In THEANO_FLAGS, you can use device=cpu rather than device=gpu0.
thank you, but when i replace it by cpu, i have some new problems, if we just use cpu , does it mean that we do not need to download Cuda ? so do we need to change the codes ? cause i think in the code it uses some functions of GPU, the error messages are here :
[win_size] = 384
[average] = False
[model_name] = outdoor_64
[batch_size] = 64
[interp] = linear
[d_weight] = 0.0
[shadow] = False
[model_type] = dcgan_theano
[framework] = theano
[n_iters] = 40
[model_file] = ./models/outdoor_64.dcgan_theano
[morph_steps] = 16
[top_k] = 16
LOADING...
n_layers= 3
load model from ./models/outdoor_64.dcgan_theano
3.22 seconds to load theano models
Traceback (most recent call last):
File "iGAN_main.py", line 43, in
model = model_class.Model(model_name=args.model_name, model_file=args.model_file)
File "/home/walter/Downloads/iGAN-master/model_def/dcgan_theano.py", line 26, in init
self._gen = self.def_gen(self.gen_params, self.gen_batchnorm, n_layers=self.n_layers, n_f=self.n_f, nc=self.nc)
File "/home/walter/Downloads/iGAN-master/model_def/dcgan_theano.py", line 39, in def_gen
gx = gen_test(z, gen_params, gen_pl, n_layers=n_layers, n_f=n_f, nc=nc, use_tanh=False)
File "/home/walter/Downloads/iGAN-master/model_def/dcgan_theano.py", line 225, in gen_test
hout = relu(batchnorm(deconv(hin, w, subsample=(2, 2), border_mode=(2, 2)), u=u, s=s, g=g, b=b))
File "/home/walter/Downloads/iGAN-master/lib/ops.py", line 101, in deconv
d_img = GpuDnnConvGradI()(kerns, img, out, desc)
File "/usr/local/lib/python3.5/dist-packages/theano/sandbox/cuda/dnn.py", line 859, in init
if version() < (5000, 5000):
File "/usr/local/lib/python3.5/dist-packages/theano/sandbox/cuda/init.py", line 424, in dnn_version
dnn_available.msg)
Exception: ("We can't determine the cudnn version as it is not available", 'CUDA not available')
In
THEANO_FLAGS
, you can usedevice=cpu
rather thandevice=gpu0
.