def run():
pack, GBNs = get_pack()
cloned, _ = clone_model(pack.net)
BASE_FLOPS, BASE_PARAM = analyse_model(cloned.module,
torch.randn(1, 3, 32, 32).cuda())
print('%.3f MFLOPS' % (BASE_FLOPS / 1e6))
print('%.3f M' % (BASE_PARAM / 1e6))
del cloned
prune(pack, GBNs, BASE_FLOPS, BASE_PARAM)
_ = Conv2dObserver.transform(pack.net.module)
pack.net.module.classifier = FinalLinearObserver(
pack.net.module.classifier)
Meltable.observe(pack, 0.001)
Meltable.melt_all(pack.net)
pack.optimizer = optim.SGD(pack.net.parameters(),
lr=1,
momentum=cfg.train.momentum,
weight_decay=cfg.train.weight_decay,
nesterov=cfg.train.nesterov)
_ = finetune(pack,
lr_min=cfg.gbn.lr_min,
lr_max=cfg.gbn.lr_max,
T=cfg.gbn.finetune_epoch)
def eval_prune(pack):
cloned, _ = clone_model(pack.net)
_ = Conv2dObserver.transform(cloned.module)
cloned.module.classifier = FinalLinearObserver(cloned.module.classifier)
cloned_pack = dotdict(pack.copy())
cloned_pack.net = cloned
Meltable.observe(cloned_pack, 0.001)
Meltable.melt_all(cloned_pack.net)
flops, params = analyse_model(cloned_pack.net.module, torch.randn(1, 3, 32, 32).cuda())
del cloned
del cloned_pack
return flops, params
def eval_prune(pack):
cloned, _ = clone_model(pack.net)
_ = Conv2dObserver.transform(cloned.module)
# cloned.module.classifier[0] = FinalLinearObserver(cloned.module.classifier[0])
cloned.module.fc = FinalLinearObserver(cloned.module.fc)
cloned_pack = dotdict(pack.copy())
cloned_pack.net = cloned
Meltable.observe(cloned_pack, 0.001)
Meltable.melt_all(cloned_pack.net)
flops = []
params = []
for res in cfg.model.resolution:
f, p = analyse_model(cloned_pack.net.module,
torch.randn(1, 3, res, res).cuda())
flops.append(f)
params.append(p)
del cloned
del cloned_pack
return flops, params
# group_id = uuid.uuid1()
# if len(masks) > 1:
# for mk in masks:
# mk.set_groupid(group_id)
# masks = []
bottleneck_set_group(pack.net)
def clone_model(net):
model = get_model()
gbns = GatedBatchNorm2d.transform(model.module)
model.load_state_dict(net.state_dict())
return model, gbns
cloned, _ = clone_model(pack.net)
BASE_FLOPS, BASE_PARAM = analyse_model(cloned.module, torch.randn(1, 3, cfg.model.resolution, cfg.model.resolution).cuda())
print('%.3f MFLOPS' % (BASE_FLOPS / 1e6))
print('%.3f M' % (BASE_PARAM / 1e6))
del cloned
def eval_prune(pack):
cloned, _ = clone_model(pack.net)
_ = Conv2dObserver.transform(cloned.module)
# cloned.module.classifier[0] = FinalLinearObserver(cloned.module.classifier[0])
cloned.module.fc = FinalLinearObserver(cloned.module.fc)
cloned_pack = dotdict(pack.copy())
cloned_pack.net = cloned
Meltable.observe(cloned_pack, 0.001)
Meltable.melt_all(cloned_pack.net)
flops, params = analyse_model(cloned_pack.net.module, torch.randn(1, 3, cfg.model.resolution, cfg.model.resolution).cuda())
del cloned
# masks = []
bottleneck_set_group(pack.net)
def clone_model(net):
model = get_model()
gbns = GatedBatchNorm2d.transform(model.module)
model.load_state_dict(net.state_dict())
return model, gbns
cloned, _ = clone_model(pack.net)
BASE_FLOPS, BASE_PARAM = [], []
for res in cfg.model.resolution:
f, p = analyse_model(cloned.module, torch.randn(1, 3, res, res).cuda())
BASE_FLOPS.append(f)
BASE_PARAM.append(p)
print('%.3f MFLOPS' % (f / 1e6))
print('%.3f M' % (p / 1e6))
del cloned
def eval_prune(pack):
cloned, _ = clone_model(pack.net)
_ = Conv2dObserver.transform(cloned.module)
# cloned.module.classifier[0] = FinalLinearObserver(cloned.module.classifier[0])
cloned.module.fc = FinalLinearObserver(cloned.module.fc)
cloned_pack = dotdict(pack.copy())
cloned_pack.net = cloned
Meltable.observe(cloned_pack, 0.001)