1
0
mirror of https://github.com/osmarks/meme-search-engine.git synced 2024-09-21 10:09:36 +00:00
Commit Graph

38 Commits

Author SHA1 Message Date
d8c147df52 Predefined embedding modes in search 2024-05-22 20:17:13 +01:00
14387a61a3 refactor configuration 2024-05-22 19:02:34 +01:00
ffc3d648a6 basic monitoring implementation 2024-05-22 18:49:32 +01:00
ce590298a7 concurrent index queries and fix database typo yet again 2024-05-22 18:25:50 +01:00
349fe802f7 meme interpretability 2024-05-22 16:18:45 +01:00
bd426a30ba Port meme acquisition pipeline to new API, database
Also fix a really stupid oversight in crawling code.
2024-05-22 15:43:56 +01:00
30b1b72712 I really should test database queries better 2024-05-22 14:35:29 +01:00
9455438bab frontend fixes 2024-05-21 20:17:23 +01:00
24fbc0dd1f apparently, that quality value is too low 2024-05-21 20:09:28 +01:00
63a9b3d9a6 the consequences of my own actions 2024-05-21 12:39:04 +01:00
e705a9db21 I hate precedence 2024-05-21 12:33:32 +01:00
b7010b41dd oops 2024-05-21 01:58:50 +01:00
7cb42e028f Rewrite entire application (well, backend) in Rust and also Go
I decided I wanted to integrate the experimental OCR thing better, so I rewrote in Go and also integrated the thumbnailer.
However, Go is a bad langauge and I only used it out of spite.
It turned out to have a very hard-to-fix memory leak due to some unclear interaction between libvips and both sets of bindings I tried, so I had Claude-3 transpile it to Rust then spent a while fixing the several mistakes it made and making tweaks.
The new Rust version works, although I need to actually do something with the OCR data and make the index queryable concurrently.
2024-05-21 00:09:04 +01:00
fa863c2075 "release" unfinished scripts and miscellaneous JSON files 2024-05-18 14:34:30 +01:00
caa8306ff7 oops 2024-05-18 13:21:13 +01:00
6491e02e88 preliminary work on OCR 2024-05-18 00:39:05 +01:00
a3574674d0 "documentation" 2024-04-27 17:33:24 +01:00
2447e134ef There were more memes than anticipated. 2024-04-27 17:24:01 +01:00
80db16d02a full pipeline 2024-04-22 18:44:29 +01:00
7bae095384 accidentally lost some manual labels, oops, etc 2024-04-22 13:54:07 +01:00
cebb4f9d00 better evals 2024-04-22 13:43:06 +01:00
58ce70bb5e meme rater model code (documentation "later") 2024-04-21 23:50:48 +01:00
0b0261f625 preliminary meme rater work 2024-04-20 16:55:11 +01:00
e9a7493343 stop requiring internal aiosqlite patch 2024-01-25 00:01:02 +00:00
e3ffc426b7 Actually delete missing files 2024-01-02 14:12:26 +00:00
4626f53bcb Return to OpenCLIP 2023-11-13 17:31:43 +00:00
74bb1bc343 thumbnailer system 2023-10-27 15:50:21 +01:00
5b5ef271aa ""documentation"" 2023-10-09 12:35:26 +01:00
68a14d7da9 unignore device 2023-10-08 22:54:06 +01:00
20fcc9317f forgot the README 2023-10-08 22:52:57 +01:00
46fca3eb7f faster indexing, SigLIP models 2023-10-08 22:52:17 +01:00
2c9ce67ab2 I really should do better testing 2023-09-30 21:10:27 +01:00
c235217751 read from query strings 2023-09-30 15:56:04 +01:00
8abe384cd1 fix concurrency bug 2023-09-29 18:35:42 +01:00
120edc1a47 oops 2023-09-29 18:34:06 +01:00
33f4382146 also typo 2023-09-28 17:32:41 +01:00
aa6163a784 typo 2023-09-28 17:30:44 +01:00
9d89e6e4f5 initial commit 2023-09-28 17:30:20 +01:00