From 7b575fd6f3f61ed4c129c6f255bcac55b1489756 Mon Sep 17 00:00:00 2001 From: Mu Li Date: Thu, 26 Nov 2015 22:27:47 -0500 Subject: [PATCH] [example] remove model downloader --- example/neural-style/model_vgg19.py | 14 -------------- 1 file changed, 14 deletions(-) diff --git a/example/neural-style/model_vgg19.py b/example/neural-style/model_vgg19.py index 4086e6a27863..14d12a10a1af 100644 --- a/example/neural-style/model_vgg19.py +++ b/example/neural-style/model_vgg19.py @@ -5,19 +5,6 @@ ConvExecutor = namedtuple('ConvExecutor', ['executor', 'data', 'data_grad', 'style', 'content']) -def _prepare_model(): - if os.path.exists("./model/vgg19-0001.params"): - logging.info("find vgg19 model") - return - if not os.path.isdir("./model"): - os.system("mkdir model") - os.chdir("model") - file_path = os.path.abspath() - sys.path.append(os.path.join(curr_path, "../../python")) - tool = os.path.dirname(__file__) + "../../tools/caffe_converter/run.sh" - os.sytem(tool + " vgg19") - os.chdir("..") - def get_model(input_size, ctx): # declare symbol data = mx.sym.Variable("data") @@ -63,7 +50,6 @@ def get_model(input_size, ctx): arg_dict = dict(zip(arg_names, [mx.nd.zeros(shape, ctx=ctx) for shape in arg_shapes])) grad_dict = dict(zip(arg_names, [mx.nd.zeros(shape, ctx=ctx) for shape in arg_shapes])) # init with pretrained weight - # _prepare_model() pretrained = mx.nd.load("./model/vgg19.params") for name in arg_names: if name == "data":