I am asking from a cloud architect's perspective. I could not find it spelled out in the documentation: Do AI Platform custom containers only support Python ML frameworks, or are other languages also supported?
Whether it makes sense to not use Python is a different topic ...
As long as you can build your binaries/libs into a Docker container images, they can be run on AI Platform.
Related
I have a task to use SonarQube.
My build are done using Google Cloud Build. How can I integrate SonarQube with Google Cloud Build
Thanks for your help
You can use custom builders. At the end, each build step is a container image:
Cloud builders are container images with common languages and tools installed in them. You can configure Cloud Build to run a specific command within the context of these builders.
There GCP documentation provides a guide on how to create a custom build. However, notice that it's inteded to be general and doesn't include any specific functionality that you might require. Nevertheless, is a great starting point for understanding how the custom builders work and create your own.
Aside from this approach, there is a community builder for Sonarqube that you can use as reference or might even suit your needs.
Edit:
In case your question is about code analysis with Sonarqube. The community builder is still relevant as it allows you to run static code analysis for your project from sonarcloud.io.
I am a new user of GCE. According to this documentation here:
https://docs.docker.com/install/linux/docker-ce/centos/#install-using-the-repository
To install Docker Engine - Community, you need a maintained version of CentOS 7. Archived versions aren’t supported or tested. Are all the OS versions offered by the GCP maintained or archived?
Public images are provided and maintained by Google, open-source communities, and third-party vendors.
Community-supported images are not directly supported by Compute Engine. It is up to the project community to ensure that images work with Compute Engine features and that security updates are maintained. Community-supported images are provided as-is by the project communities that build and maintain them.
So, yes if it is a public image you want to use, it is maintained. If you want to use a custom image they are not directly supported.
To check the Public images offered by Google Cloud Platform you can go in Compute Engine > Images.
I have finally arrived in the cloud to put my NLP work to the next level, but I am a bit overwhelmed with all the possibilities I have. So I am coming to you for advice.
Currently I see three possibilities:
SageMaker
Jupyter Notebooks are great
It's quick and simple
saves a lot of time spent on managing everything, you can very easily get the model into production
costs more
no version control
Cloud9
EC2(-AMI)
Well, that's where I am for now. I really like SageMaker, although I don't like the lack of version control (at least I haven't found anything for now).
Cloud9 seems just to be an IDE to an EC2 instance.. I haven't found any comparisons of Cloud9 vs SageMaker for Machine Learning. Maybe because Cloud9 is not advertised as an ML solution. But it seems to be an option.
What is your take on that question? What have I missed? What would you advise me to go for? What is your workflow and why?
I am looking for an easy work environment where I can quickly test my models, exactly. And it won't be only me working on it, it's a team effort.
Since you are working as a team I would recommend to use sagemaker with custom docker images. That way you have complete freedom over your algorithm. The docker images are stored in ecr. Here you can upload many versions of the same image and tag them to keep control of the different versions(which you build from a git repo).
Sagemaker also gives the execution role to inside the docker image. So you still have full access to other aws resources (if the execution role has the right permissions)
https://github.com/awslabs/amazon-sagemaker-examples/blob/master/advanced_functionality/scikit_bring_your_own/scikit_bring_your_own.ipynb
In my opinion this is a good example to start because it shows how sagemaker is interacting with your image.
Some notes on other solutions:
The problem of every other solution you posted is you want to build and execute on the same machine. Sure you can do this but keep in mind, that gpu instances are expensive and therefore you might only switch to the cloud when the code is ready to run.
Some other notes
Jupyter Notebooks in general are not made for collaborative programming. I think they want to change this with jupyter lab but this is still in development and sagemaker only use the notebook at the moment.
EC2 is cheaper as sagemaker but you have to do more work. Especially if you want to run your model as docker images. Also with sagemaker you can easily build an endpoint for model inference which would be even more complex to realize with ec2.
Cloud 9 I never used this service and but on first glance it seems good to develop on, but the question remains if you want to do this on a gpu machine. Because you're using ec2 as instance you have the same advantage/disadvantage.
One thing I'd like to call out first is SageMaker notebook is not the only IDE environment in which you can interact with other components of SageMaker such as training and hosting. In fact you can make API calls to SageMaker training/hosting through Cloud9 or any IDEs you've installed on EC2 or even your laptop, as long as you have AWS SDK or SageMaker Python SDK installed.
Regarding the choice of the IDE, it's really up to your particular needs. SageMaker notebook is Jupyter based (now also supports JupyterLab beta), ML focused, and fully managed. Hundreds of Python packages that are commonly used in ML, as well as Tensorflow, Keras, MxNet, SageMaker Python SDK, etc., are preinstalled and automatically maintained for you. It also integrates more closely with other components of SageMaker as one can imagine.
Cloud9 is a managed IDE too but it is for general purpose rather than ML specific. If you want to use Jupyter on cloud9 it requires extra work from your side. It does not preinstall and maintain the version of common ML/DL related packages like SageMaker notebook does.
I want to implement a tool that uses the technologies - Jenkins, Docker, Docker Swarm, and AWS - to achieve a deployment tool that our team of developers can use to manage instances and deploys.
Please recommend what technologies should we (both administrators and developers) be using, what needs to be built and what sorts of machines must be having.
Any help here would be much appreciated.
Your question is too generic to provide a specific answer, as there are different approaches to implement what you are trying to achieve. IMHO the best approach would be to talk with your existing dev team & administrators and come up with a solution which all parties find easy to manage and maintain container based environment rather than specifying several specific technologies.
Each tool you have mentioned has different capabilities and also there are other tools that provide the same features which would be more ideal for your situation. (Thats why proper understanding between Devs and admins are necessary on what you really want to achieve.) .
Since you have asked about what kind of machines you must be having (I suppose this is on AWS env) try Core OS on AWS instances. CoreOS (Container Linux) will be the best option to manage and run your container based environments. [About CoreOS]
Jenkins can run in a docker container and issue docker commands to deploy new docker containers that reside in the same swarm as jenkins. You also need to hook into a software repo like git. Jenkins Blue Ocean is something you could look at for pipe-lining your dev->build->test->deploy->maintain pipes. Also, Travis-ci, github, JIRA, and Dockerhub are useful components to what you are trying to achieve.
I am running a Django App inside GCP. My idea was to call a python script from "view.py" for some machine learning algorithm and then display the result on the page.
But now I understand that running a machine learning library like Scikit-learn on GAE will not be possible (read Tim's answer here and this thread).
But suppose I need to still do this, I believe there are 2 ways possible, but I am not sure weather my guess is right or wrong
1) As the Google-Datalab provides the entire anaconda like distribution, if we have any datalab api which can be called from a python file in the Django app, I can achieve my goal ?
2) If I can install the scikit-learn library on any compute engine on GCP and somehow send it the request to run my code and then return the output back to the python file in the Django app ?
I am very new to client-server and cloud computing on the whole, so please provide examples (if possible) for any suggestion/ pointer for the help.
Regards,
I believe what you want is to use the App Engine Flex environment rather than the standard App Engine environment.
App Engine Flex uses a compute engine VM for running your code, so it does not have the library limitations that standard App Engine has.
Specifically, you'll need to add a 'requirements.txt' file to specify the version of scikit-learn that you want installed, and then add a 'vm: true' clause to your app.yaml file.
sklearn is now supported on ML Engine.
So, another alternative now is to use online prediction on Cloud ML Engine, and deploy your scikit-learn model as a web service.
Here is a fully worked out example of using fully-managed scikit-learn training, online prediction and hyperparameter tuning:
https://github.com/GoogleCloudPlatform/training-data-analyst/blob/master/blogs/sklearn/babyweight_skl.ipynb