Cohee
ab52af4fb5
Add support for Koboldcpp tokenization endpoint
2023-08-24 20:19:57 +03:00
Mike Weldon
1d1109e43b
Set generate_until_sentence for NovelAI
...
* Set generate_until_sentence true for NovelAI
* Add a Story String file for NovelAI with persona before character
which works better
* Remove hardcoded dinkus for chat_start since it is in the Story String
2023-08-23 18:04:56 -07:00
Cohee
4aa31fcba9
Add fallback option for OpenRouter
2023-08-24 03:21:17 +03:00
Cohee
41cc86af9f
Add example extension for chat variables. Allow registering custom text processing functions for extensions
2023-08-22 17:46:37 +03:00
based
ba925f388c
added more options to the scale request
2023-08-22 21:29:18 +10:00
based
06902279a9
merge
2023-08-22 21:17:18 +10:00
Cohee
897ad28477
Apply preset manager for instruct presets
2023-08-21 00:56:53 +03:00
Cohee
a27bef8b12
Merge branch 'staging' into qolfeatures
2023-08-20 18:47:43 +03:00
Cohee
efa0f12349
Fix prompt manager issues
2023-08-20 16:25:16 +03:00
based
7ecc0295dc
scale cookie method
2023-08-20 20:55:37 +10:00
Cohee
6fb278266b
Merge pull request #977 from city-unit/feature/exorcism
...
Feature/exorcism
2023-08-20 12:37:56 +03:00
Cohee
94754af013
Reformat code
2023-08-20 12:37:38 +03:00
city-unit
e8c996ff61
Move to char+timestamp for img naming.
2023-08-20 01:41:58 -04:00
50h100a
83eb3449f3
localhost -> 127.0.0.1.
...
Thank you websockets, very cool.
2023-08-20 00:30:34 -04:00
city-unit
3ce8343016
Add comments and stuff
2023-08-20 00:15:57 -04:00
city-unit
e76001a600
Merge branch 'staging' of https://github.com/city-unit/SillyTavern into feature/exorcism
2023-08-19 23:58:11 -04:00
city-unit
2f09efcd7f
Exorcised base64 image stuff
2023-08-19 23:01:09 -04:00
Cohee
bab368e3d2
Fix auto-load not triggering from hotswaps
2023-08-19 22:22:24 +03:00
based
ce40780ed1
Added option to exclude assistant prefix to Claude prompt (for jailbreaks already appending Assistant:)
2023-08-20 03:09:50 +10:00
Cohee1207
5a68cd61a1
Remove unnecessary log
2023-08-19 18:58:37 +03:00
Cohee1207
2cd2318298
Code clean-up
2023-08-19 18:52:06 +03:00
Cohee
438c78574f
Merge pull request #971 from kingbased/jurassic
...
AI21 Adapter
2023-08-19 18:48:15 +03:00
based
0f21eabb6e
AI21 Adapter + Tokenization implementation
2023-08-20 01:20:42 +10:00
Cohee1207
a9849d1883
More uploads cleanup
2023-08-19 17:50:16 +03:00
Cohee1207
6471ce2074
Clean-up uploads folder when possible
2023-08-19 17:43:56 +03:00
Cohee1207
991ff98eaa
Performance optimizations for character edit / chat opening
2023-08-19 15:58:17 +03:00
deffcolony
9839ea5b07
adds green color for url line
...
This will make sure the link is more visible in console
2023-08-18 16:48:23 +02:00
Cohee
f4dcd89ef6
Merge branch 'staging' of http://github.com/SillyTavern/SillyTavern into staging
2023-08-18 12:11:32 +03:00
Cohee
52431e246a
Make file write operations sync
2023-08-18 12:11:18 +03:00
50h100a
8c191c4115
groundwork for WebUI-like streaming with mancer
2023-08-17 18:29:40 -04:00
Cohee
e502354cec
Merge pull request #956 from 50h100a/streaming-fix
...
Fix (some) websocket shenanigans in WebUI streaming
2023-08-17 19:42:55 +03:00
Cohee
9c614529ca
Make all file write operations atomic
2023-08-17 15:20:02 +03:00
Cohee
0b412033ba
Streamline and simply Novel logit bias. Make it a part of preset
2023-08-17 13:05:22 +03:00
50h100a
13299e5c39
deal with websocket shenanigans
...
(and also clarify what's being printed out sometimes)
2023-08-17 05:52:32 -04:00
Mike Weldon
722801bb50
Add logit biases for NovelAI
2023-08-16 22:14:04 -07:00
Mike Weldon
feb523bd01
NovelAI Kayra 1.1 update
...
* Updated some presets and added Cosmic Cube
* Change defaults for NovelAI to select Clio on cold start
* Automatically change the preset to an appropriate default whenever you change the model
* Removed deprecated Top G sampler
2023-08-15 18:52:29 -07:00
Cohee
ea7268febd
Replace tokenizer packages with agnai versions (not crash dumping to console)
2023-08-15 18:08:34 +03:00
Cohee
57b9a94c17
Add banned tokens for Novel
2023-08-15 15:51:14 +03:00
Mike Weldon
41ec7e5600
Many NovelAI fixes from dev guidance
...
* Remove AI Module "Autoselect" and make the auto-instruct work for all modules. This is how NAI is supposed to work.
* Log the response from the API.
* Move the AI Module setting up to the top of the settings window since it isn't part of the preset.
* Refactor phrase_rep_pen to use the actual API strings.
* Clamp the maximum token length to 150 before we call the API.
* Clamp the minimum token length in the UX from 1 to 150.
* Fix bug where the preamble was not initialized on cold start.
* Get rid of extra newline before dinkus.
* Make always_force_name2 default true.
2023-08-14 19:35:21 -07:00
Mike Weldon
a0573413b2
Add new NAI samplers and update presets
2023-08-11 17:58:40 -07:00
spacegeek69
4b48861d13
[OpenRouter] Switch to token per dollar for prices
2023-08-10 21:13:24 +02:00
Cohee
23e869f56a
Implicit variables to explicit
2023-08-10 20:06:18 +03:00
spacegeek69
0ede3894e2
[OpenRouter] Show API prices on connect in console
2023-08-10 11:21:58 +02:00
Cohee
25bd2541f7
Fix server crash in error handler
2023-08-08 23:11:27 +03:00
Cohee
c94eae6eb6
Improve OpenAI/OpenRouter error parsing
2023-08-08 20:07:41 +03:00
Cohee
9a7654598e
Don't crash the server on trying to display corrupted past chat
2023-08-08 17:56:13 +03:00
Cohee
5f97a52d58
#895 Move AI Horde connection to related endpoints
2023-08-07 19:34:10 +03:00
Cohee
151f4d322c
#883 Add option to disable CSRF tokens
2023-08-06 16:42:15 +03:00
Cohee
74b973c571
Fix Novel error handling
2023-08-05 21:15:49 +03:00
Cohee
cd8a24a712
Fix bulk edit
2023-08-04 14:41:00 +03:00
