Rooba
/
lightning
mirror of https://github.com/Lightning-AI/lightning.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

fix warning (#3800)

This commit is contained in:
Jirka Borovec 2020-10-02 19:51:02 +02:00 committed by GitHub
parent 0c12065efd
commit 22efce8f40
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
1 changed files with 4 additions and 14 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

18
pytorch_lightning/overrides/data_parallel.py
View File

@ -206,26 +206,16 @@ class LightningDistributedDataParallel(DistributedDataParallel):
self.reducer.prepare_for_backward([]) self.reducer.prepare_for_backward([])
if output is None: if output is None:
warn_missing_output(fx_called) warn_missing_output(f'{fx_called} returned None. Did you forget to re')
m = f'{fx_called} returned None. Did you forget to re'
return output return output
def warn_missing_output(fx_called): def warn_missing_output(fx_called):
if fx_called == 'training_step': if fx_called == 'training_step':
m = """ warning_cache.warn("Your training_step returned None. You should instead do:\n"
Your training_step returned None. You should instead do: "`return loss`\n or\n `return TrainResult`")
return loss
or
return TrainResult
"""
elif fx_called in ['validation_step', 'test_step']: elif fx_called in ['validation_step', 'test_step']:
m = f""" warning_cache.warn(f"Your {fx_called} returned None. You should instead do:\n `return EvalResult")
Your {fx_called} returned None. You should instead do:
return EvalResult
"""
warning_cache.warn(m)
def parallel_apply(modules, inputs, kwargs_tup=None, devices=None): # pragma: no-cover def parallel_apply(modules, inputs, kwargs_tup=None, devices=None): # pragma: no-cover