Feat/stable models - #5

Closed

b-re-w wants to merge 2857 commits into

feat/stable_models

This pull request is big! We're only showing the most recent 250 commits

Commits on Dec 30, 2025

more silent_token
aluminumbox
committed
e047900
more silent_token
aluminumbox
committed
f030a3f
update tokens
aluminumbox
committed
8fc255f
update tokens
aluminumbox
committed
a2abd89
update tokens
aluminumbox
committed
df30970

Commits on Dec 31, 2025

Merge branch 'main' into main
aluminumbox
authored
d292002
Merge branch 'main' into main
aluminumbox
authored
db0fadb
Merge branch 'main' into main
aluminumbox
authored
089d681
Merge pull request #1640 from Jzz1943/main
aluminumbox
authored
358718a
Merge pull request #1640 from Jzz1943/main
aluminumbox
authored
9267cbb
Merge pull request #1640 from Jzz1943/main
aluminumbox
authored
d4059a4
update readme
aluminumbox
committed
d731cd3
update readme
aluminumbox
committed
0c18489
update readme
aluminumbox
committed
9b3a2a1
Merge pull request #1758 from orbisai0security/fix/V-005-pickle-deserialization
aluminumbox
authored
8981fe4
Merge pull request #1758 from orbisai0security/fix/V-005-pickle-deserialization
aluminumbox
authored
f981f60
Merge pull request #1758 from orbisai0security/fix/V-005-pickle-deserialization
aluminumbox
authored
3cda348

Commits on Jan 5, 2026

Commits on Jan 7, 2026

fix sequence logic
aluminumbox
committed
34354bd
fix sequence logic
aluminumbox
committed
1be8229
fix sequence logic
aluminumbox
committed
8fb7b13
fix padding
aluminumbox
committed
94b719f
fix padding
aluminumbox
committed
68b9fcc
fix padding
aluminumbox
committed
6dd2afe

Commits on Jan 9, 2026

Update README with citation and license information
ZhenghuaBao
authored
c313f54
Format LICENSE file for better readability
ZhenghuaBao
authored
e27d2b8

Commits on Jan 12, 2026

add japanese
aluminumbox
committed
bffbee5
add japanese
aluminumbox
committed
59fcae7
add japanese
aluminumbox
committed
0fc3198

Commits on Jan 15, 2026

Update model name from Chroma-4B to Chroma 1.0
ZhenghuaBao
authored
c3d16f5

Commits on Jan 17, 2026

Merge pull request #1250 from raivisdejus/add-latvian-community-model
raivisdejus
authored
d677fc3

Commits on Jan 18, 2026

update higgs audio v2.5 in readme (#169)
iamtonymwt
authored
2b7309e

Commits on Jan 19, 2026

Update README.md
ZhenghuaBao
authored
63115f1
fix fm train bug
aluminumbox
committed
11a14ff
fix fm train bug
aluminumbox
committed
341186c
fix fm train bug
aluminumbox
committed
ed370e8
Fix Technical Report link in README
ZhenghuaBao
authored
6b6a65e
Fix link for Technical Report in README.md
ZhenghuaBao
authored
a56e001
Enhance README with getting started section
ZhenghuaBao
authored
9f4db25
Update Chroma section to link to Voice Agents
ZhenghuaBao
authored
74cfe0e
Update contact information in README
ZhenghuaBao
authored
ab30dab

Commits on Jan 21, 2026

Adding support for hf:// links on CLI (#1252)
raivisdejus
authored
6337b41
Fix epoch updates count logic as drop_last
SWivid
authored
0f0a8c1
Formatting
SWivid
committed
8ba0857
Update README.md
yi
authored
a3b62e4

Commits on Jan 22, 2026

change prepare_csv_wavs from relative path to absolute path and get duration info with soundfile and torchaudio
ZhikangNiu
committed
544889d
fix many tensorboard writer and only log in main_process
ZhikangNiu
committed
b4f4645
doc: Remove outdated playground
ZhenghuaBao
authored
eb92f88
fig: Update model architecture
ZhenghuaBao
committed
6eecadd
feat: compact latest transformers
chentanyu
committed
ac678b1
add tqdm in convert text to pinyin
ZhikangNiu
committed
8a93ac1
update: example
kaishen-Dotc
committed
afcbef3
update: torch requirements
kaishen-Dotc
committed
2cf4bc2
Initial commit
wangxiongts
committed
b048ed9
Merge pull request #1256 from ZhikangNiu/main
SWivid
authored
947c6d5
fix: depandency updates
terrencechentr
committed
db7c5d2
fix: hf speech_tokenizer download bug.
wangxiongts
committed
4ed3a72

Commits on Jan 23, 2026

Update: readme quick start
sailorjs0804
authored
9cc9591
fix and update: fix finetuning/prepare_data.py and update citation
wangxiongts
committed
a432575
update: modify default non_streaming_mode
wangxiongts
committed
838a849
fixup wrong flags
vasqu
committed
9ad399d
fix loading logic
vasqu
committed
4136752

Commits on Jan 24, 2026

Merge pull request #15 from vasqu/fix-fa-flags
wangxiongts
authored
7e23766
update: update readme
wangxiongts
committed
1582846
Optimizing 14GB Model on 4GB VRAM
agentifyanchor
committed
8d7257d

Commits on Jan 25, 2026

Increase default max_duration from 4096 to 65536 in cfm.py (#1260)
SWivid
authored
b4abb3b
update: update readme
wangxiongts
committed
f46b482

Commits on Jan 26, 2026

Commits on Jan 28, 2026

doc: Uploaded Demo
ZhenghuaBao
authored
a5e0269
Update eval README with ctranslate2 installation instructions
SWivid
authored
4794309
online feature
aluminumbox
committed
ee6a5e7
online feature
aluminumbox
committed
da98124
online feature
aluminumbox
committed
adef3a1
Update package versions in requirements.txt
agentifyanchor
authored
1ad4bd1

Commits on Jan 29, 2026

Commits on Jan 30, 2026

Commits on Feb 2, 2026

Update run_asr_wer method in utils_eval.py for compat with jiwer>=4.0.0
SWivid
committed
0c085cb

Commits on Feb 4, 2026

fix: adjust finetune script for stable training.
wangxiongts
committed
a8bdff9
fix ras
aluminumbox
committed
02a279a
fix ras
aluminumbox
committed
fa8fcf1
fix ras
aluminumbox
committed
c1a5b77
Merge pull request #1814 from hexisyztem/main
aluminumbox
authored
9307563
Merge pull request #1814 from hexisyztem/main
aluminumbox
authored
ac752ea
Merge pull request #1814 from hexisyztem/main
aluminumbox
authored
ce24b7c

Commits on Feb 5, 2026

fix: padding bugs in 12Hz tokenizer decode.
wangxiongts
committed
28e2d8b

Commits on Feb 6, 2026

fix: padding value process bug in tokenizer decode
wangxiongts
committed
b369e89

Commits on Feb 11, 2026

keep offline embedding/token extraction for compatibale
aluminumbox
committed
9b3b127
keep offline embedding/token extraction for compatibale
aluminumbox
committed
1a543cf
keep offline embedding/token extraction for compatibale
aluminumbox
committed
935e52b
simply code
aluminumbox
committed
0714299
simply code
aluminumbox
committed
a965211
simply code
aluminumbox
committed
1a82bdd
fix vllm yaml version
aluminumbox
committed
ec05f94
fix vllm yaml version
aluminumbox
committed
4eb32be
fix vllm yaml version
aluminumbox
committed
cba07ad

Commits on Feb 14, 2026

Commits on Feb 16, 2026

Make wandb project/run_name/resume_id configurable via Hydra yaml, backward compatible with defaults

ZhikangNiu
and
claude
committed
7be59ae
Optimize DiT text embedding with batched per-sample seq handling
QingyuLiu0521
committed
c190f69
Apply ruff formatting
QingyuLiu0521
committed
bb09d83
Merge pull request #1266 from ZhikangNiu/main
SWivid
authored
df34adf
Unify seq_len naming in DiT get_input_embed
QingyuLiu0521
committed
2ac4c5c
Merge pull request #1267 from QingyuLiu0521/qyl/pr-dit-only
SWivid
authored
d909862
Bump version from 1.1.15 to 1.1.16
SWivid
authored
4b4e0c5

Commits on Feb 23, 2026

feat:add mmdit flash attn support
ZhikangNiu
committed
68ee421

Commits on Feb 25, 2026

when use flash_attn, log_sample should under autocast context
ZhikangNiu
committed
a9c46bb

Commits on Feb 26, 2026

Merge pull request #1269 from ZhikangNiu/main
SWivid
authored
2a67653

Commits on Mar 4, 2026

Commits on Mar 7, 2026

Remove pydantic<=2.10.6 restriction to fit latest gradio version
SWivid
committed
bd6fee8

Commits on Mar 16, 2026

add cosyvoice3
root
committed
df9a222
add cosyvoice3
root
committed
ed9be05
add cosyvoice3
root
committed
a26f2ec
rename files
root
committed
3fdb5f5
rename files
root
committed
ffabefc
rename files
root
committed
c365957
update results
root
committed
ddcb033
update results
root
committed
58fcff5
update results
root
committed
3b9c7bc
update results
root
committed
b5358f0
update results
root
committed
7027b11
update results
root
committed
6bfbf52
fix lint
yuekaizhang
committed
c700a16
fix lint
yuekaizhang
committed
092ddeb
fix lint
yuekaizhang
committed
d151834
Merge pull request #1850 from yuekaizhang/cosy3_pr
aluminumbox
authored
afbc130
Merge pull request #1850 from yuekaizhang/cosy3_pr
aluminumbox
authored
abcd47b
Merge pull request #1850 from yuekaizhang/cosy3_pr
aluminumbox
authored
893044a
Add Arabic model details to SHARED.md (#1279)
karimouda
authored
c6b603a

Commits on Mar 17, 2026

fix finetuning bug
wangxiongts
authored
e58a068

Commits on Mar 23, 2026

F5TTS v1 Small + LibriTTS training config
ZhikangNiu
committed
06f4fe4
Merge pull request #1280 from ZhikangNiu/main
SWivid
authored
7248a60

Commits on Apr 1, 2026

Init commit
zhu-han
committed
cb5d9fc

Commits on Apr 2, 2026

update paper
zhu-han
committed
8464824
update link
zhu-han
committed
a414406

Commits on Apr 3, 2026

Fix/mps cloning interactive (#13)

Talpik
and
vladimir.talpa
authored
490700f
Fix single gpu fine-tuning (#23)
zhu-han
authored
a07899b

Commits on Apr 4, 2026

Commits on Apr 7, 2026

fix torchaudio backend
zhu-han
committed
4b62ec1
fix text processing for non-verbal and special symbols
zhu-han
committed
06aed41
add no_asr and instruct options to demo
zhu-han
committed
4af4797
relax pytorch version requirements
zhu-han
committed
7d2ce24
Add tips in readme
zhu-han
committed
8db07d6
Merge pull request #61 from zhu-han/fix_20260407
zhu-han
authored
c751d2b

Commits on Apr 9, 2026

fix: instruct in infer_batch
Pastells
committed
3c48642
Merge pull request #72 from Pastells/fix/instruct-batch
zhu-han
authored
b86ff09

Commits on Apr 11, 2026

fix: batch_inference without ref_text or ref_audio_path (#70)
Pastells
authored
b4e72bb

Commits on Apr 13, 2026

docs: add omnivoice-server to community projects (#80)
maemreyo
authored
96efd9c
fix infer_batch.py for mixed modes
zhu-han
committed
cc9af56
use soundfile+librosa instead of torchaudio to avoid issues on some devices
zhu-han
committed
d71a3c0
replace Chinese parentheses with English ones in text
zhu-han
committed
bf2f87b
add more tips in README
zhu-han
committed
faff117
add google colab examples
zhu-han
committed
c696a8f
Bump version to 0.1.4
zhu-han
committed
45463a9
restore community projects
zhu-han
committed
eae5c3a
docs: add omnivoice-rs to community projects (#94)
FerrisMind
authored
0417096
set load_asr=True in colab notebook
zhu-han
committed
8d69b02
remove language_name docs as it is not used by the model
zhu-han
committed
8e9a721

Commits on Apr 14, 2026

add issue templates
zhu-han
committed
6329250
switch to torchaudio resampling for consistency with training
zhu-han
committed
b7fdfec

Commits on Apr 16, 2026

Fixes #1287 #1288
SWivid
committed
02d5a13
Bump version from 1.1.18 to 1.1.19
SWivid
authored
299cb54

Commits on Apr 17, 2026

Commits on Apr 20, 2026

Commits on Apr 22, 2026

Update README
zhu-han
committed
a2fd9d2
Support training with SDPA
zhu-han
committed
bb438ba
update docs
zhu-han
committed
8256496

Commits on Apr 28, 2026

add more tips
zhu-han
committed
5f9033a
Bump version to 0.1.5
zhu-han
committed
9dc27ce

Commits on Apr 30, 2026

Revise disclaimer wording in README
pkufool
authored
c5c015c
Merge pull request #136 from k2-fsa/pkufool-patch-1
pkufool
authored
90da463

Commits on May 6, 2026

docs: Add Auris to community projects (#145)
nikhilprasanth
authored
b3bbbdd

Commits on May 12, 2026

fix: path traversal in finetune Gradio handlers (closes #1293)
AAtomical
committed
0da0d2f

Commits on May 13, 2026

Merge pull request #1294 from AAtomical/fix/path-traversal-finetune-gradio
SWivid
authored
80b19f1

Commits on May 18, 2026

Commits on May 19, 2026

Add Intel XPU (Intel Arc GPU) inference support (#55)
smashingtags
authored
c1667b3

Commits on May 25, 2026

Commits on May 28, 2026

Add OmniVoice-MLX to community projects (#165)

ailuntx
and
ailuntz
authored
071ce7e
Add tts-audiobook-tool project to community list (#167)

zeropointnine
and
zhu-han
authored
8b40215
Update docs
zhu-han
committed
3ce000a

Commits on Jun 5, 2026

docs: point README to Higgs Audio v3 (HF + API), archive v2/v2.5 guide (#180)

SilinMeng0510
and
claude
authored
1330a67

Commits on Jun 11, 2026

Fix END_PUNCTUATION dropping members due to unescaped triple quote (#181) (#182)

MinDongRyul
and
macos
authored
734d99a

Commits on Jun 24, 2026

Add LA Studio to community projects (#195)
dduongtrandai
authored
451f23d

Commits on Jul 3, 2026

Expose pad_duration and fade_duration to let users control fade-in/out and silence padding
zhu-han
committed
f77501c

Commits on Jul 5, 2026

v1.1.21 gradio>=6.15.0
SWivid
committed
9ec2124

Commits on Jul 6, 2026

Add CI workflows (PyPI publish, pre-commit) and format codebase with ruff
zhu-han
committed
4ab2c75
Bump version to 0.2.0
zhu-han
committed
996f73b

Commits on Jul 9, 2026

Add audio.cpp to community projects (#209)
0xShug0
authored
c33fd18
Update formatting rules
zhu-han
committed
d037d48

Commits on Jul 11, 2026

Fix flex_attention silently running fp32 under bf16 mixed precision (#212)
shuaills
authored
2876c6e

Commits on Jul 13, 2026

Honor asr_model_name when the ASR model is loaded lazily (#221)
HenryVarro666
authored
8102b02

Commits on Jul 16, 2026

Add stale issue workflow (#228)
zhu-han
authored
0bbb883

Commits on Jul 18, 2026

docs: add LocalText2Voice to community projects (#230)
estebanstifli
authored
6cb3544

Commits on Jul 23, 2026

fix: allocate fix_duration across chunks to avoid N× duration blow-up
foco
committed
d1f810e
Merge pull request #1309 from Wing900/fix/fix-duration-chunk-allocation
SWivid
authored
2606bcc
v1.1.22
SWivid
authored
b71ccc5

Commits on Jul 30, 2026

feat: FlashInfer-accelerated inference (2-2.9x lossless speedup) (#239)

authored
17988dc

Commits on Aug 5, 2026

feat: add optional LoRA finetuning support via PEFT (#243)
kawshikbuet17
authored
f94ba6a

Commits on Aug 19, 2026

Add voicestudio/models/__init__.py registry, refresh status tracking table
b-re-w
committed
e1570e8
Graft upstream git history for dia
b-re-w
committed
f432c86
Graft upstream git history for higgs_audio_v2
b-re-w
committed
22e81ab
Graft upstream git history for parler_tts
b-re-w
committed
cabef6e
Graft upstream git history for qwen3_tts
b-re-w
committed
0494cf0
Graft upstream git history for f5_tts
b-re-w
committed
240e6fb
Graft upstream git history for cosyvoice_v1
b-re-w
committed
55014de
Graft upstream git history for cosyvoice_v2
b-re-w
committed
a5b5cfa
Graft upstream git history for cosyvoice_v3
b-re-w
committed
129bd89
Graft upstream git history for omnivoice
b-re-w
committed
2ec0592
Graft upstream git history for spark_tts
b-re-w
committed
2916707
Graft upstream git history for chroma
b-re-w
committed
bb6aa27
Graft upstream git history for prompt_tts_pp
b-re-w
committed
d426a5c