-
### 🐛 Describe the bug
Capturing CUDA graphs fails with a somewhat unspecific error when using `nn.Embedding` (and back-propagating through it) with batch sizes larger than 3072. I assume that this i…
-
The niiVue graphs currently have a heavy-weight feeling. It would be much more elegant to have"
- thinner lines for plotted lines and axes.
- eliminate the grid (generally the most elegant) or make…
-
Similar to `kDLCPUPinned`, CPU shared memory is allocated and managed quite differently from normal CPU memory. In DGL, we use shared memory to store large graphs so it is quite common to see operatio…
-
[Graph contrastive learning with augmentations](https://proceedings.neurips.cc/paper_files/paper/2020/hash/3fe230348e9a12c13120749e3f9fa4cd-Abstract.html)
```bib
@article{you2020graph,
title={Gra…
-
on relatively large graphs (n > ~10k) with colocated points, the clique expansion function can get really slow. After cutting it short a few times, I think [this for-loop](https://github.com/pysal/lib…
-
[Graph neural bayesian optimization for virtual screening](https://openreview.net/forum?id=t3PzfH98Mq)
```bib
@inproceedings{wang2023graph,
title={Graph neural bayesian optimization for virtual…
-
### Is your feature request related to a problem? Please describe.
We are trying to evaluate how different cell detection algorithms find cells. And so want to find the minimum distance pairing betwe…
-
[Graph backdoor](https://www.usenix.org/conference/usenixsecurity21/presentation/xi)
```bib
@inproceedings{xi2021graph,
title={Graph backdoor},
author={Xi, Zhaohan and Pang, Ren and Ji, Shouli…
-
`test_feed_to_graph_path` itself is the slowest test by far. Create benchmarks and identify which steps are slowest. Find ways to speed up operations and get graph creation process to be as fast as po…
kuanb updated
6 years ago
-
### TLDR
**Rough Proposal**: In order to overlap graph communication with scheduling and computation, we should enable **partial** task-graph execution. Further details are discussed below, but the…