I have anaconda3 and cuda 9 installed on my linux machine (aws to be precise), when I check caffe2 installation as in the official documentation it shows me success message, but when I check the GPU with the following script python2 -c 'from caffe2.python import workspace; print(workspace.NumCudaDevices())' I see the following result Logging before InitGoogleLogging() is written to STDERR
E0729 20:58:01.995159 2016 common_gpu.cc:42] Found an unknown error - this may be due to an incorrectly set up environment, e.g. changing env variable CUDA_VISIBLE_DEVICES. What I'm I doing wrong.
Your problem translates an issue with your installation.
Usually this problem is solved by one of these actions:
Try restarting your computer (It never hurts to acutally try)
Fully reinstall NVIDIA Drivers + CUDA + CuDNN (check env vars)
Fully reinstall Caffe2 following the official documentation.
Debuging a faulty installation can be really cumbersome. If you are running on Linux, I would recommend you trying Docker + Nvidia driver:
Official Tutorial: https://caffe2.ai/docs/docker-setup.html
Nvidia Docker: https://github.com/NVIDIA/nvidia-docker
Related
I have installed Ubuntu-18.04 on my Windows 10 (Build 19041 which meets the requirements for WSL2) and am trying to convert this distro into WSL2. I have been following this guide.
To do so, I run the following command in Windows PowerShell:
wsl --set-version Ubuntu-18.04 2
But I get the following error after a few minutes:
This operation returned because the timeout period expired.
I have tried different networks (fearing it was a firewall issue) but to no avail. Any idea how to fix this?
Make sure you have the linux kernel update package installed. Then, run wsl --set-default-version 2 to set the default WSL version to WSL2.
When trying to configure a Remote Python Interpreter in Pycharm using Docker I get the following error:
com.github.dockerjava.api.excepion.DockerClientException: Enabled TLS
verification (DOCKER_TLS_VERIFY=1) but certificate path
(DOCKER_CERT_PATH) '/Users/me/.docker/machine/machines/default'
doesn't exist.
I've $export DOCKER_TLS_VERIFY=0 but with no difference.
I've manually created '/Users/me/.docker/machine/machines/default' but with no joy.
I've deinstalled and reinstalled both Docker and Pycharm but still get the same error.
I'm on a Mac OSX 10.12.1
Pycharm 2016.3
Full disclosure: I use vagrant and virtual box on my mac too.
Any pointers would be appreciated.
Kevin
I fixed it as follows:
Obtain socat (if not already installed)
e.g. brew install socat
Then: socat TCP-LISTEN:2375,reuseaddr,fork UNIX-CONNECT:/var/run/docker.sock
Then in pycharm/docker config:
Api URL: tcp://localhost:2375
Many thanks to the following found on:
https://youtrack.jetbrains.com/issue/IDEA-153973
If you have been using docker-machine for Mac to support PyCharm or the "socat" hack to access Docker API - It is no longer necessary with PyCharm EAP (https://www.jetbrains.com/pycharm/nextversion/) — You can set your remote debugger API_URL directly to: unix:/var/run/docker.sock (supporting debug mode, and the environment runs a lot faster, when not using VirtualBox).
There is a no-questions-asked 30 day trial of EAP before that update goes live (some time this month)
I am running PostgreSQL 9.6 on a Windows 7 laptop that is tightly managed by corporate IT. I do not have admin privileges on this laptop, but can do "many" things, including software installation, through elevated rights granted through BeyondTrust.
I use this database to import a number of medium-sized datasets from various sources and perform some in-depth analysis and reporting. These datasets have a crazy variation of encoding and formatting which creates challenges while importing them. For one dataset, I first import it into MySQL, strip offending characters from the data, export it back out to CSV, and finally import this cleaned-up data into PostgreSQL.
As an experiment, I installed PostgreSQL 9.6 and pg_loader on my home iMac running macOS 10.12, and it is able to import all of the datasets with no problem.
Given all this as a backdrop, I have a need to use pg_loader to import various datasets. However, the website does not offer either a Windows installer or instructions for compiling it on Windows.
Can anyone point me in the direction of one of the following, with the list going from most desirable to least desirable?
Windows installer of pg_loader
Windows installer of a comparable alternative to pg_loader
Detailed instructions for compiling pg_loader on Windows
Suggestions for using Python to clean up the data prior to import (caveat: I have very little experience with Python)
You can install the Windows Subsystem for Linux, install a Linux distro, then install pgloader with apt or other package manager -
$ sudo apt install pgloader
$ pgloader --version
pgloader version "3.4.1"
compiled with SBCL 1.3.3.debian
$ cd /mnt/c/Users/bburns/Desktop
$ pgloader ....
This solution may come a little late to the original asker, but for those still searching for a windows solution. I can confirm that I was able to get pgloader working using Docker for Windows.
My environment is a local MySQL install on a Windows 10 PC. I installed Docker for Windows. I then installed PostgreSQL via docker:
docker run --name some-postgres -p 5432:5432 -e POSTGRES_PASSWORD=mysecretpassword -d postgres
Finally, I ran pgloader via docker to pull the schema & data from mysql to postgres using the following:
docker run --rm --name pgloader dimitri/pgloader:latest pgloader
mysql://root:root#host.docker.internal/my_db_name
pgsql://postgres:mysecretpassword#host.docker.internal/postgres
This created a schema called "my_db_name" on my postgres instance. The "root:root" and "postgres:mysecretpassword" are my username:password details. This is obviously a production ready environment! The other useful bit "host.docker.internal" is docker's reference to the parent host environment. Also, I ignored the SBCL warning.
I'm still processing the results. Most everything looks like it migrated successfully. I'm missing triggers (unsupported) and some foreign keys (bug?). Anyways, I hope this might be helpful to others.
I am installing predictionIO from source code. I have downloaded and done the predictionIO installation successfully. I am now trying to install the dependencies (Spark, Elasticsearch, HBase) but I am running into errors for each of them. Below are the issues I am facing when I execute pio status:
1 - Unable to locate a proper Apache Spark installation
2 - It is also unable to find metadata files.
I have not changed any default settings. I'm using windows 8.1. On localhost, I have running IIS. On 127.0.0.1:8888 I run ipython notebook.
Please help on how I can get predictionIO up and running on my machine.
Thanks
If you are on Windows, you can install with Vagrant.
http://docs.prediction.io/community/projects/#vagrant-installation-for-predictionio
I believe the discussion has moved to the google group.
https://groups.google.com/forum/#!searchin/predictionio-user/SAS/predictionio-user/ZamBr1ZaQ3o/fyNkXh3zsv0J
https://groups.google.com/forum/#!searchin/predictionio-user/SAS/predictionio-user/0awaASUR8lE/JkLtPeRrNt4J
This is the relevant thread.
https://groups.google.com/forum/#!searchin/predictionio-user/SAS/predictionio-user/0awaASUR8lE/JkLtPeRrNt4J
Moreover, PredictionIO docs had a few errors. Below are some of them and their corrected versions.
1 - Actual line: PATH=$PATH:/home/yourname/predictionio/bin; export PATH
Corrected Version PATH=$PATH:/home/yourname/PredictionIO/bin; export PATH
2 - Actual Line: $ pio eventserver
Corrected Version: $ pio eventserver --ip 0.0.0.0
3 - Actual Line pio template get PredictionIO/templates-scala-parallel-recommendation MyRecommendation
Corrected Version pio template get PredictionIO/template-scala-parallel-recommendation MyRecommendation
Trying to install Vagrant on Windows 7 X64 with installer from Vagrant - vagrant_1.7.0.msi
Installation goes fine, however unable to use vagrant. On any command, for example vagrant -v the below is shown:
Failed to find Vagrant!
What I did:
Googled. Only 1 link with such an error on github regarding to installer. Not sure if this is relevant.
Checked that vagrant in PATH. Yes it is. Otherwise it would not start at all.
Reinstalled, and repaired - few times.
All the above with no luck. Any idea where to dig?
Thank you.
just to add my own version of this problem too:
i installed version 1.7.0
error message after a "vagrant up": "Failed to find Vagrant!"
checked paths, other environment variables in win7, should work..
so i uninstalled it and installed an older version (1.6.5)
and yes, the older version solved the problem.
i got it from their official website: https://www.vagrantup.com/download-archive/v1.6.5.html
i guess the problem was in this file:
https://github.com/mitchellh/vagrant-installers/blob/master/substrate/launcher/main.go
Same Issue on Windows 7 64.
Downgraded back to Vagrant 1.6.5
Have you gone through the installation instructions on vagrant side listed here :
https://docs.vagrantup.com/v2/installation/index.html
This is what I see listed at their site:
The installer will automatically add vagrant to your system path so that it is available in terminals. If it is not found, please try logging out and logging back in to your system (this is particularly necessary sometimes for Windows).
Please try running it from Git Bash too and let me know if it works for you.
And if you still run into issues, try using an older version of vagrant https://www.vagrantup.com/download-archive/v1.6.5.html
Vagrant 1.7.1 should fix this. https://github.com/mitchellh/vagrant/issues/4924#issuecomment-66808666