Commit Graph

312 Commits

Author SHA1 Message Date
Zach Nussbaum
d1b64d7eed refactor: imports 2023-05-04 03:16:26 +00:00
Zach Nussbaum
0402d3e28a feat: barebones pythiaseek 2023-05-03 22:05:55 +00:00
Zach Nussbaum
2771f96cb6 refactor: move to folder 2023-05-03 21:52:40 +00:00
Zach Nussbaum
bd6e471555 feat: cosine alpha schedule 2023-05-03 21:39:14 +00:00
Zach Nussbaum
27a9b2b10c fix: option for no schedule 2023-05-03 21:39:05 +00:00
Zach Nussbaum
aa6763daa8 fix: update config 2023-05-03 21:33:52 +00:00
Zach Nussbaum
3fa80f8c09 fix: remove schedule 2023-05-03 21:33:29 +00:00
Zach Nussbaum
06228d9b67 Merge branch 'junior' of https://github.com/nomic-ai/gpt4all into junior 2023-05-03 01:52:09 +00:00
Zach Nussbaum
d61cd55772 fix: alpha, projection 2023-05-02 19:38:09 +00:00
Zach Nussbaum
2c8e1096c5
Merge pull request #472 from berkantay/main
Update README.md
2023-05-02 10:15:40 -04:00
Zach Nussbaum
00f04360d2 fix: config for index building 2023-05-01 21:39:55 +00:00
Zach Nussbaum
0f61cd8b42 fix: retrieval dataset only has train split 2023-05-01 21:39:40 +00:00
Zach Nussbaum
3736eda56a feat: eval for retrieval 2023-05-01 21:39:21 +00:00
Zach Nussbaum
1b3f18bef2 fix: import path 2023-05-01 21:39:09 +00:00
Zach Nussbaum
0c0a56acab feat: data preprocessing 2023-05-01 21:38:46 +00:00
Zach Nussbaum
c9dd9152c3 feat: model def + metrics 2023-05-01 21:38:36 +00:00
Zach Nussbaum
48e07be9e9 feat: training script 2023-05-01 21:38:23 +00:00
Zach Nussbaum
80d810322a fix: lr schedule 2023-05-01 21:38:01 +00:00
Berkant
aefea2e713
Update README.md
README.md typo fix.
2023-04-30 01:07:14 +03:00
Zach Nussbaum
8a917ad4e1 chore: create data folder 2023-04-25 21:28:56 +00:00
Zach Nussbaum
a58d1eb3bd refactor: move file around 2023-04-25 21:28:42 +00:00
Zach Nussbaum
da5ce0a181 chore: ignore large arrow files 2023-04-25 20:36:04 +00:00
Zach Nussbaum
b0f92b610e refactor: clean up embed texts 2023-04-25 20:34:49 +00:00
Zach Nussbaum
c20379f7e9 refactor: clean up prep index 2023-04-25 20:34:38 +00:00
Zach Nussbaum
f2161f7e59 docs: prep index barebones 2023-04-25 20:34:26 +00:00
Zach Nussbaum
586a8abc06 fix: allow for print for fns that are used in both dist and single 2023-04-25 20:34:12 +00:00
Zach Nussbaum
7832707c37 fix: index -> id 2023-04-25 20:33:49 +00:00
AT
b00d338c1e
Update README.md 2023-04-25 09:02:34 -04:00
AT
a7ada4e4b0
Update README.md 2023-04-25 09:02:11 -04:00
AT
8b10e533bb
Update README.md 2023-04-25 09:01:31 -04:00
Zach Nussbaum
869829a065 nearly have the neighbor caching working, but combinging info into final dataset is challenging 2023-04-23 22:28:57 +00:00
Zach Nussbaum
cc3cc3f7e9 we now get the neighbors embeddings from the disk index 2023-04-23 21:04:43 +00:00
Zach Nussbaum
240feae277 added initial files for dataset prep and ingestion for gpt4all jr 2023-04-23 20:02:22 +00:00
Zach Nussbaum
84acbc8225 chore: ignore swp files 2023-04-23 17:34:49 +00:00
Zach Nussbaum
9bc88fb33d feat: add options to specify different datasets 2023-04-22 19:53:47 +00:00
Zach Nussbaum
8cb3c8de35 chore: reqs, ignore, readme 2023-04-22 19:40:06 +00:00
Zach Nussbaum
4eeab60306 feat: build knn index 2023-04-22 19:39:46 +00:00
Zach Nussbaum
a2b1f99838 feat: distributed eval of embedder 2023-04-22 19:39:30 +00:00
Zach Nussbaum
4fb19d67b5 feat: tokenize texts into chunks 2023-04-22 19:38:51 +00:00
Zach Nussbaum
2dae153c68 feat: sbert abstractor 2023-04-22 19:38:34 +00:00
Zach Nussbaum
4671f4e82f chore: pull out common dist print fn 2023-04-22 19:37:30 +00:00
Zach Nussbaum
e255e0a805 fix: batched xattn 2023-04-21 21:54:47 +00:00
Zach Nussbaum
ca66d12d89 fix: remove causal cross attn mask 2023-04-21 14:23:33 +00:00
Zach Nussbaum
e62baf87f8 fix: seed 2023-04-21 04:19:37 +00:00
Zach Nussbaum
aa814757fc fix: testing works 2023-04-21 04:18:16 +00:00
Zach Nussbaum
df79fd64b0 fix: forward works! 2023-04-21 02:50:27 +00:00
zanussbaum
09cddbedc0 feat: models wip 2023-04-20 16:04:41 -04:00
zanussbaum
97a1cd0539 chore: delete unused 2023-04-19 18:13:24 -04:00
zanussbaum
29c7ac7f73 refactor: clean up directory structure 2023-04-19 18:12:03 -04:00
zanussbaum
a514492d52 chore: ignore ds store 2023-04-19 18:11:02 -04:00