Commit Graph

214 Commits

Author SHA1 Message Date
nai-degen 068e7a834f fixes AWS legacy models for non-streaming requests 2024-03-04 21:22:43 -06:00
nai-degen 736803ad92 enables opus by default 2024-03-04 21:11:32 -06:00
nai-degen 51ffca480a adds AWS Claude Chat Completions and Claude 3 Sonnet support 2024-03-04 16:25:06 -06:00
nai-degen 90ddcac55b makes claude3 compat model customizable via environment variable 2024-03-04 14:21:55 -06:00
nai-degen 1edc93dc72 adds claude-opus model family 2024-03-04 14:08:59 -06:00
nai-degen f6c124c1d3 fixes issue with preamble-required claude keys and anthropic chat 2024-03-04 14:00:25 -06:00
nai-degen 90a053d0e0 detects and removes over-quota claude keys from keypool 2024-03-04 13:42:29 -06:00
khanon db318ec237 Implement Anthropic Chat Completions endpoint and Claude 3 (khanon/oai-reverse-proxy!64) 2024-03-04 19:06:46 +00:00
nai-degen b90abbda88 spoofs response for SillyTavern test messages 2024-02-28 15:57:18 -06:00
nai-degen 93cee1db9b removes claude v1 from AWS keychecker as it has been retired 2024-02-27 15:52:09 -06:00
nai-degen bd15728743 uses explicitly set keyprovider rather than inferring via requested model 2024-02-27 10:56:50 -06:00
nai-degen 627559b729 updates mistral modelids 2024-02-26 23:55:03 -06:00
nai-degen 5e19e2756a adds mistral-large model family, untested 2024-02-26 18:12:08 -06:00
nai-degen 59bda40bbc handles google streaming json response format variation 2024-02-19 00:12:09 -06:00
nai-degen 68d829bceb adds Claude over-quota detection 2024-02-17 15:56:22 -06:00
nai-degen 3498584a1f removes forceModel on Google AI endpoint 2024-02-15 11:41:34 -06:00
nai-degen 21d61da62b increases max image payload size for gpt4v 2024-02-12 21:59:48 -06:00
nai-degen 35dc0f4826 fixes 'Premature close' caused by fucked up AWS unmarshaller errors 2024-02-10 14:47:14 -06:00
devvnull 0ce4582f3b Improve "\n\nHuman" prefix requirement detection for Anthropic (khanon/oai-reverse-proxy!63) 2024-02-08 16:28:11 +00:00
nai-degen bbee056114 fixes Force Key Recheck admin function for azure/aws 2024-02-07 19:54:40 -06:00
nai-degen ecc804887b uses EventStreamMarshaller from AWS SDK to hopefully handle split messages 2024-02-05 19:56:41 -06:00
nai-degen a8fd3c7240 fixes AWS Claude throttlingException handling 2024-02-04 20:48:20 -06:00
nai-degen 40240601f5 refactors SSEStreamAdapter to fix leaking decoder streams 2024-02-04 18:38:06 -06:00
nai-degen 98cea2da02 replaces eventstream lib to (hopefully) fix interrupted AWS streams 2024-02-04 17:18:28 -06:00
nai-degen c88f47d0ed fixes middleware order breaking /proxy endpoint 2024-02-04 16:21:44 -06:00
nai-degen 43106d9c7f tracks Risu userid rather than IP address on usertokens 2024-02-04 14:14:36 -06:00
nai-degen fe429a7610 adds SERVICE_INFO_PASSWORD to gate infopage behind a password 2024-02-04 14:04:46 -06:00
nai-degen 235510e588 fixes incorrect AWS Claude 2.1 max context limit 2024-02-01 20:40:15 -06:00
nai-degen 7eb6eb90ad moves api schema validators from transform-outbound-payload into shared 2024-01-29 19:38:22 -06:00
nai-degen 924db33f7e attempts to auto-convert Mistral prompts for its more strict rules 2024-01-28 17:42:23 -06:00
nai-degen c9791acd85 makes gpt4-v input validation less strict to accomodate Risu 2024-01-27 13:24:11 -06:00
nai-degen e871b8ecf1 removes logprobs default value since it breaks gpt-4-vision 2024-01-27 12:19:24 -06:00
nai-degen e6dc4475e6 fixes max context size for nu-gpt4-turbo 2024-01-25 14:07:42 -06:00
nai-degen 5e646b1c86 adds gpt-4-0125-preview and gpt-4-turbo-preview alias 2024-01-25 13:27:03 -06:00
nai-degen 02a54bf4e3 fixes azure openai logprobs (actually tested this time) 2024-01-25 01:17:18 -06:00
nai-degen 79b2e5b6fd adds very basic support for OpenAI function calling 2024-01-24 16:42:26 -06:00
nai-degen 935a633325 fixes typo in Azure logprob adjustment 2024-01-24 16:03:47 -06:00
nai-degen 4a4b60ebcd handles Azure deviation from OpenAI spec on logprobs param 2024-01-24 16:01:19 -06:00
nai-degen ad465be363 fixes logprobs schema validation for turbo instruct endpoint 2024-01-24 14:31:10 -06:00
nai-degen c7a351baa8 adds support for requesting logprobs from OpenAI Chat Completions API 2024-01-24 11:46:09 -06:00
nai-degen e813cd9d22 default claude 2.1 instead of 1.3 in openai compat endpoint since 1.3 is not accessible on all keys 2024-01-18 04:14:15 -06:00
nai-degen 4c2a2c1e6c improves handle-streamed-response comments/docs [skip-ci] 2024-01-18 04:14:15 -06:00
nai-degen 81f1e2bc37 fixes broken GET models endpoint for openai/mistral 2024-01-14 05:33:24 -06:00
nai-degen c2a686f229 Revert "reduces max request body size for now"
This reverts commit 4ffa7fb12b.
2024-01-13 18:12:16 -06:00
twinkletoes 96a0f94041 Fix Mistral safe_prompt schema property (khanon/oai-reverse-proxy!60) 2024-01-14 00:11:39 +00:00
nai-degen 7b0892ddae fixes unawaited call to async enqueue 2024-01-07 16:23:53 -06:00
nai-degen 7f92565739 SSE queueing adjustments, untested 2024-01-07 16:19:22 -06:00
nai-degen 4ffa7fb12b reduces max request body size for now 2024-01-07 13:03:24 -06:00
nai-degen 8dc7464381 strips extraneous properties on zod schemas 2024-01-07 13:00:48 -06:00
twinkletoes 4a823b216f Mistral AI support (khanon/oai-reverse-proxy!58) 2023-12-25 18:33:16 +00:00