Please make sure that this is a feature request. As per our GitHub Policy, we only address code/doc bugs, performance issues, feature requests and build/installation issues on GitHub. tag:feature_template

System information

• TensorFlow version (you are using): 1.15
• Are you willing to contribute it (Yes/No): No

It should be a core layer but it doesn't show up on the website, nor the search function picks it up:

https://keras.io/layers/core/

I can use it without warnings, so I wouldn't expect it to be deprecated.

The test_logistic::test_dtype_match fails on 32 bit Windows as was reported in MacPython/scikit-learn-wheels#7 (comment) The implication is that for LogisticRegression, float32 and float64 input may result in somewhat different models depending on the processor architecture (and OS), which is troublesome.

This applies to Python 3.5, 3.6 but not 2.7, 3.4

https://github.com/aymericdamien/TensorFlow-Examples/blob/master/examples/3_NeuralNetworks/bidirectional_rnn.py#L80

return tf.matmul(outputs[-1], weights['out']) + biases['out']

the shape of + left : (?,10)
the shape of + right : (10)

Why does it work ?

Currently, torch.gather does not broadcast. For example:

t = torch.tensor([[1,2],[1,2]])
torch.gather(t, 1, torch.tensor([[0,1,0],[0,1,0]]))

gives

tensor([[ 1,  2,  1],
        [ 1,  2,  1]])

But

t = torch.tensor([[1,2],[1,2]])
torch.gather(t, 1, torch.tensor([[0,1,0]]))

and

t = torch.tensor([[1,2]])
torch.gather(t, 1, torch.ten

Page 2
Error in pixaGetCount: pixa not defined
Error in pixaGetCount: pixa not defined
Loaded 41/41 pages (0-41) of document /tmp/tmp.tY7p2Ue5TC/san/san.Baloo.exp0.lstmf
Bad box coordinates in boxfile string! विताः । नानाशस्त्रप्रहरणाः सर्वे 1576 3968 2121 4022 0
Bad box coordinates in boxfile string! रिदेवना ॥ २-२८॥ आश्चर्य 1526 2958 1995 3016 1
Bad box coordinates in boxfile string! त

It would be really nice if you could add an example which shows how the face encodings can be stored in a database and how to efficiently query them.

There should be unit tests that verify the behavior of #2909 is correct. In particular, tests should check that Caffe crashes when shared params have the different lr_mults or decay_mults, and proceeds when shared params have matching *_mults. (Other unit tests already verify the latter for certain cases, but are obviously non-exhaustive since they also passed with the previous buggy behav

They're not exported, but since we refer to them in deprecation warnings, they should probably be documented.

This should really help to keep a track of papers read so far. I would love to fork the repo and keep on checking the boxes in my local fork.

For example: Have a look at this section. People fork this repo and check the boxes as they finish reading each section.

according to https://github.com/dmlc/xgboost/blob/a725272e19748a5f3ddd7eeeff7c766660e38a06/src/objective/multiclass_obj.cc#L94-L120

we should not have probability col in multi:softmax case and instead only rawPrediction column and prediction

otherwise, probability would be filled with prediction and no real probability would be presented

Hey, I see that there are no tutorial notebooks for implementing machine learning algorithms and neural networks in PyTorch in this repo yet. PyTorch is gaining a lot of traction lately and is really going to be one of the most popular frameworks due to its dynamic computational graph & eager execution.

I would like to add such tutorial notebooks in PyTorch.

Right now if a validation error is thrown, all the user gets is:
packages/cntk/cntk_py.py", line 1584, in train_minibatch
return _cntk_py.Trainer_train_minibatch(self, *args)
RuntimeError: AsMatrix: Sparse tensors are not supported unless they are 1D or 2D matrices.

This lacks the information what operation caused the error. It should at least show
"While validating [operator instance]

I am having difficulty in running this package as a Webservice. Would appreciate if we could provide any kind of documentation on implementing an API to get the keypoints from an image. Our aim is to able to deploy this API as an Azure Function and also know if it is feasible.

Does the option to output n-best POS tags (or even all tags with score > 0) exist with some hidden option or in some branch?
If not, would such an additional function be welcome or should I just hack it into my own fork? Technically it should be very easy (basically just replace the arg_max_if_true in spacy/tagger.pyx#L211 with an appropriate collection of results), but I have a feeling it wi

I recently added "back to top" button to README. What other features would make it easier to browse? Please write your recommendation.

It may be good to provide pure Python implementation of Gradient Descent (instead of SciPy one) for Logistic Regression just for the learning purposes.