关于GCU、沐曦GPGPU、MLU、0卡V100资源4月7日恢复上架的公告>>> 关于共建具身智能开源数据集的倡议>>> 关于云脑任务中统一路径访问方式的公告>>> 关于将启智集群GPU资源迁移至智算集群的公告>>>
zhanghongyi
  • Joined on Dec 14, 2021
  • Organization
Public Activity Repositories Datasets Starred Repositories
3
 Following
0
 Followers
1
 Achievement Badge
1
Loading Heatmap…

zhanghongyi synced commits to master at zhanghongyi/dgl from mirror

  • b68a2fc069 Revert "[GraphBolt] Quick fix for GCC 8" (#7328)
  • d853a010e0 [GraphBolt] Enable `DistributedItemSampler` test only on CPU. (#7323)
  • 9a340c77b4 [GraphBolt][CUDA] Add missing thrust include for later CCCL versions. (#7325)
  • 0d62d45210 [Misc] Add Triston to jenkins allowlist. (#7327) Co-authored-by: Ubuntu <ubuntu@ip-172-31-28-63.ap-northeast-1.compute.internal>
  • a4e19691d5 [Determinism] Enable environment var to use cusparse spmm deterministic algorithm (#7310) Co-authored-by: Hongzhi (Steve), Chen <chenhongzhi.nkcs@gmail.com>
  • Compare 18 commits »

1 hour ago

zhanghongyi synced commits to gnn-llm-model-integration at zhanghongyi/pytorch_geometric from mirror

  • ad3a841283 Merge branch 'gnn-llm-model-integration' of https://github.com/pyg-team/pytorch_geometric into gnn-llm-model-integration
  • dac44f0b31 cleanup
  • 4d37e0c9cf [pre-commit.ci] auto fixes from pre-commit.com hooks for more information, see https://pre-commit.ci
  • c37df9aeb4 Merge branch 'gnn-llm-model-integration' of https://github.com/pyg-team/pytorch_geometric into gnn-llm-model-integration
  • b6d194858b cleanup
  • Compare 17 commits »

1 day ago

zhanghongyi synced commits to master at zhanghongyi/pytorch_geometric from mirror

  • e2a69a15ba Fix typo in docstring of sequential.py (#9204) There was a small error in the docstring of sequential. The ordering of the (model, function header) tuple was inconsistent.
  • 82aad03c0c Ignore `edge_label` and `edge_label_index` in `ToSparseTensor()` transform (#9199) Fixes https://github.com/pyg-team/pytorch_geometric/issues/9196
  • fcb9c0f0b5 Do not set jinja `propagate` method in `MessagePassing.__init__` twice (#9198)
  • Compare 3 commits »

2 days ago

zhanghongyi synced commits to gnn-llm-model-integration at zhanghongyi/pytorch_geometric from mirror

2 days ago

zhanghongyi synced commits to master at zhanghongyi/dgl from mirror

  • 7612af0f18 [GraphBolt] Remove old version negative sampler. (#7302) Co-authored-by: Ubuntu <ubuntu@ip-172-31-0-133.us-west-2.compute.internal>
  • 21aeed86e3 [GraphBolt] Update docstring of `itemset.py`. (#7303) Co-authored-by: Ubuntu <ubuntu@ip-172-31-0-133.us-west-2.compute.internal>
  • 15ace31a16 [Misc] Using the aliases of builtin types like np.int is deprecated (#7300)
  • Compare 3 commits »

3 days ago

zhanghongyi synced commits to gnn-llm-model-integration at zhanghongyi/pytorch_geometric from mirror

4 days ago

zhanghongyi synced commits to master at zhanghongyi/dgl from mirror

  • 2ff3006c35 [CUDA][Bug] CSR transpose bug in CUDA 12 (#7295)
  • 20e5e26697 [GraphBolt] Remove old version exclude edges. (#7299) Co-authored-by: Ubuntu <ubuntu@ip-172-31-0-133.us-west-2.compute.internal>
  • 279ebb1563 [GraphBolt][CUDA] Link prediction `--mode=cuda-cuda` fix. (#7294)
  • Compare 3 commits »

6 days ago

zhanghongyi synced commits to master at zhanghongyi/pytorch_geometric from mirror

  • 1cb7b04c01 Fix `EgoData` processing in `SnapDataset` in case file names are unsorted; conver to sparse representation if necessary (#9195) Fixes #9176
  • ab6852c77b Fix `torch.sparse.check_sparse_tensor_invariants()` check (#9194)
  • bbf0e1d6e8 style: update ruff `quote-style` to single (#9163) Ref: #9153 Also fixes `ruff` warning, viz. ```lang-bash warning: The top-level linter settings are deprecated in favour of their counterparts in the `lint` section. Please update the following options in `pyproject.toml`: - 'ignore' -> 'lint.ignore' - 'select' -> 'lint.select' ``` This PR doesn't include the actual changes `ruff` might yield, just the change to configuration. Request for Review: @rusty1s
  • 20b8f9e04f Align `pyproject.toml` and PyPi package name (#9193) Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>
  • 679c2319f9 Remove future warnings generated in `argva_node_clustering.py` example (#9187) With this fix, the next warning that appears for each epoch: ``` /usr/local/lib/python3.10/dist-packages/sklearn/cluster/_kmeans.py:870: FutureWarning: The default value of `n_init` will change from 10 to 'auto' in 1.4. Set the value of `n_init` explicitly to suppress the warning warnings.warn( ``` will no longer be generated. --------- Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>
  • Compare 10 commits »

1 week ago

zhanghongyi synced commits to gnn-llm-model-integration at zhanghongyi/pytorch_geometric from mirror

1 week ago

zhanghongyi synced commits to master at zhanghongyi/dgl from mirror

  • f0f424f70c [GraphBolt][CUDA] Better `TORCH_CUDA_ARCH_LIST` initialization. (#7293)
  • c264a4109a [Misc] rename a variable (#7260)
  • 0bd79a2bd9 [GraphBolt] Update name used in `LegacyDataset`. (#7283) Co-authored-by: Ubuntu <ubuntu@ip-172-31-0-133.us-west-2.compute.internal>
  • Compare 3 commits »

1 week ago

zhanghongyi synced commits to master at zhanghongyi/pytorch_geometric from mirror

  • e213c297bb Use `message_and_aggregate` in `MessagePassing` for `EdgeIndex` (#9131) Also enforce sort by `row` in spmm --------- Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com> Co-authored-by: rusty1s <matthias.fey@tu-dortmund.de>

1 week ago

zhanghongyi synced commits to gnn-llm-model-integration at zhanghongyi/pytorch_geometric from mirror

1 week ago

zhanghongyi synced commits to g-retriever-w-demo at zhanghongyi/pytorch_geometric from mirror

  • b87eac67b2 Merge branch 'master' into g-retriever-w-demo
  • e213c297bb Use `message_and_aggregate` in `MessagePassing` for `EdgeIndex` (#9131) Also enforce sort by `row` in spmm --------- Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com> Co-authored-by: rusty1s <matthias.fey@tu-dortmund.de>
  • Compare 2 commits »

1 week ago

zhanghongyi synced commits to master at zhanghongyi/dgl from mirror

  • d9caeaaafc [GraphBolt] Update dataset names that exposed to users in examples. (#7280) Co-authored-by: Ubuntu <ubuntu@ip-172-31-0-133.us-west-2.compute.internal>
  • e9661d3b36 [CUDA] Add Ada architecture to DGL architecture list. (#7282)
  • 7b470eb0fa [GraphBolt] Quick fix for GCC 8 (#7279)
  • e847fc44ff [GraphBolt] Update notebooks about how to construct ondisk datasets. (#7268) Co-authored-by: Ubuntu <ubuntu@ip-172-31-0-133.us-west-2.compute.internal>
  • 78df81015a [GraphBolt] Add optimized `unique_and_compact_batched`. (#7239)
  • Compare 10 commits »

1 week ago

zhanghongyi synced commits to gnn-llm-model-integration at zhanghongyi/pytorch_geometric from mirror

1 week ago

zhanghongyi synced commits to master at zhanghongyi/dgl from mirror

  • 62aca92da5 [Dev] add contributor to CI authority (#7267) Co-authored-by: Ubuntu <ubuntu@ip-172-31-10-164.us-west-2.compute.internal>
  • c2ffe3bfe4 [GraphBolt] Update notebook examples. (#7266) Co-authored-by: Ubuntu <ubuntu@ip-172-31-0-133.us-west-2.compute.internal>
  • d3483fe1f1 [GraphBolt][CUDA] Optimize hetero sampling. (#7223)
  • 2a00cd3da4 [Misc] Update ondisk_dataset_heterograph.ipynb (#7261)
  • 12f8348075 [GraphBolt] PyG advanced example `torch.compile()` bug workaround. (#7259)
  • Compare 6 commits »

1 week ago

zhanghongyi synced commits to master at zhanghongyi/pytorch_geometric from mirror

  • 38bb5f2937 Flush template file before closing (#9151)
  • 2db8aaf5f3 feat: add `CornellTemporalHyperGraphDatasets` (#9090) Reference: #8501 #7312 Request for Review: @rusty1s @wsad1 This PR aims to add HyperGraph datasets consisting of timestamped simplices where each simplex is a set of nodes. Released with the paper [Simplicial Closure and higher-order link prediction](https://huggingface.co/papers/1802.06916). Notes: > "[In compliance with PEP 8 and PEP 257, Ruff prefers double quotes for triple quoted strings and docstrings even when using quote-style = "single".](https://docs.astral.sh/ruff/settings/#format_quote-style)". Thus even with pre-commit passing there are changes which convert `'` to `"`. Should probably look into setting the `quote-style` property to `"single"`, if single quotes are desired. That way it can be enforced via pre-commit as well. --- _**Update 26th March 2024**: There are multiple datasets which follow a similar processing and loading workflow. Therefore, I have refactored to make a single `CornellTemporalHyperGraphDataset` class. I'm still encountering errors with the test on windows, would appreciate some help on this_ --------- Co-authored-by: rusty1s <matthias.fey@tu-dortmund.de>
  • 2aee5e0942 Fix bug in `to_scipy_sparse_matrix` (#9146) Setting the default device in PyTorch to "cuda" can cause a bug in the `to_scipy_sparse_matrix` function. Code to reproduce: ```python import torch from torch_geometric.utils import to_scipy_sparse_matrix torch.set_default_device("cuda") edge_index = torch.tensor([ [0, 1, 1, 2, 2, 3], [1, 0, 2, 1, 3, 2], ]) to_scipy_sparse_matrix(edge_index) ``` All tensor input to the function is always copied to the CPU for further processing. However, if `edge_attr` argument is **not** given to the `to_scipy_sparse_matrix` function, a tensor of ones is created instead. No device argument is specified currently for the creation of this tensor, causing it to be created on the GPU if that is the default setting. Almost directly after, `edge_attr.numpy()` is called, causing the following error: ``` TypeError: can't convert cuda:0 device type tensor to numpy. Use Tensor.cpu() to copy the tensor to host memory first. ``` This pull request simply adds the `device="cpu"` argument to the creation of the `edge_attr` tensor to prevent this.
  • 96d723f16a Fix `MetaPath2Vec` in case the last node is isolated (#9145) Fixes #9138
  • Compare 4 commits »

1 week ago

zhanghongyi synced commits to g-retriever at zhanghongyi/pytorch_geometric from mirror

  • 49b3bfd49a Merge branch 'master' into g-retriever
  • 38bb5f2937 Flush template file before closing (#9151)
  • 2db8aaf5f3 feat: add `CornellTemporalHyperGraphDatasets` (#9090) Reference: #8501 #7312 Request for Review: @rusty1s @wsad1 This PR aims to add HyperGraph datasets consisting of timestamped simplices where each simplex is a set of nodes. Released with the paper [Simplicial Closure and higher-order link prediction](https://huggingface.co/papers/1802.06916). Notes: > "[In compliance with PEP 8 and PEP 257, Ruff prefers double quotes for triple quoted strings and docstrings even when using quote-style = "single".](https://docs.astral.sh/ruff/settings/#format_quote-style)". Thus even with pre-commit passing there are changes which convert `'` to `"`. Should probably look into setting the `quote-style` property to `"single"`, if single quotes are desired. That way it can be enforced via pre-commit as well. --- _**Update 26th March 2024**: There are multiple datasets which follow a similar processing and loading workflow. Therefore, I have refactored to make a single `CornellTemporalHyperGraphDataset` class. I'm still encountering errors with the test on windows, would appreciate some help on this_ --------- Co-authored-by: rusty1s <matthias.fey@tu-dortmund.de>
  • e58eeb484a Merge branch 'master' into g-retriever
  • 2aee5e0942 Fix bug in `to_scipy_sparse_matrix` (#9146) Setting the default device in PyTorch to "cuda" can cause a bug in the `to_scipy_sparse_matrix` function. Code to reproduce: ```python import torch from torch_geometric.utils import to_scipy_sparse_matrix torch.set_default_device("cuda") edge_index = torch.tensor([ [0, 1, 1, 2, 2, 3], [1, 0, 2, 1, 3, 2], ]) to_scipy_sparse_matrix(edge_index) ``` All tensor input to the function is always copied to the CPU for further processing. However, if `edge_attr` argument is **not** given to the `to_scipy_sparse_matrix` function, a tensor of ones is created instead. No device argument is specified currently for the creation of this tensor, causing it to be created on the GPU if that is the default setting. Almost directly after, `edge_attr.numpy()` is called, causing the following error: ``` TypeError: can't convert cuda:0 device type tensor to numpy. Use Tensor.cpu() to copy the tensor to host memory first. ``` This pull request simply adds the `device="cpu"` argument to the creation of the `edge_attr` tensor to prevent this.
  • Compare 6 commits »

1 week ago

zhanghongyi synced commits to master at zhanghongyi/dgl from mirror

  • b743cdef71 [GraphBolt] Hack for original `seed_nodes` and `node_pairs`. (#7248) Co-authored-by: Ubuntu <ubuntu@ip-172-31-0-133.us-west-2.compute.internal>
  • 3df6e30180 [GraphBolt] Update `__len__()` and `__getitem__()` of `ItemSet` (#7253)
  • Compare 2 commits »

2 weeks ago

zhanghongyi synced commits to master at zhanghongyi/pytorch_geometric from mirror

  • cbfd1dbcc8 Fix parallelism for `OnDiskDataset` (#9140) Co-authored-by: rusty1s <matthias.fey@tu-dortmund.de>
  • 24ec11d1c0 [pre-commit.ci] pre-commit suggestions (#9137) <!--pre-commit.ci start--> updates: - [github.com/adrienverge/yamllint.git: v1.33.0 → v1.35.1](https://github.com/adrienverge/yamllint.git/compare/v1.33.0...v1.35.1) - [github.com/PyCQA/flake8: 6.1.0 → 7.0.0](https://github.com/PyCQA/flake8/compare/6.1.0...7.0.0) - [github.com/astral-sh/ruff-pre-commit: v0.1.9 → v0.3.5](https://github.com/astral-sh/ruff-pre-commit/compare/v0.1.9...v0.3.5) <!--pre-commit.ci end--> --------- Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com> Co-authored-by: rusty1s <matthias.fey@tu-dortmund.de>
  • Compare 2 commits »

2 weeks ago

Please read the following content carefully:

Dear OpenI User

Thank you for your continuous support to the Openl Qizhi Community AI Collaboration Platform. In order to protect your usage rights and ensure network security, we updated the Openl Qizhi Community AI Collaboration Platform Usage Agreement in January 2024. The updated agreement specifies that users are prohibited from using intranet penetration tools. After you click "Agree and continue", you can continue to use our services. Thank you for your cooperation and understanding.

For more agreement content, please refer to the《Openl Qizhi Community AI Collaboration Platform Usage Agreement》

Agree and continue
Disagree, exit