Closed bosmont closed 7 years ago
Thanks for the bug report @bosmont , and apologies for the late response. I just merged a PR that pins the keras version to a version we support (1.1.0). Can you try this now, after reinstalling MHCflurry and also setting KERAS_BACKEND=theano
?
In the latest release (just pushed) we also now support tensorflow. Closing for now.
hi,
I installed mhcflurry by following the instructions in the README.md:
pip install mhcflurry
mhcflurry-downloads fetch
when running
nosetests .
I got following error:
The similar error when running prediction command:
mhcflurry-predict --alleles HLA-A0201 HLA-A0301 --peptides SIINFEKL SIINFEKD SIINFEKQ
Using TensorFlow backend. Predicting for 2 alleles and 3 peptides = 6 predictions /home/ghost/anaconda2/lib/python2.7/site-packages/tensorflow/python/ops/gradients_impl.py:91: UserWarning: Converting sparse IndexedSlices to a dense Tensor of unknown shape. This may consume a large amount of memory. "Converting sparse IndexedSlices to a dense Tensor of unknown shape. " allele,peptide,mhcflurry_prediction HLA-A0201,SIINFEKL,10672.3476562 HLA-A0201,SIINFEKD,26042.7167969 HLA-A0201,SIINFEKQ,26375.7949219 HLA-A0301,SIINFEKL,25532.703125 HLA-A0301,SIINFEKD,24997.8769531 HLA-A0301,SIINFEKQ,28262.828125 Exception AttributeError: "'NoneType' object has no attribute 'TF_DeleteStatus'" in <bound method Session.del of <tensorflow.python.client.session.Session object at 0x7f2ad9c57090>> ignored
I googled around, but did not find any obvious solution. Any help will be very much appreciated.