docs: add 'Using Shared Server For The Team' #56
Conversation
| @@ -0,0 +1,94 @@ | |||
| # Using Shared Server For The Team | |||
There was a problem hiding this comment.
It took me a while to recall what's this about. Shared Server is not very obvious name for this scenario. Let's use a straightforward name, something like: Multiple Data Scientists on a Single Machine
| @@ -0,0 +1,94 @@ | |||
| # Using Shared Server For The Team | |||
|
|
|||
| The key principle that every data science team should aim for is to make | |||
There was a problem hiding this comment.
Let's try to start with the description of the scenario first:
It's pretty common to see that teams prefer using one single shared machine to run their experiments ...
There was a problem hiding this comment.
Then, we need to mention briefly, why is is happening:
- Better resource utilization - I can utilize multiple GPUs, for example
- Probably, less expensive
- Data locality
- something I am missing?
| data is super valuable as it enables your colleagues and you to take advantage | ||
| of the data processing that you've already done, and understand the context | ||
| behind results you create and surface potential caveats. In order to share | ||
| common data, many teams setup a shared server for the whole team. With DVC, |
There was a problem hiding this comment.
May be we should omit or simplify the description above? It's too long, I'm not sure it's to the point of this use case.
| $ git push | ||
| ``` | ||
|
|
||
| And now you can just as easilly get his work appear in your workspace by: |
There was a problem hiding this comment.
Should we mention abvout garbage collection, and that we are workin on LRU cache strategy to keep the size limited?
There was a problem hiding this comment.
I don't think so, it is beyond the scope for this article.
|
Good stuff. I would simplify title and intro. I'll try to come with a picture to this. |
|
@shcheklein I have actually used this https://www.forbes.com/sites/quora/2017/04/04/what-are-best-practices-for-collaboration-between-data-scientists/#babcee4335ee for the introduction to this use case and its naming. |
Fixes iterative#44 Signed-off-by: Ruslan Kuprieiev <[email protected]>
|
👍 let's merge, and polish online |
Fixes #44
Signed-off-by: Ruslan Kuprieiev [email protected]