RuntimeError: CUDA out of memory. Tried to allocate 2.46 GiB (GPU 0; 6.00 GiB total capacity; 3.99 GiB already allocated; 39.00 MiB free; 4.17 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation.
batch_size很小,
碎片化也已经设置(下图)
结果没有一点改善
求救!!!
----------
好像只能增强硬件设施啦
让朋友一两万的电脑给我跑了 没一点毛病