Error information when I run the, gsimclr.py --DS ENZYMES --lr 0.01 --local --num-gc-layers 3 --aug random4 --seed 0 about graphcl

[yyou1996]

Hi @Austinzhenghua,

Thanks for your feedback. Does torch_geometric==1.7.2 not work for you? You can take a try version 1.6.0/1.6.1 for this experiment.

[Austinzhenghua]

Hi. can I have your we-chat to ask you some more detailed questions? hua zheng ***@***.*** 签名由 网易邮箱大师 定制 On 06/29/2021 21:38, Yuning You wrote: Hi @Austinzhenghua, Thanks for your feedback. Does torch_geometric==1.7.2 not work for you? You can take a try version 1.6.0/1.6.1 for this experiment. — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or unsubscribe.

[yyou1996]

Just for a test, are you capable to run this https://github.com/fanyun-sun/InfoGraph/tree/master/unsupervised which the unsupervised_TU experiment is built on?

[Austinzhenghua]

Just for a test, are you capable to run this https://github.com/fanyun-sun/InfoGraph/tree/master/unsupervised which the unsupervised_TU experiment is built on?

Yes, I can run this algorithm, but it seems it didn't use GPU to train. The error above did cause by the version of torch_geometric. Can you run it in your computrer? Thanks a lot!

[Austinzhenghua]

Traceback (most recent call last):
File "/home/zhenghua/pythoncode/unsupervised_TU_zh/gsimclr.py", line 189, in
emb, y = model.encoder.get_embeddings(dataloader_eval)
File "/home/zhenghua/pythoncode/unsupervised_TU_zh/gin.py", line 77, in get_embeddings
x, _ = self.forward(x, edge_index, batch)
File "/home/zhenghua/pythoncode/unsupervised_TU_zh/gin.py", line 52, in forward
x = F.relu(self.convs[i](x, edge_index))
File "/home/zhenghua/.conda/envs/graphcontra/lib/python3.6/site-packages/torch/nn/modules/module.py", line 532, in call
result = self.forward(*input, **kwargs)
File "/home/zhenghua/.conda/envs/graphcontra/lib/python3.6/site-packages/torch_geometric/nn/conv/gin_conv.py", line 63, in forward
out = self.propagate(edge_index, x=x, size=size)
File "/home/zhenghua/.conda/envs/graphcontra/lib/python3.6/site-packages/torch_geometric/nn/conv/message_passing.py", line 233, in propagate
kwargs)
File "/home/zhenghua/.conda/envs/graphcontra/lib/python3.6/site-packages/torch_geometric/nn/conv/message_passing.py", line 158, in collect
j if arg[-2:] == '_j' else i)
File "/home/zhenghua/.conda/envs/graphcontra/lib/python3.6/site-packages/torch_geometric/nn/conv/message_passing.py", line 127, in lift
return src.index_select(self.node_dim, index)
RuntimeError: index out of range: Tried to access index 4324 out of table with 4323 rows. at /opt/conda/conda-bld/pytorch_1579027003190/work/aten/src/TH/generic/THTensorEvenMoreMath.cpp:418

I run it on the CPU get this error.

[Austinzhenghua]

I find the shape of x is different from your algorithm and infograph. the first one is infograph.

[yyou1996]

It works well on my machine. What is the command u use? Please take a look at readme https://github.com/Shen-Lab/GraphCL/tree/master/unsupervised_TU#readme.

[ztk1996]

Traceback (most recent call last):
File "/home/zhenghua/pythoncode/unsupervised_TU_zh/gsimclr.py", line 189, in
emb, y = model.encoder.get_embeddings(dataloader_eval)
File "/home/zhenghua/pythoncode/unsupervised_TU_zh/gin.py", line 77, in get_embeddings
x, _ = self.forward(x, edge_index, batch)
File "/home/zhenghua/pythoncode/unsupervised_TU_zh/gin.py", line 52, in forward
x = F.relu(self.convs[i](x, edge_index))
File "/home/zhenghua/.conda/envs/graphcontra/lib/python3.6/site-packages/torch/nn/modules/module.py", line 532, in call
result = self.forward(*input, **kwargs)
File "/home/zhenghua/.conda/envs/graphcontra/lib/python3.6/site-packages/torch_geometric/nn/conv/gin_conv.py", line 63, in forward
out = self.propagate(edge_index, x=x, size=size)
File "/home/zhenghua/.conda/envs/graphcontra/lib/python3.6/site-packages/torch_geometric/nn/conv/message_passing.py", line 233, in propagate
kwargs)
File "/home/zhenghua/.conda/envs/graphcontra/lib/python3.6/site-packages/torch_geometric/nn/conv/message_passing.py", line 158, in collect
j if arg[-2:] == '_j' else i)
File "/home/zhenghua/.conda/envs/graphcontra/lib/python3.6/site-packages/torch_geometric/nn/conv/message_passing.py", line 127, in lift
return src.index_select(self.node_dim, index)
RuntimeError: index out of range: Tried to access index 4324 out of table with 4323 rows. at /opt/conda/conda-bld/pytorch_1579027003190/work/aten/src/TH/generic/THTensorEvenMoreMath.cpp:418

I run it on the CPU get this error.

I have the same error. Have you fixed it?

[yyou1996]

Hi @ztk1996,

I remember I tested the command and it worked ok in my machine. Would you also share your environment and the command you run?

[ztk1996]

Hi @ztk1996,

I remember I tested the command and it worked ok in my machine. Would you also share your environment and the command you run?

Thanks for your reply. Error information when I run "./go.sh 1 AIDS subgraph" on CPU is as follows.

• for seed in 0 1 2 3 4
• CUDA_VISIBLE_DEVICES=1
• python gsimclr.py --DS AIDS --lr 0.01 --local --num-gc-layers 3 --aug subgraph --seed 0
  dataset length: 2000
  1
  ================
  lr: 0.01
  num_features: 1
  hidden_dim: 32
  num_gc_layers: 3
  ================
  Traceback (most recent call last):
  File "gsimclr.py", line 188, in
  emb, y = model.encoder.get_embeddings(dataloader_eval)
  File "/home/zt/GraphCL/unsupervised_TU/gin.py", line 89, in get_embeddings
  x, _ = self.forward(x, edge_index, batch)
  File "/home/zt/GraphCL/unsupervised_TU/gin.py", line 62, in forward
  x = F.relu(self.convs[i](x, edge_index))
  File "/home/zt/.conda/envs/GraphCL-test/lib/python3.8/site-packages/torch/nn/modules/module.py", line 889, in _call_impl
  result = self.forward(*input, **kwargs)
  File "/home/zt/.conda/envs/GraphCL-test/lib/python3.8/site-packages/torch_geometric/nn/conv/gin_conv.py", line 64, in forward
  out = self.propagate(edge_index, x=x, size=size)
  File "/home/zt/.conda/envs/GraphCL-test/lib/python3.8/site-packages/torch_geometric/nn/conv/message_passing.py", line 233, in propagate
  coll_dict = self.collect(self.user_args, edge_index, size,
  File "/home/zt/.conda/envs/GraphCL-test/lib/python3.8/site-packages/torch_geometric/nn/conv/message_passing.py", line 157, in collect
  data = self.lift(data, edge_index,
  File "/home/zt/.conda/envs/GraphCL-test/lib/python3.8/site-packages/torch_geometric/nn/conv/message_passing.py", line 127, in lift
  return src.index_select(self.node_dim, index)
  IndexError: index out of range in self

torch: 1.7.0
torch-geometric: 1.7.2

[yyou1996]

@ztk1996

Please take a try to run with torch-geometric==1.6.0 and on GPU. Since both of you use torch-geometric>=1.7.0 and on CPU, I guess it might be the source of error.

[ztk1996]

@ztk1996

Please take a try to run with torch-geometric==1.6.0 and on GPU. Since both of you use torch-geometric>=1.7.0 and on CPU, I guess it might be the source of error.

I try to run with torch_geometric==1.6.0, pytorch==1.7.0 and on GPU. And the error information is as follows.

• for seed in 0 1 2 3 4
• CUDA_VISIBLE_DEVICES=0
• python gsimclr.py --DS AIDS --lr 0.01 --local --num-gc-layers 3 --aug subgraph --seed 0
  dataset length: 2000
  1
  ================
  lr: 0.01
  num_features: 1
  hidden_dim: 32
  num_gc_layers: 3
  ================
  Traceback (most recent call last):
  File "gsimclr.py", line 188, in
  emb, y = model.encoder.get_embeddings(dataloader_eval)
  File "/home/zt/GraphCL/unsupervised_TU/gin.py", line 89, in get_embeddings
  x, _ = self.forward(x, edge_index, batch)
  File "/home/zt/GraphCL/unsupervised_TU/gin.py", line 62, in forward
  x = F.relu(self.convs[i](x, edge_index))
  File "/home/zt/.conda/envs/PYG160/lib/python3.7/site-packages/torch/nn/modules/module.py", line 727, in _call_impl
  result = self.forward(*input, **kwargs)
  File "/home/zt/.conda/envs/PYG160/lib/python3.7/site-packages/torch_geometric/nn/conv/gin_conv.py", line 69, in forward
  return self.nn(out)
  File "/home/zt/.conda/envs/PYG160/lib/python3.7/site-packages/torch/nn/modules/module.py", line 727, in _call_impl
  result = self.forward(*input, **kwargs)
  File "/home/zt/.conda/envs/PYG160/lib/python3.7/site-packages/torch/nn/modules/container.py", line 117, in forward
  input = module(input)
  File "/home/zt/.conda/envs/PYG160/lib/python3.7/site-packages/torch/nn/modules/module.py", line 727, in _call_impl
  result = self.forward(*input, **kwargs)
  File "/home/zt/.conda/envs/PYG160/lib/python3.7/site-packages/torch/nn/modules/linear.py", line 93, in forward
  return F.linear(input, self.weight, self.bias)
  File "/home/zt/.conda/envs/PYG160/lib/python3.7/site-packages/torch/nn/functional.py", line 1690, in linear
  ret = torch.addmm(bias, input, weight.t())
  RuntimeError: CUDA error: CUBLAS_STATUS_NOT_INITIALIZED when calling cublasCreate(handle)
  /opt/conda/conda-bld/pytorch_1603729047590/work/aten/src/ATen/native/cuda/Indexing.cu:658: indexSelectLargeIndex: block: [48,0,0], thread: [89,0,0] Assertion srcIndex < srcSelectDimSize failed.
  /opt/conda/conda-bld/pytorch_1603729047590/work/aten/src/ATen/native/cuda/Indexing.cu:658: indexSelectLargeIndex: block: [48,0,0], thread: [90,0,0] Assertion srcIndex < srcSelectDimSize failed.
  /opt/conda/conda-bld/pytorch_1603729047590/work/aten/src/ATen/native/cuda/Indexing.cu:658: indexSelectLargeIndex: block: [48,0,0], thread: [91,0,0] Assertion srcIndex < srcSelectDimSize failed.
  /opt/conda/conda-bld/pytorch_1603729047590/work/aten/src/ATen/native/cuda/Indexing.cu:658: indexSelectLargeIndex: block: [48,0,0], thread: [92,0,0] Assertion srcIndex < srcSelectDimSize failed.
  /opt/conda/conda-bld/pytorch_1603729047590/work/aten/src/ATen/native/cuda/Indexing.cu:658: indexSelectLargeIndex: block: [48,0,0], thread: [93,0,0] Assertion srcIndex < srcSelectDimSize failed.
  /opt/conda/conda-bld/pytorch_1603729047590/work/aten/src/ATen/native/cuda/Indexing.cu:658: indexSelectLargeIndex: block: [48,0,0], thread: [94,0,0] Assertion srcIndex < srcSelectDimSize failed.
  /opt/conda/conda-bld/pytorch_1603729047590/work/aten/src/ATen/native/cuda/Indexing.cu:658: indexSelectLargeIndex: block: [48,0,0], thread: [95,0,0] Assertion srcIndex < srcSelectDimSize failed.
  /opt/conda/conda-bld/pytorch_1603729047590/work/aten/src/ATen/native/cuda/Indexing.cu:658: indexSelectLargeIndex: block: [48,0,0], thread: [96,0,0] Assertion srcIndex < srcSelectDimSize failed.
  /opt/conda/conda-bld/pytorch_1603729047590/work/aten/src/ATen/native/cuda/Indexing.cu:658: indexSelectLargeIndex: block: [48,0,0], thread: [97,0,0] Assertion srcIndex < srcSelectDimSize failed.
  /opt/conda/conda-bld/pytorch_1603729047590/work/aten/src/ATen/native/cuda/Indexing.cu:658: indexSelectLargeIndex: block: [48,0,0], thread: [98,0,0] Assertion srcIndex < srcSelectDimSize failed.
  /opt/conda/conda-bld/pytorch_1603729047590/work/aten/src/ATen/native/cuda/Indexing.cu:658: indexSelectLargeIndex: block: [48,0,0], thread: [99,0,0] Assertion srcIndex < srcSelectDimSize failed.
  /opt/conda/conda-bld/pytorch_1603729047590/work/aten/src/ATen/native/cuda/Indexing.cu:658: indexSelectLargeIndex: block: [48,0,0], thread: [100,0,0] Assertion srcIndex < srcSelectDimSize failed.
  /opt/conda/conda-bld/pytorch_1603729047590/work/aten/src/ATen/native/cuda/Indexing.cu:658: indexSelectLargeIndex: block: [48,0,0], thread: [101,0,0] Assertion srcIndex < srcSelectDimSize failed.
  /opt/conda/conda-bld/pytorch_1603729047590/work/aten/src/ATen/native/cuda/Indexing.cu:658: indexSelectLargeIndex: block: [48,0,0], thread: [112,0,0] Assertion srcIndex < srcSelectDimSize failed.
  /opt/conda/conda-bld/pytorch_1603729047590/work/aten/src/ATen/native/cuda/Indexing.cu:658: indexSelectLargeIndex: block: [48,0,0], thread: [113,0,0] Assertion srcIndex < srcSelectDimSize failed.
  /opt/conda/conda-bld/pytorch_1603729047590/work/aten/src/ATen/native/cuda/Indexing.cu:658: indexSelectLargeIndex: block: [48,0,0], thread: [114,0,0] Assertion srcIndex < srcSelectDimSize failed.
  /opt/conda/conda-bld/pytorch_1603729047590/work/aten/src/ATen/native/cuda/Indexing.cu:658: indexSelectLargeIndex: block: [48,0,0], thread: [115,0,0] Assertion srcIndex < srcSelectDimSize failed.
  /opt/conda/conda-bld/pytorch_1603729047590/work/aten/src/ATen/native/cuda/Indexing.cu:658: indexSelectLargeIndex: block: [48,0,0], thread: [116,0,0] Assertion srcIndex < srcSelectDimSize failed.
  /opt/conda/conda-bld/pytorch_1603729047590/work/aten/src/ATen/native/cuda/Indexing.cu:658: indexSelectLargeIndex: block: [48,0,0], thread: [117,0,0] Assertion srcIndex < srcSelectDimSize failed.
  /opt/conda/conda-bld/pytorch_1603729047590/work/aten/src/ATen/native/cuda/Indexing.cu:658: indexSelectLargeIndex: block: [48,0,0], thread: [118,0,0] Assertion srcIndex < srcSelectDimSize failed.
  /opt/conda/conda-bld/pytorch_1603729047590/work/aten/src/ATen/native/cuda/Indexing.cu:658: indexSelectLargeIndex: block: [48,0,0], thread: [119,0,0] Assertion srcIndex < srcSelectDimSize failed.

Besides, when I run with torch_geometric==1.6.0, pytorch==1.7.0 and on CPU. The error information is the same as run with torch_geometric==1.7.2.

[yyou1996]

@ztk1996

My impression is that the version of torch_geometric and pytorch should be consistent (https://github.com/rusty1s/pytorch_geometric)? If using torch_geometric==1.6 I would also use pytorch==1.6. Please notify me if this also not works. Thanks.