Fixed lwf bug for alpha and prev classes per task dict

[AndreaCossu]

LwF now correctly uses one alpha per experience if alpha is given as a list.
The classes seen so far for each task were managed by a dictionary. Sometimes the key was a string, sometimes an int, leading to inconsistency. It now works.

[ContinualAI-bot]

Oh no! It seems there are some PEP8 errors! 😕
Don't worry, you can fix them! 💪
Here's a report about the errors and where you can find them:

avalanche/training/regularization.py:9:1: E302 expected 2 blank lines, found 1
1       E302 expected 2 blank lines, found 1

[coveralls]

Pull Request Test Coverage Report for Build 3282140647

• 14 of 14 (100.0%) changed or added relevant lines in 3 files are covered.
• No unchanged relevant lines lost coverage.
• Overall coverage increased (+0.004%) to 73.431%

---

Totals
Change from base Build 3242453389:	0.004%
Covered Lines:	13114
Relevant Lines:	17859

---

💛 - Coveralls