Using diferent dataset

Using diferent dataset

Hi,

 I am trying to use a different dataset for training.The dataset contains object as a root folder within that there are objects such as bus,cars etc which contains images.The Structure is Object folder is compressed within that we have sub folders as cars,buses containing all the images.When I am trying to upload the dataset the log gives me this error .

 2017-02-16 15:42:37.950 [http-nio-8080-exec-7] DEBUG c.i.a.d.c.r.f.BigFileUploadController - uploaded file is not valid, less than 2 directories

Zone: 

Thread Topic: 

Help Me
3 posts / 0 new
Last post
For more complete information about compiler optimizations, see our Optimization Notice.

Hi,

The DL-SDK expects the objects to be on the root of the archive.

Please move the "bus,cars,etc" to the root of the archive.

Regards,

Yevgeni.

Yes I was able to create the datasets but when creating a model I get error.Here I have a quick question the number of images in each subfolder should be many or can I do with few images.I have created my own model but at times I get accuracy of 62.5% and when I am trying to test there is request timeout.Yes when I am accumulating images all of them should be of fixed size? Can you point to other datasets where I can use the png files as used by MNIST dataset easily?

Leave a Comment

Please sign in to add a comment. Not a member? Join today