img_path = "/home/sz/hard_ex_segmentation/HEI-MED/data4/test/img/"
label_path = "/home/sz/hard_ex_segmentation/HEI-MED/data4/test/label/"
img_root = "/home/sz/hard_ex_segmentation/HEI-MED/data4/train_crop/img/"
label_root = "/home/sz/hard_ex_segmentation/HEI-MED/data4/train_crop/label/"
dataset = datatest.get_dataset(img_root, label_root)
train_data = mx.gluon.data.DataLoader(dataset,
batch_size,
shuffle=True,
last_batch='rollover',
num_workers=8)
ctx = [mx.gpu(4), mx.gpu(5), mx.gpu(6), mx.gpu(7)]
net = multi_att_net.set_network()
net.collect_params().reset_ctx(ctx)
net.load_parameters('pretrain2_double.params', ctx=ctx)
def get_batch(batch, ctx):
if isinstance(batch, mx.io.DataBatch):
data = batch.data[0]
label = batch.data[1]
else:
data, label = batch
return (gluon.utils.split_and_load(data, ctx),
gluon.utils.split_and_load(label, ctx), data.shape[0])
loss = dice_entropy_loss.seg_loss()
import cv2
import os
import random
import matplotlib.pyplot as plt
from mxnet import nd
from mxnet import image
from skimage.measure import label as la
ctx1=mx.gpu(4)
ctx2=mx.gpu(5)
ctx3=mx.gpu(6)
ctx4=mx.gpu(7)
net_vessel=multi_att_net.set_network()
net_vessel.collect_params().reset_ctx(ctx1)
net_vessel.load_parameters("/home/sz/hard_ex_segmentation/vessel.params",ctx=ctx1)
net_hard=multi_att_net.set_network()
net_hard.collect_params().reset_ctx(ctx2)
net_hard.load_parameters("/home/sz/hard_ex_segmentation/IDRID_HE.params",ctx=ctx2)
net_mic=multi_att_net.set_network()
net_mic.collect_params().reset_ctx(ctx3)
net_mic.load_parameters("/home/sz/hard_ex_segmentation/micro.params",ctx=ctx3)
net_heam=multi_att_net.set_network()
net_heam.collect_params().reset_ctx(ctx4)
net_heam.load_parameters("/home/sz/hard_ex_segmentation/heam.params",ctx=ctx4)