https://github.com/pytorch/pytorch/pull/34093/files

cc @ezyang @anjali411 @dylanbespalko

When opening Alacritty with --class option it changes the window instance instead the class.
For example, if I do alacritty --class foo and I check the window properties I got:

Instance: foo
Class: Alacritty
Title: Alacritty

--title option is working as expected, although.

Infos

OS: Funtoo Linux
Window Manager: dwm
Alacritty: alacritty 0.5.0-dev (03bce99)

This is an issue on the Quickstart web page. Sorry if this is not the right place for this type of issue report.

The page seems confusing regarding Maven configurations. It has a section called "Using DL4J In Your Own Projects: Configuring the POM.xml File", which sounds like the usual Maven dependencies information that so man

The problem as far as I understand is that libfoo.so's only knowledge about libcuda.so is its SONAME, which is libcuda.so.1, so a binary wanting to link in libfoo.so needs to know who to find libcuda.so.1 in order to verify the available symbols. Generally, ld's suggestion is to use -rpath-link to provide the path to the library, but the nvidia docker repository doesn't come with the symlink to li

What is wrong?

The documentation for kernel constants is very scattered and incomplete. The constants feature is awesome, it needs more love...

How important is this (1-5)?

√(-2)

This class could be used instead of cd file https://catboost.ai/docs/concepts/input-data_column-descfile.html when creating Pool from filez. The class should have init function, methods load and save, and Pool init method should be able to use object of this class instead of cd file during initialization.

https://docs-cupy.chainer.org/en/stable/reference/sparse.html#sparse-matrices
says

CuPy supports sparse matrices using cuSPARSE. These matrices have the same interfaces of SciPy’s sparse matrices.

However, the constructor

csr_matrix((data, (row_ind, col_ind)), [shape=(M, N)])

from <https://docs.scipy.org/doc/scipy/reference/generated/scipy.sparse.csr_matrix.html#scipy-spar

  File "/root/miniconda3/bin/pipeline", line 11, in <module>
    sys.exit(_main())
  File "/root/miniconda3/lib/python3.7/site-packages/cli_pipeline/cli_pipeline.py", line 5734, in _main
    _fire.Fire()
  File "/root/miniconda3/lib/python3.7/site-packages/fire/core.py", line 127, in Fire
    component_trace = _Fire(component, args, context, name)
  Fil

Hi all,
So I'm using DIGITS version: 6.1.1 with Caffe version: 0.17.0 and Caffe flavor: NVIDIA
I've successfully used NVIDIA/Digits many times but when using compatible mode I cant replicate results in opencv or caffe BVLC.

This is the output I get on Digits:

but this is what I

We have this performance issue in Vulkan backend that gfx-rs API is based on iterators, and Ash expects slices, so we need some extremely lightweight way of producing a slice, sometimes without knowing anything about the number of elements. Basically, we want alloca, but it's not available in Rust.

However there is a simple crate that does something similar - https://github.com/NotIntMan/inpl

It tells you to get version 3.0 of the SDK, which doesn't have libwrapper.so, so you get an unhelpful failure to find halide_hexagon_remote_load_library (because init_hexagon_runtime doesn't check if host_lib is non-null). This is hard to debug, because host_lib is null not because libhalide_hexagon_host.so isn't found or isn't in the path (it is!) but because a dependent library - libwrapper.so -

We would like to forward a particular 'key' column which is part of the features to appear alongside the predictions - this is to be able to identify to which set of features a particular prediction belongs to. Here is an example of predictions output using the tensorflow.contrib.estimator.multi_class_head:

{"classes": ["0", "1", "2", "3", "4", "5", "6", "7", "8", "9"],
 "scores": [0.068196

Is your feature request related to a problem? Please describe.
According to the Arrow spec:

Bitmaps are to be initialized to be all unset at allocation time (this includes padding).

This would imply that bits outside the range [0, size) should always be zero. However, in cuDF/libcudf, we take a more conservative approach and say that bits outside [0,size) are undefined in order to a

  -s --gpu-select   : Column separated list of GPU IDs to monitor
  -i --gpu-ignore   : Column separated list of GPU IDs to ignore

should be "colon" instead of "column" (to match man page)

DeepSpeed's data loader will use DistributedSampler by default unless another is provided:

https://github.com/microsoft/DeepSpeed/blob/001abe2362d9edba062070fb05df40925f54cb3e/deepspeed/pt/deepspeed_dataloader.py#L43

If DeepSpeed is configured with model parallelism, or called from a library with a sub-group of the world processes, the default behavior of DistributedSampler is incorrect

https://omnisci.github.io/omniscidb/quickstart/start.html - links are broken: “COPY FROM”, “SQLImporter”, etc.

Hi,

I try to understand Deepdetect right now, starting with the Plattforms Docker container.
It looks great on pictures, but I have a hard time right now using it :)

My Problem: The docs seems to step over important points, like using JupyterLab. All examples shows the finished Custom masks, but how do I get them?

Is there something missing in the docs?

Example: https://www.deepdetec