rixty/SillyTavern
1
0
Fork 0
mirror of https://github.com/SillyTavern/SillyTavern.git synced 2025-04-17 04:07:21 +02:00
Code Issues Packages Projects Releases Wiki Activity
8,677 Commits 12 Branches 96 Tags
Commit Graph

158 Commits

Author SHA1 Message Date
subzero5544
a503f58d0c
Adding reverse proxy support to DeepSeek chat completion (#3328) 
* added reverse proxy settings to deepseek chat completion

* Update chat-completions.js

* Update chat-completions.js

* Update chat-completions.js

* Update chat-completions.js

* Update chat-completions.js

* Update chat-completions.js

* Unify API key requirement

---------

Co-authored-by: Cohee <18619528+Cohee1207@users.noreply.github.com>
 2025-01-23 09:02:44 +02:00
Cohee
d7bb92be54 deepseek reasoner 
Closes #3322
 2025-01-20 23:31:40 +02:00
Cohee
b66c4d7206 Merge branch 'staging' into geminifeat2 2025-01-06 19:20:58 +02:00
Cohee
9062b6bc4d Whitespace clean-up 2025-01-06 19:20:04 +02:00
tincansimagine
2103e6238c Fix: Apply OFF safety settings for gemini-2.0-flash-exp model 2025-01-07 02:17:09 +09:00
Fizz~
b6e7f4f493
Change Gemini system instruction parameter key to systemInstruction 
Refer to https://github.com/googleapis/python-genai/blob/main/google/genai/models.py#L772, it seems like both work(?) but this is probably the more correct option if the newer library is using it
 2025-01-06 11:44:11 -05:00
Cohee
72e0f8ccc5 Fix available CC models log header 2025-01-01 16:28:47 +02:00
Cohee
cdb31699d4 Expose new post-processing as "Semi-strict" 2024-12-29 21:20:15 +02:00
Cohee
c9db7a1574 Don't add user placeholders 2024-12-29 21:14:27 +02:00
Cohee
0ded442bd3 Fix status check 2024-12-29 20:39:47 +02:00
Cohee
4c7d160d41 DeepSeek 
Closes #3233
 2024-12-29 20:38:13 +02:00
Cohee
662f0e9c73 Gemini: Thought toggle 
Closes #3220
 2024-12-29 18:23:07 +02:00
Cohee
39cfb35c1a Gemini: Fix cross-chunk parsing of multipart replies 2024-12-27 23:15:09 +02:00
Cohee
a82c05a8ac Gemini thinking: Specify context size, system prompt and vision support 2024-12-27 22:39:26 +02:00
Cohee
7adc6d38e2 OpenRouter: Add control for middle-out transform 
Closes #3033
 2024-12-24 21:51:47 +02:00
Cohee
73614f2f8d Refactor prompt converters with group names awareness 2024-12-20 23:30:57 +02:00
Cohee
00363cc206 Merge branch 'release' into staging 2024-12-17 21:44:28 +02:00
Cohee
dbc78c1c42 Fix OpenRouter caching at depth with null model selection 2024-12-17 21:40:33 +02:00
M0cho
3f253f42f2 Update: [constants.js] Remove BISON_SAFETY 2024-12-16 12:45:30 +09:00
M0cho
43feffdfae Update: [chat-completions.js] Update sendMakerSuiteRequest function 
Removed branching logic for differences in JSON request body between PaLM and Gemini, following the removal of PaLM from Google AI Studio.
 2024-12-15 13:06:48 +09:00
M0cho
a64c8ade9d Support Gemini 2.0 Flash-exp 2024-12-12 06:31:27 +09:00
Cohee
22ea5f3c3c Lint fix 2024-12-06 21:30:05 +02:00
M0cho
073b76a693 Support Gemini-exp-1206 2024-12-07 02:19:15 +09:00
Cohee
9382845dee Claude: remove user filler from prompt converter 2024-11-24 19:05:41 +02:00
Cohee
70c45fb001 Merge branch 'staging' into fix/connRefusedErrMsg 2024-11-24 01:41:15 +02:00
Cohee
85ca08a2ea Settings for new gemini 2024-11-22 17:50:33 +00:00
ceruleandeep
8de551fc94 Return 502 with error description when connection to remote CC API fails 
If chat-completions/generate returns an error, throw the error message

Reformat display of exceptions during SD prompt text generation
 2024-11-22 11:55:27 +11:00
Cohee
c3b416b4cf Claude: fix prefill tool calling hack with at-depth caching 2024-11-19 20:48:02 +02:00
Honey Tree
c3caa1699e Moved most of code to prompt converters 2024-11-18 08:06:27 -03:00
Honey Tree
befe5a7171 Adding Claude caching support to OpenRouter as well 2024-11-17 15:53:02 -03:00
Honey Tree
5397614347 Defaulting to -1 rather than boolean false 2024-11-17 14:01:22 -03:00
Honey Tree
ac33e4d668 cache_control positioning fix 2024-11-17 09:21:20 -03:00
Honey Tree
73dabd8905 Simple implementation of caching at depth that should be useful for most use cases 2024-11-17 08:32:36 -03:00
Cohee
c9d2b609f1 match => includes 2024-11-16 15:41:41 +02:00
M0cho
30bca8e39b Supports GEMINI EXP 2024-11-15 05:26:10 +09:00
Cohee
8da98faa99 MistralAI: Include frequency/presence penalties in requests 2024-11-08 11:09:57 +02:00
Cohee
2daff0e736 Console log full CC error response 2024-11-05 16:40:05 +00:00
dylan
4b5f485bd5 Add NanoGPT as chat completions provider 2024-10-26 16:57:49 +13:00
Cohee
4fcad0752f [chore] Fix type errors 2024-10-11 21:33:36 +03:00
Cohee
90eef29982 Merge branch 'staging' into es-modules 2024-10-11 20:30:23 +03:00
Cohee
a0889a15fd Explicitly import node process 2024-10-11 10:43:29 +03:00
Cohee
d52b4fbbde Node: Migrate to ES Modules 2024-10-10 22:37:22 +03:00
Honey Tree
ac05840b81 Remove a trailing slash from the API URL in case it gets added in by URL.toString 2024-10-10 14:58:03 -03:00
Cohee
c2935ba4a7 Migrate Cohere to /v2/chat 
Closes #2916
 2024-10-08 23:50:09 +03:00
Cohee
991e10c75f Claude: remove tools from prompt if tools not supported or not provided 2024-10-06 23:02:38 +03:00
Cohee
6185974e17 Claude: Use multi-part system prompt, cache tools 2024-10-05 18:04:08 +03:00
Cohee
3b4a455ef8 Split Custom OAI prompt post-processing modes 2024-10-05 16:09:39 +03:00
Cohee
c3c10a629e Claude: new prompt converter + non-streaming tools 2024-10-04 03:41:25 +03:00
Cohee
559f1b81f7 Remove tool calling for Cohere v1 2024-10-04 02:11:46 +03:00
Cohee
576352817e Merge branch 'staging' into tool-calling 2024-10-03 22:36:02 +03:00