issues
search
lhwcv
/
mlsd_pytorch
Pytorch implementation of "M-LSD: Towards Light-weight and Real-time Line Segment Detection"
Apache License 2.0
189
stars
37
forks
source link
How to convert the model to onnx format?
#8
Open
magicChos
opened
2 years ago
magicChos
commented
2 years ago
my conversion seems wrong.
thanks very much!
SuperChay
commented
2 years ago
convert the model to onnx
```java import torch from models.mbv2_mlsd_large import MobileV2_MLSD_Large def pth2onnx(): model_path = '../models/mlsd_large_512_fp32.pth' model = MobileV2_MLSD_Large().cuda().eval() device = torch.device('cuda' if torch.cuda.is_available() else 'cpu') model.load_state_dict(torch.load(model_path, map_location=device), strict=True) dummy_input = torch.randn(1, 4, 512, 512, device=device) torch.onnx.export(model, dummy_input, "../models/test.onnx", verbose=True, opset_version=11) if __name__ == '__main__': pth2onnx() ```
my conversion seems wrong.
thanks very much!