SSD PyTorch - RuntimeError: received 0 items of ancdata

Hi All,

I am trying to train SSD from this public repo on PASCAL VOC dataset fallowing the instructions in the repo.

When I try with num_workers = 8 and batch_size = 128 it works fine, but when I increase num_workers to 12, it is throwing the fallowing error.

I read online out of memory might be issue, but I have plenty of RAM available. Anyone else facing similar issue? Can any one help me solve it?

    Loading base network...
    Initializing weights...
    Loading Dataset...
    Training SSD on VOC0712
    Traceback (most recent call last):
      File "", line 232, in <module>
      File "", line 171, in train
        images, targets = next(batch_iterator)
      File "/home/XXXX/.virtualenvs/pytorch/lib/python3.5/site-packages/torch/utils/data/", line 206, in __next__
        idx, batch = self.data_queue.get()
      File "/usr/lib/python3.5/multiprocessing/", line 345, in get
        return ForkingPickler.loads(res)
      File "/home/XXXX/.virtualenvs/pytorch/lib/python3.5/site-packages/torch/multiprocessing/", line 70, in rebuild_storage_fd
        fd = df.detach()
      File "/usr/lib/python3.5/multiprocessing/", line 58, in detach
        return reduction.recv_handle(conn)
      File "/usr/lib/python3.5/multiprocessing/", line 181, in recv_handle
        return recvfds(s, 1)[0]
      File "/usr/lib/python3.5/multiprocessing/", line 160, in recvfds
      RuntimeError: received 0 items of ancdata   

System configuration : 4 Nvidia 1080 Ti’s, 384 GB RAM, 40 CPU cores.
Environment: PyTorch 0.1.12_2, CUDA 8.0.61 with patch 2, cuDNN 6.0

I have seen this, but couldn’t find a solution there.