Cohee
a6a7810be2
Regex: fix "Alter Outgoing Prompt" applying when unchecked ( #3380 )
...
* Regex: rework ephemerality options.
* Fix ephemeral regex being executed when unwanted
* Revert to old code style
2025-01-29 11:39:32 +02:00
Cohee
9e8fd3f5a0
Merge pull request #3382 from SillyTavern/staging
...
Staging into reasoning-regex
2025-01-29 11:38:50 +02:00
Cohee
4f456b2b80
TTS: Call module worker when audio stops playing
...
#3371
2025-01-28 23:14:26 +02:00
Cohee
6dde068e71
Merge pull request #3371 from fearmear/release
...
Decrease TTS generation delay by splitting a message on a new line
2025-01-28 23:10:12 +02:00
Cohee
2aa1dd41f5
eslint
2025-01-28 23:07:59 +02:00
Cohee
26feb448a3
Merge pull request #3368 from SillyTavern/profile-stopstrings
...
Add command and profile for custom stop strings
2025-01-28 23:03:09 +02:00
Dzmitry Kazlouski
63e7acb87b
Make this feature togglable in extensions > "Narrate by paragraphs (when not streaming)"
2025-01-28 05:50:25 +03:00
Small Eggs
145136059e
Fix tts.skip_tags's regex to match newlines
...
The extension_settings.tts.skip_tags setting is meant to skip sending
tags and their content to the TTS API provider. The original regular
expression matched content inside tags with ".*?". Unfortunately,
Javascript's engine does *not* match newlines on the "." without the /s
flag.
The /s flag was added in ES2018. To be more compatible, the regex
has been changed to "[\s\S]+?". This gives similar performance (instead
of using capture groups) and matches all content within a tag, as the
original regex intended.
2025-01-27 17:43:24 -08:00
Dzmitry Kazlouski
283ceb6bbf
Decrease TTS generation delay by splitting a message on a new line
2025-01-28 02:51:05 +03:00
Cohee
6fc342d446
Add regex processing for reasoning blocks
2025-01-27 23:06:07 +02:00
Cohee
fad4e4e75e
Add command and profile for custom stop strings
2025-01-27 22:30:35 +02:00
sirius422
11882827c7
Add new Gemini thinking model and its alias, specify context size and vision support
2025-01-22 18:37:43 +08:00
Cohee
4ebff0ffcc
Merge branch 'staging' into ruRuJan
2025-01-13 16:53:26 +00:00
Cohee
b177affd81
Merge pull request #3289 from zerofata/vector-db-search-cmd
...
Add returnChunks and resultSize args to /db-search
2025-01-11 23:05:06 +02:00
zerofata
89c984830b
rename resultSize to count
2025-01-12 07:20:36 +13:00
Yokayo
1d5cf8d25c
Work on translation
2025-01-12 00:42:58 +07:00
zerofata
fc5debc660
replaced returnChunks with return arg to use existing helper functions
2025-01-11 17:42:19 +13:00
Wolfsblvt
b89d41a701
Refactor prepare/redraw for consistency
2025-01-11 02:31:14 +01:00
Wolfsblvt
3455da404a
Add QR set duplicate button
2025-01-11 02:22:10 +01:00
Wolfsblvt
0f45ebda03
Refactor QR set popups to new popup
2025-01-11 01:59:38 +01:00
Wolfsblvt
4c1c62a56b
Add QR set rename button
2025-01-11 01:41:46 +01:00
zerofata
7f5ce54b1f
Add returnChunks and resultSize args to /db-search
2025-01-11 11:32:05 +13:00
Cohee
efeedc1274
Revert gallery selector changes
2025-01-09 22:39:40 +02:00
Wolfsblvt
a8b5f8a95f
Add /css-var slash command
...
- Add /css-var slash command to set CSS variables on some elements
- Update gallery div id to something actually unique ("#draggable_gallery")
2025-01-09 18:37:50 +01:00
Cohee
23779fe565
Merge branch 'release' into staging
2025-01-08 21:20:50 +02:00
erew123
5ea8309a30
Change check from ! not to is == V2
2025-01-08 00:01:36 +00:00
erew123
de8ecc6903
Correct skipping RVC on AllTalk V1 Server
...
Silly bug that caused the extension when used with the legacy V1 AllTalk server, would cause it to skip downloading Narrator voices.
Literally changed "v2" to "v1" on line 391
No other changes.
2025-01-07 23:20:53 +00:00
Cohee
d7b3a56c3d
chore: await before returning
2025-01-07 20:07:41 +02:00
Cohee
81841ca2a6
WebLLM: use current tokenizer if not available
2025-01-07 20:01:59 +02:00
Cohee
989f5a8692
Image Gen: don't generate a new swipe seed if already using a random seed
2025-01-05 00:49:23 +02:00
Cohee
77bee453ab
Vectors: Add only custom boundary chunking
2025-01-03 20:44:25 +02:00
Cohee
02900520c9
chore: Add index property to HashedMessage
2025-01-03 03:09:16 +02:00
Cohee
eff4da42c5
Vector storage: skip for quiet prompts
2025-01-03 03:04:45 +02:00
Cohee
2f5f9a437d
Allow running generate interceptors on quiet prompts
2025-01-03 01:10:14 +02:00
Cohee
d0e1c3287a
NovelAI: Fix scheduler auto-assign
2025-01-02 22:58:06 +02:00
Cohee
e0af5ff353
Default to undefined for no VAE selected
2024-12-31 00:13:26 +02:00
Cohee
8447c7e637
Fix VAE for SD forge
2024-12-31 00:12:25 +02:00
Cohee
f416b6cd54
Revert "Pass VAE name to SD forge"
...
This reverts commit 3a297561095df5139d2ff34142c919e0e7e0edad.
2024-12-30 23:47:57 +02:00
Cohee
3a29756109
Pass VAE name to SD forge
2024-12-30 01:12:37 +02:00
Cohee
e1b6eab9eb
Fix SD forge generation with hires
2024-12-30 00:09:37 +02:00
Cohee
c3ad87cea1
ComfyUI: negate clip_skip placeholder value
2024-12-28 16:05:58 +02:00
Cohee
149de7d257
Image Gen: make CFG step 0.1
2024-12-28 02:28:58 +02:00
Cohee
59772bfd27
Merge pull request #3229 from Sevenyine/staging
...
Add new Google AI Studio model
2024-12-27 22:40:04 +02:00
Cohee
a82c05a8ac
Gemini thinking: Specify context size, system prompt and vision support
2024-12-27 22:39:26 +02:00
Cohee
29acc2e2db
Add Deepl endpoint selection
...
Fixes #3224
2024-12-27 13:02:06 +00:00
Cohee
c42a7739d6
NovelAI: Fix backward compatibility with scheduler setting
2024-12-27 00:24:53 +02:00
Cohee
b5eb2b5150
NovelAI: Limit maximum steps to 50
2024-12-26 22:51:58 +02:00
Cohee
27f8768141
NovelAI: Make Euler Ancestral a default sampler
2024-12-26 22:44:15 +02:00
Cohee
844d8bdee7
NovelAI: Disable SMEA/DYN for V4
2024-12-26 22:42:01 +02:00
Cohee
f92f8f4a84
NovelAI: Remove deprecated SD models
2024-12-26 22:24:37 +02:00
