4227968dfa
Allow using JSON schema with llamacpp server
2024-05-18 18:50:48 +03:00
c7d75b7789
llamacpp broke
2024-05-12 21:41:07 +03:00
27ccc6b090
Minor stylistic changes
2024-05-11 11:38:22 +03:00
62faddac8d
Textgen: Add banned_strings
...
TabbyAPI supports the ability to ban the presence of strings during
a generation. Add this support in SillyTavern by handling lines
enclosed in quotes as a special case.
Signed-off-by: kingbri <bdashore3@proton.me>
2024-05-11 00:58:29 -04:00
c73bfbd7b0
Safari bruh moment
2024-05-06 21:21:03 +03:00
7063fce2af
Selectable openrouter providers
2024-05-06 19:26:20 +03:00
05db2552b3
Fix Top K disabled state for Infermatic.
...
Also an icon.
2024-05-04 02:37:05 +03:00
7bfd666321
Add llama 3 tokenizer
2024-05-03 23:59:39 +03:00
7b87f44518
Clean-up API-specific settings
2024-05-03 20:02:13 +03:00
2bd239fe81
Initial vLLM support
2024-05-02 22:40:40 +00:00
022c180b62
Lint and clean-up
2024-04-15 00:39:15 +03:00
3e60919289
Specify LLM prompt in case JSON schema is not supported
2024-04-14 17:13:54 +03:00
b8b49f0012
TextgenSettings: Fix JSON schema fallback
...
Did not fall back if the provided string was empty, resulting in
errors
Signed-off-by: kingbri <bdashore3@proton.me>
2024-04-09 22:15:00 -04:00
51b3b8bfaa
Add smooth streaming
2024-04-02 14:56:15 +03:00
04edf32ef0
Do not send dynatemp to backends if disabled
2024-04-02 11:29:49 +03:00
0b76e1d350
Fix schema not loading from presets. Fix ESLint warnings
2024-04-02 11:23:29 +03:00
5210db5679
Format
...
Signed-off-by: kingbri <bdashore3@proton.me>
2024-04-02 01:01:59 -04:00
4f0322351e
Sampling: Add ability to send JSON schemas
...
TabbyAPI supports the ability to send JSON schemas with prompts in
addition to EBNF strings supported by outlines. Add an extra box
for TabbyAPI only.
Signed-off-by: kingbri <bdashore3@proton.me>
2024-04-02 00:59:21 -04:00
a3ec0938c5
KoboldCpp grammar fix
2024-03-29 17:28:28 +02:00
9bd3a526aa
Fix for unique swipes
...
Fix for unique swipes in Aphrodite
2024-03-26 23:57:24 -04:00
8b092adc14
Use mode enum to toggle dynatemp behavior.
2024-03-25 12:25:03 -04:00
df805d692b
Fix some code
2024-03-24 21:42:27 +02:00
6f7e7b85ab
For Mancer:
...
- Allow logprobs (works)
- Allow multiswipe (not yet)
- Adjust visible samplers
Fix: 0 logprob is 100% chance, handle accordingly.
2024-03-24 14:45:37 -04:00
c8f84bd413
Textgen setting refactors
2024-03-19 01:38:55 +02:00
dc74f546d5
Merge pull request #1875 from kalomaze/cubic-curve
...
smoothing_curve UI support
2024-03-17 02:20:51 +02:00
a0279b636b
Remove dead code
2024-03-08 08:41:54 +02:00
2cdfda9d69
Actually use getCurrentDreamGenModelTokenizer
2024-03-08 08:40:03 +02:00
bc8d41b530
Implement suggestions
2024-03-07 17:28:38 +01:00
5c410986a4
Add support for DreamGen API.
...
API docs: https://dreamgen.com/docs/models/opus/v1
API keys: https://dreamgen.com/account/api-keys
I decided to base this on the text-completion API since it's more
flexible with SillyTavern's prompt formating capabilities.
This also means that custom context and instruct settings are required.
Will add documentation in a followup PR.
2024-03-07 12:25:48 +01:00
95c49029f7
Add aphrodite model selector
2024-03-01 23:02:43 +02:00
45776de1d5
Smoothing curve support for ooba
2024-03-01 00:06:34 -06:00
76669ff8bb
add max_tokens_second
2024-02-29 00:55:25 +01:00
f962ad5c02
Add OpenRouter as a text completion source
2024-02-25 22:47:07 +02:00
fc289126fa
Add event type for text completion generation request settings ready
2024-02-24 21:45:33 +02:00
d140b8d5be
Parse non-streaming tabby logprobs
2024-02-24 20:10:53 +02:00
8848818d67
Fix dynatemp neutralization
2024-02-24 15:32:12 +02:00
299bd9d563
Merge branch 'staging' into llamacpp-sampler-order
2024-02-24 15:10:58 +02:00
13aebc623a
Merge pull request #1854 from deciare/llamacpp-probs
...
Request and display token probabilities from llama.cpp backend
2024-02-24 15:06:28 +02:00
9287ff18de
Fix for non-streaming
2024-02-24 14:50:06 +02:00
9eba076ae4
Sampler order for llama.cpp server backend
2024-02-23 23:01:04 -05:00
936fbac6c5
Merge remote-tracking branch 'origin/staging' into llamacpp-probs
2024-02-23 17:45:54 -05:00
344b9eedbc
Request token probabilities from llama.cpp backend
...
llama.cpp server token probabilities are given as values ranging from
0 to 1 instead of as logarithms.
2024-02-23 14:01:46 -05:00
f569424f3e
Merge branch 'staging' into InfermaticAI
2024-02-22 08:32:10 -06:00
0ccdfe4bb7
Fix duped line
2024-02-22 02:45:35 +02:00
0c1cf9ff2e
Send sampler priority as array
2024-02-21 00:53:54 +02:00
32ee58e5e6
fix kcpp order reset
2024-02-19 18:12:56 -06:00
b5887960b6
Merge branch 'release' into InfermaticAI
2024-02-16 08:53:04 -06:00
9fed7ed742
Make the neutralize option turn off quad sampling
2024-02-10 17:57:00 -06:00
70deb11d27
implement jank js + drag n drop sampler priority
2024-02-10 16:32:46 -06:00
818029288e
Remove sending it if it's Aphrodite or TabbyAPI
2024-02-10 15:22:24 -06:00
2065f95edc
Sampler priority support
2024-02-10 14:57:41 -06:00
d8d4732614
Add koboldcpp param aliases
2024-02-09 15:32:40 +02:00
530455979f
Textgen: Add tabby to logprobs
...
Tabby now supports logprob returns. Add support to the backend gate.
Signed-off-by: kingbri <bdashore3@proton.me>
2024-02-08 20:59:54 -05:00
f372b2ac16
Textgen: Fix dynatemp defaults
...
The default min and max temp values should be 1
Signed-off-by: kingbri <bdashore3@proton.me>
2024-02-08 20:57:40 -05:00
f1cadfde2b
Fix llamacpp server
2024-02-04 17:41:46 +02:00
3e34123b31
Textgen: Fix sampling param requests
...
1. TabbyAPI was not being included in temperature_last
2. DynaTemp parameters were still being sent even though the checkbox
was off. Cleanup the parameters if dynatemp is off for backends that
don't take in a dynatemp boolean.
Signed-off-by: kingbri <bdashore3@proton.me>
2024-02-04 01:44:07 -05:00
6d0d92aaff
Add grammar and dynatemp for aphrodite
2024-02-02 20:06:46 +02:00
da68d35ca5
Remove Noisy Sampling
...
Leaves DynaTemp Exponent and Quadratic Sampling.
2024-01-30 21:50:50 -05:00
92aae0da04
Enter the Kalo Zone
2024-01-28 15:14:21 -05:00
8af3939f09
fix: logprobs for Aphrodite
...
Since Aphrodite follows the same logprobs style as ooba, we can enable it easily.
2024-01-26 07:15:39 +00:00
60044c18a4
Implement Token Probabilities UI using logprobs
2024-01-25 18:34:46 -06:00
f29f934c6b
Progress
2024-01-24 06:59:27 -06:00
570d5a30bd
[skip ci] Lint fix
2024-01-20 20:40:40 +02:00
5cb4dbc280
Trim TG types after split
2024-01-15 20:45:21 +02:00
22e605168b
Properly reset seed with neutralize samplers
2024-01-15 20:44:07 +02:00
c6487fa9cb
make sampler nuke button work with zensliders
2024-01-14 18:33:40 +09:00
80b39bdac1
add neutralize samplers button for text completion slider panel
2024-01-14 18:07:43 +09:00
44bd57f763
filter ooba-only samplers out of non-ooba text completion slider panels
2024-01-14 17:26:22 +09:00
1367642eb4
Fix maxTokens reference for n_predict alias
2024-01-12 12:20:03 +02:00
8659bd57b7
Add koboldcpp rep pen aliases
2024-01-12 12:12:53 +02:00
ec63cd8b6d
Dynatemp Range
...
Kobold Dynatemp range set when deactivated, now will work properly
2024-01-09 00:54:20 -05:00
5ad980cf99
Fix for realzies
2024-01-09 00:02:53 -05:00
04a5d8390d
Dynatemp UI v3.1
...
fixes for html positioning as well as api settings.
2024-01-08 23:58:06 -05:00
b334acec23
DynaTemp UI v3
...
Backend bros won't play along now we gotta make them hold hands.
2024-01-08 22:20:27 -05:00
2bea90a264
Merge pull request #1633 from AAbushady/new-dynatemp-ui
...
DynaTemp UI v2
2024-01-06 00:33:25 +02:00
224dcc83a1
Update DynaTemp Sliders
...
More inline with changes made that will be implemented in KCPP main.
2024-01-05 14:36:39 -05:00
5f93c30a96
#1627 Bypass status check and custom model for textgen type
2024-01-05 19:15:07 +02:00
86d715cc16
Improve tokenizer detection
2024-01-05 16:17:06 +02:00
d66be24518
DynaTemp UI v2
...
With DynaTemp moving into main KCPP now ST can have DynaTemp too!
2024-01-05 02:08:44 -05:00
47cb017a45
#1589 Add 'cache_prompt' for l.cpp
2023-12-25 02:42:03 +02:00
67dd52c21b
#1309 Ollama text completion backend
2023-12-19 16:38:11 +02:00
edd737e8bd
#371 Add llama.cpp inference server support
2023-12-18 22:38:28 +02:00
6e8104873e
#1569 Add logit bias for text completions
2023-12-18 18:57:10 +02:00
b0d9f14534
Re-add Together as a text completion source
2023-12-17 23:38:03 +02:00
5b3c96df50
Rename /textgenerationwebui endpoint
...
I'd like to migrate over to using "textgen" to mean text-generation APIs
in general, so I've renamed the /textgenerationwebui/* endpoints to
/backends/text-completions/*.
2023-12-12 16:40:14 -05:00
b0e7b73a32
Fix streaming processor error handler hooks
2023-12-08 02:01:08 +02:00
cdcd913805
Don't stream events if the API returned a 4xx code
2023-12-07 18:00:36 -05:00
5540c165cf
Refactor server-sent events parsing
...
Create one server-sent events stream class which implements the entire
spec (different line endings, chunking, etc) and use it in all the
streaming generators.
2023-12-07 18:00:36 -05:00
0a092629ce
Fix undefined variable reference
2023-12-03 23:24:39 +02:00
24ccef3aba
Fix "settings" variable being shadowed
2023-12-03 15:44:36 -05:00
9c33ddbafc
Make textgen settings type checks more concise
2023-12-03 14:56:01 -05:00
047c897ead
Remove is[API] functions
...
Just use an equality comparison. It's a bit longer, but only because
"textgenerationwebui_settings" is a long identifier.
2023-12-03 14:56:01 -05:00
ba54e3dea0
Replaces is_[api] params with api_type param
...
These were 5 mutually-exclusive booleans, which can be replaced with one
param that takes on 5 values, one for each API type.
2023-12-03 14:56:01 -05:00
939e938ba2
Disallow multiswipe for quiet gens
2023-12-03 20:56:25 +02:00
1786b0d340
#1403 Add Aphrodite multi-swipe
2023-12-03 20:40:09 +02:00
64a3564892
lint: Comma dangle
2023-12-02 22:06:57 +02:00
c63cd87cc0
lint: Require semicolons
2023-12-02 21:11:06 +02:00
a37f874e38
Require single quotes
2023-12-02 13:04:51 -05:00
60ac1aa1c7
Enable no-empty lint
2023-12-02 10:32:07 -05:00
61908935f5
Stop string for user-continue. Trim spaces after name2
2023-11-22 16:16:48 +02:00
Cohee
kingbri
sasha0552
Alexander Abushady
50h100a
50h100a
DreamGenX
kalomaze
gabriel dhimoila
Deciare
NWilson
AlpinDale
khanon
RossAscends
Alexander Abushady
valadaptive