kwaroran
c3d7210cc3
Refactor caching logic in sw.js and transformers.ts
2024-01-06 21:18:36 +09:00
kwaroran
afd085bd93
Fix tfCache initialization and assignment
2024-01-06 20:50:49 +09:00
kwaroran
46ee706399
Add initialization function for Transformers and update model paths
2024-01-06 20:49:34 +09:00
kwaroran
85cd1cbc65
Add 404 response for 'tf' resource in sw.js
2024-01-06 20:32:19 +09:00
kwaroran
da77f12f7e
Update version to 1.71.2
2024-01-06 20:25:31 +09:00
kwaroran
a9872b3d5d
Update localModelPath in transformers.ts
2024-01-06 20:25:03 +09:00
kwaroran
1114d5b61b
Update version to 1.71.1
2024-01-06 19:46:19 +09:00
kwaroran
229aeced06
Merge branch 'main' of https://github.com/kwaroran/RisuAI
2024-01-06 19:45:57 +09:00
kwaroran
5a305a8f10
Fix sw register temp
2024-01-06 19:45:50 +09:00
kwaroran
cba3ff802c
Add selectable tokenizer supports on Ooba ( #281 )
...
# PR Checklist
- [ ] Did you check if it works normally in all models? *ignore this
when it dosen't uses models*
- [ ] Did you check if it works normally in all of web, local and node
hosted versions? if it dosen't, did you blocked it in those versions?
- [ ] Did you added a type def?
# Description
I write simple changes on code, which allow user to choose tokenizers.
As I write on https://github.com/kwaroran/RisuAI/issues/280 , differences
in tokenizers makes error when use mistral based models.
As I'm not good at javascript, I simply implement this work by write
name of tokenizer model, and select one on tokenizer.ts file.
I test it on my node RisuAI and I send long context to my own server.
As result, ooba returned 15858 as prompt tokens.
And as I test on official tokenizer implementations, it shows 1k
differences between llama tokenizer and mistral tokenizer.
So I think adding this option will help users use oobabooga with less
error.
2024-01-06 19:16:58 +09:00
kwaroran
8ba0d10d56
Update version to 1.71.0
2024-01-06 19:13:57 +09:00
kwaroran
70f25968b5
add transformers remoteHost again
2024-01-06 19:12:18 +09:00
kwaroran
497e0c6dfb
Update service worker and remote host URLs
2024-01-06 19:10:21 +09:00
kwaroran
f4d6fcb38c
Update characterCards.ts with VITS export support
2024-01-06 07:05:20 +09:00
kwaroran
7344e566f4
Add model selection for VitsModel
2024-01-06 06:45:18 +09:00
kwaroran
66c6511684
Refactor sw.js to support custom headers in registerCache function
2024-01-06 01:28:00 +09:00
kwaroran
0ca4ec3695
Add VITS support
2024-01-06 00:28:06 +09:00
kwaroran
cfdd60bead
Remove TransformersBodyType
2024-01-05 23:58:23 +09:00
justpain02
ff4c67b993
Add selectable tokenizer supports on Ooba
2024-01-05 23:57:59 +09:00
kwaroran
56edb14e44
Update transformers import and add config parameter to runTransformers function
2024-01-05 23:54:01 +09:00
kwaroran
5ae3f35df6
Refactor Transformers
2024-01-05 23:51:37 +09:00
kwaroran
b6787b93a7
Fix extractor initialization in runEmbedding function
2024-01-05 23:33:27 +09:00
kwaroran
b4272e9e69
Update @xenova/transformers to version 2.13.4
2024-01-05 23:31:03 +09:00
kwaroran
dcebac658f
[fix] embedding over 512 tokens
2024-01-05 23:29:57 +09:00
kwaroran
809d8706cf
Merge branch 'main' of https://github.com/kwaroran/RisuAI
2024-01-05 19:58:43 +09:00
kwaroran
8ebaaca59b
Update version to 1.70.3
2024-01-05 19:58:37 +09:00
kwaroran
846441f04a
fix translator customscript
2024-01-05 19:58:19 +09:00
kwaroran
ca44a6c27b
Update README.md
2024-01-04 22:45:45 +09:00
kwaroran
1d6e896914
Update README.md
2024-01-04 22:44:59 +09:00
kwaroran
f373363038
Update version to 1.70.2
2024-01-04 17:48:53 +09:00
kwaroran
ec26af35d9
Merge branch 'main' of https://github.com/kwaroran/RisuAI
2024-01-04 17:48:28 +09:00
kwaroran
28b08d1c6c
Update msgpackr version to 1.10.1
2024-01-04 17:48:23 +09:00
kwaroran
0b3c702f62
[fix] writer
2024-01-04 17:47:23 +09:00
kwaroran
1c58b9654e
Update README.md
2024-01-04 17:12:02 +09:00
kwaroran
495c4df698
[ver] 1.70.1
2024-01-04 17:09:43 +09:00
kwaroran
e5c65d7b09
[fix] welcome freezing
2024-01-04 17:09:07 +09:00
kwaroran
7c77b55e39
[fix] localwriter
2024-01-04 05:30:06 +09:00
kwaroran
a890a2acb7
[ver] 1.70.0
2024-01-04 03:58:07 +09:00
kwaroran
20bc35d194
[feat] change default italic for high contrast
2024-01-04 03:57:04 +09:00
kwaroran
5093fd9d0b
[fix] high constrast by default
2024-01-04 03:54:13 +09:00
kwaroran
7428fed704
[feat] new loading screen
2024-01-04 03:52:44 +09:00
kwaroran
34a4422af0
[feat] add horde default key
2024-01-04 03:49:21 +09:00
kwaroran
36c12de39f
[patchnote] 1.70.0
2024-01-04 03:48:06 +09:00
kwaroran
7aae1aa139
[feat] translations
2024-01-04 03:45:08 +09:00
kwaroran
4f77504d3f
[feat] add translate replace regex
2024-01-04 03:37:05 +09:00
kwaroran
5f2275eb22
[feat] better image saving
2024-01-04 03:18:11 +09:00
kwaroran
a13fdced2e
[fix] openai max context
2024-01-04 02:36:45 +09:00
kwaroran
036b23dbdc
[fix] block native
2024-01-04 02:35:08 +09:00
kwaroran
499e91ce6f
[fix] mobile saving
2024-01-04 02:34:22 +09:00
kwaroran
32e3060b19
[feat] better openrouter support
2024-01-04 02:32:53 +09:00
