bad argument #1 to '?' (must be strictly positive at /tmp/luarocks_torch-scm-1-2561/torch7/lib/TH/generic/THTensorMath.c:1467)

[SeekPoint]

rzai@rzai00:/prj/plantvillage-challenge/crowdai_train_tar$ ll
total 20
drwxrwxr-x 4 rzai rzai 4096 12月 23 11:38 ./
drwxrwxr-x 4 rzai rzai 4096 12月 23 11:35 ../
-rw-rw-r-- 1 rzai rzai 74 12月 23 11:38 divide_dat.sh
drwxrwxr-x 40 rzai rzai 4096 12月 23 11:38 train/
drwxrwxr-x 40 rzai rzai 4096 12月 23 11:38 val/
rzai@rzai00:/prj/plantvillage-challenge/crowdai_train_tar$
rzai@rzai00:/prj/plantvillage-challenge/crowdai_train_tar$ cd ..
rzai@rzai00:/prj/plantvillage-challenge$ th main.lua -model alexnet -data crowdai_train_tar
Creating new alexnet model
=> Training epoch # 1
[======================================== 21917/21917 ================================>] Tot: 4m18s | Step: 9ms
train Loss: 2.3574 Accuracy: 35.75 time: 259.87 s
==> Validating
/home/rzai/torch/install/bin/luajit: bad argument #1 to '?' (must be strictly positive at /tmp/luarocks_torch-scm-1-2561/torch7/lib/TH/generic/THTensorMath.c:1467)
stack traceback:
[C]: at 0x7f405a1308b0
[C]: in function 'randperm'
...zai/prj/plantvillage-challenge/datasets/plantvillage.lua:36: in function 'valGenerator'
/home/rzai/prj/plantvillage-challenge/train.lua:87: in function 'validate'
main.lua:72: in main chunk
[C]: in function 'dofile'
...rzai/torch/install/lib/luarocks/rocks/trepl/scm-1/bin/th:145: in main chunk
[C]: at 0x00406670
rzai@rzai00:~/prj/plantvillage-challenge$

[pankajkumar]

Hello @lovejasmine ,

I am also getting the same error while trying to train my model on torch and openNMT, Can you please share how did your fixed it?

[SeekPoint]

@pankajkumar I just want have a quick eval on this project, not fix it

[pankajkumar]

No issues, I am trying out different things. Will share here if I am able to fix it. Any helpful eval are welcome too.

[pankajkumar]

Update: It has been fixed by specifying correct src_seq_length and tgt_seq_length in my case