Hi Dieter and Carlos,
When using cuDNN, check the support matrix to ensure you have compatible NVIDIA HW, CUDA toolkit, and driver (https://docs.nvidia.com/deeplearning/sdk/cudnn-support-matrix/index.html). This can also be accessed at the download site for cuDNN. The CUDA toolkit is designed to be functionally backward compatible. For general situations, using the newest release of the CUDA toolkit is fine to do with existing code bases.
Let me know if there are any questions on this or other topics. Thank you.
-Jon
Jonathan Lefman
Developer Relations Manager - Healthcare and Life Sciences
NVIDIA
-----Original Message-----
From: Collaborative Computational Project in Electron cryo-Microscopy <[log in to unmask]> On Behalf Of Carlos Oscar S. Sorzano
Sent: Sunday, December 16, 2018 4:16 PM
To: [log in to unmask]
Subject: Re: [ccpem] Deep Consensus and Scipion/Xmipp
Dear Dieter,
DeepConsensus has been integrated into Scipion but it has not been included in the scipion current release yet. At some moment in the near future, the release of the packages will be different from the release of Scipion. The different packages will be plugins and you will be notified when new plugins or newer versions are available.
At the moment, in order to use it you, have to switch to co_tensor_new development branch. I am afraid that it cannot be executed in command line.
If you want to install scipion version with DeepConsensus integrated, you just need to use the following commands ( It will be almost automatic once the release will include DeepConsensus).
WARNING, in order to use GPU acceleration you will need CUDA-8.0 and
CUDNN-6
or CUDA-9.0 and CUDNN-7 (https://developer.nvidia.com/cudnn).
Basically, to install CUDNN you just have to uncompress and copy to CUDA path subfolds lib64 and lib the contained fields
#Install scipion and xmipp to try deepConsensus before release
Suppose we want to install scipion in /home/ruben/app/scipion and xmipp in /home/ruben/app/xmipp_bundle
#First install xmipp
cd /home/ruben/app/
mkdir xmipp_bundle
cd xmipp_bundle
wget https://raw.githubusercontent.com/I2PC/xmipp/devel/xmipp -O xmipp chmod a+x xmipp ./xmipp all N=8 br=co_tensor_new #Edit xmipp.conf. to use CUDA e.g. with gedit or nano nano xmipp.conf
CUDA=True
#Then, install scipion
cd /home/ruben/app/
git clone https://github.com/I2PC/scipion cd scipion git checkout co_tensor_new_new ./scipion config #Edit ./config/scipion.conf. to use CUDA e.g. with gedit or nano nano config.scipion.conf
CUDA_LIB = /home/ruben/app/cudas/cuda-8-CDNN_6/lib64
CUDA_BIN = /home/ruben/app/cudas/cuda-8-CDNN_6/bin
NVCC_INCLUDE = /home/ruben/app/cudas/cuda-8-CDNN_6/include
CUDA=True
#Tell scipion where xmipp is
./scipion installp -p /home/ruben/app/xmipp_bundle/src/scipion-em-xmipp
--devel --noBin
ln -s /home/ruben/app/xmipp_bundle/src/scipion-em-xmipp/xmipp3
/home/ruben/app/scipion/software/lib/python2.7/site-packages/
#Edit ~/.config/scipion/scipion.conf to tell where xmipp binaries are nano -w ~/.config/scipion/scipion.conf
XMIPP_HOME = /home/ruben/app/xmipp_bundle/build
#Finally, install deepConsensus
./scipion installb deepLearnigToolkit
Now you can create scipion project by
./scipion
The GUI will be launched
In order to use deepConsensus, import the micrographs, the particles, possibly pick with another picker and fill the protocol form for deepConsensus.
To find it easily you can use ctrl+f
and type DeepConsensus
If you have any further problem, please do not hesitate to tell us.
Kind regards, Ruben and Carlos Oscar
El 16/12/2018 a las 8:26, Dieter Blaas escribió:
> Dear Carlos and all,
>
> I was wondering how to use DeepConsensus for the elimination of
> already picked 'bad' virus particles instead of cleaning the data set
> (in relion) with multiple Class3D runs and cutting off
> _rlnLogLikeliContribution outlayers. Since the paper on Deep Consensus
> states that the program had been integrated into Scipion I upgraded my
> installation following the description
> (https://github.com/I2PC/scipion/wiki/How-to-Install). However, it
> still reports "version-1.1.beta". Even after a completely new install
> the version did not change. Can anybody please indicate whether it is
> possible to run DeepConsensus on a relion star file and if so how? I'd
> very much like to use it in the old-fashioned way from command line :-).
>
> Thanks, bw Dieter
>
> ----------------------------------------------------------------------
> --
> Dieter Blaas,
> Max F. Perutz Laboratories
> Medical University of Vienna
> Inst. Med. Biochem., Vienna Biocenter (VBC), Dr. Bohr Gasse 9/3,
> A-1030 Vienna, Austria,
> Tel: 0043 1 4277 61630,
> Fax: 0043 1 4277 9616,
> e-mail: [log in to unmask]
> ----------------------------------------------------------------------
> --
>
--
------------------------------------------------------------------------
Carlos Oscar Sánchez Sorzano e-mail: [log in to unmask]
Biocomputing unit http://i2pc.es/coss
National Center of Biotechnology (CSIC)
c/Darwin, 3
Campus Universidad Autónoma (Cantoblanco) Tlf: 34-91-585 4510
28049 MADRID (SPAIN) Fax: 34-91-585 4506
------------------------------------------------------------------------
########################################################################
To unsubscribe from the CCPEM list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=CCPEM&A=1
-----------------------------------------------------------------------------------
This email message is for the sole use of the intended recipient(s) and may contain
confidential information. Any unauthorized review, use, disclosure or distribution
is prohibited. If you are not the intended recipient, please contact the sender by
reply email and destroy all copies of the original message.
-----------------------------------------------------------------------------------
########################################################################
To unsubscribe from the CCPEM list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=CCPEM&A=1
|