Commit graph

480 commits

Author SHA1 Message Date
Krrish Dholakia
35cf6ef0a1 batch completions for vllm now works too 2023-09-06 19:26:19 -07:00
ishaan-jaff
2880a7b6b4 allow users to pass custom timing for replicate 2023-09-06 18:32:40 -07:00
ishaan-jaff
fc7ad0c245 bump v 2023-09-06 18:14:58 -07:00
ishaan-jaff
8b3b682000 add replicate pricing 2023-09-06 18:14:34 -07:00
Krrish Dholakia
4cfcabd919 adding support for vllm 2023-09-06 18:07:44 -07:00
ishaan-jaff
bd77d5ac21 docs update 2023-09-06 17:16:24 -07:00
ishaan-jaff
1ba6b6761b show pricing for tg ai completion 2023-09-06 17:10:49 -07:00
ishaan-jaff
04f8b20651 fix linting errors @krrishdholakia 2023-09-06 16:38:42 -07:00
ishaan-jaff
c5151aa573 move model_prices to root github 2023-09-06 16:31:44 -07:00
ishaan-jaff
bab27634a8 rename max_tokens.json 2023-09-06 16:28:17 -07:00
ishaan-jaff
fbd67bc24c add experimental together_computer cost calc 2023-09-06 16:08:44 -07:00
Krrish Dholakia
0ace48d719 update custom prompt template function 2023-09-06 13:14:36 -07:00
ishaan-jaff
dc0c084813 improve replicate usage 2023-09-06 12:29:34 -07:00
Krrish Dholakia
4bcfa71e11 commenting out flaky circle ci test 2023-09-06 12:07:47 -07:00
ishaan-jaff
601a6a4a92 better documentation on __init__ 2023-09-06 11:31:52 -07:00
Krrish Dholakia
91d2a57970 update linting 2023-09-06 11:30:40 -07:00
ishaan-jaff
cfe35580e2 add support for flant5 2023-09-06 11:28:43 -07:00
Krrish Dholakia
44f71aa321 logging replicate response logs 2023-09-06 11:28:40 -07:00
ishaan-jaff
1da6026622 add flan + vicuna + fix replicate errors 2023-09-06 11:23:58 -07:00
Krrish Dholakia
48ee4a08ac updates 2023-09-06 11:21:48 -07:00
ishaan-jaff
af60b2ba77 add vicuna translation 2023-09-06 11:14:24 -07:00
Krrish Dholakia
afcd6b28cc bump version 2023-09-06 11:05:11 -07:00
ishaan-jaff
0ddda7c035 send optional_params for llama2-70b chat replicate 2023-09-06 11:01:39 -07:00
Krrish Dholakia
d236d68fa4 fix linting issues 2023-09-06 10:41:52 -07:00
ishaan-jaff
d4c4a138ca add replicate support for max_tokens 2023-09-06 10:38:21 -07:00
Krrish Dholakia
ef43141554 updates to exception mapping 2023-09-06 10:36:22 -07:00
ishaan-jaff
d0b16892e0 fix replicate supported llms 2023-09-06 10:28:03 -07:00
ishaan-jaff
bc9b629726 add Replicate Error class 2023-09-06 10:25:40 -07:00
ishaan-jaff
1c61b7b229 add replicate streaming 2023-09-06 10:23:13 -07:00
ishaan-jaff
c45b132675 use replicate http requests instead 2023-09-06 09:43:05 -07:00
Krrish Dholakia
3d6836417e adding prompt template for falcon 180b 2023-09-06 08:44:13 -07:00
Krrish Dholakia
b4a9699138 update docs on together ai 2023-09-06 08:26:05 -07:00
ishaan-jaff
fb3983d058 fix timeouts + tests + bump v 2023-09-05 17:17:58 -07:00
ishaan-jaff
0bbcba269d fix linting 2023-09-05 17:11:31 -07:00
ishaan-jaff
8ecef03f63 custom api base 2023-09-05 16:50:32 -07:00
ishaan-jaff
262bb07ade clean install and import - vertexai 2023-09-05 16:10:07 -07:00
ishaan-jaff
2250d1375e clean up azure implementation 2023-09-05 15:31:59 -07:00
ishaan-jaff
2a36f06763 remove install_and_import remove petals 2023-09-05 15:06:24 -07:00
ishaan-jaff
079ec7064b add completion types 2023-09-05 14:57:03 -07:00
Krrish Dholakia
074f6dbfaf fixing linting errors 2023-09-05 14:52:57 -07:00
ishaan-jaff
846aeea4f9 types completion() 2023-09-05 14:49:53 -07:00
Krrish Dholakia
9a82bfa4d2 test parallelism on circle ci 2023-09-05 14:47:50 -07:00
ishaan-jaff
794964993d add types to completion() 2023-09-05 14:42:10 -07:00
Krrish Dholakia
faa78ad543 bump version 2023-09-05 14:30:03 -07:00
Krrish Dholakia
af33a85043 only use tgai's prompt template for llama2 instruct models 2023-09-05 12:25:52 -07:00
Krrish Dholakia
8845938b31 adding support for custom prompt templates to together ai 2023-09-05 12:20:09 -07:00
Krrish Dholakia
64f3d3c56e prompt formatting for together ai llama2 models 2023-09-05 11:57:13 -07:00
Krrish Dholakia
dc130efac6 update streaming docs to show it working for async completion calls 2023-09-05 09:18:37 -07:00
Krrish Dholakia
5870dabae1 update init with comments 2023-09-05 09:14:57 -07:00
Krrish Dholakia
e2c143dfbc test async streaming 2023-09-04 15:42:24 -07:00