10 commits

Author	SHA1	Message	Date
Rowe Wilson Frederisk Holme	ed0695ce3a	Fix #119	2022-11-25 00:19:22 +08:00
Fred de Gier	2a91018792	Resolve merge conflicts and fix issues with setup.sh	2022-10-20 16:09:12 +02:00
Parth Thakkar	01f1cbb629	Add python backend support ... - Modify dockerfile to include bitsandbytes, transformers and latest version of pytorch - Minor modifications in utils/codegen.py so that same client works with FT and Py-backend - Minor modifications in launch.sh (no need to name models by GPU) - Add installation script for adding a new python model (with super simple config_template) - Modify setup.sh so that it aworks with both FT and Python backend models Signed-off-by: Parth Thakkar <thakkarparth007@gmail.com>	2022-10-16 22:05:09 -05:00
Fred de Gier	de71bb6ff5	Resole merge conflicts	2022-10-03 14:27:32 +02:00
Fred de Gier	87f4f53e27	Simplify config and port handling	2022-10-03 14:13:10 +02:00
Rowe Wilson Frederisk Holme	5ffbaa8178	Maintain compatibility with previous versions	2022-09-22 04:12:44 +08:00
Rowe Wilson Frederisk Holme	9cf72d3666	Fix uvicorn undefined error ... This also alleviates #61.	2022-09-22 02:28:47 +08:00
Fred de Gier	8895b74238	Rewrite API to FastAPI, separate API from CodeGen, remove dev settings	2022-09-12 12:59:37 +02:00
Brendan Dolan-Gavitt	0fa041a9ea	Bump Triton docker to moyix/triton_with_ft:22.09 ... This version of the Docker image supports a wider range of cards (down to SM 6.0), so it should be compatible with older cards. The corresponding commit to build the image is a34df0f053 . Fixes #30 .	2022-09-02 18:39:37 -04:00
Brendan Dolan-Gavitt	a23a1478f7	Initial commit	2022-08-02 21:47:27 -04:00