-
A unit test in pyTorch that runs the config `convfp16 -n 4097 -c 2 -H 512 -W 512 -k 2 -y 8 -x 8 -p 0 -q 0 -u 8 -v 8 -l 1 -j 1 -m conv -g 1 -F 1 -t 1` gives a seg fault. The log from the test with MIOP…
-
Moving the discussion from [here](https://github.com/coreos/rkt/pull/2735), /cc @sdboyer @sgotti
> > I wish it was easier to distinguish direct deps from transitive deps. I mistakenly tried to remove…
lucab updated
8 years ago
-
export MIOPEN_ENABLE_LOGGING=1
export MIOPEN_ENABLE_LOGGING_CMD=1
export MIOPEN_LOG_LEVEL=6
```
MIOpen(HIP): Info [get_device_name] Raw device name: gfx1030
MIOpen(HIP): Info [Handle] stream: 0, …
-
**Describe the bug**
In Recur.java at line 696 there is a loop to quickly move to the start of the period. If the distance to the start is too long, crossing daylight saving, it will corrupt the star…
-
I really thought I'd solved the endgame algorithm even if it was slow. This endgame does not solve:
```
macondo> endgame 10
plies 10, deepening true, simpleEval false, pruningDisabled false
…
-
Vulnerable Library - lint-gradle-26.5.3.jar
Path to dependency file: /android/app/build.gradle
Path to vulnerable library: /home/wss-scanner/.gradle/caches/modules-2/files-2.1/org.bouncycastle/bcpr…
-
Summary: deterministic selection of deterministic cuDNN convolution algorithms (added in TensorFlow version 1.14) has been broken since version 2.5. This issue explains how it was broken, provides a w…
-
## Hi, there seems to be a bug when I tried to train the transformer model. It also happened when I tried to use Marian-dev.
The log error shows:
[2020-05-03 12:35:12] [config] after-batches: 0…
-
Introduction
----------------
This is an issue to discuss how eigensolvers could be added to Ginkgo.
In general, eigensolvers compute an [eigendecomposition](https://en.wikipedia.org/wiki/Eigen…
-
I've been struggling with some unexpected results in my computations with polynomials over the symbolic ring. Eventually, I came to the conclusion that it should not be possible to *coerce* polynomi…