def main():
model_bin, model_xml = get_ir_paths(args.model, args.bin)
with open(model_bin, "rb") as f:
buffer = f.read()
ignored = args.ignore.split(",") + get_ignored_layers(
model_xml, args.sparsify_first_conv, args.sparsify_fc)
all_parameters = find_all_parameters(buffer, model_xml)
total_params = 0
zero_params = 0
for name, param in all_parameters.items():
if name.split('.')[0] in ignored:
continue
total_params += param.data.size
zero_params += (param.data == 0).sum()
print("initial sparsity: {:.2f}%".format(
(zero_params / total_params) * 100))
if args.sparsity_level:
sparsify(all_parameters, ignored, args.normed_threshold)
if args.output:
print("saving new ir to: {}".format(args.output))
with open(args.output, "wb") as f:
f.write(generate_new_bin(buffer, all_parameters))
def main():
model_bin, model_xml = get_ir_paths(args.model, args.bin)
config = NNCFConfig.from_json(args.config)
input_infos_list = create_input_infos(config)
image_size = input_infos_list[0].shape[-1]
size = int(image_size / 0.875)
print('IE version: {}'.format(get_version()))
# NOTE: importing torch after loading IE to plugin to avoid issue with built-in MKLDNN of PyTorch
plugin = IEPlugin(device='CPU', plugin_dirs=args.cpu_plugin_dir)
plugin.add_cpu_extension(
os.path.join(args.cpu_plugin_dir, "libcpu_extension.so"))
net = IENetwork(model=model_xml, weights=model_bin)
exec_net = getExecNet(plugin, net)
from torch.utils.data import DataLoader
import torchvision.datasets as datasets
import torchvision.transforms as transforms
val_loader = DataLoader(datasets.ImageFolder(
args.data,
transforms.Compose([
transforms.Resize(size),
transforms.CenterCrop(image_size),
transforms.ToTensor(),
transforms.Normalize(mean=[0.485, 0.456, 0.406],
std=[0.229, 0.224, 0.225])
])),
batch_size=1,
shuffle=False,
num_workers=4,
pin_memory=True)
if not os.path.exists(args.output_dir):
os.makedirs(args.output_dir)
config['log_dir'] = args.output_dir
infer_fn = partial(infer_ie_model, net=net)
validate_general(val_loader, exec_net, infer_fn)
validate_torch_model(os.path.join(args.output_dir, "PTH"),
config=config,
num_layers=args.num_layers,
dump=args.dump,
val_loader=val_loader,
cuda=args.cuda)
def main():
model_bin, model_xml = get_ir_paths(args.model, args.bin)
if not os.path.exists(args.output_dir):
os.makedirs(args.output_dir)
ir_weights = collect_IR_weights(os.path.join(args.output_dir, "IR"), model_xml, model_bin, args.num_layers)
config = json.load(open(args.config))
torch_weights = collect_torch_weights(os.path.join(args.output_dir, "PTH"), config, args.num_layers)
assert len(ir_weights) == len(torch_weights), '{} vs {}'.format(len(ir_weights), len(torch_weights))
print("Maximum of absolute difference - IR vs Torch")
max_max = []
for (k1, v1), (k2, v2) in zip(ir_weights.items(), torch_weights.items()):
max_diff = abs(v1 - v2).max()
max_max.append(max_diff)
print("{0:.5} - max diff [{1:}] vs [{2:}]".format(max_diff, k1, k2))
print("Global maximum: {0:.5}".format(np.max(max_max)))