15 Commits

Author SHA1 Message Date
joachim-danswer
463340b8a1
Reduce ranking scores for short chunks without actual information (#4098)
* remove title for slack

* initial working code

* simplification

* improvements

* name change to information_content_model

* avoid boost_score > 1.0

* nit

* EL comments and improvements

Improvements:
  - proper import of information content model from cache or HF
  - warm up for information content model

Other:
  - EL PR review comments

* nit

* requirements version update

* fixed docker file

* new home for model_server configs

* default off

* small updates

* YS comments - pt 1

* renaming to chunk_boost & chunk table def

* saving and deleting chunk stats in new table

* saving and updating chunk stats

* improved dict score update

* create columns for individual boost factors

* RK comments

* Update migration

* manual import reordering
2025-03-13 17:35:45 +00:00
pablodanswer
21ec5ed795 welcome to onyx 2024-12-13 09:56:10 -08:00
Shukant Pal
362156f97e
Model inference for connector classifier on queries (#2137) 2024-09-08 14:46:00 -07:00
Yuhong Sun
3b035d791e
Fix Model Server (#2191) 2024-08-20 17:57:09 -07:00
Yuhong Sun
5ab4d94d94
Logging Level Update (#2165) 2024-08-18 21:53:40 -07:00
hagen-danswer
61b5bd569b
Reworked chunking to support mega chunks (#2032) 2024-08-14 22:18:53 -07:00
Yuhong Sun
1c10f54294
GPU Model Server (#2135) 2024-08-14 11:04:28 -07:00
Yuhong Sun
7dcc42aa95
Intent Model Update (#2069) 2024-08-08 20:45:53 -07:00
hagen-danswer
1be1959d80
Changed default local model to nomic (#1943) 2024-07-31 18:54:02 -07:00
Yuhong Sun
b59912884b
Fix Model Server (#1320) 2024-04-10 23:13:22 -07:00
Yuhong Sun
2db906b7a2
Always Use Model Server (#1306) 2024-04-07 21:25:06 -07:00
Chris Weaver
e94fd8b022
Remove un-needed imports (#999) 2024-01-25 12:10:19 -08:00
Yuhong Sun
2e6be57880
Model Server CVEs (#898) 2023-12-29 21:14:08 -08:00
Yuhong Sun
c3cf9134bb
Telemetry Revision (#868) 2023-12-24 17:39:37 -08:00
Yuhong Sun
7433dddac3
Model Server (#695)
Provides the ability to pull out the NLP models into a separate model server which can then be hosted on a GPU instance if desired.
2023-11-06 16:36:09 -08:00