AttributeError: 'Trainer' object has no attribute 'run_evaluation'

[hassiahk]

I am getting the below error when running trainer.fit:
AttributeError: 'Trainer' object has no attribute 'run_evaluation'

Full traceback:

Traceback (most recent call last):
  File "sdr_main.py", line 81, in <module>
    main()
  File "sdr_main.py", line 28, in main
    main_train(model_class_pointer, hyperparams,parser)
  File "sdr_main.py", line 73, in main_train
    trainer.fit(model)
  File "/usr/local/lib/python3.7/dist-packages/pytorch_lightning/trainer/trainer.py", line 741, in fit
    self._fit_impl, model, train_dataloaders, val_dataloaders, datamodule, ckpt_path
  File "/usr/local/lib/python3.7/dist-packages/pytorch_lightning/trainer/trainer.py", line 685, in _call_and_handle_interrupt
    return trainer_fn(*args, **kwargs)
  File "/usr/local/lib/python3.7/dist-packages/pytorch_lightning/trainer/trainer.py", line 777, in _fit_impl
    self._run(model, ckpt_path=ckpt_path)
  File "/usr/local/lib/python3.7/dist-packages/pytorch_lightning/trainer/trainer.py", line 1199, in _run
    self._dispatch()
  File "/usr/local/lib/python3.7/dist-packages/pytorch_lightning/trainer/trainer.py", line 1279, in _dispatch
    self.training_type_plugin.start_training(self)
  File "/usr/local/lib/python3.7/dist-packages/pytorch_lightning/plugins/training_type/training_type_plugin.py", line 202, in start_training
    self._results = trainer.run_stage()
  File "/usr/local/lib/python3.7/dist-packages/pytorch_lightning/trainer/trainer.py", line 1289, in run_stage
    return self._run_train()
  File "/usr/local/lib/python3.7/dist-packages/pytorch_lightning/trainer/trainer.py", line 1319, in _run_train
    self.fit_loop.run()
  File "/usr/local/lib/python3.7/dist-packages/pytorch_lightning/loops/base.py", line 140, in run
    self.on_run_start(*args, **kwargs)
  File "/usr/local/lib/python3.7/dist-packages/pytorch_lightning/loops/fit_loop.py", line 200, in on_run_start
    self.trainer.call_hook("on_train_start")
  File "/usr/local/lib/python3.7/dist-packages/pytorch_lightning/trainer/trainer.py", line 1495, in call_hook
    callback_fx(*args, **kwargs)
  File "/usr/local/lib/python3.7/dist-packages/pytorch_lightning/trainer/callback_hook.py", line 138, in on_train_start
    callback.on_train_start(self, self.lightning_module)
  File "/content/SDR/utils/pytorch_lightning_utils/callbacks.py", line 10, in on_train_start
    return trainer.run_evaluation()
AttributeError: 'Trainer' object has no attribute 'run_evaluation'

My Trainer object:

trainer = pytorch_lightning.Trainer(
    num_sanity_val_steps=2,
    gradient_clip_val=hparams.max_grad_norm,
    callbacks=[RunValidationOnStart()],
    checkpoint_callback=ModelCheckpoint(
        save_top_k=3,
        save_last=True,
        mode="min" if "acc" not in hparams.metric_to_track else "max",
        monitor=hparams.metric_to_track,
        dirpath=model.hparams.hparams_dir,
        filename="{epoch}",
        verbose=True,
    ),
    logger=logger,
    max_epochs=hparams.max_epochs,
    gpus=hparams.gpus,
    strategy="dp",
    limit_val_batches=hparams.limit_val_batches,
    limit_train_batches=hparams.limit_train_batches,
    limit_test_batches=hparams.limit_test_batches,
    check_val_every_n_epoch=hparams.check_val_every_n_epoch,
    profiler=SimpleProfiler(),
    accumulate_grad_batches=hparams.accumulate_grad_batches,
    reload_dataloaders_every_epoch=True,
    resume_from_checkpoint=hparams.resume_from_checkpoint,
)

Any idea on how to fix this? My pytorch-lightning version is 1.5.10

[akihironitta]

  File "/content/SDR/utils/pytorch_lightning_utils/callbacks.py", line 10, in on_train_start
    return trainer.run_evaluation()
AttributeError: 'Trainer' object has no attribute 'run_evaluation'

@hassiahk What are you trying to evaluate there? Could I see what your callback looks like?

[hassiahk]

@akihironitta The only defined callback is RunValidationOnStart(). I have added my Trainer object above if you need anymore info.

class RunValidationOnStart(Callback):
    def __init__(self):
        pass

    def on_train_start(self, trainer: Trainer, pl_module):
        return trainer.run_evaluation()

[rohitgr7]

hey !

this was removed in the previous release.

You can try:

trainer.validating = True
trainer.reset_val_dataloader()
trainer.val_loop.run()
trainer.training = True

[rohitgr7]

try: trainer.fit_loop.epoch_loop.val_loop.run()

[hassiahk]

Thanks! That totally worked. :)

[ananthsub]

@hassiahk Why not call trainer.validate() first before calling trainer.fit() ?

[rohitgr7]

well, that was what I was thinking, but I wasn't sure what else the callback is doing.
but if all you need is to validate your model before fit @hassiahk , then @ananthsub solution is better :)

[hassiahk]

Thanks for your suggestions! I guess I can but right now I am only trying to run someone's code without any errors. They seemed to have used PL 1.1.8 and I am trying to use PL 1.5.10, so just fixing the errors for the time being.