Train problems

[YjDai]

CUDA_VISIBLE_DEVICES='0' python -u train.py --network y1 --loss margin_distillation --dataset emore

I have a problem:

Traceback (most recent call last): File "train.py", line 717, in main() File "train.py", line 714, in main train_net(args) File "train.py", line 703, in train_net epoch_end_callback = epoch_cb ) File "/usr/local/lib/python2.7/dist-packages/mxnet/module/base_module.py", line 498, in fit for_training=True, force_rebind=force_rebind) File "/usr/local/lib/python2.7/dist-packages/mxnet/module/module.py", line 429, in bind state_names=self._state_names) File "/usr/local/lib/python2.7/dist-packages/mxnet/module/executor_group.py", line 280, in init self.bind_exec(data_shapes, label_shapes, shared_group) File "/usr/local/lib/python2.7/dist-packages/mxnet/module/executor_group.py", line 376, in bind_exec shared_group)) File "/usr/local/lib/python2.7/dist-packages/mxnet/module/executor_group.py", line 670, in _bind_ith_exec shared_buffer=shared_data_arrays, **input_shapes) File "/usr/local/lib/python2.7/dist-packages/mxnet/symbol/symbol.py", line 1782, in simple_bind raise RuntimeError(error_msg) RuntimeError: simple_bind error. Arguments: data: (256, 3, 112, 112) softmax_label: (256, 2) Error in operator slice_axis0: [17:36:53] src/operator/tensor/./matrix_op-inl.h:1295: Check failed: *end <= axis_size: Invalid end for end=129 as axis_size is 2

[david-svitov]

How you obtained emore dataset? Did you perform all steps from "Data preparation" instruction in README.md file?

[doitslow]

How you obtained emore dataset? Did you perform all steps from "Data preparation" instruction in README.md file?

In terms of the last step, can you give an example command? Thanks!

[david-svitov]

We fixed Data preparation instruction in README.md. Please try new one.

[XiXiRuPan]

CUDA_VISIBLE_DEVICES='0' python -u train.py --network y1 --loss margin_distillation --dataset emore

I have a problem:

Traceback (most recent call last): File "train.py", line 717, in main() File "train.py", line 714, in main train_net(args) File "train.py", line 703, in train_net epoch_end_callback = epoch_cb ) File "/usr/local/lib/python2.7/dist-packages/mxnet/module/base_module.py", line 498, in fit for_training=True, force_rebind=force_rebind) File "/usr/local/lib/python2.7/dist-packages/mxnet/module/module.py", line 429, in bind state_names=self._state_names) File "/usr/local/lib/python2.7/dist-packages/mxnet/module/executor_group.py", line 280, in init self.bind_exec(data_shapes, label_shapes, shared_group) File "/usr/local/lib/python2.7/dist-packages/mxnet/module/executor_group.py", line 376, in bind_exec shared_group)) File "/usr/local/lib/python2.7/dist-packages/mxnet/module/executor_group.py", line 670, in _bind_ith_exec shared_buffer=shared_data_arrays, **input_shapes) File "/usr/local/lib/python2.7/dist-packages/mxnet/symbol/symbol.py", line 1782, in simple_bind raise RuntimeError(error_msg) RuntimeError: simple_bind error. Arguments: data: (256, 3, 112, 112) softmax_label: (256, 2) Error in operator slice_axis0: [17:36:53] src/operator/tensor/./matrix_op-inl.h:1295: Check failed: *end <= axis_size: Invalid end for end=129 as axis_size is 2

I have the same problem. Have you solved？

[david-svitov]

I repeated the dataset creation procedure from README.md. And faced no problem with training then on the obtained dataset. I consider the next options as possible sources of the problem:

1. You use '--dataset emore'. Please check that you use the correct path in 'config.py' or use '--dataset emore_soft' as proposed in the instruction;
2. You use Python 2.7. I tested code with Python 3.6 only.
3. If you still have this problem I prepared a small training dataset for you: https://drive.google.com/file/d/1qDRkI_H0RI_MIHghjvOO63ETDK4UcpJt/view?usp=sharing If you have this problem with this small dataset and Python 3.6, please let me know.