Running main.lua for training gives reshape function exception

[rohanchabra]

I am able to run the test set correctly. But training gives me such error. The images are in the correct folder. What might be the problem?

./main.lua kitti slow -a train_trkitti
slow -a train_tr
luajit: ./main.lua:378: inconsistent tensor size, expected tensor [389 x 1 x 350 x 1242] and src [] to have the same number of elements, but got 169098300 and 0 elements respectively at /home/rohan140290/torch/pkg/torch/lib/TH/generic/THTensorCopy.c:86
stack traceback:
[C]: in function 'reshape'
./main.lua:378: in function 'fromfile'
./main.lua:428: in main chunk
[C]: at 0x00405d50

It seems like torch is not able to parse the file correctly.
x = torch.FloatTensor(torch.FloatStorage(fname)) :- Seems to have issues.

I instead tried to do this:-

for i = 1,#dim do
s = s * dim[i]
end

x = torch.FloatTensor(torch.FloatStorage(s))
torch.DiskFile(fname,'r'):binary():readFloat(x:storage())

This seems to work for me.

[strivejin]

I change the code just like you do :
"I instead tried to do this:-

for i = 1,#dim do
s = s * dim[i]
end

x = torch.FloatTensor(torch.FloatStorage(s))
torch.DiskFile(fname,'r'):binary():readFloat(x:storage())"
but it still don't work ,do you know why?

kitti slow -a train_tr
luajit: ./main.lua:369: attempt to perform arithmetic on global 's' (a nil value)
stack traceback:
./main.lua:369: in function 'fromfile'
./main.lua:434: in main chunk
[C]: at 0x00405d50

[strivejin]

@rohanchabra

[ComVisDinh]

I have the same problem. Do you have a solution for this?

[strivejin]

you should give s a initial value，like this ：

for i = 1,#dim do
s=1
s = s * dim[i]
end

@ComVisDinh

[ComVisDinh]

It works. Thank you 👍

[sooyeonshin]

your problem fixed?
I have same issue, so I change the code but still show error message at reshape function.

if type == 'float32' then
for i = 1,#dim do
s = 1
s = s * dim[i]
end
x = torch.FloatTensor(torch.FloatStorage(s))
torch.DiskFile(fname,'r'):binary():readFloat(x:storage())
elseif type == 'int32' then
for i = 1,#dim do
s = 1
s = s * dim[i]
end
x = torch.IntTensor(torch.IntStorage(s))
torch.DiskFile(fname,'r'):binary():readInt(x:storage())

elseif type == 'int64' then
for i = 1,#dim do
s = 1
s = s * dim[i]
end
x = torch.LongTensor(torch.LongStorage(s))
torch.DiskFile(fname,'r'):binary():readLong(x:storage())

else
print(fname, type)
assert(false)
end

x = x:reshape(torch.LongStorage(dim))
return x

-------------------------------------------------------------------------error message-----------------------------------
luajit: ./main.lua:396: inconsistent tensor size, expected tensor [389 x 1 x 350 x 1242] and src [1242] to have the same number of elements, but got 169098300 and 1242 elements respectively at /home/vclab-ubuntu/torch/pkg/torch/lib/TH/generic/THTensorCopy.c:86
stack traceback:
[C]: in function 'reshape'
./main.lua:396: in function 'fromfile'
./main.lua:446: in main chunk
[C]: at 0x00405d50

[ComVisDinh]

Variable s should be outside of a FOR loop. If it is inside, it always resets to value 1 for each repeat.

if type == 'float32' then
--x = torch.FloatTensor(torch.FloatStorage(fname))

  s=1
  for i = 1,#dim do	 
 s = s * dim[i]
  end

  x = torch.FloatTensor(torch.FloatStorage(s))
  torch.DiskFile(fname,'r'):binary():readFloat(x:storage())

elseif type == 'int32' then
print('inside --------------------- 1 ')
--x = torch.IntTensor(torch.IntStorage(fname))

  s=1
  for i = 1,#dim do	 
 s = s * dim[i]
  end

  x = torch.IntTensor(torch.IntStorage(s))
  torch.DiskFile(fname,'r'):binary():readInt(x:storage())

elseif type == 'int64' then
print('inside --------------------- 3 ')
--x = torch.LongTensor(torch.LongStorage(fname))

  s=1
  for i = 1,#dim do	 
 s = s * dim[i]
  end

  x = torch.LongTensor(torch.LongStorage(s))
  torch.DiskFile(fname,'r'):binary():readLong(x:storage())

else