Krasimir Angelov
|
bb4218433f
|
- tweak the tokenizer in pgf_lookup_sentence to threat .!?,: as separate tokens
+ bugfix which causes crashes
|
2018-02-22 11:35:54 +01:00 |
|
Krasimir Angelov
|
4a1d19086a
|
added one more include to define alloca for Windows
|
2017-12-09 23:06:34 +01:00 |
|
Krasimir Angelov
|
527d97fdd2
|
fix the type signatures for the linearizer callbacks
|
2017-09-08 22:58:42 +02:00 |
|
Krasimir Angelov
|
a8eeb49767
|
the linearizer API now allows to detect metavariables. This is used for instancein the lookup where the tokens produced from the metavariables are interpreted as distinct from all other tokens.
|
2017-08-07 16:39:19 +02:00 |
|
Krasimir Angelov
|
a4c19875ed
|
fix the sentence lookup for cases when the top category is undefined.
|
2017-08-07 14:47:20 +02:00 |
|
krasimir
|
3464e912c1
|
filtering during the lookup nicely integrates with concretization
|
2017-06-27 19:30:43 +00:00 |
|
krasimir
|
e8726b1cfa
|
faster lookup
|
2017-06-23 17:52:43 +00:00 |
|
krasimir
|
2af8ffb269
|
fix in the lookup debuger
|
2017-06-01 09:08:42 +00:00 |
|
krasimir
|
20a4887311
|
reintroduce counts as a prefilter before applying cosine similarity
|
2017-06-01 09:06:35 +00:00 |
|
krasimir
|
2036606d61
|
more debug information in the lookup
|
2017-06-01 08:27:02 +00:00 |
|
krasimir
|
2a8d2806e4
|
lookup for syntagmatic words
|
2017-05-30 12:35:02 +00:00 |
|
krasimir
|
bf3557cf90
|
finally a smoothed and more precise ranking for lookups
|
2017-05-30 11:42:32 +00:00 |
|
krasimir
|
8de69b0c31
|
an optimized string kernel
|
2017-05-24 19:44:32 +00:00 |
|
krasimir
|
cb67a24c99
|
an optimized expression extraction in the lookup
|
2017-05-23 21:06:17 +00:00 |
|
krasimir
|
622274180a
|
a naive buf working ranking in the sentence lookup
|
2017-05-23 14:07:55 +00:00 |
|
krasimir
|
08f9f6fbaf
|
bugfix in the stack handling during lookup
|
2017-05-16 10:55:00 +00:00 |
|
krasimir
|
578f84dd4f
|
switch off the debugging mode in the sentence lookup
|
2017-05-12 11:13:15 +00:00 |
|
krasimir
|
7db25c5a74
|
bugfix in the sentence lookup
|
2017-05-12 10:51:49 +00:00 |
|
krasimir
|
577b7662d8
|
the first prototype for sentence lookup
|
2017-05-12 09:31:47 +00:00 |
|
krasimir
|
7dd171cdcc
|
optimize memory allocation in the lookup by using worker pools
|
2017-05-04 08:29:51 +00:00 |
|
krasimir
|
a05194a350
|
remove duplicate entries from the lookup index
|
2017-05-04 08:21:29 +00:00 |
|
krasimir
|
73e9eb934a
|
merging spines
|
2017-05-04 07:55:11 +00:00 |
|
krasimir
|
d0269e8513
|
first steps for sentence lookup in the C runtime
|
2017-05-03 16:28:20 +00:00 |
|