google / paxml

Pax is a Jax-based machine learning framework for training large scale models. Pax allows for advanced and fully configurable experimentation and parallelization, and has demonstrated industry leading model flop utilization rates.
Apache License 2.0
458 stars 69 forks source link

Installing paxml from source failed due to dependency problem #25

Closed yhtang closed 1 year ago

yhtang commented 1 year ago

This is using the latest development version (see full log here.):

#8 79.88 + git clone https://github.com/google/paxml.git /opt/paxml
#8 79.89 Cloning into '/opt/paxml'...
#8 80.87 + pushd /opt/paxml
#8 80.87 + git checkout HEAD
#8 80.87 /opt/paxml /
#8 80.89 Your branch is up to date with 'origin/main'.
#8 80.89 + pip install -e '.[gpu]'
...
#8 94.75 ERROR: Cannot install paxml and paxml[gpu]==1.0.0 because these package versions have conflicting dependencies.
#8 94.75 
#8 94.75 The conflict is caused by:
#8 94.75     paxml[gpu] 1.0.0 depends on tensorflow~=2.9.2
#8 94.75     tensorflow-text 2.9.0 depends on tensorflow<2.10 and >=2.9.0; platform_machine != "arm64" or platform_system != "Darwin"
#8 94.75     lingvo 0.12.1 depends on tensorflow==2.9
yhtang commented 1 year ago

adding @ashors1 @nouiz @mjsML @terrykong for viz

A9isha commented 1 year ago

Thanks @yhtang for opening this issue.

We are working on a fix. For immediately unblocking you, could you please use the commit "bb5029a1729aec3d94474550253a0343b94a3391" which is from a week back on May 15 in case you don't need a feature from last week

phoenix-meadowlark commented 1 year ago

An alternative workaround is to set seqio==0.0.15 in requirements.in.

yhtang commented 1 year ago

@A9isha @phoenix-meadowlark Thanks for the tips!

A9isha commented 1 year ago

The issue has been resolved now. You should be able to install from head now. Please let us know if you are hitting any issue still.