waifu-collective
/
toolbox
2
1
Fork 0
Code Issues 2 Pull Requests Releases Wiki Activity

Commit Graph

  • 50ae8816a1 refactor: archive the old repo master 11b 2023-01-08 17:31:20 -0300
  • beec9ba31f chore: update gitignore since training code is no longer here 11b 2023-01-08 17:09:48 -0300
  • 23eb4a6ab2 refactor: move non-data related stuff to other repositories in the org 11b 2023-01-08 16:31:37 -0300
  • 7d385ec13c chore: add packages required by the SODA dataset 11b 2023-01-08 15:49:58 -0300
  • ea162de2e0 feat: add SODA dataset TearGosling 2023-01-08 12:15:15 -0600
  • eb997a3d3f chore: point CAI dumper userscript to the GitHub repo instead 11b 2023-01-08 12:16:38 -0300
  • 9a3719127c refactor: delete old training code 11b 2023-01-08 11:46:07 -0300
  • 4a1784f8a1 feat: add options to control text generation on the UI feat/generation-options-on-gradio-ui 11b 2022-12-25 15:41:10 -0300
  • 5e34b105dc feat: alternative way of handling and augmenting episode data (wip) 11b 2023-01-04 09:05:51 -0300
  • 46a552ad28 chore: add link to roadmap on the README 11b 2023-01-01 11:51:46 -0300
  • 1409bafd2b chore: update ROADMAP 11b 2023-01-01 11:50:30 -0300
  • 53494a6567 chore: fix linter/style problems 11b 2023-01-01 11:50:23 -0300
  • e4594338d2 feat: changes to log and discard some not-so-great data 11b 2023-01-01 11:34:31 -0300
  • 9f55ecfca7 feat: attempt to detect looping in CAI logs and discard from final dataset 11b 2023-01-01 11:32:57 -0300
  • aebd405bbd feat: proper checkpoint resume in CLM fine-tune script 11b 2022-12-27 13:21:20 -0300
  • e99277ec52 feat: log LR in CLM fine-tune script 11b 2022-12-27 13:21:00 -0300
  • 96b41dee60 feat: improve handling of special tokens in the Kajiwoto dataset 11b 2022-12-27 12:46:57 -0300
  • b95b30cf88 feat: implement arg to skip over episodes when debugging data build 11b 2022-12-27 12:46:36 -0300
  • 3e798f6767 fix: rename folder so import actually works 11b 2022-12-26 20:44:35 -0300
  • 93e283daee feat: implement utility to convert ColossalAI checkpoints to HF pre-trained model 11b 2022-12-26 20:43:01 -0300
  • b79ac657a4 fix: haru's sft being incompatible with the ColossalAI fine-tune script 11b 2022-12-26 20:42:48 -0300
  • 5dbde00d27 feat: bring down target word count per episode 11b 2022-12-26 17:31:28 -0300
  • bcbf0910b4 feat: add supervised fine-tuning code based on haru's work 11b 2022-12-26 17:31:00 -0300
  • 60e649f57a feat: some minor filtering to hopefully improve CAI data 11b 2022-12-26 12:04:04 -0300
  • 4f794489ac feat: add support for fine-tuning GPT-NeoX-based models, save optimizer and LR scheduler to checkpoint 11b 2022-12-25 15:42:59 -0300
  • 186df60691 feat: update inference code for pythia/cai data-based models 11b 2022-12-25 15:37:34 -0300
  • 3bfb623f26 fix: human/bot messages being incorrectly labeled as eachother 11b 2022-12-24 17:58:33 -0300
  • 5b26097905 feat: implement Gradio UI for proper model inference (WIP) 11b 2022-12-24 12:12:55 -0300
  • cef8f54fc4 fix: ignore invalid CAI JSON dumps 11b 2022-12-23 16:45:09 -0300
  • d91367e902 chore: update module list in build_dataset.py 11b 2022-12-23 16:37:58 -0300
  • a16673ebe0 refactor: adjust Kajiwoto modules to use the proper prompt constants 11b 2022-12-23 16:37:47 -0300
  • 60e0a21a3c chore: add pdbpp for better debugging experience 11b 2022-12-23 16:28:37 -0300
  • 3d6def871d refactor: use LIGHT as PDM instead of VDM, ignore actions 11b 2022-12-23 16:28:20 -0300
  • 1f273f13f3 chore: bump pdm version 11b 2022-12-23 16:21:02 -0300
  • e0552639fa feat: update CAI dataset/module to handle userscript dumps and use definitions 11b 2022-12-23 16:20:53 -0300
  • aef9289678 chore: update ROADMAP to add links about contributing with CAI dumps 11b 2022-12-23 10:59:58 -0300
  • 69aeea85b9 chore: reorganize CAI dumper README 11b 2022-12-21 20:05:15 -0300
  • 7087f39d5a fix: cai dumper crashing if chat had no messages 11b 2022-12-21 20:05:01 -0300
  • d6e05e6e5b chore: add changelog to the CAI dumper 11b 2022-12-21 16:24:35 -0300
  • e612386424 fix: handle edge-case regarding extra whitespace on char name 11b 2022-12-21 16:15:49 -0300
  • d638bb5625 feat: update userscript to allow dumping of definitions as well 11b 2022-12-21 15:39:42 -0300
  • ecf2e65e76 fix: anonymization within message text 11b 2022-12-21 14:19:27 -0300
  • 21ebd5834e chore: fix path links in the README 11b 2022-12-21 13:51:20 -0300
  • a8dfd396cc fix: more aggressive anonymization within message text 11b 2022-12-21 13:41:19 -0300
  • 6bc2a03ff9 chore: update ROADMAP 11b 2022-12-20 21:52:28 -0300
  • 1ddd991471 chore: clarify what DHT means in the Discord module 11b 2022-12-20 21:44:16 -0300
  • cec59a5511 docs: update CAI dumper README 11b 2022-12-20 21:43:57 -0300
  • 4f78bb73cb feat: implement userscript to dump CAI chats/basic bot info 11b 2022-12-20 21:34:39 -0300
  • ecd4efe3ce chore: run isort 11b 2022-12-20 17:55:17 -0300
  • 009c837439 feat: implement Discord dialogue module 11b 2022-12-20 17:54:20 -0300
  • b42131191a fix: don't write file when printing the dataset for debugging 11b 2022-12-20 17:41:08 -0300
  • 45f495a4dd docs: add basic roadmap 11b 2022-12-20 17:40:48 -0300
  • 30ff3751de feat: allow fine-tuning of non-OPT LMs 11b 2022-12-18 22:24:43 -0300
  • 25ae9da046 fix: tensorboard getting messed up after epoch end 11b 2022-12-18 22:24:26 -0300
  • fb3dc4ba22 chore: fix typo in README 11b 2022-12-18 17:36:01 -0300
  • 877b1bc56b feat: barebones inference script for testing checkpoints 11b 2022-12-18 17:26:32 -0300
  • f82b4ea913 feat: super early CAI dataset and module for testing 11b 2022-12-18 17:26:16 -0300
  • 657cbe1d61 chore: update gitignore to handle checkpoints 11b 2022-12-18 17:25:50 -0300
  • b0d2d80ac3 feat: minor improvements to the fine-tune script 11b 2022-12-18 17:25:36 -0300
  • 925f5767ec feat: fine-tuning scripts and instructions 11b 2022-12-17 21:46:32 -0300
  • 6fbd660a67 feat: implement script to build final data file 11b 2022-12-17 21:37:27 -0300
  • 8df2d87355 feat: Kajiwoto dataset and modules 11b 2022-12-17 21:36:33 -0300
  • a076746f9d fixup! feat: initial commit 11b 2022-12-17 21:32:46 -0300
  • f5552cde74 feat: add the LIGHT dataset and VDM 11b 2022-12-17 21:32:34 -0300
  • cb1d3dd68e chore: enable strict mode on mypy, add isort check 11b 2022-12-17 21:31:41 -0300
  • ad3c378d58 feat: initial commit 11b 2022-12-17 17:52:34 -0300