Commits

Breadcrumbs

History for

onmain

User selector

All users

All time

Commit History

Commits on May 30, 2023

• Fix temb attention (#3607)

  Show description for c0f867a
  patrickvonplaten

  authored
  c0f867a

  Copy full SHA for c0f867a

Commits on May 29, 2023

• remove print statements from attention processor. (#3592)

  sayakpaul

  authored
  c6ae883

  Copy full SHA for c6ae883

Commits on May 26, 2023

• [docs] Add AttnProcessor to docs (#3474)

  Show description for 7948db8
  stevhliu

  authored
  7948db8

  Copy full SHA for 7948db8

• renamed variable to input_ and output_ (#3507)

  Show description for ffa33d6
  vikasmech

  authored
  ffa33d6

  Copy full SHA for ffa33d6

• fix broken change for vq pipeline (#3563)

  Show description for e5215de
  

  yiyixuxu

  and

  yiyixuxu

  authored
  e5215de

  Copy full SHA for e5215de

Commits on May 25, 2023

• Add Kandinsky 2.1 (#3308)

  Show description for 03b7a84
  

  6 peopleauthored
  03b7a84

  Copy full SHA for 03b7a84

Commits on May 24, 2023

• explicit broadcasts for assignments (#3535)

  williamberman

  authored
  db56f8a

  Copy full SHA for db56f8a

Commits on May 23, 2023

• fix attention mask pad check (#3531)

  williamberman

  authored
  c13dbd5

  Copy full SHA for c13dbd5

Commits on May 22, 2023

• Support for cross-attention bias / mask (#2634)

  Show description for 64bf5d3
  Birch-san

  authored
  64bf5d3

  Copy full SHA for 64bf5d3

• feat: allow disk offload for diffuser models (#3285)

  Show description for f3d570c
  

  40 peopleauthored
  f3d570c

  Copy full SHA for f3d570c

Commits on May 21, 2023

• [Attention processor] Better warning message when shifting to `AttnProcessor2_0` (#3457)

  Show description for 4bbc51d
  sayakpaul

  authored
  4bbc51d

  Copy full SHA for 4bbc51d

Commits on May 19, 2023

• [{Up,Down}sample1d] explicit view kernel size as number elements in flattened indices (#3479)

  Show description for 85eff63
  williamberman

  authored
  85eff63

  Copy full SHA for 85eff63

Commits on May 17, 2023

• Update full dreambooth script to work with IF (#3425)

  williamberman

  authored
  c9f939b

  Copy full SHA for c9f939b

• Reduce peak VRAM by releasing large attention tensors (as soon as they're unnecessary) (#3463)

  Show description for bd78f63
  cmdr2

  authored
  bd78f63

  Copy full SHA for bd78f63

• Fix gradient checkpointing bugs in freezing part of models (requires_grad=False) (#3404)

  Show description for 15f1bab
  

  IrisRainbowNeko

  and

  patrickvonplaten

  authored
  15f1bab

  Copy full SHA for 15f1bab

Commits on May 16, 2023

• fix tiled vae blend extent range (#3384)

  Show description for 92ea5ba
  superlabs-dev

  authored
  92ea5ba

  Copy full SHA for 92ea5ba

• [Docs] Fix incomplete docstring for resnet.py (#3438)

  Show description for 754fac8
  Laurent2916

  authored
  754fac8

  Copy full SHA for 754fac8

Commits on May 12, 2023

• attention refactor: the trilogy (#3387)

  Show description for 909742d
  williamberman

  authored
  909742d

  Copy full SHA for 909742d

Commits on May 11, 2023

• Support ControlNet v1.1 shuffle properly (#3340)

  Show description for 01c056f
  

  takuma104

  and

  patrickvonplaten

  authored
  01c056f

  Copy full SHA for 01c056f

Commits on May 10, 2023

• add: a warning message when using xformers in a PT 2.0 env. (#3365)

  Show description for 94a0c64
  

  sayakpaul

  and

  patrickvonplaten

  authored
  94a0c64

  Copy full SHA for 94a0c64

Commits on May 9, 2023

• [docs] Improve safetensors docstring (#3368)

  Show description for 26832aa
  stevhliu

  authored
  26832aa

  Copy full SHA for 26832aa

• if dreambooth lora (#3360)

  Show description for a757b2d
  williamberman

  authored
  a757b2d

  Copy full SHA for a757b2d

Commits on May 5, 2023

• Add upsample_size to AttnUpBlock2D, AttnDownBlock2D (#3275)

  Show description for 36f43ea
  

  Will Rice

  authored
  36f43ea

  Copy full SHA for 36f43ea

Commits on May 2, 2023

• [Torch 2.0 compile] Fix more torch compile breaks (#3313)

  Show description for 5c7a35a
  

  patrickvonplaten

  and

  Chillee

  authored
  5c7a35a

  Copy full SHA for 5c7a35a

Commits on May 1, 2023

• Torch compile graph fix (#3286)

  Show description for 0e82fb1
  patrickvonplaten

  authored
  0e82fb1

  Copy full SHA for 0e82fb1

Commits on Apr 28, 2023

• Allow disabling torch 2_0 attention (#3273)

  Show description for 4d35d7f
  patrickvonplaten

  authored
  4d35d7f

  Copy full SHA for 4d35d7f

Commits on Apr 26, 2023

• Allow fp16 attn for x4 upscaler (#3239)

  Show description for abbf3c1
  patrickvonplaten

  authored
  abbf3c1

  Copy full SHA for abbf3c1

Commits on Apr 25, 2023

• add model (#3230)

  Show description for e51f19a
  

  patrickvonplaten

  and

  williamberman

  authored
  e51f19a

  Copy full SHA for e51f19a

Commits on Apr 24, 2023

• [Bug fix] Fix batch size attention head size mismatch (#3214)

  patrickvonplaten

  authored
  c5933c9

  Copy full SHA for c5933c9

Commits on Apr 22, 2023

• Make sure VAE attention works with Torch 2_0 (#3200)

  Show description for 425192f
  patrickvonplaten

  authored
  425192f

  Copy full SHA for 425192f

Commits on Apr 21, 2023

• make `from_flax` work for controlnet (#3161)

  Show description for bc0392a
  

  yiyixuxu

  and

  patrickvonplaten

  authored
  bc0392a

  Copy full SHA for bc0392a

Commits on Apr 20, 2023

• adding custom diffusion training to diffusers examples (#3031)

  Show description for 3979aac
  

  5 peopleauthored
  3979aac

  Copy full SHA for 3979aac

Commits on Apr 19, 2023

• Correct `Transformer2DModel.forward` docstring (#3074)

  Show description for c8fdfe4
  offchan42

  authored
  c8fdfe4

  Copy full SHA for c8fdfe4

Commits on Apr 18, 2023

• class labels timestep embeddings projection dtype cast (#3137)

  Show description for fc18839
  williamberman

  authored
  fc18839

  Copy full SHA for fc18839

• Add unet act fn to other model components (#3136)

  Show description for f0c74e9
  williamberman

  authored
  f0c74e9

  Copy full SHA for f0c74e9