Xiaodong Deng

I still have some doubts. In the following example, the results of autodiff and my manual grad are inconsistent. Perhaps the autodiff for reduce operations needs to be re implemented?...

> I met No.3 error，loss increases sharply during training. do you have any solution to fix it? refer to my [code](https://github.com/Dinngger/Jeuralangelo/blob/98ee26c11ad5392bc040b133a33c1f38e5e690bd/hash_encoder/grid_encode.py#L120), it might help you.

Another bug found. ```python import taichi as ti ti.init(ti.cuda) @ti.kernel def mytest(): for _ in range(1): x = ti.Vector.zero(ti.f32, 1) x[0] = 1 res = ti.Vector.zero(ti.f32, 1) for i in...