Commit Graph

35 Commits

Author SHA1 Message Date
Yuval
b775d15f2e
fix: Updated AI21 models from J1 to J2 (#104)
also, added ai21 example to README

Co-authored-by: Yuval Belfer <yuvalb@ai21.com>
2023-07-01 21:10:56 -07:00
Laurel Orr
6324e0fe43
feat: streaming support completions (#99) 2023-05-21 15:50:03 -07:00
Laurel Orr
147436c9b2
feat: unify run_chat and run (#92) 2023-05-16 22:52:41 -07:00
Laurel Orr
8548329be9
feat: added run_chat for chat models (#88) 2023-05-03 10:09:51 -07:00
Laurel Orr
afe0fc5a1d
feat: added run_chat for chat models (#87) 2023-05-02 10:08:10 -07:00
Laurel Orr
d7401c6ec5
fix: added pydantic types to response (#84) 2023-04-24 10:10:47 -07:00
Laurel Orr
4602fb919b
chore: update readme (#82) 2023-04-18 11:59:21 -06:00
Laurel Orr
db963cf4a7 fix: added client pool support (#81)
* fix: added client pool support

* Added async across client pool
2023-04-16 23:07:35 -06:00
Laurel Orr
40de0e7f59 feat: openai embedding support (#75) 2023-04-08 22:55:04 -07:00
Laurel Orr
395ac06a95 feat: async support, openai chatgpt, batch cache fix (#68) 2023-03-11 20:05:41 -08:00
Laurel Orr
ace3ad4324 chore: fix manifest imports (#56) 2023-01-30 15:48:13 -08:00
Laurel Orr
e351bd5315 Update README.md citation 2023-01-27 15:36:29 -08:00
Laurel Orr
56eae406ce feat: chatgpt client added (#47) 2023-01-08 14:58:12 -08:00
Laurel Orr
6f5b64f0df Laurel/diffusion (#40)
* Sketch of diffusers added

* [WIP] Array caching implemented with end2end diffusion working

* [WIP] Make initial pass on CLIP model

* [WIP] Get endpoint running for CLIP

* Add support for clip images

* [chore] merge main

* chore: fix xxhash install

Co-authored-by: Sabri Eyuboglu <eyuboglu@stanford.edu>
2022-12-24 15:17:02 -08:00
Laurel Orr
e86b8e81bf [feat] batch run; remove prompt batch_run zoo and opt; refactor to us… (#39)
* [feat] batch run; remove prompt batch_run zoo and opt; refactor to use standard request inputs

* [chore] test rounding fix
2022-11-11 00:19:27 -08:00
Laurel Orr
a4cd201b8e Deepspeed (#38)
* [feat] deepspeed and batch support huggingface api

* [chore] add deepspeed to readme

* [chore] fix hf api test
2022-11-07 21:01:14 -08:00
Laurel
8b423d6962 v0.0.1 release 2022-11-08 04:39:49 +00:00
Laurel
37a232ef7d [chore] fix readme install 2022-11-04 23:43:26 +00:00
Laurel Orr
c5f00b9f73 Request cache refactor (#37)
* [wip] cache import clean up and removed api imports

* [chore] add timeout to requests

* [chore] toma heartbeats to logger
2022-11-04 16:02:54 -07:00
yichern
e42f335b14 Fix typo in README (run_batch returns list) (#27)
In `result_object = manifest.run_batch(prompt, ["Laurel", "Avanika"], return_response=True)`, the return type of `run_batch` is a list of response objects, not a response object. This PR fixes a bug in the demo code to print only after looping through the response objects.

Co-authored-by: Laurel Orr <57237365+lorr1@users.noreply.github.com>
2022-10-14 16:57:03 -07:00
yichern
c351d7fa5e Fix typo in README (batch_run -> run_batch) (#26)
As in title!
2022-10-14 16:55:39 -07:00
Ben Newman
9f313c6f12 Fix typo in README ("Boom" -> "Bloom") (#24) 2022-10-14 16:37:31 -07:00
Laurel Orr
a68ee63c18 [docs] update readme.md 2022-10-06 01:05:32 +00:00
Laurel Orr
a846b72be0 Session id None 2022-10-05 23:33:21 +00:00
Laurel Orr
d610e3c800 Laurel/dev (#20)
* fix: added perc gpu memory reduction flag app

* fix: removed poetry and standardized api inputs

* Removed poetry and use setup.py
* Addressing https://github.com/HazyResearch/manifest/issues/16
2022-06-17 23:25:02 -07:00
Laurel Orr
99bcc76fde Laurel/multigpuhf (#15)
* feat: added user session logging

* fix: removed raw queries in tests

* build: move torch to dev dependency

* fix: adding accelerate to api

Addresses https://github.com/HazyResearch/manifest/issues/14

* build: fix torch github
2022-06-16 15:07:55 -07:00
Laurel Orr
d29d10de28 feat: added user session logging (#13)
* feat: added user session logging

* fix: removed raw queries in tests

* build: move torch to dev dependency
2022-06-14 18:11:02 -07:00
Laurel Orr
c04aa1e62c fix: ai21 prompt added with logprobs 2022-06-11 19:31:40 +00:00
Laurel Orr
b34e2a714b Fix readme and separate dev 2022-06-11 18:22:16 +00:00
Laurel Orr
a82c8c89df [feature] better naming models in cache 2022-05-26 23:27:42 -07:00
Laurel Orr
894042c370 [feature] added opt; better kwargs for errors 2022-05-26 00:05:49 -07:00
Laurel Orr
bf16963f3d [feature] redis DB, flask API, tests 2022-05-25 20:53:54 -07:00
ANarayan
15e712b060 remove circular imports in client, manifest; add prompt serialization and dill dep 2022-05-24 14:36:10 -07:00
Laurel Orr
ed301be2a7 First main commit 2022-05-24 00:29:17 -07:00
Laurel Orr
6ada1f2d2c Initial commit 2022-05-21 15:15:42 -07:00