What is the meaning of 1000 in the final return of the loss function?

[wansiwei0521]

Preference-Planning-Deep-IRL/loss_fn.py

Lines 6 to 14 in 5b496d7

	def maxentirl_loss(learner, expert, reward_func, device):
	learner_torch = torch.FloatTensor(learner).to(device)
	expert_torch = torch.FloatTensor(expert).to(device)
	learner_r = reward_func.r(learner_torch).view(-1)
	expert_r = reward_func.r(expert_torch).view(-1)
	return 1000 * (learner_r.mean() - expert_r.mean())

[tonylitianyu]

I was testing scaling up the loss to see whether it will speed up the converging process or not. It did not change too much and I forgot to delete it. Sorry about the confusion.

[wansiwei0521]

Thank you very much for your reply, this project is very helpful to me.