Commits

Branch selector

 main

User selector

All users

Datepicker

All time

Commit History

Commits on Nov 18, 2025

• fix svd init memory (#208)

  Glaceon-Hyy

  authored
  d82861c

  Copy full SHA for d82861c

• set module device to skip weight init (#207)

  Show description for 93c34f3
  qzzz95

  authored
  93c34f3

  Copy full SHA for 93c34f3

Commits on Nov 14, 2025

• support load lora from state dict (#206)

  qzzz95

  authored
  f24e582

  Copy full SHA for f24e582

• Fix circular dependence (#205)

  Show description for e3cf908
  qzzz95

  authored
  e3cf908

  Copy full SHA for e3cf908

Commits on Nov 12, 2025

• support svd quant (#202)

  Show description for ae4faeb
  Glaceon-Hyy

  authored
  ae4faeb

  Copy full SHA for ae4faeb

• auto enable vsa (#203)

  akaitsuki-ii

  authored
  f7119c8

  Copy full SHA for f7119c8

Commits on Nov 7, 2025

• Enable aiter attention for rocm (#198)

  Show description for 6b2825e
  guangzlu

  authored
  6b2825e

  Copy full SHA for 6b2825e

Commits on Nov 6, 2025

• Fix/import vsa (#200)

  Show description for 1f969ae
  akaitsuki-ii

  authored
  1f969ae

  Copy full SHA for 1f969ae

Commits on Nov 5, 2025

• Fix/qwen image (#197)

  Show description for 0590444
  akaitsuki-ii

  authored
  0590444

  Copy full SHA for 0590444

• video sparse attention (#190)

  Show description for 4ae8f2c
  akaitsuki-ii

  authored
  4ae8f2c

  Copy full SHA for 4ae8f2c

Commits on Nov 4, 2025

• load encoder optional (#196)

  Show description for 5e86c87
  

  qzzz95

  and

  gemini-code-assist[bot]

  authored
  5e86c87

  Copy full SHA for 5e86c87

Commits on Nov 3, 2025

• fix mask dtype differ from latent (#195)

  Show description for 90c73be
  

  qzzz95

  and

  Glaceon-Hyy

  authored
  90c73be

  Copy full SHA for 90c73be

Commits on Oct 30, 2025

• Fix Wan2.2 low noise model load LoRA bug (#188)

  continue-revolution

  authored
  96f5d4e

  Copy full SHA for 96f5d4e

Commits on Oct 29, 2025

• define qwen image edit system prompt (#194)

  Show description for 88cd350
  

  qzzz95

  and

  Glaceon-Hyy

  authored
  88cd350

  Copy full SHA for 88cd350

Commits on Oct 28, 2025

• fix redux multiple ref images (#192)

  qzzz95

  authored
  5b91139

  Copy full SHA for 5b91139

• fix compile repeated blocks (#191)

  akaitsuki-ii

  authored
  a1133d6

  Copy full SHA for a1133d6

• torch.compile with dynamic=False (#185)

  Show description for 6f8073c
  

  akaitsuki-ii

  and

  Glaceon-Hyy

  authored
  6f8073c

  Copy full SHA for 6f8073c

• enable FSDP for qwen vl (#184)

  Show description for f925211
  akaitsuki-ii

  authored
  f925211

  Copy full SHA for f925211

Commits on Oct 27, 2025

• convert diffusers lora key (#189)

  Show description for d7032d0
  

  qzzz95

  and

  Glaceon-Hyy

  authored
  d7032d0

  Copy full SHA for d7032d0

Commits on Oct 22, 2025

• supports sequence parallel and use custom image size for Qwen Image (#186)

  Show description for d0763b2
  akaitsuki-ii

  authored
  d0763b2

  Copy full SHA for d0763b2

Commits on Oct 11, 2025

• fix key convert for lora kohya lora (#183)

  Show description for ece6fec
  qzzz95

  authored
  ece6fec

  Copy full SHA for ece6fec

Commits on Sep 30, 2025

• Feature/qwen edit plus (#180)

  Show description for e9521cf
  

  Glaceon-Hyy

  and

  akaitsuki-ii

  authored
  e9521cf

  Copy full SHA for e9521cf

Commits on Sep 28, 2025

• Feature/qwen image control (#176)

  Show description for 0acf4cb
  Glaceon-Hyy

  authored
  0acf4cb

  Copy full SHA for 0acf4cb

Commits on Sep 26, 2025

• supports flux kontext with multiple input images (#173)

  akaitsuki-ii

  authored
  33d00d0

  Copy full SHA for 33d00d0

Commits on Sep 22, 2025

• bug fix

  

  tenderness-git

  authored and

  sir1st-inc

  committed
  9821529

  Copy full SHA for 9821529

• use utf8 (#178)

  tenderness-git

  authored
  29967d6

  Copy full SHA for 29967d6

Commits on Sep 19, 2025

• supports flash attn 3 fp8 (#174)

  akaitsuki-ii

  authored
  28338b9

  Copy full SHA for 28338b9

Commits on Sep 18, 2025

• ADD Option: OFFLINE fetch modelscope model (#172)

  Show description for 39cc5c5
  

  qzzz95

  and

  gemini-code-assist[bot]

  authored
  39cc5c5

  Copy full SHA for 39cc5c5

Commits on Sep 17, 2025

• progress

  

  tenderness-git

  authored and

  sir1st-inc

  committed
  ccf97d8

  Copy full SHA for ccf97d8

Commits on Sep 16, 2025

• ADD update_weights for flux and qwen_image (#168)

  Show description for 7f7d72d
  

  qzzz95

  and

  gemini-code-assist[bot]

  authored
  7f7d72d

  Copy full SHA for 7f7d72d

• fix wan umt5 state dict converter (#170)

  akaitsuki-ii

  authored
  e068d10

  Copy full SHA for e068d10

Commits on Sep 11, 2025

• init all dit module with device and dtype for speed up (#164)

  qzzz95

  authored
  63188ac

  Copy full SHA for 63188ac

Commits on Sep 9, 2025

• Wan Speech2Video (#162)

  Show description for ca8a9a5
  continue-revolution

  authored
  ca8a9a5

  Copy full SHA for ca8a9a5

Commits on Sep 4, 2025

• no fa3 with attention mask (#163)

  Show description for b74ba8d
  

  akaitsuki-ii

  and

  qufei.qf

  authored
  b74ba8d

  Copy full SHA for b74ba8d

Commits on Sep 1, 2025

• remove redundant empty_cache in parallel forward (#161)

  akaitsuki-ii

  authored
  665f74a

  Copy full SHA for 665f74a

Pagination

Previous

Next