What does "ilsvrc12_train.txt" entail?

[BJTUJia]

Dear author, thanks a lot for your contribution! However, I can't find the file ilsvrc12_train.txt anywhere. I guess it's the training images list. I am new to deep learning and not familiar with Imagenet either. Can you kindly attach the files?

[topiaruss]

for a tar of tars of images: http://www.image-net.org/challenges/LSVRC/2012/nonpub-downloads Untar the tars, then untar those tars to get the images, record the imagenames.

Roughly this: After download you will have ISVRC2012_img_train.tar $ mkdir ISVRC2012_img_train $ cd ISVRC2012_img_train $ tar xvf ../ISVRC2012_img_train.tar $ find . -name *tar -exec tar xvf {} ; > ilsvrc12_train.txt

that may take a while. At the end you will have all the filenames in the .txt file.

You should then run produce_small_data.py. I got this working in python3 on the master branch of my fork. There is no need for a class label in the txt file, since we are doing self-supervised learning.

[chho-work]

Sir, thanks for sharing the code! I have the same question about "ilsvrc12_train.txt". I am not using the original Imanaget1K dataset, but a custom Imagenet dataset with fewer classes and images. So, I didn't download the tar files as you mentioned above. Suppose I have a train dir with 100K jpg image files, all in one single dir. The file names are similar to these: ILSVRC2012_val_00000907.JPEG ILSVRC2012_val_00002769.JPEG
n02086240_105.JPEG n02086240_126.JPEG etc...

Then, the "custom_ilsvrc12_train.txt" that I need to create should contain a list of 100K file names. Each file name should be listed in each line without coma or space. Is this correct? Your prompt response will be appreciated! Thanking you in advance!

[siyuan2018]

Sir, thanks for sharing the code! I have the same question about "ilsvrc12_train.txt". I am not using the original Imanaget1K dataset, but a custom Imagenet dataset with fewer classes and images. So, I didn't download the tar files as you mentioned above. Suppose I have a train dir with 100K jpg image files, all in one single dir. The file names are similar to these: ILSVRC2012_val_00000907.JPEG ILSVRC2012_val_00002769.JPEG n02086240_105.JPEG n02086240_126.JPEG etc...

Then, the "custom_ilsvrc12_train.txt" that I need to create should contain a list of 100K file names. Each file name should be listed in each line without coma or space. Is this correct? Your prompt response will be appreciated! Thanking you in advance!

Hi, may I ask how you solved the problem? I have the same question here. Thank you!

[Aiiiiii]

Dear author, thanks a lot for your contribution! However, I can't find the file ilsvrc12_train.txt anywhere. I guess it's the training images list. I am new to deep learning and not familiar with Imagenet either. Can you kindly attach the files?

Any update for this?