Open TortoiseHam opened 4 years ago
When a pytorch model which has multiple inputs passing through the same initial layers (ex. Siamese network), hiddenlayer draws this as 2 separate network branches rather than one branch with 2 inputs. An example network which exhibits this issue can be found at: https://github.com/fastestimator/fastestimator/blob/master/apphub/one_shot_learning/siamese_network/siamese_torch.py
When a pytorch model which has multiple inputs passing through the same initial layers (ex. Siamese network), hiddenlayer draws this as 2 separate network branches rather than one branch with 2 inputs. An example network which exhibits this issue can be found at: https://github.com/fastestimator/fastestimator/blob/master/apphub/one_shot_learning/siamese_network/siamese_torch.py