imagenet_example cannot run

[borisdayma]

🐛 Bug

imagenet_example cannot be executed

To Reproduce

Steps to reproduce the behavior:

1. Download imagenet
2. cd pl_examples/full_examples/imagenet
3. python imagenet_example.py --data-path imagenet_path

Error:

Traceback (most recent call last):
  File "imagenet_example.py", line 248, in <module>
    main(get_args())
  File "imagenet_example.py", line 229, in main
    model = ImageNetLightningModel(hparams)
TypeError: Can't instantiate abstract class ImageNetLightningModel with abstract methods forward

Expected behavior

The example should run.

Environment

PyTorch version: 1.3.1
Is debug build: No
CUDA used to build PyTorch: 10.1.243

OS: Ubuntu 18.04.3 LTS
GCC version: (Ubuntu 7.4.0-1ubuntu1~18.04.1) 7.4.0
CMake version: Could not collect

Python version: 3.7
Is CUDA available: Yes
CUDA runtime version: Could not collect
GPU models and configuration: GPU 0: GeForce RTX 2080 Ti
Nvidia driver version: 440.44
cuDNN version: Could not collect

Versions of relevant libraries:
[pip3] numpy==1.18.1
[pip3] pytorch-lightning==0.6.0
[pip3] torch==1.3.1
[pip3] torchvision==0.4.2
[conda] blas 1.0 mkl
[conda] mkl 2019.4 243
[conda] mkl-service 2.3.0 py37he904b0f_0
[conda] mkl_fft 1.0.14 py37ha843d7b_0
[conda] mkl_random 1.1.0 py37hd6b4f25_0

Additional context

I suspect that there was breaking changes in the API and that the example is obsolete as the method forward is now expected.

[williamFalcon]

@neggert @Borda let's remove forward from being abstract?

[HarshSharma12]

I did a local fix for now by updating line 39 from output = self.model(images) to output = self.forward(images) and my forward function is just calling self.model

def forward(self, x):
        return self.model(x)

Also following issues also exist

1. Argument 'seed' must not be None as it breaks tensorboard writer.
2. Argument 'data' might have been updated to 'data_path' recently but there are references to hparams.data in the dataloaders.

After fixing all three, I was able to run the example.

I can create a PR if required.

[williamFalcon]

PR please!

[Rishabh-Choudhry]

I still get this issue when using conda installation of pytorch-lightning

[akihironitta]

@Rishabh-Choudhry Your issue could be a different issue since this issue is quite outdated. Could you create a new issue with your detail?

[gbeltramo]

I get the same error if I try to run the "Train a model (basic)" https://pytorch-lightning.readthedocs.io/en/latest/model/train_model_basic.html code after installing pytorch-lightning with conda install pytorch-lightning -c conda-forge

To fix the problem I had to run

conda remove pytorch-lightning
pip install pytorch-lightning