search

Linfeng-Tang / SeAFusion

The code of " Image fusion in the loop of high-level vision tasks: A semantic-aware real-time infrared and visible image fusion network"
MIT License
184 stars 34 forks source link

about train #27

Open limyoonahh opened 9 months ago

limyoonahh commented 9 months ago

在运行train.py程序时，一直提示“CUDA out of memory. Tried to allocate 900.00 MiB. GPU 0 has a total capacty of 14.75 GiB of which 62.81 MiB is free. Process 118187 has 14.68 GiB memory in use. Of the allocated memory 14.52 GiB is allocated by PyTorch, and 36.59 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation”，通过翻阅以前的问题，我尝试了将pin-memory设置为false或将train_fusion中的batch_size改为args.batch_size还是出现上述问题，请问这是因为显存不足引起的问题吗，我使用的谷歌colab的T4 GPU

Linfeng-Tang commented 9 months ago

由于之前的程序不太规范 设置batch-szie的参数没有被很好地利用，你可以重新从Github push train.py文件或者尝试重新修改dataloader中设置batch-szie的参数来解决这个问题。希望能解决这个问题哈。

limyoonahh commented 8 months ago

谢谢