osmarks/meme-search-engine
1
0
Fork 0
mirror of https://github.com/osmarks/meme-search-engine.git synced 2024-11-13 07:19:54 +00:00
Code Issues Releases Wiki Activity

Commit Graph

  • 8097ce8d91 improve dump processing and misc performance fixes master osmarks 2024-11-11 19:43:07 +0000
  • c277b49dc1 fix resumption, oops osmarks 2024-11-07 20:43:26 +0000
  • b9bb629e6f performance improvements osmarks 2024-11-07 16:52:58 +0000
  • 7fa14d45ae improve observability and fix up Reddit dump for full-scale run osmarks 2024-11-02 19:38:05 +0000
  • 1d0ff95955 Sparse autoencoder testing osmarks 2024-10-05 17:22:44 +0100
  • fc6d0c9409 Fix crawler (rate limit changes) osmarks 2024-07-15 18:48:55 +0100
  • 43ff9215fb File metadata storage. The backend now knows how big images are, so the frontend can size images correctly before they're loaded. This should significantly improve the UI on slow connections. Also fix bug where videos weren't erased from the index properly. osmarks 2024-06-26 20:02:12 +0100
  • 1ab254ff1d Adjust index storage for memory efficiency and fix SQLite interface type confusion osmarks 2024-06-25 08:23:30 +0100
  • e7adf738f6 Fix typo, in the sense of application-killing bug. osmarks 2024-05-31 00:35:16 +0100
  • 747058e254 misc fixes - thumbnails/OCR off was broken - problematic video files caused segfaults (I blame ffmpeg for this) osmarks 2024-05-30 19:05:54 +0100
  • 3257521068 Video search osmarks 2024-05-30 15:58:31 +0100
  • 74d91d52e5 probably I should do better testing osmarks 2024-05-29 21:25:50 +0100
  • 5eae8674ce video parsing basics osmarks 2024-05-28 22:28:41 +0100
  • 129b769a56 hackily patch horrifyingly nondeterministic-but-fast image encoder in osmarks 2024-05-27 20:21:44 +0100
  • d4e136b6a7 AITemplate builds of the image encoder work, at great personal cost osmarks 2024-05-27 19:05:25 +0100
  • a8329e43fc more progress on Reddit osmarks 2024-05-27 15:22:28 +0100
  • f8d68d9d54 WIP Reddit dump loader osmarks 2024-05-24 17:47:18 +0100
  • 978aadda6a Improved UI for sliders osmarks 2024-05-22 20:26:23 +0100
  • d8c147df52 Predefined embedding modes in search osmarks 2024-05-22 20:17:13 +0100
  • 14387a61a3 refactor configuration osmarks 2024-05-22 19:02:34 +0100
  • ffc3d648a6 basic monitoring implementation osmarks 2024-05-22 18:49:32 +0100
  • ce590298a7 concurrent index queries and fix database typo yet again osmarks 2024-05-22 18:25:50 +0100
  • 349fe802f7 meme interpretability osmarks 2024-05-22 16:18:45 +0100
  • bd426a30ba Port meme acquisition pipeline to new API, database Also fix a really stupid oversight in crawling code. osmarks 2024-05-22 15:43:56 +0100
  • 30b1b72712 I really should test database queries better osmarks 2024-05-22 14:35:29 +0100
  • 9455438bab frontend fixes osmarks 2024-05-21 20:17:23 +0100
  • 24fbc0dd1f apparently, that quality value is too low osmarks 2024-05-21 20:09:28 +0100
  • 63a9b3d9a6 the consequences of my own actions osmarks 2024-05-21 12:39:04 +0100
  • e705a9db21 I hate precedence osmarks 2024-05-21 12:33:32 +0100
  • b7010b41dd oops osmarks 2024-05-21 01:58:50 +0100
  • 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. osmarks 2024-05-21 00:09:04 +0100
  • fa863c2075 "release" unfinished scripts and miscellaneous JSON files osmarks 2024-05-18 14:34:30 +0100
  • caa8306ff7 oops osmarks 2024-05-18 13:21:13 +0100
  • 6491e02e88 preliminary work on OCR osmarks 2024-05-18 00:39:05 +0100
  • a3574674d0 "documentation" osmarks 2024-04-27 17:33:24 +0100
  • 2447e134ef There were more memes than anticipated. osmarks 2024-04-27 17:24:01 +0100
  • 80db16d02a full pipeline osmarks 2024-04-22 18:44:29 +0100
  • 7bae095384 accidentally lost some manual labels, oops, etc osmarks 2024-04-22 13:54:07 +0100
  • cebb4f9d00 better evals osmarks 2024-04-22 13:43:06 +0100
  • 58ce70bb5e meme rater model code (documentation "later") osmarks 2024-04-21 23:50:48 +0100
  • 0b0261f625 preliminary meme rater work osmarks 2024-04-20 16:55:11 +0100
  • e9a7493343 stop requiring internal aiosqlite patch osmarks 2024-01-25 00:01:02 +0000
  • e3ffc426b7 Actually delete missing files osmarks 2024-01-02 14:12:26 +0000
  • 4626f53bcb Return to OpenCLIP osmarks 2023-11-13 17:31:43 +0000
  • 74bb1bc343 thumbnailer system osmarks 2023-10-27 15:50:21 +0100
  • 5b5ef271aa ""documentation"" osmarks 2023-10-09 12:35:26 +0100
  • 68a14d7da9 unignore device osmarks 2023-10-08 22:54:06 +0100
  • 20fcc9317f forgot the README osmarks 2023-10-08 22:52:57 +0100
  • 46fca3eb7f faster indexing, SigLIP models osmarks 2023-10-08 22:52:17 +0100
  • 2c9ce67ab2 I really should do better testing osmarks 2023-09-30 21:10:27 +0100
  • c235217751 read from query strings osmarks 2023-09-30 15:56:04 +0100
  • 8abe384cd1 fix concurrency bug osmarks 2023-09-29 18:35:42 +0100
  • 120edc1a47 oops osmarks 2023-09-29 18:34:06 +0100
  • 33f4382146 also typo osmarks 2023-09-28 17:32:41 +0100
  • aa6163a784 typo osmarks 2023-09-28 17:30:44 +0100
  • 9d89e6e4f5 initial commit osmarks 2023-09-28 17:30:20 +0100