Problem with Pytorch tutorial for transfer a model from Pytorch to caffe2

vision
1

I just download the tutorial code from Pytorch tutorial for transfer model with onnx,
and I try to run it, I found the assertionError, which should mean the Pytorch output be different with caffe2, I am confusing about that…

AssertionError:
Arrays are not almost equal to 3 decimals

(mismatch 99.8875070861678%)
x: array([[[[ 0.383, 0.601, 0.876, …, -0.167, -0.151, -0.149],
[ 0.273, 0.587, 0.949, …, -0.094, -0.168, -0.16 ],
[ 0.147, 0.363, 0.637, …, 0.077, -0.107, -0.186],…
y: array([[[[ 0.383, 1.055, 0.883, …, 0.099, -0.151, -0.149],
[-0.121, 0.061, 0.719, …, 0.133, 0.46 , -0.245],
[ 0.337, -0.409, 0.126, …, 0.161, 1.509, 0.033],…

2

Having same issue :confused:

AssertionError: 
Arrays are not almost equal to 3 decimals

(mismatch 99.88639987244898%)
 x: array([ 0.271,  0.347,  0.455, ..., -0.064, -0.148, -0.038], dtype=float32)
 y: array([ 0.271,  0.473,  0.358, ...,  0.193,  0.505, -0.038], dtype=float32)

here is the full stack trace:

---------------------------------------------------------------------------
AssertionError                            Traceback (most recent call last)
<ipython-input-6-cfccf8d5c977> in <module>
     22 
     23 # Verify the numerical correctness upto 3 decimal places
---> 24 np.testing.assert_almost_equal(torch_out.data.cpu().numpy(), c2_out, decimal=3)
     25 
     26 print("Exported model has been executed on Caffe2 backend, and the result looks good!")

/usr/local/miniconda3/envs/pytorch_onnx/lib/python3.7/site-packages/numpy/testing/_private/utils.py in assert_almost_equal(actual, desired, decimal, err_msg, verbose)
    566     if isinstance(actual, (ndarray, tuple, list)) \
    567             or isinstance(desired, (ndarray, tuple, list)):
--> 568         return assert_array_almost_equal(actual, desired, decimal, err_msg)
    569     try:
    570         # If one of desired/actual is not finite, handle it specially here:

/usr/local/miniconda3/envs/pytorch_onnx/lib/python3.7/site-packages/numpy/testing/_private/utils.py in assert_array_almost_equal(x, y, decimal, err_msg, verbose)
    971     assert_array_compare(compare, x, y, err_msg=err_msg, verbose=verbose,
    972              header=('Arrays are not almost equal to %d decimals' % decimal),
--> 973              precision=decimal)
    974 
    975 

/usr/local/miniconda3/envs/pytorch_onnx/lib/python3.7/site-packages/numpy/testing/_private/utils.py in assert_array_compare(comparison, x, y, err_msg, verbose, header, precision, equal_nan, equal_inf)
    787                                 verbose=verbose, header=header,
    788                                 names=('x', 'y'), precision=precision)
--> 789             raise AssertionError(msg)
    790     except ValueError:
    791         import traceback

My dependecy versions: python 3.7.1, onnx 1.3.0, pytorch-nightly 1.0.0.dev20181222