language model 0.0025277700000000004
language models 0.00240211
same model 0.0019059160000000001
small data 0.001817144
word counts 0.001794396
large model 0.001779161
estimate language 0.0017061580000000002
language modeling 0.0017039870000000001
language pairs 0.001665033
data structure 0.001645545
language pair 0.001644722
last word 0.001641587
interpolation model 0.00163272
trie data 0.001627948
ent data 0.0016192379999999998
strained data 0.001616667
unpruned language 0.0016142790000000001
route data 0.001610077
data flow 0.001610077
model figure 0.001543408
small models 0.001510864
unknown word 0.001451946
penultimate word 0.001450181
unpruned model 0.001434669
language 0.00135369
different backoff 0.001348135
context order 0.001321025
translation task 0.001317209
machine translation 0.001298399
final probability 0.001244465
model 0.00117408
same streaming 0.001166192
translation experiments 0.001153867
other layer 0.001145798
same file 0.001144988
same ram 0.001127631
pseudo probability 0.001101111
test machine 0.0010914940000000001
same pass 0.001081768
test set 0.001067007
smoothing methods 0.001057295
models 0.00104842
single value 0.001048326
backoff file 0.001031357
other toolkits 0.001031283
other cases 0.00102075
same level 0.0010138740000000001
suffix order 9.96972E-4
efficient algorithm 9.86593E-4
last task 9.5145E-4
lexicographic order 9.50955E-4
single machine 9.48066E-4
first pass 9.289669999999999E-4
adjusted count 9.27081E-4
count zero 9.2246E-4
count pruning 9.2246E-4
secondary backoff 9.15029E-4
crawl corpus 9.143980000000001E-4
interpolation step 9.11426E-4
exact integer 9.10881E-4
stupid backoff 9.031060000000001E-4
feature weights 8.976470000000001E-4
gram counts 8.95765E-4
input file 8.94462E-4
local disk 8.9395E-4
memory mapping 8.91714E-4
virtual memory 8.86921E-4
smoothing statistics 8.773100000000001E-4
previous work 8.76597E-4
uniform distribution 8.76143E-4
adjusting counts 8.721620000000001E-4
large amounts 8.693780000000001E-4
adjusted counts 8.67692E-4
counts divisionsumming 8.67587E-4
estimation comparison 8.67546E-4
low perplexity 8.64003E-4
counts the 8.62919E-4
work figure 8.625060000000001E-4
cpu time 8.6176E-4
memory limit 8.61229E-4
vocabulary size 8.55478E-4
memory usage 8.550599999999999E-4
memory budgeting 8.45458E-4
estimation pipeline 8.44515E-4
pipeline estimation 8.44515E-4
wall time 8.32974E-4
query time 8.19031E-4
local scores 8.189580000000001E-4
translation 8.17466E-4
probability 8.12852E-4
estimation code 8.09287E-4
related work 8.06694E-4
words 8.06317E-4
much work 7.991790000000001E-4
compression methods 7.99059E-4
bleu scores 7.98425E-4
ditional feature 7.95741E-4
lion tokens 7.899250000000001E-4
hash table 7.88648E-4
future work 7.88363E-4
