add Action-LSTM

[hans]

Available in the rnng-incremental repo. Peng's instructions:

you can find the model params for ActionLSTM here: /om/user/pqian/shared/ntparse_gen_D0.3_2_256_256_256_256-pid18099.params

And here is the command line to run evaluation:

build/nt-parser/nt-parser-gen-action-only --dynet-mem 2000  -x -T train_gen.oracle \
 -v $1 -f $2 --clusters clusters-train-berk.txt --input_dim 256  --action_dim 256 \
--lstm_input_dim 256 --hidden_dim 256 \
-m ntparse_gen_D0.3_2_256_256_256_256-pid18099.params

Likewise, you should be able to set up beam_size, fasttrack_beam_size and word_beam_size through the command line. The surprisal is computed on base 2. Usually the value for beam_size >> word_beam_size. The default value is the following one used in all previous experiments: beam_size=100, fasttrack_beam_size=5, word_beam_size=10.

[hans]

JRNN auto eval