Does lightning output not go through stdout? NeptuneLogger logs it only partially.

[NiklasKappel]

The neptune logger by default logs everything that goes to sdtout and stderr between the calls to run = neptune.init_run() and run.stop(). But when I run the following mwe

import lightning as L
import torch
from pytorch_lightning.loggers import NeptuneLogger
from torch.utils.data import DataLoader, Dataset


class MyDataset(Dataset):
    def __init__(self):
        super().__init__()
        self.data = torch.rand(100, 3)

    def __len__(self):
        return len(self.data)

    def __getitem__(self, idx):
        return self.data[idx]


class MyDataModule(L.LightningDataModule):
    def __init__(self, num_workers=7):
        super().__init__()
        self.num_workers = num_workers

    def setup(self, stage=None):
        del stage  # unused
        self.train_dataset = MyDataset()

    def train_dataloader(self):
        return DataLoader(self.train_dataset, num_workers=self.num_workers)


class MyModel(L.LightningModule):
    def __init__(self):
        super().__init__()
        self.layer = torch.nn.Linear(3, 1)

    def forward(self, x):  # type: ignore
        return self.layer(x)

    def training_step(self, batch):  # type: ignore
        loss = self(batch).mean()
        self.log("train_loss", loss)
        return loss

    def configure_optimizers(self):
        return torch.optim.Adam(self.parameters())


def main():
    model = MyModel()
    data_module = MyDataModule()
    neptune_logger = NeptuneLogger(
        project="niklaskappel/test",
        api_key="asdf",
    )
    trainer = L.Trainer(logger=neptune_logger, max_epochs=1)  # type: ignore
    trainer.fit(model, data_module)


if __name__ == "__main__":
    main()

the output on the console I see is

GPU available: False, used: False
TPU available: False, using: 0 TPU cores
IPU available: False, using: 0 IPUs
HPU available: False, using: 0 HPUs
https://app.neptune.ai/niklaskappel/test/e/TES-1
Info (NVML): NVML Shared Library Not Found. GPU usage metrics may not be reported. For more information, see https://docs.neptune.ai/help/nvml_error/

  | Name  | Type   | Params
---------------------------------
0 | layer | Linear | 4     
---------------------------------
4         Trainable params
0         Non-trainable params
4         Total params
0.000     Total estimated model params size (MB)
Epoch 0: 100%|████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 100/100 [00:00<00:00, 245.75it/s, v_num=ES-1]`Trainer.fit` stopped: `max_epochs=1` reached.
Epoch 0: 100%|████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 100/100 [00:00<00:00, 156.28it/s, v_num=ES-1]
Shutting down background jobs, please wait a moment...
Done!
Waiting for the remaining 13 operations to synchronize with Neptune. Do not kill this process.
All 13 operations synced, thanks for waiting!
Explore the metadata in the Neptune app:
https://app.neptune.ai/niklaskappel/test/e/TES-1/metadata

but in the neptune app, only the following is logged for stdout

2024/01/26 11:26:00	Training: | | 0/? [00:00<?, ?it/s]
2024/01/26 11:26:00	Training: 0%| | 0/100 [00:00<?, ?it/s]
2024/01/26 11:26:00	Epoch 0: 0%| | 0/100 [00:00<?, ?it/s]
2024/01/26 11:26:00	Epoch 0: 1%|█▋ | 1/100 [00:00<00:01, 53.08it/s]
2024/01/26 11:26:00	Epoch 0: 1%|█▌ | 1/100 [00:00<00:01, 51.60it/s, v_num=ES-1]
2024/01/26 11:26:00	Epoch 0: 2%|███ | 2/100 [00:00<00:01, 84.87it/s, v_num=ES-1]
2024/01/26 11:26:00	Epoch 0: 2%|███ | 2/100 [00:00<00:01, 82.99it/s, v_num=ES-1]
# many more progress bars
2024/01/26 11:26:01	Epoch 0: 100%|████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 100/100 [00:00<00:00, 156.28it/s, v_num=ES-1]
2024/01/26 11:26:01	
2024/01/26 11:26:01	Shutting down background jobs, please wait a moment... 

and for stderr

2024/01/26 11:25:59	Info (NVML): NVML Shared Library Not Found. GPU usage metrics may not be reported. For more information, see https://docs.neptune.ai/help/nvml_error/

Clearly, the e.g. the model summery is missing from the stdout log, even though the logger must be active before, as can be seen from the stderr log. Is lightning somehow special about how it prints output to the console?