Chapter3: a strange error when I run Listing 3.7

[DemonsHunter]

In the Listing 3.7, we use both memory replay and target network to improve the stablility.

However, in the memory loop:

if len(replay) > batch_size:
minibatch = random.sample(replay, batch_size)
...
action_batch = torch.Tensor([a for (s1,a,r,s2,d) in minibatch])

The compiler tells me this error:

---> 42 action_batch = torch.Tensor([a for (s1,a,r,s2,d) in minibatch])
too many dimensions 'str'

I suppose that when we memory, the action is represented by a characteristic. There, nevertheless, corresponding number are needed.

So I propose to make a reverse action set to fill this transform.