run train_mnist error

[gzy-gxq]

When I run the train_mnist.py in the example, firstly, it generates this error: if computation =='cpu': UnboundLocalError: local variable'computation' referenced before assignment When I modify the definitioncomputation ='cpu' model_name='pop' The following error occurs : RuntimeError: Search process using CPU terminated but no result exists!
The reason is that the machine_0_finish.txt file does not exist. Can you provide a solution? Thank you !

[viebboy]

What command did you use to run "train_mnist.py"?

[gilsoneng]

Hello, I got the same error.
I attached the script
teste_pygop.zip

---

RuntimeError Traceback (most recent call last)
in
10 params['convergence_measure'] = 'mean_squared_error'
11
---> 12 performance, progressive_history, finetune_history = model.fit(params, data_func, data_arguments)

~/.local/lib/python3.6/site-packages/pygop-0.2.3-py3.6.egg/GOP/models/pop.py in fit(self, params, train_func, train_data, val_func, val_data, test_func, test_data, verbose)
149 test_func,
150 test_data,
--> 151 verbose)
152
153 if verbose:

~/.local/lib/python3.6/site-packages/pygop-0.2.3-py3.6.egg/GOP/models/pop.py in progressive_learn(self, params, train_func, train_data, val_func, val_data, test_func, test_data, verbose)
228 val_data,
229 test_func,
--> 230 test_data)
231
232 if verbose:

~/.local/lib/python3.6/site-packages/pygop-0.2.3-py3.6.egg/GOP/utility/gop_utils.py in search_cpu(params, train_states, train_func, train_data, val_func, val_data, test_func, test_data)
637
638 if not os.path.exists(os.path.join(path, 'machine_0_finish.txt')):
--> 639 raise RuntimeError('Search process using CPU terminated but no result exists!')
640
641 with open(os.path.join(path, 'machine_0.pickle'), 'rb') as fid:

RuntimeError: Search process using CPU terminated but no result exists!