Code execution is slow.

[hyosong86]

Thank you for sharing your great work.
Loading initial model parameters takes approximately 10 minutes.
Execution speed also takes about 2.5 seconds for a 0004x3.png (680x448) image.

GPU A6000, torch 1.13.1 opencv-python 4.5.1.48 onnxruntime 1.10.0 onnxruntime-gpu 1.10.0
I ran it in the above environment.
Are there any possible causes?

[Xzy765039540]

Thank you for sharing your great work. Loading initial model parameters takes approximately 10 minutes. Execution speed also takes about 2.5 seconds for a 0004x3.png (680x448) image.

GPU A6000, torch 1.13.1 opencv-python 4.5.1.48 onnxruntime 1.10.0 onnxruntime-gpu 1.10.0 I ran it in the above environment. Are there any possible causes?

Hey,
Did you fixed this problem? I meet this problem as well.

[Ysnower]

@Xzy765039540 @hyosong86 I haven't encountered this problem,you can try it

onnx_model.zip
`import cv2
import onnxruntime as ort
import numpy as np
import time

onnx_model_path = "./out.onnx"
load_time = time.time()
ort_session = ort.InferenceSession(onnx_model_path, providers=['CPUExecutionProvider'])
print("load onnx time:{}s".format(time.time()-load_time))
onnx_input_name = ort_session.get_inputs()[0].name
onnx_outputs_names = ort_session.get_outputs()[0].name

img_path = "0004x3.png" # 680x468 pix
img = cv2.imread(img_path)
img = np.asarray(img, np.float32)/255.0
img = img.transpose((2, 0, 1))
img = img[np.newaxis, :, :, :]
start = time.time()
for i in range(100):
onnx_result = ort_session.run([onnx_outputs_names], input_feed={onnx_input_name: img})[0]
print("avg infer time:{}s".format((time.time()-start)/100))
onnx_result = onnx_result.squeeze()*255
onnx_result = onnx_result.transpose((1, 2, 0))
cv2.imwrite("out_images/"+img_path.split('/')[-1], onnx_result)`