Commits

Breadcrumbs

History for

onpatch-1

User selector

All users

All time

Commit History

Commits on Jun 21, 2024

• [LoRA] get rid of the legacy lora remnants and make our codebase lighter (#8623)

  Show description for 8eb1731
  sayakpaul

  authored
  8eb1731

  Copy full SHA for 8eb1731

• a few fix for shard checkpoints (#8656)

  Show description for c71c19c
  

  YiYi Xu

  authored
  c71c19c

  Copy full SHA for c71c19c

Commits on Jun 20, 2024

• [LoRA SD3] add support for lora fusion in sd3 (#8616)

  Show description for 668e34c
  sayakpaul

  authored
  668e34c

  Copy full SHA for 668e34c

Commits on Jun 19, 2024

• Support SD3 ControlNet and Multi-ControlNet. (#8566)

  Show description for e5564d4
  

  wangqixun

  and

  haofanwang

  authored
  e5564d4

  Copy full SHA for e5564d4

Commits on Jun 18, 2024

• Fix gradient checkpointing issue for Stable Diffusion 3 (#8542)

  Show description for 3376252
  

  3 peopleauthored
  3376252

  Copy full SHA for 3376252

• add sd1.5 compatibility to controlnet-xs and fix unused_parameters error during training (#8606)

  Show description for 16170c6
  SamMaoYS

  authored
  16170c6

  Copy full SHA for 16170c6

• self.upsample = Upsample1D (#8580)

  Show description for 4408047
  kkj15dk

  authored
  4408047

  Copy full SHA for 4408047

• [Core] Add `shift_factor` to SD3 tiny autoencoder (#8618)

  Show description for cd30820
  sayakpaul

  authored
  cd30820

  Copy full SHA for cd30820

• Fix sharding when no device_map is passed (#8531)

  Show description for 96399c3
  

  SunMarc

  and

  sayakpaul

  authored
  96399c3

  Copy full SHA for 96399c3

Commits on Jun 13, 2024

• Fix duplicate variable assignments in SD3's JointAttnProcessor (#8516)

  Show description for 896fb6d
  Prgckwb

  authored
  896fb6d

  Copy full SHA for 896fb6d

Commits on Jun 12, 2024

• fix warning log for Transformer SD3 (#8496)

  Show description for 6cf0be5
  sayakpaul

  authored
  6cf0be5

  Copy full SHA for 6cf0be5

• fix dual transformer2d import (#8491)

  Show description for ec068f9
  sayakpaul

  authored
  ec068f9

  Copy full SHA for ec068f9

• Add Stable Diffusion 3 (#8483)

  Show description for 04717fd
  

  3 peopleauthored
  04717fd

  Copy full SHA for 04717fd

• 🤫 Quiet IP Adapter Mask Warning (#8475)

  Show description for 1066fe4
  

  ghunkins

  and

  sayakpaul

  authored
  1066fe4

  Copy full SHA for 1066fe4

• change max_shard_size to 10GB (#8445)

  Show description for d38f69e
  

  sayakpaul

  and

  Wauplin

  authored
  d38f69e

  Copy full SHA for d38f69e

Commits on Jun 10, 2024

• 🔧 chore: use modeling_outputs.Transformer2DModelOutput (#8436)

  Show description for 1d9a6a8
  

  Jannchie

  and

  sayakpaul

  authored
  1d9a6a8

  Copy full SHA for 1d9a6a8

Commits on Jun 7, 2024

• [Core] support saving and loading of sharded checkpoints (#7830)

  Show description for 7d88711
  

  sayakpaul

  and

  Wauplin

  authored
  7d88711

  Copy full SHA for 7d88711

Commits on Jun 6, 2024

• [Core] fix: legacy model mapping (#8416)

  Show description for a3faf3f
  sayakpaul

  authored
  a3faf3f

  Copy full SHA for a3faf3f

Commits on Jun 5, 2024

• [LoRA] Remove legacy LoRA code and related adjustments (#8316)

  Show description for a0542c1
  sayakpaul

  authored
  a0542c1

  Copy full SHA for a0542c1

• [Hunyuan] feat: support chunked ff. (#8397)

  Show description for a8ad666
  sayakpaul

  authored
  a8ad666

  Copy full SHA for a8ad666

• [Hunyuan DiT] feat: enable fusing qkv projections when doing attention (#8396)

  Show description for 14f7b54
  sayakpaul

  authored
  14f7b54

  Copy full SHA for 14f7b54

Commits on Jun 4, 2024

• [Transformer2DModel] Handle `norm_type` safely while remapping (#8370)

  Show description for 6ddbf62
  

  sayakpaul

  and

  DN6

  authored
  6ddbf62

  Copy full SHA for 6ddbf62

• Fix AsymmetricAutoencoderKL forward (#8378)

  townwish4git

  authored
  6be43bd

  Copy full SHA for 6be43bd

Commits on Jun 1, 2024

• Tencent Hunyuan Team: add HunyuanDiT related updates (#8240)

  Show description for 4136044
  

  3 peopleauthored
  4136044

  Copy full SHA for 4136044

Commits on May 31, 2024

• [Core] Introduce class variants for `Transformer2DModel` (#7647)

  Show description for 983dec3
  sayakpaul

  authored
  983dec3

  Copy full SHA for 983dec3

Commits on May 29, 2024

• Fix Copying Mechanism typo/bug (#8232)

  Show description for a2ecce2
  tolgacangoz

  authored
  a2ecce2

  Copy full SHA for a2ecce2

• [Core] Refactor `IPAdapterPlusImageProjection` a bit (#7994)

  Show description for ba1bfac
  sayakpaul

  authored
  ba1bfac

  Copy full SHA for ba1bfac

• move `vqmodel` to `models.autoencoders`. (#8292)

  Show description for 5edd0b3
  sayakpaul

  authored
  5edd0b3

  Copy full SHA for 5edd0b3

• [Post release 0.28.0] remove deprecated blocks. (#8291)

  Show description for 3a28e36
  sayakpaul

  authored
  3a28e36

  Copy full SHA for 3a28e36

Commits on May 20, 2024

• Make VAE compatible to torch.compile() (#7984)

  Show description for 6529ee6
  

  rootonchair

  and

  yiyixuxu

  authored
  6529ee6

  Copy full SHA for 6529ee6

Commits on May 16, 2024

• refactor: Refactored code by Merging `isinstance` calls (#7710)

  Show description for 2afea72
  

  3 peopleauthored
  2afea72

  Copy full SHA for 2afea72

Commits on May 15, 2024

• Adding VQGAN Training script (#5483)

  Show description for d27e996
  

  3 peopleauthored
  d27e996

  Copy full SHA for d27e996

Commits on May 14, 2024

• [Core] separate the loading utilities in modeling similar to pipelines. (#7943)

  Show description for e0e8c58
  sayakpaul

  authored
  e0e8c58

  Copy full SHA for e0e8c58

Commits on May 13, 2024

• fix AnimateDiff creation with a unet loaded with IP Adapter (#7791)

  Show description for 44aa9e5
  

  fabiorigano

  and

  DN6

  authored
  44aa9e5

  Copy full SHA for 44aa9e5

Commits on May 10, 2024

• #7535 Update FloatTensor type hints to Tensor (#7883)

  Show description for be4afa0
  vanakema

  authored
  be4afa0

  Copy full SHA for be4afa0