Jayce

> %env OBJC_DISABLE_INITIALIZE_FORK_SAFETY=YES @till-m Sorry for the late reply. I have tried the `env` magic, but unfortunately encounter the same error.

> (1): Pandarallel basically doubles the amount of needed memory, as stated in [the documentation](https://nalepae.github.io/pandarallel/): > > > pandarallel gets around this limitation by using all cores of your computer....

`AssertionError: CUDA_HOME does not exist, unable to compile CUDA op(s)` 有人遇到一样的问题吗？安装了`pytorch==2.1.0` + `cuda11`，竟然说找不到CUDA…… 跑的知乎上的demo ``` import torch from lightseq.training.ops.pytorch.transformer_encoder_layer import LSTransformerEncoderLayer def train(model, inputs, masks): inputs = inputs.to(device="cuda:0") masks =...

hello everyone, I am facing the same issue with TF2.4.1 + cuda11.0 + P40 GPU. The GPU utilization is only 40%. I have tried the `XLA`, but unfortunately it cant...