Jetson memory issues
grant opened this issue · comments
Grant Timmerman commented
Looks like there is not enough memory on the Jetson board. Here's the log:
Loading XML file ../blog-code/training_xml_web//Marcia_Turca.xml
storage_map=getattr(self.fn, 'storage_map', None))
File "/usr/local/lib/python2.7/dist-packages/theano/gof/link.py", line 314, in raise_with_op
reraise(exc_type, exc_value, exc_trace)
File "/usr/local/lib/python2.7/dist-packages/theano/compile/function_module.py", line 859, in __call__
outputs = self.fn()
MemoryError: alloc failed
Apply node that caused the error: Alloc(TensorConstant{0.0}, Elemwise{add,no_inplace}.0, Elemwise{switch,no_inplace}.0, Elemwise{Composite{Switch(EQ(i0, i1), i2, i0)}}.0)
Toposort index: 226
Inputs types: [TensorType(float64, scalar), TensorType(int64, scalar), TensorType(int64, scalar), TensorType(int64, scalar)]
Inputs shapes: [(), (), (), ()]
Inputs strides: [(), (), (), ()]
Inputs values: [array(0.0), array(128L, dtype=int64), array(780L, dtype=int64), array(600L, dtype=int64)]
Outputs clients: [[forall_inplace,cpu,grad_of_scan_fn}(Elemwise{maximum,no_inplace}.0, Elemwise{mul}.0, Alloc.0, Alloc.0, Alloc.0, Subtensor{int64:int64:int64}.0, Subtensor{int64:int64:int64}.0, Subtensor{int64:int64:int64}.0, Subtensor{int64:int64:int64}.0, Alloc.0, Alloc.0, Alloc.0, Alloc.0, Alloc.0, Alloc.0, Alloc.0, Alloc.0, Alloc.0, Alloc.0, Alloc.0, Alloc.0, Alloc.0, Alloc.0, Alloc.0, Alloc.0, Alloc.0, Alloc.0, Layer.linear_matrix, Layer.linear_matrix, Layer.linear_matrix, Layer.linear_matrix, Layer.linear_matrix, Layer.linear_matrix, Layer.linear_matrix, Layer.linear_matrix, Elemwise{Cast{float64}}.0, InplaceDimShuffle{1,0}.0, InplaceDimShuffle{1,0}.0, InplaceDimShuffle{1,0}.0, InplaceDimShuffle{1,0}.0, InplaceDimShuffle{1,0}.0, InplaceDimShuffle{1,0}.0, InplaceDimShuffle{1,0}.0, InplaceDimShuffle{1,0}.0, InplaceDimShuffle{0,x}.0, InplaceDimShuffle{0,x}.0, InplaceDimShuffle{0,x}.0, InplaceDimShuffle{0,x}.0, InplaceDimShuffle{0,x}.0, InplaceDimShuffle{0,x}.0, InplaceDimShuffle{0,x}.0, InplaceDimShuffle{0,x}.0, Alloc.0, Alloc.0, Alloc.0, Alloc.0, Alloc.0, Alloc.0, Alloc.0, Alloc.0)]]
Backtrace when the node is created:
File "/usr/local/lib/python2.7/dist-packages/theano/gradient.py", line 1113, in access_term_cache
input_grads = node.op.grad(inputs, new_output_grads)
HINT: Use the Theano flag 'exception_verbosity=high' for a debugprint and storage map footprint of this apply node.
Training process nnn terminated!!!!
Grant Timmerman commented
According to free -m
there's 3Gb of memory available on rnnmusic@jetson2.cs.washington.edu
:
free -m
total used free shared buffers cached
Mem: 3853 707 3145 8 110 213
-/+ buffers/cache: 383 3470
Swap: 0 0 0