This is in response to #7264, since we decided that PR is not the right way to do it.

I think we decided on: move the non-browser-specific tests out of the browser-specific test files.

So if the "chrome only" tests pass when using Firefox for instance, they should be moved to unit tests or into a common spec file of some kind.

Commands that are described within the on-line documentation aren't where they are described to be.

Issue Description

When I run the example cartpole with the default parameters, it can not converge to the max reward 200, I wonder what went wrong.

Version Information

Please indicate relevant versions, including, if relevant:

• Deeplearning4j versi

The ParseTreeWalker class has a typo the runtime documentation of the enterRule method. Here is the link to the docs: https://www.antlr.org/api/Java/index.html

Bash completion capabilities are nicely structured into different bash functions in https://github.com/facebook/buck/blob/master/scripts/buck-completion.bash though.
This means completion functions like _buck_completion_try_build can be easily reused by users/developers for other custom functions beyond buck.

However the way logging currently works for bash completion,
there is `local log

Describe the bug
I'm trying to export with « Custom tabular exporter… » and choosing « Upload to -> A new Google spreadsheet », and in some rows, if a value is missing, then the cell value disapear and the cell on the right shift left.

To Reproduce
Steps to reproduce the behavior:

1. « Custom tabular exporter… »
2. choose « Upload to -> A new Google spreadsheet »
3. See the result

In the Evaluation doc in wiki, it incorrectly says Maidenhead (mls) can form words. Since MLS alternates pairs of letters and pairs numerals, it could only form EL33tt haxor words, and at least for the experienced user, or with a syntax checking interface, no letter/number ambiguity.

(MLS does have seams between fields and between squares so that adjacent places may be disimilar, but even OL

What is the issue?

neural engine support for PHP SDK

How can someone reproduce this issue (if applicable)?

1. Use https://github.com/awsdocs/aws-doc-sdk-examples/blob/master/php/example_code/polly/SynthesizeSpeech.php

change client region to 'region' => 'us-east-1'

1. pass in this array into synthesizeSpeech ()

'Text' => 'Hello world'
'OutputFormat' => 'mp3',
'VoiceId

Issue Description

Please describe your issue, along with:

• expected behavior
• encountered behavior

Version Information

Please indicate relevant versions, including, if relevant:

• Deeplearning4j version
• platform information (OS, etc)
• CUDA version, if used
• NVIDIA driver version, if in use

Contributing

If you'd like to help us fix the issue by contributi

Current MxNet example is just show how to do prediction, please add some example code for creating model & training, thanks.

MxNet has full featured cpp examples, such as
https://github.com/dmlc/MXNet.cpp/blob/master/example/mlp.cpp

ISSUE TYPE

• Bug Report

COMPONENT NAME

https://github.com/bartdag/py4j/blob/216432d859de41441f0d1a0d55b31b5d8d09dd28/py4j-python/src/py4j/java_gateway.py#L222

Currently defaults to "java", which means the logic for looking at JAVA_HOME is not hit by default. Seems odd for clients to specify java_path=None to get the nice behaviour of looking at JAVA_HOME and falling back to PATH.

As it stands there are a bunch of gotchas with this project. From certain versions of packages not working (we eventually learn it is related to the fact they aren't pulling from Pypi) to having a very tight coupling to pivy-importer (which is admitted to be more of a throwaway as linkedin use an internal package) I feel the documentation needs an overall.

Without clear working examples to get