search

quic / aimet

AIMET is a library that provides advanced quantization and compression techniques for trained neural network models.
https://quic.github.io/aimet-pages/index.html
Other
2.16k stars 384 forks source link

I cannot install aimet-tf-cpu on ubuntu 20.04! #3200

Open miladdona opened 4 months ago

miladdona commented 4 months ago

Hello,

I want to use AIMET with Tensorflow on a system (server) with Ubuntu 20.04. I followed the instruction here. Based on this link I can setup the environment using the pre-built docker images using the instructions in this link. I prefer to use GPU as well but the link mentions, I need to have nvidia-docker to use GPU and the instructions to install nvidia-docker is in this link, but in this link they mentioned that the nvidia-docker wrapper is no longer supported. So I decided to setup the AIMET first with CPU and after that fix the problem with GPU support inside of the docker. I followed the instructions here and set up the docker using AIMET_VARIANT="tf-cpu" . after that inside of the docker I built the AIMET code based on the instructions in this link. After a successful build I tried to install the package using this instructions 

cd $WORKSPACE/aimet/build
make install

but I got this error:

../../../Examples/Examples
CMake Error: Unable to read from file '../../../Examples/Examples': Can't lstat ../../../Examples/Examples
CMake Error: Problem creating tar: /data/mkokhazadeh/aimet/aimet/build/staging/universal/Examples.tar.gz
-- Up-to-date: /data/mkokhazadeh/aimet/aimet/build/staging/universal/Docs

Could you please help me in this to install it? And I will be thankful if you guide me to install tf-gpu as well!

Thanks.

k1m0io commented 4 months ago

Hey I got the same error but I ignored it since it is only related to Examples. I managed to build it from source and install it and it works fine for me. Do you really need that examples folder?

miladdona commented 4 months ago

Thanks for the answer. Actually I don't need examples folder, I want to sets my models. How did you manage to build it from source and install it? Could you please let me know?

Also did you install it for CPU only or GPU?

Thanks