Hello,
I am writing to inform you that there is a problem between the avalanche-rl and DQNStrategy.
Unfortunately, in avalanche-rl when a strategy uses EWC plugin, below error will be raise:
Error for Compute EWC penalty and add it to the loss.
exp_counter = strategy.clock.train_exp_counter
if exp_counter == 0:
return
'DQNStrategy' object has no attribute 'clock'.
It needs to add 'clock' attribute in 'DQNStrategy'. I will appreciate if you update this bug in the avalanhe-rl.
avalanche-lib==0.3.1
Hi @Ai360n, I'm having the same problem right now, but if you don't need the clock and you just need an alternative way to use the experience_counter, then just use the `training_exp_counter` attribute of the `RLBaseStrategy` class, which has the same meaning of train_exp_counter.
So inside ewc.py of avalanche just change your line to `exp_counter = strategy.training_exp_counter` (in both before_backward and after_training_exp methods).
Of course in order to do this you should have installed avalanche 0.3.1 locally with `pip install -e .` option, in order to use your local modifies as the library.