14-10-14 Dongxu Zhang

Accomplished this week

• Built a raw vocabulary(166K) for lm from 700G corpora
• compiled rwthlm tool and read some papers about rnn, lstm-rnn.

Planned for next week

• finish vocabulary task.
• use rwthlm to train lm on a 200MB corpora. Figure out the use of parameters and get some feel about how to tune it.