search

hendrycks / outlier-exposure

Deep Anomaly Detection with Outlier Exposure (ICLR 2019)
Apache License 2.0
541 stars 107 forks source link

80 Million Tiny Images subset for download? #7

Closed robintibor closed 4 years ago

robintibor commented 4 years ago

Thanks for sharing the code. I wonder if you can somehow make available a useful subset of 80 million tiny images for faster experimentation? Since you write in the paper "In fact, experiments in this paper often used around 1% of the images in the 80 Million Tiny Images dataset since we only briefly fine-tuned the models."

And then one would not have to download >200GB :) Would be great.

hendrycks commented 4 years ago

One way is to simply download the dataset, and take the first 50k images (select the first 3072*50000 bytes).

On Wed, Nov 6, 2019 at 1:44 AM robintibor notifications@github.com wrote:

Thanks for sharing the code. I wonder if you can somehow make available a useful subset of 80 million tiny images for faster experimentation? Since you write in the paper "In fact, experiments in this paper often used around 1% of the images in the 80 Million Tiny Images dataset since we only briefly fine-tuned the models."

And then one would not have to download >200GB :) Would be great.

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/hendrycks/outlier-exposure/issues/7?email_source=notifications&email_token=ACZBITVO6YGGTD4YWZND3NTQSKGX7A5CNFSM4JJSIDCKYY3PNVWWK3TUL52HS4DFUVEXG43VMWVGG33NNVSW45C7NFSM4HXFOV5A, or unsubscribe https://github.com/notifications/unsubscribe-auth/ACZBITWRKJSTOST37IIFIJTQSKGX7ANCNFSM4JJSIDCA .