You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@mxnet.apache.org by GitBox <gi...@apache.org> on 2017/11/15 20:48:24 UTC

[GitHub] szha opened a new issue #8671: Discussion and troubleshooting on PyPI (pip) installation

szha opened a new issue #8671: Discussion and troubleshooting on PyPI (pip) installation
URL: https://github.com/apache/incubator-mxnet/issues/8671
 
 
   PyPI for [Apache MXNet](http://mxnet.apache.org) installation has been available for a while and has been the most important channel for distribution. I'm opening this issue to collect feedbacks, suggestions and issue reports for releases and nightly pre-releases.
   
   ## Current Status
   Supported variants:
   - [mxnet](https://pypi.python.org/pypi/mxnet/) on `Windows`, `Linux`, and `Mac OSX`, with CPU-only without MKL2017 support.
   - [mxnet-mkl](https://pypi.python.org/pypi/mxnet-mkl/) on both `Linux` and `Mac OSX`, with CPU-only `MKL2017` support (with MKL experimental).
   - [mxnet-cu75](https://pypi.python.org/pypi/mxnet-cu75/) on `Linux`, supports `CUDA-7.5`.
   - [mxnet-cu80](https://pypi.python.org/pypi/mxnet-cu80/) on `Windows` and `Linux`, supports `CUDA-8.0`.
   - [mxnet-cu90](https://pypi.python.org/pypi/mxnet-cu90/) on `Windows` and `Linux`, supports `CUDA-9.0`.
   - [mxnet-cu75mkl](https://pypi.python.org/pypi/mxnet-cu75mkl/) on `Linux`, supports `CUDA-7.5` and `MKL2017` (with MKL experimental).
   - [mxnet-cu80mkl](https://pypi.python.org/pypi/mxnet-cu80mkl/) on `Linux`, supports `CUDA-8.0` and `MKL2017` (with MKL experimental).
   - [mxnet-cu90mkl](https://pypi.python.org/pypi/mxnet-cu90mkl/) on `Linux`, supports `CUDA-9.0` and `MKL2017` (with MKL experimental).
   
   (Note: MKL-DNN integration is in progress will likely replace MKL2017.)
   
   ## Usage
   ### Release
   For release, users can install it via
   ```bash
   MXNET_VARIANT=mxnet-cu90 MXNET_VERSION=0.12.0 pip install --upgrade ${MXNET_VARIANT}==${MXNET_VERSION}
   ```
   
   Be sure to update `pip` (`setuptools`) to the most recent version, by `pip install --upgrade pip`.
   
   ### Nightly
   For nightly pre-release, users can install it via `pip install --pre --upgrade`. Be sure to update `pip` (`setuptools`) to the most recent version, by `pip install --upgrade pip`.
   
   ## Report issue
   For bug submissions on issues found in the pre-releases, please attach:
   - MXNet pip package version
   - Environment report:
   ```
   What to do:
   1. Download the diagnosis script from https://raw.githubusercontent.com/apache/incubator-mxnet/master/tools/diagnose.py
   2. Run the script using `python diagnose.py` and paste its output here.
   
   ```
   - Problem code
     - Code snippet to reproduce the issue
     - Code should use dummy data (e.g. `mx.nd.random.uniform`), unless bug is only triggered with specific data value, in which case that specific value should be attached.

----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
users@infra.apache.org


With regards,
Apache Git Services