emailweixu

@oars, this problem persists even with tf2.0 rc1. Now it's not just the problem of ml_layers. EnocdingNetwork also has this problem since EncodingNetwork creates the layers itself, Do you have...

@rPortelas Actually, I have reasons to believe that zero score for Freeway is expected. If you play Freeway yourself, you can see that it needs consistent exploration for one direction...

@rPortelas I know both EfficientZero and MuZero reported reasonable performance on Freeway. The original MuZero is not opensourced so I cannot re-run the experiments and cannot know for sure. But...

@rPortelas did you try the "raw" version you mentioned in #21 on Freeway?

Got same problem with Julia 1.6. ``` julia> a=Torch.Tensor(3,4, dev=-1) 3×4 Torch.Tensor{Float32, 2}: 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 julia> a[1,1] ERROR: getindex not...

Thanks for looking at this issue. But current behavior make default_ip=ti.i64 unusable since it's unnatural and very easy to forget the explicit type cast ti.i32(0).

Fair enough. I have never built taichi before. It sounds like it requires some effort. So I will not try it since I know how to circumvent it for now....

> Alright. But still, out of curiosity, what platform are you running on? (in case I set up CI or something) It's ubuntu 22.04, python 3.10, cuda 12.4

Thanks for the response. This solves the issue. It would be nice if this is clarified in the documentation regarding reduction. (https://docs.taichi-lang.org/docs/performance).

Thanks for the replay. I understand that matrix argument is immutable. But I think `mat[:, 0] += vec` shouldn't run without giving any error because the use of `+=` is...