rusty1s / himp-gnn Goto Github PK

Using PyG 1.5.0 or 1.6.1

got this error during python run:

Traceback (most recent call last):
File "train_tox21.py", line 113, in
if val_perf > best_val_perf:
TypeError: '>' not supported between instances of 'dict' and 'int'

same for Hiv:

Traceback (most recent call last):
File "train_hiv.py", line 113, in
if val_perf > best_val_perf:
TypeError: '>' not supported between instances of 'dict' and 'int'

Hi,

Thanks for the awesome work! When I ran python train_ogbpcba.py

I got the following error:

Namespace(device=0, dropout=0.25, epochs=150, hidden_channels=300, no_inter_message_passing=False, num_layers=3)
Downloading https://snap.stanford.edu/ogb/data/graphproppred/csv_mol_download/pcba.zip
Downloaded 0.04 GB: 100%|███████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 39/39 [00:02<00:00, 14.08it/s]
Extracting data/pcba.zip
Processing...
Loading necessary files...
This might take a while.
Processing graphs...
100%|███████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 437929/437929 [00:02<00:00, 148731.74it/s]
Converting graphs into PyG objects...
100%|███████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 437929/437929 [00:02<00:00, 148342.39it/s]
Traceback (most recent call last):
File "train_ogbpcba.py", line 36, in
dataset = PygGraphPropPredDataset(name, 'data', pre_transform=transform)
File "/home/muhan/anaconda3/envs/seal/lib/python3.8/site-packages/ogb/graphproppred/dataset_pyg.py", line 43, in init
super(PygGraphPropPredDataset, self).init(self.root, transform, pre_transform)
File "/home/muhan/anaconda3/envs/seal/lib/python3.8/site-packages/torch_geometric/data/in_memory_dataset.py", line 53, in init
super(InMemoryDataset, self).init(root, transform, pre_transform,
File "/home/muhan/anaconda3/envs/seal/lib/python3.8/site-packages/torch_geometric/data/dataset.py", line 93, in init
self._process()
File "/home/muhan/anaconda3/envs/seal/lib/python3.8/site-packages/torch_geometric/data/dataset.py", line 166, in _process
self.process()
File "/home/muhan/anaconda3/envs/seal/lib/python3.8/site-packages/ogb/graphproppred/dataset_pyg.py", line 130, in process
data, slices = self.collate(data_list)
File "/home/muhan/anaconda3/envs/seal/lib/python3.8/site-packages/torch_geometric/data/in_memory_dataset.py", line 115, in collate
data[key] = torch.cat(data[key],
RuntimeError: Expected object of scalar type long int but got scalar type float for sequence element 55970.

Could you have a look? I am using torch-geometric 1.5.0, pytorch 1.5.1 and ogb 1.2.1.

Hi,

When I am running train_tox21.py, it shows TypeError: inc() takes 3 positional arguments but 4 were given.
Could you help me with this?

Traceback (most recent call last):
File "/home/znyu/projects/Explainer/train_himp.py", line 109, in
loss = train(epoch)
File "/home/znyu/projects/Explainer/train_himp.py", line 60, in train
for data in train_loader:
File "/home/znyu/anaconda3/envs/pyg202/lib/python3.9/site-packages/torch/utils/data/dataloader.py", line 521, in next
data = self._next_data()
File "/home/znyu/anaconda3/envs/pyg202/lib/python3.9/site-packages/torch/utils/data/dataloader.py", line 561, in _next_data
data = self._dataset_fetcher.fetch(index) # may raise StopIteration
File "/home/znyu/anaconda3/envs/pyg202/lib/python3.9/site-packages/torch/utils/data/_utils/fetch.py", line 52, in fetch
return self.collate_fn(data)
File "/home/znyu/anaconda3/envs/pyg202/lib/python3.9/site-packages/torch_geometric/loader/dataloader.py", line 39, in call
return self.collate(batch)
File "/home/znyu/anaconda3/envs/pyg202/lib/python3.9/site-packages/torch_geometric/loader/dataloader.py", line 19, in collate
return Batch.from_data_list(batch, self.follow_batch,
File "/home/znyu/anaconda3/envs/pyg202/lib/python3.9/site-packages/torch_geometric/data/batch.py", line 69, in from_data_list
batch, slice_dict, inc_dict = collate(
File "/home/znyu/anaconda3/envs/pyg202/lib/python3.9/site-packages/torch_geometric/data/collate.py", line 85, in collate
value, slices, incs = _collate(attr, values, data_list, stores,
File "/home/znyu/anaconda3/envs/pyg202/lib/python3.9/site-packages/torch_geometric/data/collate.py", line 156, in _collate
incs = get_incs(key, values, data_list, stores)
File "/home/znyu/anaconda3/envs/pyg202/lib/python3.9/site-packages/torch_geometric/data/collate.py", line 207, in get_incs
repeats = [
File "/home/znyu/anaconda3/envs/pyg202/lib/python3.9/site-packages/torch_geometric/data/collate.py", line 208, in
data.inc(key, value, store)
TypeError: inc() takes 3 positional arguments but 4 were given

I see in the tox21 example , the evaluation metric is the roc_auc_score on the multilabel binary classification task. However, the final activation function in the model is linear. Does this output interact well with the roc_auc_score? I believe the sklearn implementation of the metric expects an array of shape (n_samples, n_classes) where the values in the array correspond to probability estimates for the labels in each class. As I understand, the current output of the model does not correspond to a probability estimate (linear activation), so does this pose an issue for getting the correct roc_auc value? Would a better implementation be with a sigmoid activation function at the output layer?
Thanks

Is there a reason why your last three lines haven't just been added to your nn in self.atom_convs[i]? It seems like it's basically continuing a sequence that we could have wrapped up in the original nn. Please let me know if I'm missing something.

Hi,
Here is the output:
Traceback (most recent call last):
File "train_muv.py", line 113, in
if val_perf > best_val_perf:
TypeError: '>' not supported between instances of 'dict' and 'int'

and the bug looks like from this line:

Can you provide an example on regression task please ?

• ESOL for example

BR

We noticed that the readout function in the original paper is concatenation, while the one in code implementation is addition.

When I try to reproduce experiments as stated, I invariably get an error:

RuntimeError: received 0 items of ancdata

The error is somewhat erratic, e.g. in train_zinc_subset.py I get it after 15 to 50 epoch in the first run. It seems to have to do with some kind of memory of other saturation I cannot detect.
Other examples in Python Geometric work just fine, so I think it is something specific to the code in himp-gnn.
I am using Python 3.6.9 under WSL2 Ubuntu 18.04.

Hello Matthias,

I was trying to run the command python train_ogbhiv.py, but yield the error below.

Traceback (most recent call last):
  File "himp-molhiv.py", line 96, in <module>
    loss = train(epoch)
  File "himp-molhiv.py", line 61, in train
    out = model(data)
  File "/cmlscratch/kong/anaconda3/envs/graph/lib/python3.8/site-packages/torch/nn/modules/module.py", line 550, in __call__
    result = self.forward(*input, **kwargs)
  File "/nfshomes/kong/projects/ogb/examples/graphproppred/mol/himp-molhiv/model.py", line 134, in forward
    x_clique = self.clique_encoder(data.x_clique.squeeze())
AttributeError: 'Batch' object has no attribute 'x_clique'

Would you please have a look?

BTW I'm using PyG==1.6.1, OGB==1.2.3

What are the order of the coarse-to-fine information flow and the fine-to-coarse information flow?
Which information flow is put ahead of another ?

I am on recent pytorch, so this is kind of to be expected, but just a heads up.

Versions:

dependencies:
  - pytorch=1.6.0=py3.7_cuda10.1.243_cudnn7.6.3_0
pip:
    - torch-cluster==1.5.6
    - torch-geometric==1.6.0
    - torch-scatter==2.0.5
    - torch-sparse==0.6.6
    - torch-spline-conv==1.2.0

Error:

Traceback (most recent call last):
  File "train_ogbhiv.py", line 96, in <module>
    loss = train(epoch)
  File "train_ogbhiv.py", line 58, in train
    for data in train_loader:
  File "/home/sooheon_k_gmail_com/anaconda3/envs/dl/lib/python3.7/site-packages/torch/utils/data/dataloader.py", line 363, in __next__
    data = self._next_data()
  File "/home/sooheon_k_gmail_com/anaconda3/envs/dl/lib/python3.7/site-packages/torch/utils/data/dataloader.py", line 989, in _next_data
    return self._process_data(data)
  File "/home/sooheon_k_gmail_com/anaconda3/envs/dl/lib/python3.7/site-packages/torch/utils/data/dataloader.py", line 1014, in _process_data
    data.reraise()
  File "/home/sooheon_k_gmail_com/anaconda3/envs/dl/lib/python3.7/site-packages/torch/_utils.py", line 395, in reraise
    raise self.exc_type(msg)
RuntimeError: Caught RuntimeError in DataLoader worker process 0.
Original Traceback (most recent call last):
  File "/home/sooheon_k_gmail_com/anaconda3/envs/dl/lib/python3.7/site-packages/torch/utils/data/_utils/worker.py", line 185, in _worker_loop
    data = fetcher.fetch(index)
  File "/home/sooheon_k_gmail_com/anaconda3/envs/dl/lib/python3.7/site-packages/torch/utils/data/_utils/fetch.py", line 47, in fetch
    return self.collate_fn(data)
  File "/home/sooheon_k_gmail_com/anaconda3/envs/dl/lib/python3.7/site-packages/torch_geometric/data/dataloader.py", line 34, in __call__
    return self.collate(batch)
  File "/home/sooheon_k_gmail_com/anaconda3/envs/dl/lib/python3.7/site-packages/torch_geometric/data/dataloader.py", line 15, in collate
    return Batch.from_data_list(batch, self.follow_batch)
  File "/home/sooheon_k_gmail_com/anaconda3/envs/dl/lib/python3.7/site-packages/torch_geometric/data/batch.py", line 53, in from_data_list
    item = item + cum if cum != 0 else item
RuntimeError: Boolean value of Tensor with more than one value is ambiguous

Hi Matthias,

Thanks for providing the implementation to your interesting paper. I've tried running the train_ogbhiv.py script but I'm encountering an error.

This is the output of python train_ogbhiv.py:

Namespace(device=0, dropout=0.5, epochs=50, hidden_channels=64, no_inter_message_passing=False, num_layers=2)

Run 1:

Traceback (most recent call last):
  File "train_ogbhiv.py", line 96, in <module>
    loss = train(epoch)
  File "train_ogbhiv.py", line 58, in train
    for data in train_loader:
  File "/nfs/staff-ssd/zuegnerd/miniconda3/envs/torch/lib/python3.8/site-packages/torch/utils/data/dataloader.py", line 345, in __next__
    data = self._next_data()
  File "/nfs/staff-ssd/zuegnerd/miniconda3/envs/torch/lib/python3.8/site-packages/torch/utils/data/dataloader.py", line 856, in _next_data
    return self._process_data(data)
  File "/nfs/staff-ssd/zuegnerd/miniconda3/envs/torch/lib/python3.8/site-packages/torch/utils/data/dataloader.py", line 881, in _process_data
    data.reraise()
  File "/nfs/staff-ssd/zuegnerd/miniconda3/envs/torch/lib/python3.8/site-packages/torch/_utils.py", line 394, in reraise
    raise self.exc_type(msg)
RuntimeError: Caught RuntimeError in DataLoader worker process 0.
Original Traceback (most recent call last):
  File "/nfs/staff-ssd/zuegnerd/miniconda3/envs/torch/lib/python3.8/site-packages/torch/utils/data/_utils/worker.py", line 178, in _worker_loop
    data = fetcher.fetch(index)
  File "/nfs/staff-ssd/zuegnerd/miniconda3/envs/torch/lib/python3.8/site-packages/torch/utils/data/_utils/fetch.py", line 47, in fetch
    return self.collate_fn(data)
  File "/nfs/staff-ssd/zuegnerd/miniconda3/envs/torch/lib/python3.8/site-packages/torch_geometric/data/dataloader.py", line 34, in __call__
    return self.collate(batch)
  File "/nfs/staff-ssd/zuegnerd/miniconda3/envs/torch/lib/python3.8/site-packages/torch_geometric/data/dataloader.py", line 15, in collate
    return Batch.from_data_list(batch, self.follow_batch)
  File "/nfs/staff-ssd/zuegnerd/miniconda3/envs/torch/lib/python3.8/site-packages/torch_geometric/data/batch.py", line 53, in from_data_list
    item = item + cum if cum != 0 else item
RuntimeError: bool value of Tensor with more than one value is ambiguous

Versions installed:
torch_geometric==1.6.0
torch==1.4.0
ogb==1.2.1

Do you have an idea what's going on and how to fix it?

Thanks again!

Daniel