Reputation: 2411
I am new to using Anaconda and as my hard drive is filling up I realized there are multiple (large) files that exist as duplicates in the following directories:
/anaconda3/pkgs/.../lib/<file>
/anaconda3/lib/<file>
Example of a file that is duplicated is libmklml.dylib
. In addition there are duplicate files such as:
tensorflow-base-1.13.1-mkl_py36hc36dc97_0.tar.bz2
tensorflow-base-1.13.1-mkl_py37h66b1bf0_0.tar.bz2
in the same directory.
My guess is that the /pkgs/ folders is the file as it was downloaded and the anaconda/lib/ directory is where the libraries are being executed. I am puzzled though why there are duplicate versions of each file.
Can I remove any of them, and if so, how? I tried conda clean -p
(as well as conda clean -a
) which seems to have removed about 1.5 Gb of files but not these type of duplicates.
Upvotes: 2
Views: 2299
Reputation: 31
On linux there is a tool fslint-gui, that finds duplicate files and can then replace the duplicates with a symlink.
Not sure how stable that is when certain packages are updated.
Upvotes: 0
Reputation: 148
conda clean -a
is enough for routinely cleanup. Anaconda keep a separated copy for installed package when you do a update. You probably can safely delete all the backups in /anaconda3/pkgs/
, but it is maybe more wisely leave it to conda
deal with.
P.S. you can try miniconda
which have only a few default packages.
Upvotes: 6