tensorflow warnings triggered by pytorch lightning module

[nimrare]

Hi all

I tried a simple pytorch lightning example (code below). It runs fine, but triggers the following tensorflow warning in my console: W tensorflow/stream_executor/platform/default/dso_loader.cc:59] Could not load dynamic library 'libcudart.so.10.1'; dlerror: libcudart.so.10.1: cannot open shared object file:

Im using:
pytorch-lightning: 1.5.10
pytorch: 1.9.0+cu102
Ubuntu 20.04.3

Do you have an idea why this warning is triggered? pytorch-lightning is also installing tensorboard in the background, as far as i know but still, this should not trigger such a cuda warning. I am a bit worried what parts of tensorflow are invoked and wreaking havoc in the background. Thanks for your help. I did not think this was worth creating an issue since the answer might be quite obvious to some of you.

The code:

import numpy as np
import pytorch_lightning as pl

from torch.utils.data import DataLoader, TensorDataset
from pytorch_lightning.callbacks.early_stopping import EarlyStopping

class PTNNClassifier(pl.LightningModule):
    
    def __init__(self, input_dim, output_dim):
        
        super().__init__()
        
        self.network = torch.nn.Sequential(
            torch.nn.Linear(input_dim, 16),
            torch.nn.ReLU(),
            torch.nn.Linear(16, 12),
            torch.nn.ReLU(),
            torch.nn.Linear(12, output_dim),
            torch.nn.Sigmoid()
        )

    def forward(self, x):
        x = self.network(x)
        return x
    
    def training_step(self, batch, batch_idx):
        x, y  = batch
        y_hat = self.network(x)
        loss  = torch.functional.F.binary_cross_entropy(y_hat, y)
        self.log("train_loss", loss, on_step=True, on_epoch=True, prog_bar=True, logger=True)
        return loss
    
    def validation_step(self, batch, batch_idx):
        x, y  = batch
        y_hat = self.network(x)
        loss  = torch.functional.F.binary_cross_entropy(y_hat, y)
        self.log("val_loss", loss)
        
    def configure_optimizers(self):
        return torch.optim.Adam(self.parameters(), lr=0.02)

from pytorch_lightning.callbacks.early_stopping import EarlyStopping
    
model = PTNNClassifier(input_dim=3, output_dim=1)

early_stop_callback = EarlyStopping(monitor="val_loss", min_delta=0.00, patience=1, verbose=False, mode="min")
trainer = pl.Trainer(max_epochs=10, callbacks=[early_stop_callback])

x_train = np.random.randn(100,3)
y_train = np.random.randn(100,1).round().clip(0,1)

x_test = np.random.randn(100,3)
y_test = np.random.randn(100,1).round().clip(0,1)

train_dataset    = TensorDataset(torch.Tensor(x_train), torch.Tensor(y_train))
train_dataloader = DataLoader(train_dataset, num_workers=16, batch_size=16)

test_dataset    = TensorDataset(torch.Tensor(x_test), torch.Tensor(y_test))
test_dataloader = DataLoader(test_dataset, num_workers=16, batch_size=16)

trainer.fit(model, train_dataloader, test_dataloader)```