biocore / mmvec

Neural networks for microbe-metabolite interaction analysis
BSD 3-Clause "New" or "Revised" License
118 stars 51 forks source link

numpy 1.17 gives error/warning #102

Closed marcomeola closed 4 years ago

marcomeola commented 4 years ago

Installed conda environment with conda create -n mmvec -c conda-forge mmvec and got that error when executing mmvec --help: miniconda3/envs/mmvec/lib/python3.7/site-packages/tensorflow/python/framework/dtypes.py:516: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'. _np_qint8 = np.dtype([("qint8", np.int8, 1)])

I solved it by downgrading to numpy 1.16.5 as suggested here: https://github.com/tensorflow/tensorflow/issues/30427

mortonjt commented 4 years ago

@marcomeola , if that is just a warning (and passes the unittests), it should be ok. Running the tutorials can serve as another sanity check (the biplots you generate should look similar to those in the tutorial).

Note that numpy 1.16.3 does have a bug in the SVD - not sure if this applies to 1.16.5 (see here). So this will mess up the results of the biplots.

FoxerLee commented 4 years ago

You may update tensorflow to 2.0.0 to solve it.

try tensorflow==2.0.0 and numpy==1.17.4

mortonjt commented 4 years ago

That will not work - Tensorflow 2.0.0 is not backwards compatible

On Mon, Dec 2, 2019, 10:28 AM Yuan Li notifications@github.com wrote:

You may update tensorflow to 2.0.0 to solve it.

try tensorflow==2.0.0 and numpy==1.17.4

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/biocore/mmvec/issues/102?email_source=notifications&email_token=AA75VXK747RQB2ZJSFZRLH3QWUSQ5A5CNFSM4JK4UBX2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEFT3OTA#issuecomment-560445260, or unsubscribe https://github.com/notifications/unsubscribe-auth/AA75VXL4K7JIQRN67BFWVKTQWUSQ5ANCNFSM4JK4UBXQ .

mortonjt commented 4 years ago

@marcomeola I'm assuming that this has been addressed, feel free to leave more details if it is still outstanding.