKTs6nHMorB0Osz5iWI31k?usp=sharing

In this link there are only "Annotations" and "JPEGImages" folders, so when I run the command:

python tools/train_net.py --num-gpus 4 --dist-url='tcp://127.0.0.1:52125' --resume --config-file ./configs/OWOD/t1/t1_train.yaml SOLVER.IMS_PER_BATCH 8 SOLVER.BASE_LR 0.01 OUTPUT_DIR "./output/t1"

I face this error:

[...]

  File "/home/mtarantino/OWOD/detectron2/data/datasets/pascal_voc.py", line 81, in load_voc_instances
    with PathManager.open(os.path.join(dirname, "ImageSets", "Main", split + ".txt")) as f

FileNotFoundError: [Errno 2] No such file or directory: 'datasets/VOC2007/ImageSets/Main/t1_train.txt'

In fact in the folder datasets there is no subfolder called VOC2007/ImageSets/Main, while there is a subfolder called OWOD_imagesets containing the different TXT files.

So I wonder what is the strategy to handle correctly the different datasets once downloaded "Annotations" and "JPEGImages" from Google drive link in order to avoid FIleNotFoundError?

Thanks in advance

[MatteoT9890]

SOLVED:

Solution:

(1) Create folder datasets/VOC2007

(2) Put Annotations and JPEGImages inside datasets/VOC2007

(3) Create folder datasets/VOC2007/ImageSets/Main

(4) Put the content of datasets/OWOD_imagesets inside datasets/VOC2007/ImageSets/Main

I suggest you to put this flow inside the README.md!

Have a nice day!

[JosephKJ]

Thanks @MatteoT9890! I have updated the README with a link to your comment.

[lxtGH]

@JosephKJ Hi! Jose. For the lack of access to google drive, I cannot download the whole dataset. Hope for your response, thanks a lot. Could you please update the download link?

[PixelChen24]

@JosephKJ Hi! Jose. For the lack of access to google drive, I cannot download the whole dataset. Hope for your response, thanks a lot. Could you please update the download link?

兄弟，我已经下载了这个数据并且存在阿里云盘了，可惜阿里云盘无法共享压缩包文件，你要是真的要的话我可以扫码授权登录你下载

[XuHu0529]

@PixelChen24 Hi, I cannot download the whole dataset. Could you share me the files? Thanks a lot!

[chenfangchenf]

@PixelChen24 Hello, could you let me download the data? Thank you very much！

[PixelChen24]

@PixelChen24 Hello, could you let me download the data? Thank you very much！

Of course, please give me your email so that we can chat via email

[PixelChen24]

@PixelChen24 Hi, I cannot download the whole dataset. Could you share me the files? Thanks a lot!

Of course, please give me your email so that we can chat via email

[JosephKJ]

Hello @lxtGH @chenfangchenf @PixelChen24 , Please use this link instead: https://drive.google.com/drive/folders/1S5L-YmIiFMAKTs6nHMorB0Osz5iWI31k?usp=sharing

(I have updated the first post too)

[chenfangchenf]

thank you very much!

…

------------------ 原始邮件 ------------------ 发件人: "JosephKJ/OWOD" ***@***.***>; 发送时间: 2023年1月3日(星期二) 晚上9:39 ***@***.***>; ***@***.******@***.***>; 主题: Re: [JosephKJ/OWOD] ImageSets folder (#59) Hello, Please use this link instead: https://drive.google.com/drive/folders/1S5L-YmIiFMAKTs6nHMorB0Osz5iWI31k?usp=sharing (I have updated the first post too) — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you commented.Message ID: ***@***.***>

[chenfangchenf]

thank you very much ! My email is  ***@***.***

…

------------------ 原始邮件 ------------------ 发件人: "JosephKJ/OWOD" ***@***.***>; 发送时间: 2023年1月3日(星期二) 晚上9:35 ***@***.***>; ***@***.******@***.***>; 主题: Re: [JosephKJ/OWOD] ImageSets folder (#59) @PixelChen24 Hello, could you let me download the data? Thank you very much！ Of course, please give me your email so that we can chat via email — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you commented.Message ID: ***@***.***>

[PixelChen24]

@PixelChen24您好，我无法下载整个数据集。你能把文件分享给我吗？多谢！

Me too
I have written a blog on how to download huge datasets on Google Drive for Chinese developers or those who have trouble accessing Google, you can follow this

[HongdaChen]

Helloooooooooo friends from china mainland!
Here is a solution for downloading data effectively:
#122

[nikoubaibai]

Use”python tools/train_net.py --num-gpus=1 --config-file=./configs/OWOD/iOD/all_20_train.yaml“to it report an error