model_3 = model_3.get_all_feat(all_feats.cuda()).cuda()
del all_feats ##clear memory
del data_loader2
del datamgr2
del x
del x_var
ds = True
freeze_backbone = params.freeze_backbone
##################################################################
pretrained_dataset = "miniImageNet"
if params.test_dataset == "ISIC":
print("Loading ISIC")
datamgr = ISIC_few_shot.SetDataManager2(image_size,
n_eposide=iter_num,
n_query=15,
**few_shot_params)
novel_loader = datamgr.get_data_loader(num_aug=params.gen_examples)
elif params.test_dataset == "EuroSAT":
print("Loading EuroSAT")
datamgr = EuroSAT_few_shot.SetDataManager2(image_size,
n_eposide=iter_num,
n_query=15,
**few_shot_params)
novel_loader = datamgr.get_data_loader(num_aug=params.gen_examples)
elif params.test_dataset == "CropDisease":
print("Loading CropDisease")
datamgr = CropDisease_few_shot.SetDataManager2(image_size,
n_eposide=iter_num,
base_datamgr = miniImageNet_few_shot.SetDataManager(
image_size, n_query=n_query, **train_few_shot_params)
base_loader = base_datamgr.get_data_loader(aug=params.train_aug)
# use unlabeled data from these novel domains for adversarial domain adaptation
# TODO: since the data is unlabeled, we need to modify data manager / data loader
if params.dataset == "ChestX":
target_datamgr = Chest_few_shot.SetDataManager(
image_size, n_query=n_query, **train_few_shot_params)
elif params.dataset == "EuroSAT":
target_datamgr = EuroSAT_few_shot.SetDataManager(
image_size, n_query=n_query, **train_few_shot_params)
elif params.dataset == "ISIC2018":
target_datamgr = ISIC_few_shot.SetDataManager(
image_size, n_query=n_query, **train_few_shot_params)
elif params.dataset == "CropDiseases":
target_datamgr = CropDisease_few_show.SetDataManager(
image_size, n_query=n_query, **train_few_shot_params)
else:
raise ValueError('Unknown dataset')
target_loader = target_datamgr.get_data_loader(novel_file,
aug=params.train_aug)
if params.adversarial or params.adaptFinetune:
# TODO: check argv
target_datamgr = SetDataManager(image_size,
n_query=n_query,
#params.save_iter = 399
if params.save_iter != -1:
modelfile = get_assigned_file(checkpoint_dir, params.save_iter)
elif params.method in ['baseline'] :
modelfile = get_resume_file(checkpoint_dir)
else:
modelfile = get_best_file(checkpoint_dir)
if params.save_iter != -1:
outfile = os.path.join( checkpoint_dir.replace("checkpoints","features"), params.dataset + "_" + str(params.save_iter)+ ".hdf5")
else:
outfile = os.path.join( checkpoint_dir.replace("checkpoints","features"), params.dataset + ".hdf5")
if params.dataset in ["ISIC"]:
datamgr = ISIC_few_shot.SimpleDataManager(image_size, batch_size = 64)
data_loader = datamgr.get_data_loader(aug = False )
elif params.dataset in ["EuroSAT"]:
datamgr = EuroSAT_few_shot.SimpleDataManager(image_size, batch_size = 64)
data_loader = datamgr.get_data_loader(aug = False )
elif params.dataset in ["CropDisease"]:
datamgr = CropDisease_few_shot.SimpleDataManager(image_size, batch_size = 64)
data_loader = datamgr.get_data_loader(aug = False )
elif params.dataset in ["ChestX"]:
datamgr = Chest_few_shot.SimpleDataManager(image_size, batch_size = 64)
data_loader = datamgr.get_data_loader(aug = False )
n_query=n_query,
mode="train",
**train_few_shot_params)
base_loader = datamgr.get_data_loader(aug=params.train_aug)
#the above line waste a lot of time
val_datamgr = miniImageNet_few_shot.SetDataManager(
image_size, n_query=n_query, mode="val", **test_few_shot_params)
val_loader = val_datamgr.get_data_loader(aug=False)
#=========== ISIC or EuroSAT ========
few_shot_params = dict(n_way=params.test_n_way,
n_support=params.n_shot)
if params.method == 'mytpnadaptisic':
print("init mytpnadaptisic")
datamgr = ISIC_few_shot.SetDataManager(image_size,
n_eposide=600,
n_query=n_query,
**few_shot_params)
elif params.method == 'mytpnadapteurosat':
print("init mytpnadapteurosat")
datamgr = EuroSAT_few_shot.SetDataManager(image_size,
n_eposide=600,
n_query=n_query,
**few_shot_params)
else:
raise ValueError('Unknown method')
novel_loader = datamgr.get_data_loader(aug=False)
#=====================================
#======= load pretrained model and ckpt =======
model = MyTPN_Adapt(model_dict[params.model], **train_few_shot_params)