“RNN test”版本间的差异

2014年11月3日 (一) 06:26的最后版本

@@ 第1行： / 第1行： @@
-==140901==
+==tool==
+*LTSM/RNN training, GPU&deep supported [http://sourceforge.net/projects/currennt/]
+* RNNLM: RNN LM toolkit [http://www.fit.vutbr.cz/~imikolov/rnnlm/]
+* RWTHLM: RNN LTSM toolkit [http://www-i6.informatik.rwth-aachen.de/web/Software/rwthlm.php]
+* nplm: NN LM, large scale data [http://nlg.isi.edu/software/nplm/]
+* RNN toolkit from microsoft [http://research.microsoft.com/en-us/projects/rnn/]
+* cslm [http://www-lium.univ-lemans.fr/~cslm/]
-=== wsj_data ===
+==paper==
+*[[14-9-30]]
+*[[2014-10-9]]
-.parameter
+==Steps==
-      rand_seed=1
+[[process dict and data]]
-      nwords=10000 # This is how many words we're putting in the vocab of the RNNLM.
+==Test==
-      hidden=320
+*[[wsj_data]]
-      class=300 # Num-classes... should be somewhat larger than sqrt of nwords.
+*[[chinese_data_gigword]]
-      direct=2000 # Number of weights that are used for "direct" connections, in millions.
+*[[jt-chinese]]
-      rnnlm_ver=rnnlm-0.3e # version of RNNLM to use
-      threads=1 # for RNNLM-HS
-      bptt=2 # length of BPTT unfolding in RNNLM
-      bptt_block=20 # length of BPTT unfolding in RNNLM
-{| border="2px"
-|+ Train Set Environment
-|-
-! Parameters  !! hidden !! class !! direct !! bbt !! bptt_block !! threads !!direct-order!!rand_seed!!nwords!!time(min)
-|-
-!set1
-| 320 || 300 || 2000 || 2 || 20 || 1 || 4 || 1 || 10000||3380(56h)
-|-}