AttributeError: 'numpy.float32' object has no attribute 'to_cpu'

[msabda...@gmail.com]

Hi all,

Could you please help me with this error? Below is the snipestof the code from which the error originates.

def get_idp_acc(model, dataset_tuple, comp_ratio, profile = None, batchsize = 128, gpu = 0):
    chainer.config.train = True
    import numpy as np
    #xp = np if gpu < 0 else cuda.cupy
    x, indices, x_zf, HtH, Hty = dataset_tuple._datasets[0], dataset_tuple._datasets[1], dataset_tuple._datasets[2], dataset_tuple._datasets[3], dataset_tuple._datasets[4]
   
    accs = 0
    BERs = 0
    model.train = False

    for j in range(0, len(x), batchsize):
        x_batch = np.array(x[j:j + batchsize])
        indices_batch = np.array(indices[j:j + batchsize])
        x_zf_batch = np.array(x_zf[j:j + batchsize])
        HtH_batch = np.array(HtH[j:j + batchsize])
        Hty_batch = np.array(Hty[j:j + batchsize])
        if profile == None:
            acc_data = model(x_batch, indices_batch, x_zf_batch, HtH_batch, Hty_batch, comp_ratio=comp_ratio,
                            ret_param = 'acc')
        else:
            acc_data = model(x_batch, indices_batch, x_zf_batch, HtH_batch, Hty_batch, comp_ratio=comp_ratio,
                            ret_param='acc', profile=profile)
        acc_data.to_cpu()
        acc = acc_data.data
        BER = 1 - acc
        accs += acc * len(x_batch)
        BERs += BER * len(x_batch)
    return (accs / len(x)) * 100.

The error is as shown below.

Traceback (most recent call last):
  File "/Users/mac/Documents/idp_detnet/examples/run_mlp.py", line 14, in <module>
    mlp.run(args)
  File "/Users/mac/Documents/idp_detnet/examples/mlp.py", line 39, in run
    acc_dict[name], BER_dict[name] = util.sweep_idp(model, test, comp_ratios, args)
  File "/Users/mac/Documents/idp_detnet/examples/util.py", line 109, in sweep_idp
    batchsize=args.batchsize, profile=profile))
  File "/Users/mac/Documents/idp_detnet/examples/util.py", line 85, in get_idp_acc
    acc_data.to_cpu()
AttributeError: 'numpy.float32' object has no attribute 'to_cpu'

Thank you in anticipation of your kind response.

Abdullahi Mohammad

[Kenichi Maehashi]

The question has been answered on:

https://stackoverflow.com/questions/57989774/attributeerror-numpy-float32-object-has-no-attribute-to-cpu