Hi guys,

I'm a computer science student and I have to make a project in Go with Proto.Actor. Because of my lack of experience (this is my first project in Go, coming from Java) and the lack of documentation on the official website I find it very difficult to understand how one can use it. It would be nice if you could provide a better documentation. Thank you!

I tried building the docs, but was met with a graphviz error. Typically this means I can spend a few hours pecking away at the dependencies until I get stable build... or someone that has it working can export their environment, and publish an environment.yml that we can use with the build instructions.
I was going off of the d2l book since that's a dep here, but their [environment.yml](https://g

Description
In some rare cases, for example, when you need to finetune a large model on a small dataset the majoring part of training loop is waiting for saving model checkpoints to a hard drive.

Proposal
Would be logically to add a CheckpointCallback with parameter save_n_best=0 to a configuration and do not store best checkpoints and instead use the latest state of the model.

usr/local/spark/python/pyspark/ml/param/init.pyc in getOrDefault(self, param)
346 return self._paramMap[param]
347 else:
--> 348 return self._defaultParamMap[param]
349
350 def extractParamMap(self, extra=None):

KeyError: Param(parent=u'ElephasEstimator_49f09b9cf514ec3ebc45', name='loss', doc='Keras loss')

In Advanced view when Activity on any of the items is set to Suspend, it will suspend that item for eternity until 'based on preferences' or 'always' is chosen.

In Simple view there is a Suspend button at the bottom, which suspends calculations for just one hour.
It does the same thing as the Snooze option in the (Windows) right-click BOINC Manager icon in the system tray.

For consistenc

In some workloads with highly compressible data we would like to trade off some computation time for more in-memory storage automatically. Dask workers store data in a MutableMapping (the superclass of dict). So in principle all we would need to do is make a MutableMapping subclass that overrides the getitem and setitem methods to compress and decompress data on demand.

This would be an i

hello world usage code

var props = Actor.FromProducer(() => new HelloActor());
var pid = Actor.Spawn(props);

but Actor code is like this

public static class Actor
{
    public static readonly Task Done;
    public static EventStream EventStream { get; }
}

so this code is not working at first.

Not sure if it's worth adding to the documentation but I found I had to add the following to the end of the fstab command otherwise the system would not boot cleanly.

,x-systemd.mount-timeout=30,_netdev

The line in fstab now looks like

mfsmount /mnt/mymnt fuse mfssubfolder=mydir,allow_other,x-systemd.mount-timeout=30,_netdev

I think it was trying to mount mooseFS before it had fini

I don't want to use current chunks loop for replication (mostly because is still unclear how it works).
What I would like is to set our cluster to replicate as much chunks as possible up to REPLICATION_BANDWIDTH_LIMIT_KBPS

What should I have to set and how ? I think that both REPLICATION_BANDWIDTH_LIMIT_KBPS and chunks loop limits are honored, so I have to set chunk loop to very high value

Winxp, need to redo Server and Client Documentation

Wiki should be updated to inform users the C# client is technically Depreciated at this time and current wiki should be converted to using the python client as its main focus ( Or make the link to the Client repo much more apparent ((probably the best option)))

Videos should be rather generic on setup Pointing users to use commands found

Add 'Best Practices For Package Developers using Futures' vignette with instructions on, for instance,

1. How to test code, e.g. run tests with plan(sequential) and plan(multisession) and don't assume that plan(multicore) is available.

2. Make sure to shut down PSOCK clusters at the end of examples and package tests, i.e. call plan(sequential) at the end (e.g. #327)

3. Avoid setti

I am by far the least qualified person to review your work, but I find it relentlessly fascinating. I want your work to gain more traction, but I think that English is not your first language. I would like to offer my help to do some editing on your FP-complete posts to make them sound more natural in English. I can't offer much, but as I read through them I could offer some edits/questions that m

The docs state that in the output path a $ is replaced by the worker id and # by the file counter.
Hash is fine $ is wrong, it's @ which is used.
This can have very subtle and unexpected consequences: If you pass "foo-$#.txt the $ will just stay but the worker id will be appended to the end of the name (before extension) which will yield files like this foo-$<counter><worker>.txt whi

If enter_data() is called with the same train_path twice in a row and the data itself hasn't changed, a new Dataset does not need to be created.

We should add a column which stores some kind of hash of the actual data. When a Dataset would be created, if the metadata and data hash are exactly the same as an existing Dataset, nothing should be added to the ModelHub database and the existing

When getting the value of a key it is not possible to get all attributes.
It might be useful to get the session for example in the case that it is needed to discover a leader.

Starting from this commit douban/paracel@c81b4ff, the deployment section of document is a little lagged with v0.3.0. Some users complained about that! To make it work, we need to update the document. Also honor different document versions for different Paracel releases.