Inputting two different graphs not working

[mandelbrot-walker]

I am trying to load two different graphs following the solution mentioned in issue 781. However, my batching is failing due to dim mismatch as mentioned here issue 807. Very much new to PyG, therefore, can't figure out why this solution is not working where as it has worked for others and still trying to wrap my head around the details. I also tried PairData mentioned in the document. However, that solution is also not working.
Here's my code and error.

class PairDataset(torch.utils.data.Dataset):
    def __init__(self, datasetA, datasetB):
        assert len(datasetA) == len(datasetB), 
        self.datasetA = datasetA
        self.datasetB = datasetB
    def __len__(self):
        return len(self.datasetA)
    
    def __getitem__(self, idx):
        return self.datasetA[idx], self.datasetB[idx]

 def collate(data_list):
    print(data_list)
    batchA = Batch.from_data_list([data[0] for data in data_list])
    batchB = Batch.from_data_list([data[1] for data in data_list])
    return batchA, batchB

 pair_dataset = PairDataset(tweets,evidence)
 train_loader = torch.utils.data.DataLoader(pair_dataset, batch_size=4, shuffle=True, collate_fn=collate)

 for batchA, batchB in train_loader:
    print("Processing batch...")

This is the data list being passed

[(Data(edge_index=[2, 10], y=0, x=[11, 2048]), Data(x=[10, 128], edge_index=[18, 2])), 
(Data(edge_index=[2, 10], y=0, x=[11, 2048]), Data(x=[10, 128], edge_index=[18, 2])), 
(Data(edge_index=[9, 2], y=1, x=[10, 2048]), Data(x=[9, 128], edge_index=[2, 16])), 
(Data(edge_index=[2, 9], y=0, x=[10, 2048]), Data(x=[9, 128], edge_index=[16, 2]))]

This is the error
RuntimeError: Sizes of tensors must match except in dimension 1. Expected size 2 but got size 9 for tensor number 2 in the list.

Also this was my implementation for PairDataset using documentation

train_list = []
for f in graph_data:
    data_tweets = f[0]
    data_evidence = f[1]
    data_tweets.edge_index = data_tweets.edge_index.t().contiguous()
    data_evidence.edge_index = data_evidence.edge_index.t().contiguous()                     

    if len(data_tweets.edge_index.size()) == 1:
        continue
    if len(data_evidence.edge_index.size()) == 1:
        continue
    
    if (len(data_tweets.edge_index[0]) > len(data_evidence.edge_index[0])):
        x_s = data_tweets
        x_t = data_evidence
        edge_index_s = data_tweets.edge_index
        edge_index_t = data_evidence.edge_index
        data = PairData(x_s=x_s, edge_index_s=edge_index_s,x_t=x_t, edge_index_t=edge_index_t)
    else:
        x_s = data_evidence
        x_t = data_tweets
        edge_index_s = data_evidence.edge_index
        edge_index_t = data_tweets.edge_index
        data = PairData(x_s=x_s, edge_index_s=edge_index_s,x_t=x_t, edge_index_t=edge_index_t)

    train_list.append(data)

train_loader = geometric_loader(train_list, batch_size=4, follow_batch=['x_s', 'x_t'])
for data in train_loader:
    print("Processing batch...")

Appreciate any suggestions on how to handle this. Thanks.