Mirrors_Framework/deepmind-research
1
0
Fork 0
mirror of https://github.com/google-deepmind/deepmind-research.git synced 2026-05-29 19:55:25 +08:00
Code Issues Actions Packages Projects Releases Wiki Activity

Update arXiv version + citation for OGB-LSC

Browse Source 
PiperOrigin-RevId: 394217992
This commit is contained in:
Petar Veličković
2021-09-01 14:42:15 +01:00
committed by Saran Tunyasuvunakool
parent 769bfdbeaf
commit 08a26772ec
3 changed files with 10 additions and 7 deletions
Download Patch File Download Diff File Expand all files Collapse all files
ogb_lsc/README.md
+2 -1
View File
@@ -5,7 +5,7 @@ This repository contains DeepMind's entry to the [PCQM4M-LSC](https://ogb.stanfo
tracks of the [OGB Large-Scale Challenge](https://ogb.stanford.edu/kddcup2021/) tracks of the [OGB Large-Scale Challenge](https://ogb.stanford.edu/kddcup2021/)
(OGB-LSC). (OGB-LSC).
For full details regarding this entry, please see our [technical report](https://storage.googleapis.com/deepmind-ogb-lsc/reports/OGB_LSC_Tech_Report.pdf). For full details regarding this entry, please see our [technical report](https://arxiv.org/abs/2107.09422).
## Code structure ## Code structure
@@ -25,5 +25,6 @@ To cite this work:
title = {Large-scale graph representation learning with very deep GNNs and title = {Large-scale graph representation learning with very deep GNNs and
self-supervision}, self-supervision},
year = {2021}, year = {2021},
journal={arXiv preprint arXiv:2107.09422},
} }
``` ```
ogb_lsc/mag/README.md
+4 -3
View File
@@ -3,7 +3,7 @@
This repository contains DeepMind's entry to the [MAG240M-LSC](https://ogb.stanford.edu/kddcup2021/mag240m/) (academic graph) track of the This repository contains DeepMind's entry to the [MAG240M-LSC](https://ogb.stanford.edu/kddcup2021/mag240m/) (academic graph) track of the
[OGB Large-Scale Challenge](https://ogb.stanford.edu/kddcup2021/) (OGB-LSC). [OGB Large-Scale Challenge](https://ogb.stanford.edu/kddcup2021/) (OGB-LSC).
For full details regarding this entry, please see our [technical report](https://storage.googleapis.com/deepmind-ogb-lsc/reports/OGB_LSC_Tech_Report.pdf). For full details regarding this entry, please see our [technical report](https://arxiv.org/abs/2107.09422).
## DeepMind MAG Team ("Academic") ## DeepMind MAG Team ("Academic")
@@ -23,7 +23,7 @@ For full details regarding this entry, please see our [technical report](https:/
## Performance ## Performance
Our final test set performance was achieved by pooling an ensemble of 10 folds. Our final test set performance was achieved by pooling an ensemble of 10 folds.
See [technical report](https://storage.googleapis.com/deepmind-ogb-lsc/reports/OGB_LSC_Tech_Report.pdf) for details. See [technical report](https://arxiv.org/abs/2107.09422) for details.
Each model was trained for < 72 hours using 4x Google Cloud TPUv4 and 1x AMD Each model was trained for < 72 hours using 4x Google Cloud TPUv4 and 1x AMD
EPYC 7B12 64-core CPU @2.25GHz. EPYC 7B12 64-core CPU @2.25GHz.
@@ -140,7 +140,8 @@ To cite this work (together with our PCQM4M-LSC entry):
title = {Large-scale graph representation learning with very deep GNNs and title = {Large-scale graph representation learning with very deep GNNs and
self-supervision}, self-supervision},
year = {2021}, year = {2021},
journal={arXiv preprint arXiv:2107.09422},
} }
``` ```
Our technical report can be found [here](https://storage.googleapis.com/deepmind-ogb-lsc/reports/OGB_LSC_Tech_Report.pdf). Our technical report can be found [here](https://arxiv.org/abs/2107.09422).
ogb_lsc/pcq/README.md
+4 -3
View File
@@ -4,7 +4,7 @@ This repository contains DeepMind's entry to the [PCQM4M-LSC](https://ogb.stanfo
track of the [OGB Large-Scale Challenge](https://ogb.stanford.edu/kddcup2021/) track of the [OGB Large-Scale Challenge](https://ogb.stanford.edu/kddcup2021/)
(OGB-LSC). (OGB-LSC).
For full details regarding this entry, please see our [technical report](https://storage.googleapis.com/deepmind-ogb-lsc/reports/OGB_LSC_Tech_Report.pdf). For full details regarding this entry, please see our [technical report](https://arxiv.org/abs/2107.09422).
## DeepMind PCQ Team ("Quantum") ## DeepMind PCQ Team ("Quantum")
@@ -25,7 +25,7 @@ For full details regarding this entry, please see our [technical report](https:/
## Performance ## Performance
Our final test set performance was achieved by pooling an ensemble of 20 models Our final test set performance was achieved by pooling an ensemble of 20 models
(10 folds x 2 seeds). See [technical report](https://storage.googleapis.com/deepmind-ogb-lsc/reports/OGB_LSC_Tech_Report.pdf) for details. (10 folds x 2 seeds). See [technical report](https://arxiv.org/abs/2107.09422) for details.
Each model was trained for < 48 hours using 4x Google Cloud TPUv4 and 1x AMD Each model was trained for < 48 hours using 4x Google Cloud TPUv4 and 1x AMD
EPYC 7B12 64-core CPU @2.25GHz. EPYC 7B12 64-core CPU @2.25GHz.
@@ -110,7 +110,8 @@ To cite this work (together with our MAG240M-LSC entry):
title = {Large-scale graph representation learning with very deep GNNs and title = {Large-scale graph representation learning with very deep GNNs and
self-supervision}, self-supervision},
year = {2021}, year = {2021},
journal={arXiv preprint arXiv:2107.09422},
} }
``` ```
Our technical report can be found [here](https://storage.googleapis.com/deepmind-ogb-lsc/reports/OGB_LSC_Tech_Report.pdf). Our technical report can be found [here](https://arxiv.org/abs/2107.09422).