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
2.08k stars 373 forks source link

aimet common didn't release？ #3147

Open 1826133674 opened 2 months ago

1826133674 commented 2 months ago

Hello, may I ask why the latest version of AIMet common has not been released together with other wheel packages? The most recent available AIMet common version is 1.30.0, which is incompatible with the Python 3.10 environment.

e-said commented 2 months ago

Hello The installation instructions have changed slightly. Did you check the latest doc?

1826133674 commented 2 months ago

Hello The installation instructions have changed slightly. Did you check the latest doc?

Yes，I have checked it。I just want to know how I should handle such an error. File "exceptions.py", line 30, in from aimet_torch.qc_quantize_op import QcQuantizeWrapper File "aimet_py310/lib/python3.10/site-packages/aimet_torch/qc_quantize_op.py", line 49, in import aimet_common.libpymo as libpymo ImportError: liblapacke.so.3: cannot open shared object file: No such file or directory

1826133674 commented 2 months ago

Hello The installation instructions have changed slightly. Did you check the latest doc?

Yes，I have checked it。I just want to know how I should handle such an error. File "exceptions.py", line 30, in from aimet_torch.qc_quantize_op import QcQuantizeWrapper File "aimet_py310/lib/python3.10/site-packages/aimet_torch/qc_quantize_op.py", line 49, in import aimet_common.libpymo as libpymo ImportError: liblapacke.so.3: cannot open shared object file: No such file or directory

This error occurred while I was attempting to reproduce the high-precision quantization of the SD large model demo tutorial using the latest version of aimet-torch-gpu-pt21-1.32.0+cu118. Here is the link to the tutorial: https://docs.qualcomm.com/bundle/publicresource/topics/80-64748-1/model_updates.html

e-said commented 1 month ago

I successfully imported the library you mentioned without any issues.

I'm installing AIMET on top of a custom docker image. If it helps, you can find below the shell script I created to install AIMET in my development environment.

#!/bin/bash
# Script to build aimet lib

export AIMET_VARIANT="aimet_torch_gpu_pt21"
export release_tag=1.32.0
export download_url="https://github.com/quic/aimet/releases/download/${release_tag}"
export wheel_file_suffix="cu118-cp310-cp310-manylinux_2_34_x86_64.whl"

# TODO aliasing to aimet python version
# /usr/bin/python3.10 -m pip install ${download_url}/AimetCommon-${AIMET_VARIANT}_${release_tag}-${wheel_file_suffix}
/usr/bin/python3.10 -m pip install ${download_url}/${AIMET_VARIANT}-${release_tag}.${wheel_file_suffix} -f https://download.pytorch.org/whl/torch_stable.html
# /usr/bin/python3.10 -m pip install ${download_url}/Aimet-${AIMET_VARIANT}_${release_tag}-${wheel_file_suffix}

# Install common debian packages:
cat /usr/local/lib/python3.10/dist-packages/aimet_torch/bin/reqs_deb_common.txt | xargs apt-get --assume-yes install
# TODO: raise ticket to aimet, reqs_deb_torch_common does not exist
cat /usr/local/lib/python3.10/dist-packages/aimet_torch/bin/reqs_deb_torch_common.txt | xargs apt-get --assume-yes install
cat /usr/local/lib/python3.10/dist-packages/aimet_torch/bin/reqs_deb_torch_gpu.txt | xargs apt-get --assume-yes install
ln -s /usr/lib/x86_64-linux-gnu/libjpeg.so /usr/lib
escorciav commented 1 month ago

@e-said could you pls point to me the comment that I should take a look & provide a sentence for context? I dunno how it applies to my particular case-