Where to put the training or testing dataset?

[JiajiaLi04]

Dear Author,

I wonder where to put the training and testing dataset? Are there any rules to name them. It is weired that you did not have the parameters in your parameter lines. Do I need to change the dataset name in the code? Hoping for your reply. Thanks!

[wmrenr]

Did you solve this problem please?

[wondervictor]

Hi @JiajiaLi04 @wangxueping78, I'm much sorry for the late reply!
Preparing the dataset is simple, and you can refer to Dataset for Detectron2 for more details.
Exactly, if you MS-COCO dataset, you just download the dataset from the official site and unzip the files.
Then you can enter into the directory of this repo by:

cd BMaskR-CNN
cd projects/BMaskR-CNN

setup a datasets folder, or use a soft link if you store the data in other place,

mkdir datasets
cd datasets
mkdir coco

then put the COCO dataset into coco directory, and then the directory tree of projects/BMaskR-CNN is like:

datasets
| -- coco
      | -- annotations
      | -- val2017
      | -- train2017

If you have any question, I'll reply as soon as possible.

[wmrenr]

Hi @JiajiaLi04 @wangxueping78, I'm much sorry for the late reply! Preparing the dataset is simple, and you can refer to Dataset for Detectron2 for more details. Exactly, if you MS-COCO dataset, you just download the dataset from the official site and unzip the files. Then you can enter into the directory of this repo by:

cd BMaskR-CNN
cd projects/BMaskR-CNN

setup a datasets folder, or use a soft link if you store the data in other place,

mkdir datasets
cd datasets
mkdir coco

then put the COCO dataset into coco directory, and then the directory tree of projects/BMaskR-CNN is like:

datasets
| -- coco
      | -- annotations
      | -- val2017
      | -- train2017

If you have any question, I'll reply as soon as possible.
Oh thank you so much!
Excuse me,should the instances_train2017.json and instances_val2017.json be placed in the annotations folder?

[wondervictor]

@wangxueping78 yep, the instances_train2017.json and instances_val2017.json should be placed into the annotations directory, that is,

datasets
| -- coco
      | -- annotations
            | -- instances_train2017.json
            | -- instances_val2017.json
      | -- val2017
      | -- train2017

[wmrenr]

oh，I get it！Thanks 发自我的iPhone

…

------------------ Original ------------------ From: Tianheng Cheng ***@***.***> Date: Thu,Mar 31,2022 11:12 PM To: hustvl/BMaskR-CNN ***@***.***> Cc: wangxueping78 ***@***.***>, Mention ***@***.***> Subject: Re: [hustvl/BMaskR-CNN] Where to put the training or testing dataset? (#22) @wangxueping78 yep, the instances_train2017.json and instances_val2017.json should be placed into the annotations directory, that is, datasets | -- coco | -- annotations | -- instances_train2017.json | -- instances_val2017.json | -- val2017 | -- train2017 — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you were mentioned.Message ID: ***@***.***>