165 Commits

Author SHA1 Message Date
ArrangingFear56
6435ea40f1 change default computing types for different devices 2024-12-10 01:23:13 +08:00
ArrangingFear56
d86a8622c3 modified server.py for ease of choosing faster-whisper-device 2024-12-10 01:03:49 +08:00
ArrangingFear56
dbec27b18a added arguments for faster-whisper usage without breaking original whisper implementation 2024-12-10 00:55:46 +08:00
Cohee
3d40acaacc Pass seed and clip skip to SD 2024-05-30 20:21:51 +03:00
Juha Jeronen
af873577f7 report number of items the embedding is being computed for
(since SentenceTransformer doesn't show that in the progress bar)
2024-01-24 16:31:47 +02:00
Juha Jeronen
09c007b96c add clarifying progress message 2024-01-24 15:44:33 +02:00
Juha Jeronen
0e506b1ce8 add vector embeddings provider 2024-01-24 15:12:15 +02:00
Juha Jeronen
1b1ae11c70 cleanliness 2024-01-24 15:12:04 +02:00
Juha Jeronen
f0a311cbfe startup message formatting 2024-01-10 17:26:24 +02:00
Juha Jeronen
84c7176a34 better to reset to defaults by a blank JSON (consistent semantics) 2024-01-10 17:15:40 +02:00
Juha Jeronen
89fc188694 add API endpoints load_emotion_templates, load_animator_settings 2024-01-10 16:57:47 +02:00
Juha Jeronen
930e1c8736 improve docstring 2024-01-10 16:57:35 +02:00
Juha Jeronen
332b4ed08d refactor: reorganize main script to improve readability 2024-01-10 15:48:01 +02:00
Juha Jeronen
f8e1efb5b6 startup message: highlight talkinghead device and model 2024-01-09 23:54:23 +02:00
Juha Jeronen
1f4f2a1ffa talkinghead now manages its threads internally 2024-01-09 23:54:13 +02:00
Juha Jeronen
bd29500c9e refactor THA3 model installer 2024-01-08 15:45:44 +02:00
Juha Jeronen
6633535e19 autopep8, and also apply some idiomatic Python style manually.
No more flake8 warnings.
2024-01-08 14:51:27 +02:00
Juha Jeronen
e94521251f API cleanliness 2024-01-08 14:46:09 +02:00
Juha Jeronen
f9804dc30c add support for "/emote xxx" in talkinghead 2024-01-02 02:26:51 +02:00
Juha Jeronen
45c2ae5ea5 add missing @require_module for talkinghead API endpoints 2024-01-02 02:26:05 +02:00
Juha Jeronen
6558edb97f Squashed commit of the following:
- Add experimental visual postproc chain (lo-fi scifi hologram)
  - TODO: not configurable yet, it's always on until I fix this
- Improve emotion preset loading logic
  - Even if an emotion preset JSON is missing, load the emotion from _defaults.json.
- Add blunder recovery (emotion preset factory reset) options to manual poser
- Fix factory-default preset name angry -> anger
- Manual poser: return nonzero exit code on init error
- Manual poser too now auto-installs THA3 models if needed
- Move TODO list into its own file, dump everything there
- Add a README for the new revised talkinghead
2023-12-25 02:08:10 +02:00
Juha Jeronen
4c6f843ff9 Talkinghead performance improvements and refactoring (#207)
* talkinghead: fix and improve THA3 manual poser

* server.py: no, don't yet use fp16 for talkinghead

* talkinghead: remove wxPython dependency from live mode

* comment out unused functions

* add TODO list

* coding style

* remove unused import

* add TODO marker

* message wordings

* fix typos in variable names

* talkinghead updates

* talkinghead updates

* Empty commit

* presentation order, sectioning

* fix the inMotion flag update

* mark a TODO

* refactor

* remove done TODO items

* mark a TODO

* comment wording

* pause animation while loading a new image

* parser doesn't belong here, not a command-line app anymore

* message wording

* use finally

* remove superfluous "global" declarations

* lots of cleanup

* remove silly sys.path entry

* improve docstring

* oops

* app.py now only serves the live mode for the plugin

* talkinghead live mode: remove ifacialmocap stuff (unused)

* improve comment

* list walking is so 1990s

* use double quotes consistently

* remove now-unused ifacialmocap-related files from the repo

* remove done TODO item

* improve docstring

* update comment

* remove now-unused function

* update comment

* improve docstring

* add TODO marker

* oops, typo

* add --talkinghead-model command-line option to server.py

Default is 'auto': float16 on GPU, float32 on CPU.

* talkinghead: auto-install THA3 models if needed

* remove tha3/models from git repo (have autodownload now)

* Add hf-hub as explicit dependency

* Add THA models to gitignore

---------

Co-authored-by: Cohee <18619528+Cohee1207@users.noreply.github.com>
2023-12-21 23:48:25 +02:00
Cohee
47a5489142 Use pipeline for text summarization 2023-12-20 01:57:38 +02:00
Cohee
423f51e3f8 Use pipeline for image captioning 2023-12-20 01:28:53 +02:00
Cohee
98535e6fa3 Add links to search results 2023-12-11 23:52:25 +02:00
Cohee
aded5b4363 Add support for DuckDuckGo search engine in web search API 2023-11-28 22:39:35 +02:00
Cohee
05cab1c918 Add websearch module 2023-11-28 20:36:02 +02:00
Cohee
51a680946f Fix chroma embedding function 2023-11-16 18:28:52 +02:00
Cohee
beb434ae78 Fix chromadb purge 2023-10-18 23:54:37 +03:00
Cohee
60045b599f Fix double memory consumption for classify module 2023-09-03 16:45:20 +03:00
Cohee
5d51e752f4 #144 Increase max content length limit 2023-09-01 14:36:36 +03:00
Cohee
d68dd41e12 Merge pull request #133 from Tony-sama/neo
RVC expression-based dynamic voice
2023-08-24 21:27:18 +03:00
joe
63528152ef Ignore Auth for TalkingHead Fix 2023-08-24 18:08:45 +09:00
Tony Ribeiro
650b05190a Added dedicated script for classify feature. Clean call to classifier module in RVC. 2023-08-21 17:15:48 +02:00
Tony Ribeiro
16d267c916 Merge branch 'SillyTavern:neo' into neo 2023-08-21 05:21:00 +02:00
deffcolony
4bc0238f5c Update server.py 2023-08-18 23:31:42 +02:00
deffcolony
7311748c1f Python Colorama for some console prints 2023-08-18 23:23:14 +02:00
Tony Ribeiro
b3dd6f8ce3 Merge branch 'SillyTavern:neo' into neo 2023-08-18 17:43:32 +02:00
deffcolony
e3b51dfbab adds green color for url line
This will make sure the link is more visible in console
2023-08-18 16:49:04 +02:00
Tony Ribeiro
fa462c1979 Prototype of RVC + emotion classification pipeline 2023-08-18 01:10:57 +02:00
Tony Ribeiro
b5fcde2611 Added RVC request function to allow upload models from ST. Added parameter --max-content-length=X (Mb) to set server max request data size accordingly if needed. 2023-08-17 04:44:49 +02:00
Tony Ribeiro
185fc1334f Added parameter --rvc-save-file to save last rvc input/output audio file into data/tmp. For research purpose and easy sharing of TTS/RVC sample files 2023-08-17 01:43:13 +02:00
Tony Ribeiro
4992ce0bd3 Added Coqui STT module request function for listing local models and generating tts using it. 2023-08-16 19:20:55 +02:00
Tony Ribeiro
da6bbf879e Added RVC module function to fix models folder organisation (if possible) 2023-08-16 16:38:59 +02:00
Tony Ribeiro
fd6ef64d80 Updated RVC voice conversion function to handle new parameters 2023-08-15 21:16:59 +02:00
Tony Ribeiro
75d48a9638 Added argument '--coqui-models=model1,model2' allowing to download given coqui-api models and keep the last one loaded ready for first tts request. 2023-08-15 17:20:23 +02:00
Tony Ribeiro
84ed0f5de3 remove debug message 2023-08-15 15:43:58 +02:00
Tony Ribeiro
a403eb0247 Corrected coqui gpu option to be actually taken into account ^^ 2023-08-15 15:43:00 +02:00
Tony Ribeiro
c800eae3e8 Cleanup remaining old coqui implementation 2023-08-14 05:32:34 +02:00
Tony Ribeiro
46fecb5b48 Merge branch 'neo' of https://github.com/Tony-sama/SillyTavern-extras into neo 2023-08-14 04:04:55 +02:00