RuntimeError about paperrobot HOT 13 CLOSED

Thank you for your interest. Which pytorch version did you use? Have you used cuda?

from paperrobot.

Thanks a lot for your kind reply.
pytorch version: torch-1.1.0-cp36-cp36m-manylinux1_x86_64.whl
I'm not sure wether cuda is used. Follow your instruction, I did not find the requirements on cuda.

from paperrobot.

I mean have you use gpu?

from paperrobot.

The default setting is to use GPU if your system has one

from paperrobot.

Maybe I need to disable the GPU option, because it takes a lot time to run the script. so, i move the script from osx to remote server

from paperrobot.

How to disable the GPU option?

from paperrobot.

Emm, I haven't tried to use cpu, I recommend to use the gpu, you can disable it by adding --gpu 0

from paperrobot.

no lucky.

Traceback (most recent call last): File "train.py", line 236, in <module> batch_o_t, teacher_forcing_ratio=1) File "/usr/local/lib/python3.6/dist-packages/torch/nn/modules/module.py", line 493, in __call__ result = self.forward(*input, **kwargs) File "/mnt/sync/ubuntu/PaperRobot-master/New paper writing/memory_generator/seq2seq.py", line 18, in forward stopwords, sflag) File "/usr/local/lib/python3.6/dist-packages/torch/nn/modules/module.py", line 493, in __call__ result = self.forward(*input, **kwargs) File "/mnt/sync/ubuntu/PaperRobot-master/New paper writing/memory_generator/Decoder.py", line 134, in forward max_source_oov, term_output, term_id, term_mask) File "/mnt/sync/ubuntu/PaperRobot-master/New paper writing/memory_generator/Decoder.py", line 68, in decode_step term_context, term_attn = self.memory(_h.unsqueeze(0), term_output, term_mask, cov_mem) File "/usr/local/lib/python3.6/dist-packages/torch/nn/modules/module.py", line 493, in __call__ result = self.forward(*input, **kwargs) File "/mnt/sync/ubuntu/PaperRobot-master/New paper writing/memory_generator/utils.py", line 32, in forward e_t = self.vt_layers[i](torch.tanh(enc_proj + dec_proj).view(batch_size * max_enc_len, -1)) RuntimeError: [enforce fail at CPUAllocator.cpp:56] posix_memalign(&data, gAlignment, nbytes) == 0. 12 vs 0

from paperrobot.

Let me test on my computer without gpu

from paperrobot.

I haven't encountered the same problem, maybe you can try a smaller batch size by setting --batch_size 10?

from paperrobot.

Lucky for me. It's running now with the setting --batch_size 10

from paperrobot.

Great, I think maybe it's out of memory

from paperrobot.

I found that the previous error is a CPU OOM message pytorch/pytorch#20618

from paperrobot.