RuntimeError: Found dtype Double but expected Float

[anthonysirico]

It seems that everytime I fix one thing another error pops up. Once again, here is the code that created the dataset:

import torch
from scipy.sparse import coo_matrix
from torch_geometric.data import Data, Dataset, download_url
import numpy as np
from sklearn.preprocessing import MinMaxScaler
import pandas as pd

def graph_data(A, X, labels):

    tg_graphs = []
    sc = MinMaxScaler()

    y_train = np.array(labels)
    y_train = sc.fit_transform(y_train.reshape((-1,1)))
    y_train = torch.from_numpy(y_train).view(-1,1)

    for i in range(len(A)):
        coo = coo_matrix(A[i])
        indices = np.vstack((coo.row, coo.col))
        x = [ord(i) for i in X[i]]
        index = torch.LongTensor(indices)
        feature = torch.tensor(x, dtype=torch.float)

        graph = Data(x=feature, edge_index=index, y=y_train[i])

        tg_graphs.append(graph)

    return tg_graphs, y_train

Here is the code that creates the model:

import torch

dataset = torch.load('data/dataset.pt')
#%%
data = dataset[0]  # Get the first graph object.

print()
print(data)
print('=============================================================')

# Gather some statistics about the first graph.
print(f'Number of nodes: {data.num_nodes}')
print(f'Number of edges: {data.num_edges}')
print(f'Average node degree: {data.num_edges / data.num_nodes:.2f}')
print(f'Has isolated nodes: {data.has_isolated_nodes()}')
print(f'Has self-loops: {data.has_self_loops()}')
print(f'Is undirected: {data.is_undirected()}')
#%%
train_dataset = dataset[:33000]
test_dataset = dataset[33000:]

print(f'Number of training graphs: {len(train_dataset)}')
print(f'Number of test graphs: {len(test_dataset)}')
#%%
from torch_geometric.loader import DataLoader

train_loader = DataLoader(train_dataset, batch_size=64, shuffle=True)
test_loader = DataLoader(test_dataset, batch_size=64, shuffle=False)

for step, data in enumerate(train_loader):
    print(f'Step {step + 1}:')
    print('=======')
    print(f'Number of graphs in the current batch: {data.num_graphs}')
    print(data)
    print()
#%%
from torch.nn import Linear
import torch.nn.functional as F
from torch_geometric.nn import GCNConv
from torch_geometric.nn import global_mean_pool

class GCN(torch.nn.Module):
    def __init__(self, hidden_channels):
        super(GCN, self).__init__()
        torch.manual_seed(12345)
        self.conv1 = GCNConv(1, hidden_channels)
        self.conv2 = GCNConv(hidden_channels, hidden_channels)
        self.conv3 = GCNConv(hidden_channels, hidden_channels)
        self.lin = Linear(hidden_channels, 1)

    def forward(self, x, edge_index, batch):
        # 1. Obtain node embeddings
        x = self.conv1(x, edge_index)
        x = x.relu()
        x = self.conv2(x, edge_index)
        x = x.relu()
        x = self.conv3(x, edge_index)

        # 2. Readout layer
        x = global_mean_pool(x, batch)  # [batch_size, hidden_channels]

        # 3. Apply a final classifier
        x = F.dropout(x, p=0.5, training=self.training)
        x = self.lin(x)

        return x

model = GCN(hidden_channels=3)
print(model)
#%%
model = GCN(hidden_channels=64)
optimizer = torch.optim.Adam(model.parameters(), lr=0.01)
criterion = torch.nn.functional.mse_loss

def train():
    model.train()

    for data in train_loader.dataset:  # Iterate in batches over the training dataset.
        out = model(data.x.reshape(-1,1), data.edge_index, data.batch)  # Perform a single forward pass.
        loss = criterion(out, data.y)  # Compute the loss.
        loss.backward()  # Derive gradients.
        optimizer.step()  # Update parameters based on gradients.
        optimizer.zero_grad()  # Clear gradients.

def test(loader):
    model.eval()

    correct = 0
    for data in loader:  # Iterate in batches over the training/test dataset.
        out = model(data.x.reshape(-1,1), data.edge_index, data.batch)
        pred = out.argmax(dim=1)  # Use the class with highest probability.
        correct += int((pred == data.y).sum())  # Check against ground-truth labels.
    return correct / len(loader.dataset)  # Derive ratio of correct predictions.
#%%
for epoch in range(1, 171):
    train()
    train_acc = test(train_loader)
    test_acc = test(test_loader)
    print(f'Epoch: {epoch:03d}, Train Acc: {train_acc:.4f}, Test Acc: {test_acc:.4f}')

Here is the error:

C:\Users\Anthony Sirico\AppData\Local\Temp\ipykernel_80248\2205290571.py:10: UserWarning: Using a target size (torch.Size([1])) that is different to the input size (torch.Size([1, 1])). This will likely lead to incorrect results due to broadcasting. Please ensure they have the same size.
  loss = criterion(out, data.y)  # Compute the loss.
RuntimeError                              Traceback (most recent call last)
Input In [7], in <cell line: 1>()
      1 for epoch in range(1, 171):
----> 2     train()
      3     train_acc = test(train_loader)
      4     test_acc = test(test_loader)

Input In [6], in train()
      9 out = model(data.x.reshape(-1,1), data.edge_index, data.batch)  # Perform a single forward pass.
     10 loss = criterion(out, data.y)  # Compute the loss.
---> 11 loss.backward()  # Derive gradients.
     12 optimizer.step()  # Update parameters based on gradients.
     13 optimizer.zero_grad()

File ~\anaconda3\lib\site-packages\torch\_tensor.py:363, in Tensor.backward(self, gradient, retain_graph, create_graph, inputs)
    354 if has_torch_function_unary(self):
    355     return handle_torch_function(
    356         Tensor.backward,
    357         (self,),
   (...)
    361         create_graph=create_graph,
    362         inputs=inputs)
--> 363 torch.autograd.backward(self, gradient, retain_graph, create_graph, inputs=inputs)

File ~\anaconda3\lib\site-packages\torch\autograd\__init__.py:173, in backward(tensors, grad_tensors, retain_graph, create_graph, grad_variables, inputs)
    168     retain_graph = create_graph
    170 # The reason we repeat same the comment below is that
    171 # some Python versions print out the first line of a multi-line function
    172 # calls in the traceback and some print out the last line
--> 173 Variable._execution_engine.run_backward(  # Calls into the C++ engine to run the backward pass
    174     tensors, grad_tensors_, retain_graph, create_graph, inputs,
    175     allow_unreachable=True, accumulate_grad=True)
RuntimeError: Found dtype Double but expected Float

I checked all the dtype for each one:
out.dtype returns torch.float32
loss.dtype returns torch.float64
data.y.dtype returns torch.64

Where could it be reading double?

[rusty1s]

Are you sure that data.y is of correct type? Since you are using torch.nn.functional.mse_loss, the dtype of out and data.y have to match. Can you confirm that this is the case?

You can also use torch.autograd.detect_anomaly to debug this error further.

[anthonysirico]

That did it.