mi804
8d8bfc7f54
minor fix
2026-02-25 19:04:10 +08:00
mi804
8e15dcd289
support ltx2 train -2
2026-02-25 18:06:02 +08:00
mi804
586ac9d8a6
support ltx-2 training
2026-02-25 17:19:57 +08:00
Hong Zhang
625b5ff16d
Apply suggestion from @gemini-code-assist[bot]
...
Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>
2026-02-24 15:26:49 +08:00
mi804
ee73a29885
qwen_image layercontrol v2
2026-02-24 15:19:16 +08:00
Zhongjie Duan
288bbc7128
Merge pull request #1299 from modelscope/firered
...
support FireRed
2026-02-15 14:18:13 +08:00
Zhongjie Duan
5002ac74dc
Update Qwen-Image.md
2026-02-15 14:15:44 +08:00
Zhongjie Duan
863a6ba597
Merge branch 'main' into firered
2026-02-15 14:12:44 +08:00
Artiprocher
b08bc1470d
support firered
2026-02-15 14:02:50 +08:00
feng0w0
96143aa26b
Merge branch 'npu_fused' of https://github.com/Feng0w0/DiffSynth-Studio into npu_fused
2026-02-13 10:06:39 +08:00
feng0w0
71cea4371c
[doc][NPU]Documentation on modifications, NPU environment installation, and additional parameter
2026-02-13 09:58:27 +08:00
Mr_Dwj
fc11fd4297
chore: remove invalid comment code
2026-02-13 09:38:14 +08:00
Mr_Dwj
bd3c5822a1
fix: WanVAE2.2 decode error
2026-02-13 01:13:08 +08:00
Mr_Dwj
96fb0f3afe
fix: unpack Resample38 output
2026-02-12 23:51:56 +08:00
Mr-Neutr0n
b68663426f
fix: preserve sign of denominator in clamp to avoid inverting gradient direction
...
The previous .clamp(min=1e-6) on (sigma_ - sigma) flips the sign when
the denominator is negative (which is the typical case since sigmas
decrease monotonically). This would invert the target and cause
training divergence.
Use torch.sign(denom) * torch.clamp(denom.abs(), min=1e-6) instead,
which prevents division by zero while preserving the correct sign.
2026-02-11 21:04:55 +05:30
Mr-Neutr0n
0e6976a0ae
fix: prevent division by zero in trajectory imitation loss at last step
2026-02-11 19:51:25 +05:30
Hong Zhang
94b57e9677
Fix readthedocs rendering ( #1290 )
...
* test latex
* test latex
* fix conf
2026-02-11 11:32:27 +08:00
Hong Zhang
3fb037d33a
Correct hyperlinks for docs
2026-02-10 20:59:47 +08:00
Hong Zhang
b3b63fef3e
Add readthedocs for diffsynth-studio
...
* add conf docs
* add conf docs
* add index
* add index
* update ref
* test root
* add en
* test relative
* redirect relative
* add document
* test_document
* test_document
2026-02-10 19:51:04 +08:00
Zhongjie Duan
f6d85f3c2e
Merge pull request #1282 from mi804/ltx-2
...
add inference script for ltx-2 lora
2026-02-10 15:13:06 +08:00
mi804
2f22e598b7
fix load lora
2026-02-10 15:06:04 +08:00
Hong Zhang
888caf8b88
Update README_zh.md
...
Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>
2026-02-10 14:48:19 +08:00
mi804
b6e39c97af
add inference script for ltx-2 lora
2026-02-10 14:32:30 +08:00
Zhongjie Duan
02124c4034
Merge pull request #1280 from modelscope/issue-fix
...
fix mix-precision issues in low-version torch
2026-02-10 11:14:12 +08:00
Artiprocher
fddc98ff16
fix mix-precision issues in low-version torch
2026-02-10 11:12:50 +08:00
Zhongjie Duan
0dfcd25cf3
Merge pull request #1278 from modelscope/issue-fix
...
update lora loading in docs
2026-02-10 10:50:18 +08:00
Artiprocher
ff10fde47f
update lora loading in docs
2026-02-10 10:48:44 +08:00
Zhongjie Duan
dc94614c80
Merge pull request #1256 from Feng0w0/npu_fused
...
[model][NPU]:Add NPU fusion operator patch to Zimage model to improve performance
2026-02-09 20:08:44 +08:00
feng0w0
e56a4d5730
[model][NPU]:Add NPU fusion operator patch to Zimage model to improve performance
2026-02-09 12:31:34 +08:00
feng0w0
3f8468893a
[model][NPU]:Add NPU fusion operator patch to Zimage model to improve performance
2026-02-09 09:51:06 +08:00
Mr-Neutr0n
6383ec358c
Fix AttributeError when pipe.dit is None
...
When using split training with 'sft:data_process' task, the DiT model
is not loaded but the attribute 'dit' exists with value None. The
existing hasattr check returns True but then accessing siglip_embedder
fails.
Add an explicit None check before accessing pipe.dit.siglip_embedder.
Fixes #1246
2026-02-07 05:23:11 +05:30
Zhongjie Duan
1b47e1dc22
Merge pull request #1272 from modelscope/zero3-fix
...
Support DeepSpeed ZeRO 3
2026-02-06 16:33:12 +08:00
Artiprocher
b0bf78e915
refine code & doc
2026-02-06 16:27:23 +08:00
Zhongjie Duan
abdf66d09e
Merge pull request #1265 from lzws/main
...
fix wanS2V bug and update readme
2026-02-06 10:22:48 +08:00
lzws
27b1fe240b
add examples
2026-02-05 17:17:10 +08:00
lzws
1635897516
update readme
2026-02-05 16:56:39 +08:00
lzws
8d172127cd
fix wans2v bug and update readme
2026-02-05 16:52:38 +08:00
feng0w0
fccb1ecdd7
Initialize qwen-image on CPU
2026-02-05 11:54:36 +08:00
Zhongjie Duan
c0f7e1db7c
Merge pull request #1261 from modelscope/examples-update
...
update examples
2026-02-05 11:11:35 +08:00
Artiprocher
53890bafa4
update examples
2026-02-05 11:10:55 +08:00
feng0w0
6886f7ba35
fix wan decoder bug
2026-02-05 10:31:41 +08:00
Zhongjie Duan
afd48cd706
Merge pull request #1259 from mi804/multi_controlnet
...
add example for multiple controlnet
2026-02-04 17:04:11 +08:00
mi804
24b68c2392
add example for multiple controlnet
2026-02-04 16:52:39 +08:00
Zhongjie Duan
280ff7cca6
Merge pull request #1229 from Feng0w0/wan_rope
...
[bugfix][NPU]:Fix bug that correctly obtains device type
2026-02-04 13:26:00 +08:00
Zhongjie Duan
b4b62e2f7c
Merge pull request #1221 from Feng0w0/usp_npu
...
[NPU]:Support USP feature in NPU
2026-02-04 13:25:24 +08:00
feng0w0
051b957adb
[model][NPU] Add NPU fusion operator patch to Zimage model to improve performance
2026-02-03 19:50:21 +08:00
feng0w0
ca9b5e64ea
[feature]:Add adaptation of all models to zero3
2026-02-03 15:44:53 +08:00
Zhongjie Duan
6d1be405b9
Merge pull request #1242 from mi804/ltx-2
...
LTX-2
2026-02-03 13:07:41 +08:00
Zhongjie Duan
25c3a3d3e2
Merge branch 'main' into ltx-2
2026-02-03 13:06:44 +08:00
mi804
49bc84f78e
add comment for tuple noise_pred
2026-02-03 10:43:25 +08:00
