LiangxuanZhao

> The version from SC2v5.x is probably `115` or something like that. So it's a matter of changing one number. You'll need MPQEditor (http://www.zezula.net/en/mpq/download.html) to edit that file (`.SC2Map` is...

> Well, MPQ would let me change it, but it wouldn't keep the change. It would just revert back to 115. I'm on a mac with wine. Weird. Ohhh it...

> Hey! Thanks for your great work! I have installed the `0.4.4` release of `torch_sparse` and it totally works out in my experiments! Maybe you could add this information to...

> try use other strategies, like `colossalai_zero2` @HuangLK I tried other strategies, but the problem still exists. Why do you think other strategies would solve this problem? Thanks~

> I find the problem(maybe), if you delete `model = model.to(torch.float16)` in the python file you can get a normal loss number, but the accuracy and distance will just slightly...

Meet the same problem. Thank you very much! @s-isaev

@panxb833 Hi! I met the same problem as you. Do you know how to solve it?

> @LuciusMos thanks a lot! It won't affect the training but indeed affect the training log. Would you like to create a PR to resolve the issue. Thank you so...

This problem is due to 560m & 7b1 bloomz model use left-padding by default, which is really weird :( You can change the padding style to right-padding to avoid this...

> > This problem is due to 560m & 7b1 bloomz model use left-padding by default, which is really weird :( You can change the padding style to right-padding to...