Commits

Commits on Dec 28, 2025

refactor: simplify logic for saving results (#1149)
leejet
authored
4ff2c8c
Copy full SHA for 4ff2c8c
chore: reformat named cache params description into single line
leejet
committed
51bd9c8
Copy full SHA for 51bd9c8
feat: support mmap for model loading (#1059)
wbruna
authored
d0d836a
Copy full SHA for d0d836a

Commits on Dec 27, 2025

refactor: move pmid condition logic into get_pmid_condition (#1148)
leejet
authored
a2d83dd
Copy full SHA for a2d83dd
fix: consistently pass 2nd-order samplers half steps as negatives (#1095)
wbruna
authored
cc10771
Copy full SHA for cc10771
fix: handle redirected UTF-8 output correctly on Windows (#1147)
leejet
authored
37c9860
Copy full SHA for 37c9860

Commits on Dec 26, 2025

feat: add __index_timestep_zero__ support (#1146)
leejet
authored
ccb6b0a
Copy full SHA for ccb6b0a
feat: add png sequence output for vid_gen (#1117)
CarlGao4
authored
df4efe2
Copy full SHA for df4efe2

Commits on Dec 24, 2025

fix: avoid crash when using taesd for preview only (#1141)
leejet
authored
860a78e
Copy full SHA for 860a78e
feat: add support for qwen image edit 2511 (#1096)
leejet
authored
a0adcfb
Copy full SHA for a0adcfb
feat: add support for more underline loras (#1135)
leejet
authored
3d5fdd7
Copy full SHA for 3d5fdd7
chore: use Ninja on Windows to speed up build process (#1120)
CarlGao4
authored
3e6c428
Copy full SHA for 3e6c428
feat: add --serve-html-path option to example server (#1123)
Liyulingyue
authored
96fcb13
Copy full SHA for 96fcb13

Commits on Dec 23, 2025

fix: correct ggml_pad_ext (#1133)
leejet
authored
3e81246
Copy full SHA for 3e81246

Commits on Dec 22, 2025

docs: update README.md
leejet
committed
98916e8
Copy full SHA for 98916e8
feat: add more caching methods (#1066)
rmatif
authored
298b110
Copy full SHA for 298b110

Commits on Dec 21, 2025

fix: add the missing }
leejet
committed
30a9113
Copy full SHA for 30a9113
fix: correct the parsing of --convert-name opotion
leejet
committed
c6937ba
Copy full SHA for c6937ba
feat: do not convert tensor names by default in convert mode (#1122)
leejet
authored
ca5b196
Copy full SHA for ca5b196
feat: add seamless texture generation support (#914)
Show description for 50ff966

authored
50ff966
Copy full SHA for 50ff966
feat: add scale_rope support (#1121)
leejet
authored
88ec9d3
Copy full SHA for 88ec9d3
feat: select vulkan device with env variable (#629)
stduhpf
authored
60abda5
Copy full SHA for 60abda5

Commits on Dec 19, 2025

feat: add support for Chroma Radiance x0 (#1091)
Show description for 23fce0b

stduhpf
and
leejet
authored
23fce0b
Copy full SHA for 23fce0b
chore: give feedback about cfg values smaller than 1 (#1088)
wbruna
authored
7c88c47
Copy full SHA for 7c88c47
docs: document usage of tae for VRAM reduction using wan (#1108)
CarlGao4
authored
1f77545
Copy full SHA for 1f77545

Commits on Dec 18, 2025

feat: add support for underline style lora of flux (#1103)
Show description for 8e9f3a4
leejet
authored
8e9f3a4
Copy full SHA for 8e9f3a4
feat: default to LCM scheduler for LCM sampling (#1109)
Show description for 78e15bd

wbruna
and
leejet
authored
78e15bd
Copy full SHA for 78e15bd
feat: add KL Optimal scheduler (#1098)
daniandtheweb
authored
97cf2ef
Copy full SHA for 97cf2ef

Commits on Dec 17, 2025

chore: remove unused debug code
leejet
committed
bda7fab
Copy full SHA for bda7fab
fix: make flash attn work with high noise diffusion model (#1111)
leejet
authored
c2e18c8
Copy full SHA for c2e18c8

Commits on Dec 16, 2025

refactor: optimize the printing of version log (#1102)
leejet
authored
c3ad6a1
Copy full SHA for c3ad6a1
feat: supports correct UTF-8 printing on windows (#1101)
leejet
authored
ebe9d26
Copy full SHA for ebe9d26
feat: add taehv support for Wan/Qwen (#937)
stduhpf
authored
9fa7f41
Copy full SHA for 9fa7f41
fix: added a clean exit in ModelLoader::load_tensors if OOM (#1097)
akleine
authored
a23262d
Copy full SHA for a23262d
chore: remove lora_model_dir parameter (#1100)
wbruna
authored
e687913
Copy full SHA for e687913