treelstm.pytorch

treelstm.pytorch copied to clipboard

Hi, I was going through another implementation of tree LSTM from https://github.com/ttpro1995/TreeLSTMSentiment where for sentiment classification, they consider the hidden state values instead of the cell state values. The original...

navid5792

Hi, I am trying to use this model for parse sentences in German with the dependency parser that is used in this code. So, the DependencyParse.java file has the following...

inigo-jauregi

Constituency trees showed slightly lower performance in the paper, but there should be people wanting to use them under the belief that phrase structures are appropriate for different purposes, including...

yukiar

In your code: ```python if best < test_pearson: best = test_pearson checkpoint = { 'model': trainer.model.state_dict(), 'optim': trainer.optimizer, 'pearson': test_pearson, 'mse': test_mse, 'args': args, 'epoch': epoch } logger.debug('==> New optimum...

feiaa

I noticed that lr in code is 0.01 while paper is 0.05 with adagrad, and I tried with 0.05 to train the model but train loss doesn't decrease at all,...

ryh95

Why zero out embeddings for special words if they are absent in vocab

2

Hi, I noticed that in `main.py`, you zero out the embeddings for special words if they are absent in vocabulary: ```python # zero out the embeddings for padding and other...

Silenthinker

Why move output to cpu?

1

I noticed that in **test** function of **trainer** module you write `output = output.data.squeeze().cpu()` Why you move `output` to cpu ? By the way in **SimilarityTreeLSTM** of **model** module `output...

ryh95

Batched

1

This brings the rewritten, more efficient model from https://github.com/zackchase/mxnet-the-straight-dope/tree/master/chapter09_natural-language-processing - On 0.2 this is about 50% faster on CPU (OMP_NUM_THREADS=1) - On master this is about 3x faster on CPU...

soumith