data monitor callbacks

[awaelchli]

Before submitting

• Was this discussed/approved via a Github issue? (no need for typos and docs improvements)
• Did you read the contributor guideline, Pull Request section?
• Did you make sure to update the docs?
• Did you write any new necessary tests?

What does this PR do?

I'd like to contribute two callbacks: TrainingDataMonitor and ModuleDataMonitor as proposed in #194
https://github.com/awaelchli/pytorch-lightning-snippets#callbacks

TrainingDataMonitor

    from pl_bolts.callbacks import TrainingDataMonitor
    from pytorch_lightning import Trainer

    # log the histograms of input data sent to LightningModule.training_step
    monitor = TrainingDataMonitor(log_every_n_steps=25)

    model = YourLightningModule()
    trainer = Trainer(callbacks=[monitor])
    trainer.fit()

ModuleDataMonitor

    from pl_bolts.callbacks import ModuleDataMonitor
    from pytorch_lightning import Trainer

    # log the in- and output histograms of LightningModule's `forward`
    monitor = ModuleDataMonitor()

    # all submodules in LightningModule
    monitor = ModuleDataMonitor(submodules=True)

    # specific submodules
    monitor = ModuleDataMonitor(submodules=["generator", "generator.conv1"])

    model = YourLightningModule()
    trainer = Trainer(callbacks=[monitor])
    trainer.fit()

[codecov]

Codecov Report

Merging #285 into master will decrease coverage by 0.73%.
The diff coverage is 90.99%.

@@            Coverage Diff             @@
##           master     #285      +/-   ##
==========================================
- Coverage   83.86%   83.12%   -0.74%     
==========================================
  Files          91       92       +1     
  Lines        4858     5066     +208     
==========================================
+ Hits         4074     4211     +137     
- Misses        784      855      +71     

Flag	Coverage Δ
#cpu	24.65% <43.24%> (+0.73%)	⬆️
#pytest	24.65% <43.24%> (+0.73%)	⬆️
#unittests	83.50% <90.99%> (+0.18%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

Impacted Files	Coverage Δ
pl_bolts/callbacks/data_monitor.py	90.90% <90.90%> (ø)
pl_bolts/callbacks/__init__.py	100.00% <100.00%> (ø)
pl_bolts/callbacks/self_supervised.py	77.77% <0.00%> (-22.23%)	⬇️
pl_bolts/models/self_supervised/amdim/datasets.py	57.69% <0.00%> (-2.69%)	⬇️
pl_bolts/models/self_supervised/cpc/cpc_module.py	23.65% <0.00%> (+3.18%)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update f48357b...40870f6. Read the comment docs.

[Borda]

in 3 tests just with different parameters?