显存被占满

[YYX]

您好，我在试运行您的代码的时候，显示我的显存被占满。报错信息如下：

RuntimeError: CUDA out of memory. Tried to allocate 290.00 MiB (GPU 0; 31.75 GiB total capacity; 29.78 GiB already allocated; 113.50 MiB free; 30.27 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF

我检索到占满内存的代码是 loss = model(image, caption, alpha=alpha, idx=idx)

想问一下该怎么解决内存占满的问题呢？谢谢！

[Dachuan Shi]

您好，如果有多卡，可以尝试使用python -m torch.distributed.run进行多卡训练。如果使用单卡，可以将超参数batch size的数值调小。

Hi, if you have multiple GPUs, please launch scripts with python -m torch.distributed.run for distributed training. If you are using a single GPU, please try a smaller value for batch size.