[ex_aflw_parser]

[speculaas]

Dear mpatacchiola,

Thanks for your help!

https://github.com/mpatacchiola/deepgaze/blob/master/examples/ex_dnn_head_pose_estimation_training/ex_aflw_parser.py

In the above file:

#Separating the label from the dataset
dataset = data[:,0:4096]
label = data[:,4096:4099]

I think length of the image in 1D should be 64*64 = 4096?
It's 1600 in the current commit and I got error when I try to run ex_dnn_head_pose_estimation_training.py :

Training set (19507, 1600) (19507, 3)
Validation set (2438, 1600) (2438, 3)
Test set (2439, 1600) (2439, 3)
Starting Graph creation...
Traceback (most recent call last):
File "ex_dnn_head_pose_estimation_training.py", line 182, in
weights_output3, biases_output3)
File "ex_dnn_head_pose_estimation_training.py", line 105, in multilayer_model
_input0_result = tf.matmul(_X, _input0) + _biases_input0
File "/home3/jimmy_extra_space/headpose/deepgaze/.env/local/lib/python2.7/site-packages/tensorflow/python/ops/math_ops.py", line 2018, in matmul
a, b, transpose_a=transpose_a, transpose_b=transpose_b, name=name)
File "/home3/jimmy_extra_space/headpose/deepgaze/.env/local/lib/python2.7/site-packages/tensorflow/python/ops/gen_math_ops.py", line 4456, in mat_mul
name=name)
File "/home3/jimmy_extra_space/headpose/deepgaze/.env/local/lib/python2.7/site-packages/tensorflow/python/framework/op_def_library.py", line 787, in _apply_op_helper
op_def=op_def)
File "/home3/jimmy_extra_space/headpose/deepgaze/.env/local/lib/python2.7/site-packages/tensorflow/python/util/deprecation.py", line 454, in new_func
return func(*args, **kwargs)
File "/home3/jimmy_extra_space/headpose/deepgaze/.env/local/lib/python2.7/site-packages/tensorflow/python/framework/ops.py", line 3155, in create_op
op_def=op_def)
File "/home3/jimmy_extra_space/headpose/deepgaze/.env/local/lib/python2.7/site-packages/tensorflow/python/framework/ops.py", line 1731, in init
control_input_ops)
File "/home3/jimmy_extra_space/headpose/deepgaze/.env/local/lib/python2.7/site-packages/tensorflow/python/framework/ops.py", line 1579, in _create_c_op
raise ValueError(str(e))
ValueError: Dimensions must be equal, but are 1600 and 4096 for 'MatMul_4' (op: 'MatMul') with input shapes: [2438,1600], [4096,256].

I changed the 1600 to 4096 and was able to proceed with ex_dnn_head_pose_estimation_training.py.

Also I did a brief search here, but didn't find such issue.
Hope this is not a duplicate issue.

BR,
JimmyYS

[NieHw]

You should resize the face pictures to 64 * 64 at first