no val_dataloader when lr_find

[jgsch]

🐛 Bug

To Reproduce

If you want to give the dataloaders as parameters during fitting (so training_step, validation_step are defined but not train_dataloader and val_dataloader), if you want to do a learning rate finder, it return you the following error : pytorch_lightning.utilities.exceptions.MisconfigurationException: You have defined 'validation_step()', but have not passed in a val_dataloader().

Code sample

import os

import torch
from torch.nn import functional as F
from torch.utils.data import DataLoader
from torchvision.datasets import MNIST
from torchvision import transforms

import pytorch_lightning as pl
from pytorch_lightning import Trainer


class LitModel(pl.LightningModule):

    def __init__(self):
        super().__init__()
        self.l1 = torch.nn.Linear(28 * 28, 10)

    def forward(self, x):
        return torch.relu(self.l1(x.view(x.size(0), -1)))

    def training_step(self, batch, batch_idx):
        x, y = batch
        y_hat = self(x)
        loss = F.cross_entropy(y_hat, y)
        tensorboard_logs = {'train_loss': loss}
        return {'loss': loss, 'log': tensorboard_logs}

    def validation_step(self, batch, batch_idx):
        x, y = batch
        y_hat = self(x)
        return {'val_loss': F.cross_entropy(y_hat, y)}

    def validation_epoch_end(self, outputs):
        avg_loss = torch.stack([x['val_loss'] for x in outputs]).mean()
        tensorboard_logs = {'val_loss': avg_loss}
        return {'val_loss': avg_loss, 'log': tensorboard_logs}

    def configure_optimizers(self):
        return torch.optim.Adam(self.parameters(), lr=0.001)


train_dataset = MNIST(os.getcwd(), train=True, download=True, transform=transforms.ToTensor())
train_loader = DataLoader(train_dataset, batch_size=32, num_workers=4, shuffle=True)

val_dataset = MNIST(os.getcwd(), train=False, download=True, transform=transforms.ToTensor())
val_loader = DataLoader(val_dataset, batch_size=32, num_workers=4, shuffle=True)

model = LitModel()
trainer = Trainer(gpus=1)
lr = trainer.lr_find(model, train_loader)

Expected behavior

Simply determines the best learning rate

Environment

* CUDA:
        - GPU:
                - GeForce RTX 2080 Ti
        - available:         True
        - version:           10.1.243
* Packages:
        - numpy:             1.17.4
        - pyTorch_debug:     False
        - pyTorch_version:   1.3.1
        - pytorch-lightning: 0.7.5
        - tensorboard:       2.0.2
        - tqdm:              4.35.0
* System:
        - OS:                Linux
        - architecture:
                - 64bit
                - ELF
        - processor:         x86_64
        - python:            3.6.9
        - version:           #1 SMP Tue Oct 29 08:30:10 EDT 2019

[SkafteNicki]

The quick fix is to make the validation set a part of your model i.e. define the val_dataloader method in the model. However, you are probably right that it should be possible to run the lr_find method without this workaround.