googler | : octocat : Get into Google for Humans

In CLion, What is the difference between Qt Console and Qt Widgets Executable project templates. My intuition is that Qt Console is CLI and Widgets is GUI, but googling hasn't confirmed this. I'm hoping this QnA will help future googlers.

This is probably really easy, but its driving me crazy.

I have the following function:

So I am fairly new to React hooks, and I am trying to use similar variables over again.

I want to have several buttons, each of which launch a different login mechanism, like so:

Hi y'all I'm trying to replicate a collapsing accordian example I found on https://getbootstrap.com/docs/4.1/components/collapse/ but for some reason they are not collapsing. In fact they start in the uncollapsed position which I think is strange because in the exmaple they start collapsed and then on the button click they collapse open. Anyways here is my code, I hope someone out there can help because I can't figure this out.

FINAL UPDATE: for any googlers that stumble upon this Here is the code that worked for me. Notice that the data-target => attr.data-target is the only way to use text binding with the data-target

I have been following this example from ML .NET tutorials: https://github.com/dotnet/samples/tree/master/machine-learning/tutorials/GitHubIssueClassification

And built my own version of this example which reads data from .xlsx (different dataset) and splits it into training and testing sets. It works well and makes correct predictions, but I can't for the life of me figure out why evaluation metrics (every parameter) always show 0 when I feed _testSet into it. When I feed _trainSet it evaluates to 1 which is expected.

Even if I set TestFraction == 0.5 it still evaluates to 0.

I can see that TVirtualStringTree has a property called AccessibleName yet, I don't see it working, that is, giving the string which is set to screen reader such as NVDA.

What I tried is setting this in the OnGetText event but that doesn't seem to produce any results (e.g. Sender->AccessibleName = "my node name";). All it reads is entire contents of the tree, not just a current node just the same as if I didn't make any change to the property, and if node is changed (using keyboard) it doesn't do anything.

What it should be reading is - Column name - column content - another column name - another column content etc. like it does with the regular TListView.

There seems to be an implementation of iAccessible interface here - https://github.com/Virtual-TreeView/Virtual-TreeView/blob/master/Source/VTAccessibility.pas

Can you share some example how to use VirtualTreeView built in accessibility to make the component readable to screen readers as it seems that such examples are really scarce?

Edit (for future googlers):

Related question which solves the C++ Builder accessiblity unit linking:

Implementing VirtualTreeView TVTDefaultAccessibleProvider in C++ Builder

Edit 2 (for future googlers):

Additionally, TVirtualStringTree also provides OnGetImageText. This event is used to provide accessible text for images used in the tree, for example if you have multiple images in a node to represent certain states you can describe these states with image labels.

We want to query gcr.io using Python. However, as of this writing, any of the below is not possible:

• Docker SDK: client.images.list() only lists local images
• there is no Google Client library for GCR
• images are stored in GCS but as one big list of digests with no metadata. Cannot tell the repo or tags
• docker registry in desktop is experimental and not available for querying GCR

The only "hack" is to execute gcloud container images list (and list-tags) xxx in a subprocess and parse to extract the info you need.

We could too parse the HTML response from the GCP console (browser) but that would be more work.

Any other idea on how to easily list GCR images in Python ?

And to Docker and Googlers, any plan on extending your Python library to interact with remote registries, *.gcr.io in particular ?

When training a model (with keras), what does the val_loss (as it's shown in the keras training log) / epoch_loss (as it's shown in tensorboard) metric measure when the model has multiple outputs? And if it matters, then also if those outputs are weighted?
And I mean multiple separate outputs (e.g. 6 tensors of shape ()), not a single output with multiple values (e.g. 1 tensor with shape (6)). ^{(Though an answer to the latter might be useful for future googlers.)}

I originally thought it would be a weighted average of all the outputs. But my observations do not agree with that guess. Below is a screencap of tensorboard showing the validation loss of each of the outputs, as well as the epoch_loss value (also on validation). As you can see, on each output the loss is increasing, yet the epoch_loss value is continually decreasing.

I have also confirmed that these metrics match the val_* metrics (val_1_loss, val_2_loss, ..., val_loss) shown in the keras training log.

Note that I'm not interested in investigating why the loss increasing or ways to fix the model or anything. I know this is a bad model. This was just one model from a hyperparameter search that I picked as a good example for the question. Though I am currently using val_loss as the hyperparameter search optimization metric, which from this data does not seem like a good idea. I can create my own metric to do a weighted mean of all the outputs, but wanted to understand what val_loss is actually measuring.

Edit for Googlers: this turned out to be caused by using an out-of-date beta release of Java 9.

I am trying to batch download zip files from this URL using java - SRTM files and it requires a username/password to download and I am using the following java code and it gives me the following exception