Valohai Platform Documentation
Please find our new knowledge base at help.valohai.com.
This repository is now archived.
:blue_book: Valohai ecosystem documentation
Home Page: https://docs.valohai.com/
Please find our new knowledge base at help.valohai.com.
This repository is now archived.
Write in advanced peruna level. Designed for someone who has some mileage with valohai-utils and/or valohai already.
Ad-hoc runs are not really mentioned in docs, they should be as they are pretty nifty feature.
The leadfeeder script
<script>
(function(){ window.ldfdr = window.ldfdr || {};
(function(d, s, ss, fs){ fs = d.getElementsByTagName(s)[0];
function ce(src){ var cs = d.createElement(s); cs.src = src;
setTimeout(function(){fs.parentNode.insertBefore(cs,fs)}, 1); }
ce(ss); })(document, 'script',
'https://lftracker.leadfeeder.com/lftracker_v1_X0n7aEZkr8NxNpPG.js'); })(); </script>
needs to be added to the global template.
If there's a Leadforensics script, this shall replace it.
This mirrors https://github.com/valohai/valohai-site/pull/89 .
{parameter:hello} {parameter-value:decoder-spec}
In fact, I wrote this already when on an earlier on-boarding.
I just need to copy-paste it in, maybe add images during that or later.
Use environment variables to select the correct input and output folders in the docs. Currently we are just using hardwritten paths.
We've had questions how to use custom Docker images.
latest
Document this.
Add the docs for this new feature.
As noted at the HelsinkiAI meetup, it's hard to realize we have an CLI and an API in addition to the GUI.
We should tell people about that front-and-center.
e.g. https://github.com/valohai/docs/blob/master/source/jupyter/index.rst has xxx
as :description:
.
valohai/valohai-yaml#29 added forced filename and keep-directories features. Those aren't yet documented.
Via 2hat
As discussed with Someturva, it would be useful to have better documentation on YAML endpoint
s (at all!), especially about the modes (WSGI or server-command).
vh exec run --help
doesn't give any help on exec inputs
vh exec run model1 --help
helps moar, but is hard to find
First one should hint the latter.
Also documentation should mention and explain that this feature even exists.
Valohai can be setup in numerous ways; we do describe that in the architecture section of the docs.
But we are only showing the SaaS installation architectural diagram.
TODO:
Page: tutorials/quickstart/quickstart-metadata.html
Feedback:
There is a typo in the doc: logMetadta
Explain in a more concise way how Valohai handles things.
Preferably one image with some text.
Valohai/Decks/Ruksi/valohai-architecture.pdf/pptx
environment-variables for valohai.yaml are not documented in https://docs.valohai.com/valohai-yaml/index.html
e.g. s3:///path/to/files/*
valohai.yaml
format (& indendation)vh lint
to validate your YAMLvalohai-utils
to reduce headacheThe environment-variables
stanza is currently undocumented.
Page: tutorials/apis/fetchProjects.html
Feedback:
typo?: python --veresion
Executions can be used to do batch predictions, which is frequently what users want but it might not be obvious that Valohai steps can be used this way.
Add some guide or tutorial about it.
Since the schema for the valohai.yaml file is available in the valohai_yaml repo, it would be nice to have the docs process automagically create documentation pages (of a kind to be defined) for them.
Automatically generate meta descriptions for all pages that don't have them set explicitly.
Moz.com recommends meat descriptions to be long enough that you can understand the content it's about but not too long. Ideally apparently between 50–300 characters.
We should document how a docker container name gets resolved and how to point to your own custom repositories.
E.g. here https://docs.valohai.com/core-concepts/executions.html under "Tip: You can use ready available Docker images e.g. from DockerHub or host your own."
Could @DrazenDodik and others interested add existing docs painpoints here? What is bad?
Based on Twohat feedback, the docs are not clear that you need to specify the STEP NAME.
It caused problems so it should be more clear.
A declarative, efficient, and flexible JavaScript library for building user interfaces.
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
The Web framework for perfectionists with deadlines.
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
We are working to build community through open source technology. NB: members must have two-factor auth.
Open source projects and samples from Microsoft.
Google ❤️ Open Source for everyone.
Alibaba Open Source for everyone
Data-Driven Documents codes.
China tencent open source team.