FPN_int.onnx → FPN_int_NHWC.onnx RENAMED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:4a172eb921a5119875bd12561e658450ca4dae95c4aa2ea350dfd603cd27f14a
3
- size 45595505
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d1cca42fe8ac3429c55aa1a28a3ecdf08cd279b6193e6487d9b18fa28cfc2c7e
3
+ size 45595686
infer_onnx.py CHANGED
@@ -19,7 +19,7 @@ from datasets.utils import colorize_mask, build_img
19
 
20
  if __name__ == "__main__":
21
  parser = argparse.ArgumentParser(description='SemanticFPN model')
22
- parser.add_argument('--onnx_path', type=str, default='FPN_int.onnx')
23
  parser.add_argument('--save_path', type=str, default='./data/demo_results/senmatic_results.png')
24
  parser.add_argument('--input_path', type=str, default='data/cityscapes/leftImg8bit/test/bonn/bonn_000000_000019_leftImg8bit.png')
25
  parser.add_argument('--ipu', action='store_true', help='use ipu')
@@ -37,8 +37,8 @@ if __name__ == "__main__":
37
  onnx_path = args.onnx_path
38
  input_img = build_img(args)
39
  session = onnxruntime.InferenceSession(onnx_path, providers=providers, provider_options=provider_options)
40
- ort_input = {session.get_inputs()[0].name: input_img.cpu().numpy()}
41
- ort_output = session.run(None, ort_input)[0]
42
  if isinstance(ort_output, (tuple, list)):
43
  ort_output = ort_output[0]
44
 
 
19
 
20
  if __name__ == "__main__":
21
  parser = argparse.ArgumentParser(description='SemanticFPN model')
22
+ parser.add_argument('--onnx_path', type=str, default='FPN_int_NHWC.onnx')
23
  parser.add_argument('--save_path', type=str, default='./data/demo_results/senmatic_results.png')
24
  parser.add_argument('--input_path', type=str, default='data/cityscapes/leftImg8bit/test/bonn/bonn_000000_000019_leftImg8bit.png')
25
  parser.add_argument('--ipu', action='store_true', help='use ipu')
 
37
  onnx_path = args.onnx_path
38
  input_img = build_img(args)
39
  session = onnxruntime.InferenceSession(onnx_path, providers=providers, provider_options=provider_options)
40
+ ort_input = {session.get_inputs()[0].name: input_img.cpu().numpy().transpose(0,2,3,1)}
41
+ ort_output = session.run(None, ort_input)[0].transpose(0,3,1,2)
42
  if isinstance(ort_output, (tuple, list)):
43
  ort_output = ort_output[0]
44
 
test_onnx.py CHANGED
@@ -22,7 +22,7 @@ class Configs():
22
  # dataset
23
 
24
  parser.add_argument('--dataset', type=str, default='citys', help='dataset name (default: citys)')
25
- parser.add_argument('--onnx_path', type=str, default='FPN_int.onnx', help='onnx path')
26
  parser.add_argument('--num-classes', type=int, default=19,
27
  help='the classes numbers (default: 19 for cityscapes)')
28
  parser.add_argument('--test-folder', type=str, default='./data/cityscapes',
@@ -78,8 +78,8 @@ def eval_miou(data,path="FPN_int.onnx", device='cpu'):
78
 
79
  for i, (image, target) in enumerate(tbar):
80
  image, target = image.to(device), target.to(device)
81
- ort_input = {session.get_inputs()[0].name: image.cpu().numpy()}
82
- ort_output = session.run(None, ort_input)[0]
83
  if isinstance(ort_output, (tuple, list)):
84
  ort_output = ort_output[0]
85
  ort_output = torch.from_numpy(ort_output).to(device)
 
22
  # dataset
23
 
24
  parser.add_argument('--dataset', type=str, default='citys', help='dataset name (default: citys)')
25
+ parser.add_argument('--onnx_path', type=str, default='FPN_int_NHWC.onnx', help='onnx path')
26
  parser.add_argument('--num-classes', type=int, default=19,
27
  help='the classes numbers (default: 19 for cityscapes)')
28
  parser.add_argument('--test-folder', type=str, default='./data/cityscapes',
 
78
 
79
  for i, (image, target) in enumerate(tbar):
80
  image, target = image.to(device), target.to(device)
81
+ ort_input = {session.get_inputs()[0].name: image.cpu().numpy().transpose(0,2,3,1)}
82
+ ort_output = session.run(None, ort_input)[0].transpose(0,3,1,2)
83
  if isinstance(ort_output, (tuple, list)):
84
  ort_output = ort_output[0]
85
  ort_output = torch.from_numpy(ort_output).to(device)