Error with `BackboneFinetuning` and `LearningRateMonitor` with `logging_interval="epoch"`

[ethanwharris]

🐛 Bug

When using the backbone finetuning callback + learning rate moniter with logging interval = epoch an error is thrown after the backbone is unfrozen.

Please reproduce using the BoringModel

import os

import torch
from torch.utils.data import DataLoader, Dataset

from pytorch_lightning import LightningModule, Trainer
from pytorch_lightning.callbacks.finetuning import BackboneFinetuning
from pytorch_lightning.callbacks.lr_monitor import LearningRateMonitor


class RandomDataset(Dataset):

    def __init__(self, size, length):
        self.len = length
        self.data = torch.randn(length, size)

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

    def __len__(self):
        return self.len


class BoringModel(LightningModule):

    def __init__(self):
        super().__init__()
        self.backbone = torch.nn.Sequential(torch.nn.Linear(32, 32), torch.nn.ReLU(True))
        self.layer = torch.nn.Linear(32, 2)

    def forward(self, x):
        return self.layer(self.backbone(x))

    def training_step(self, batch, batch_idx):
        loss = self(batch).sum()
        self.log("train_loss", loss)
        return {"loss": loss}

    def validation_step(self, batch, batch_idx):
        loss = self(batch).sum()
        self.log("valid_loss", loss)

    def test_step(self, batch, batch_idx):
        loss = self(batch).sum()
        self.log("test_loss", loss)

    def configure_optimizers(self):
        opt = torch.optim.Adam(self.layer.parameters(), lr=0.1)
        return [opt], [torch.optim.lr_scheduler.StepLR(opt, step_size=1)]


def run():
    train_data = DataLoader(RandomDataset(32, 64), batch_size=2)
    val_data = DataLoader(RandomDataset(32, 64), batch_size=2)
    test_data = DataLoader(RandomDataset(32, 64), batch_size=2)

    model = BoringModel()
    trainer = Trainer(
        default_root_dir=os.getcwd(),
        limit_train_batches=1,
        limit_val_batches=1,
        num_sanity_val_steps=0,
        max_epochs=6,
        weights_summary=None,
        callbacks=[BackboneFinetuning(unfreeze_backbone_at_epoch=4), LearningRateMonitor(logging_interval="epoch")]
    )
    trainer.fit(model, train_dataloader=train_data, val_dataloaders=val_data)
    trainer.test(model, test_dataloaders=test_data)


if __name__ == '__main__':
    run()

Error:

lr_monitor.py", line 160, in _extract_lr
    self.lrs[name].append(lr)
KeyError: 'lr-Adam/pg1'
Exception ignored in: <function tqdm.__del__ at 0x7ff5e83d7d30>
Traceback (most recent call last):
  File "/Users/ethan/miniconda3/lib/python3.8/site-packages/tqdm/std.py", line 1090, in __del__
  File "/Users/ethan/miniconda3/lib/python3.8/site-packages/tqdm/std.py", line 1303, in close
  File "/Users/ethan/miniconda3/lib/python3.8/site-packages/tqdm/std.py", line 1481, in display
  File "/Users/ethan/miniconda3/lib/python3.8/site-packages/tqdm/std.py", line 1093, in __repr__
  File "/Users/ethan/miniconda3/lib/python3.8/site-packages/tqdm/std.py", line 1443, in format_dict
TypeError: cannot unpack non-iterable NoneType object

Expected behavior

No error

Environment

PL master branch