小臣子吃大橙子
小臣子吃大橙子
The single-gpu training process works just fine for me, and the output samples are satisfactory. However when I set `--n_gpu 2` or `--n_gpu 4`, the training process will get stuck...
**Describe the Bug** **Minimal Steps/Code to Reproduce the Bug** I've followed the installation instruction in the [README](https://github.com/NVIDIA/apex#linux), but an ImportError occurs when I import `fused_layer_norm_cuda`. I think the problem may...
执行前无报错退出(小概率不退出但是terminal卡死  执行初始化时无报错退出与terminal卡死  执行到一半无报错退出(可能是demo phase结束后,也可能是自由探索的任何一轮对话之后  请问这种不稳定性可能是什么导致的