pytorch-tree-lstm
pytorch-tree-lstm copied to clipboard
unbounce
Pytorch implementation of the child-sum Tree-LSTM model
_label_node_index function in example_usage.py is giving different nodes same label because of the recursion. I think it should be something like below code to retain the last value assigned to...
Commit 66f29a44e98c7332661b57d22501107bcb193f90 modified `TreeLSTM` class, specifically, regarding `h_sum`. By using `torch.unique_consecutive` to construct `h_sum` on the fly seems to assume that if two edges have the same parent, they must...
Every node in tree has one and only one parent, except for root node (that doesn't have a parent). This is means that E = N - 1 E: number...
https://github.com/unbounce/pytorch-tree-lstm/blob/66f29a44e98c7332661b57d22501107bcb193f90/treelstm/tree_lstm.py#L92
Hi, I find this optimized version of tree-lstm interesting and I have took a loot at the code. It seems that the method “_label_node_index” does not guarantee one-to-one mapping when...
We are using Dataloader without collate_fn, which mean with one tree at a time. Inside forward function, the batch_size will be 1(becuase of node_order.shape[0]) and that will be used to...
Great work! My question is: with batch_tree_input and train_data_generator, it's possible to train a batch of trees right? Could you please provide an example how to generate the TreeDataset() in...
Hi, I think the PyTorch versions specified in the REAME are not correct. When I run `example_usage.py` with either PyTorch 0.4.0 or PyTorch 1.0.1 I have the error `module 'torch'...
Thank you for your implementation of PyTorch Tree LSTM. In your model, however, it seems like there are many in-place operations? Will this cause any problems with autograd? For example,...