Log Average Throughput and Inference Run Time

[Lilferrit]

Introduced end of run logging functionality for throughput and total inference run time. Similar functionality could be implemented using lightning.pytorch.callbacks.ThroughputMonitor for less code complexity, however this would log the throughput to the PyLightning logs (e.g. the tensorboard or CSV logs) and I'm not sure if this would be desirable behavior.

[codecov]

Codecov Report

Attention: Patch coverage is 95.12195% with 2 lines in your changes missing coverage. Please review.

Project coverage is 94.14%. Comparing base (ba4a8a1) to head (134c8e5).

Files with missing lines	Patch %	Lines
casanovo/utils.py	92.59%	2 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##              dev     #431      +/-   ##
==========================================
- Coverage   94.56%   94.14%   -0.43%     
==========================================
  Files          14       14              
  Lines        1289     1316      +27     
==========================================
+ Hits         1219     1239      +20     
- Misses         70       77       +7     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[bittremieux]

I think this code is way too complex for what it adds. What is the real benefit of having this vs just the total runtime?