yujunjun/olmocr
1
0
Fork 0
mirror of https://github.com/allenai/olmocr.git synced 2025-10-09 15:23:08 +00:00
Code Issues Packages Projects Releases Wiki Activity

Commit Graph

  • 1ba028658c
         Bump sphinx-autodoc-typehints from 1.23.3 to 3.3.0 dependabot/pip/sphinx-autodoc-typehints-3.3.0 dependabot[bot] 2025-10-08 22:11:49 +00:00
  • 95dd21b66c GRPO Documentation jakep/new_data Jake Poznanski 2025-10-07 20:40:10 +00:00
  • 7ef6020b46 Bump version to v0.3.9 for release main v0.3.9 Jake Poznanski 2025-10-07 20:00:11 +00:00
  • 9fb29bcde4 Rebuild with new runner so docker image gets made Jake Poznanski 2025-10-07 20:00:05 +00:00
  • 1f791c4a19 Changes Jake Poznanski 2025-10-07 18:29:08 +00:00
  • b99d149a43 Merge branch 'jakep/new_data' into jakep/vllm_beam jakep/vllm_beam Jake Poznanski 2025-10-07 18:18:28 +00:00
  • 727b345715 Merge fix Jake Poznanski 2025-10-07 18:16:31 +00:00
  • 983c5fbbeb Adding new runs-on bigger runner Jake Poznanski 2025-10-07 17:57:28 +00:00
  • 8ef68fde88 Merge branch 'main' into jakep/new_data Jake Poznanski 2025-10-07 17:44:54 +00:00
  • e15615aadb Model defaults Jake Poznanski 2025-10-07 17:10:45 +00:00
  • 44f6b9f0de Trying beam search idea Jake Poznanski 2025-10-07 04:00:39 +00:00
  • b81e40602d Readme score fixes Jake Poznanski 2025-10-06 22:59:00 +00:00
  • 2e3d1a0317 Comitting test script to be used in model cards for individual one-off inference Jake Poznanski 2025-10-06 22:47:06 +00:00
  • c89787183a Bump version to v0.3.8 for release v0.3.8 Jake Poznanski 2025-10-06 21:46:18 +00:00
  • e12941a608 Version bump Jake Poznanski 2025-10-06 21:46:10 +00:00
  • 7fe756fe63 Formatting Jake Poznanski 2025-10-06 21:10:32 +00:00
  • 9c7c670f1f Bump version to v0.3.7 for release v0.3.7 Jake Poznanski 2025-10-06 21:10:07 +00:00
  • 1951a849ec Version bump with new vllm Jake Poznanski 2025-10-06 21:10:00 +00:00
  • 75095f4a05
         Bump transformers from 4.53.2 to 4.57.0 dependabot/pip/transformers-4.57.0 dependabot[bot] 2025-10-06 20:34:59 +00:00
  • c75f5b98a1 Cleaning up pr 341 arguments to match with vllm 0.11, which only has V1 engine and thus always does chunked prefill. And fixes arg syntax Jake Poznanski 2025-10-06 20:26:41 +00:00
  • e202c22822 Merge branch 'vllm_0_11' Jake Poznanski 2025-10-06 20:24:26 +00:00
  • 2b70b50312
         Merge pull request #341 from charitarthchugh/charitarthchugh/vllm-defaults-speedup Jake Poznanski 2025-10-06 13:23:47 -07:00
  • 81be6f5c1f Transformers version vllm_0_11 Jake Poznanski 2025-10-06 19:52:55 +00:00
  • 9b517a02be Git lfs in docker image Jake Poznanski 2025-10-06 19:47:19 +00:00
  • 9feb41af82 New docker file approach for vllm 0.11 Jake Poznanski 2025-10-06 18:57:16 +00:00
  • 59266ed419 More readmes Jake Poznanski 2025-10-01 22:27:37 +00:00
  • 476ba212dc Bolds Jake Poznanski 2025-10-01 22:05:58 +00:00
  • bb7790a138 Bolds Jake Poznanski 2025-10-01 22:05:30 +00:00
  • 4e68b174bf New bench scores added Jake Poznanski 2025-10-01 22:04:40 +00:00
  • 8ef7f8085a isort and black Jake Poznanski 2025-09-30 17:37:10 +00:00
  • b5b1de98dd Allowing more max tokens in pipeline for new models Jake Poznanski 2025-09-29 22:12:27 +00:00
  • f4356de091 deepinfra readme improved Jake Poznanski 2025-09-29 17:56:03 +00:00
  • 8982bae756 Bump version to v0.3.6 for release v0.3.6 Jake Poznanski 2025-09-29 17:37:25 +00:00
  • fb1ef9e38a Release script fix Jake Poznanski 2025-09-29 17:37:14 +00:00
  • c587eb9050 Ugh, release script adds all files by default Jake Poznanski 2025-09-29 17:36:41 +00:00
  • 3670b2088f Bump version to v0.3.5 for release v0.3.5 Jake Poznanski 2025-09-29 17:30:04 +00:00
  • 38d0361226 Version bump Jake Poznanski 2025-09-29 17:29:58 +00:00
  • a0bc5a4690 Deepinfra readme Jake Poznanski 2025-09-29 17:29:28 +00:00
  • 0c6d889863 Adding retry code on 429 errors from exteranl providers Jake Poznanski 2025-09-29 17:26:22 +00:00
  • 9c750903ce Ignore files Jake Poznanski 2025-09-29 17:06:14 +00:00
  • f0caa188ab
         Merge pull request #344 from allenai/amanr/deepinfra Jake Poznanski 2025-09-29 10:03:29 -07:00
  • 7f4b728dcd Skip docker checks if using beaker image Jake Poznanski 2025-09-26 23:27:04 +00:00
  • f3c4073395 added Api_key argument to pipeline pytests amanr/deepinfra aman-17 2025-09-26 14:25:25 -07:00
  • 359abef654 updated pytests aman-17 2025-09-26 14:19:22 -07:00
  • 1099820bf6 Hmm, going to have to train a filtered version Jake Poznanski 2025-09-26 21:15:15 +00:00
  • 556ff26d58 fixed lint, style, ruff aman-17 2025-09-26 14:08:40 -07:00
  • e7ae5e6240 fixed style aman-17 2025-09-26 13:58:34 -07:00
  • 90589e16de Added deepinfra usage to readme aman-17 2025-09-26 13:56:34 -07:00
  • 2a5792e5ed add if else for vllm local usage bug for api argument aman-17 2025-09-26 13:29:48 -07:00
  • bb06829840 SOuping in fp32 Jake Poznanski 2025-09-26 20:03:29 +00:00
  • e1bc7b8861 fixed lint and style amanr/dotsocr aman-17 2025-09-26 19:44:14 +00:00
  • 7dc6a4b2a5 Added bench eval for dotsocr aman-17 2025-09-26 18:31:59 +00:00
  • 7fe3f65de7 added support for deepinfra aman-17 2025-09-26 11:06:51 -07:00
  • a5e73b031b
         Bump furo from 2023.7.26 to 2025.9.25 dependabot/pip/furo-2025.9.25 dependabot[bot] 2025-09-25 22:11:44 +00:00
  • fe425fde20
         Add chunked prefill and limit mm per prompt options Charitarth Chugh 2025-09-25 14:29:49 -04:00
  • b868e2c35c Basing off of vllm official docker image now jakep/vllm_0_10_2 Jake Poznanski 2025-09-24 21:48:41 +00:00
  • 1669df3596 Basing off of vllm official docker image now Jake Poznanski 2025-09-24 21:48:07 +00:00
  • 36c22279da trying vllm Jake Poznanski 2025-09-24 21:29:11 +00:00
  • 3d6e6a6a01 Setting defaults, adding in seed Jake Poznanski 2025-09-24 17:37:12 +00:00
  • 1654c760d9 Fixing grpo eos reward Jake Poznanski 2025-09-23 20:27:07 +00:00
  • b4b121b118 Testing reward eos Jake Poznanski 2025-09-23 20:10:11 +00:00
  • 01bc1ff7b6 Allow passing in beaker image to run benchmark Jake Poznanski 2025-09-23 19:48:42 +00:00
  • a00d9d172e Adding stricter math and table tests when in synthetic mode Jake Poznanski 2025-09-23 18:37:50 +00:00
  • 1197c35808 Mix contamination checker script Jake Poznanski 2025-09-23 18:17:13 +00:00
  • 9818797fbc One last try to disable timeouts in katex thing Jake Poznanski 2025-09-23 03:26:45 +00:00
  • 022d87fb7a Another attempt at playwright rendering fix for long RL jobs Jake Poznanski 2025-09-22 20:47:58 +00:00
  • 97db9cdfb1 Hmm Jake Poznanski 2025-09-22 18:46:55 +00:00
  • 4e77ecac3f Ok, should have a cleaner playwright story on multithreaded environments Jake Poznanski 2025-09-22 18:14:30 +00:00
  • 1a64728420 Checking to see if specifying number of dataloader workers helps thread issue Jake Poznanski 2025-09-22 17:08:06 +00:00
  • 3b20322eb0 Adding some more debug stuff to try to figure out playwright issue Jake Poznanski 2025-09-22 16:03:49 +00:00
  • 780bc7d934 Better context manager and cleanup of old browser instances Jake Poznanski 2025-09-21 23:38:09 +00:00
  • 7e786c79c5 Save total limit Jake Poznanski 2025-09-20 19:52:34 +00:00
  • 83d965c768 Adding contam check for olmocr-bench when making synth data Jake Poznanski 2025-09-20 03:42:42 +00:00
  • 68defa23d7 fixed dotsocr runner amanr/adding_ocr_models aman-17 2025-09-19 16:14:06 -07:00
  • 4f7623c429 fixed dotsocr runner aman-17 2025-09-19 16:14:05 -07:00
  • 796c021ab8 added dotsocr aman-17 2025-09-19 13:34:48 -07:00
  • a8ad6c12b5 Convert fix Jake Poznanski 2025-09-19 17:36:34 +00:00
  • b1242db8e2 last threading fix Jake Poznanski 2025-09-19 17:23:12 +00:00
  • 0a74746da9 Ugh Jake Poznanski 2025-09-19 17:20:28 +00:00
  • f114a7fbec Crahses Jake Poznanski 2025-09-19 17:19:12 +00:00
  • b52ac23073 Fixing non async threading Jake Poznanski 2025-09-19 17:12:22 +00:00
  • e607b53748 Keeps getting killed Jake Poznanski 2025-09-19 17:02:02 +00:00
  • cedd4a80cf Fixing paddle ocr to run fast Jake Poznanski 2025-09-19 16:57:05 +00:00
  • 8f75ea062e testing paddle Jake Poznanski 2025-09-19 16:55:01 +00:00
  • e9ab2fd1bb Adding paddlepaddle v5 runner for benchmarking Jake Poznanski 2025-09-19 16:45:53 +00:00
  • 1c703917df Synthmix ignore Jake Poznanski 2025-09-19 15:58:58 +00:00
  • 4907b1c700 Bumping rotation augment a tad Jake Poznanski 2025-09-18 20:38:02 +00:00
  • a42d8199cd Adding 1025 mix dataset, should be ready for final run Jake Poznanski 2025-09-18 19:50:40 +00:00
  • 1ac72ad169 Adding some scripts to clean data Jake Poznanski 2025-09-18 19:44:30 +00:00
  • eb8dce01f4 Hmm, things work better but still have issues importing images jakep/new_data_image_boxes Jake Poznanski 2025-09-17 20:12:05 +00:00
  • 0f04cc5c67 Merge branch 'jakep/new_data' into jakep/new_data_image_boxes Jake Poznanski 2025-09-17 19:51:56 +00:00
  • 30750f77c1 Ok, rendering smaller version of the page, since this is the max suppored by claude and it would get rescaled anyways Jake Poznanski 2025-09-17 19:51:03 +00:00
  • 54cda1662b Brings in images from original documents, but it seems worse quality Jake Poznanski 2025-09-17 18:58:04 +00:00
  • a60c84ed14 Maybe better scaling with no losing of text Jake Poznanski 2025-09-16 22:01:49 +00:00
  • 52df81873a Glob path fixes Jake Poznanski 2025-09-16 21:42:46 +00:00
  • 3b729d6770 Oops fixing random gen things Jake Poznanski 2025-09-16 21:17:12 +00:00
  • 2fa67a980e More reliable test gen Jake Poznanski 2025-09-16 19:57:00 +00:00
  • e3e09c04db Synth data fixups Jake Poznanski 2025-09-16 18:47:54 +00:00
  • 85a6032f34 Trying 3 epoch config jakep/new_data_finetune_on_bench_data Jake Poznanski 2025-09-11 18:35:18 +00:00
  • 3a908f0e84 Trying a fine tune on the grpo data mix directly Jake Poznanski 2025-09-11 17:49:54 +00:00