Commits

Breadcrumbs

History for

onmain

User selector

All users

All time

Commit History

Commits on Dec 18, 2022

• kakaobrain unCLIP (#1428)

  Show description for 2dcf64b
  

  williamberman

  and

  patrickvonplaten

  authored
  2dcf64b

  Copy full SHA for 2dcf64b

Commits on Dec 13, 2022

• Make sure all pipelines can run with batched input (#1669)

  Show description for b345c74
  

  patrickvonplaten

  and

  pcuenca

  authored
  b345c74

  Copy full SHA for b345c74

Commits on Dec 9, 2022

• do not automatically enable xformers (#1640)

  Show description for 6b68afd
  patrickvonplaten

  authored
  6b68afd

  Copy full SHA for 6b68afd

Commits on Dec 7, 2022

• Make cross-attention check more robust (#1560)

  Show description for 5e03692
  pcuenca

  authored
  5e03692

  Copy full SHA for 5e03692

• fix upcast in slice attention (#1591)

  Show description for ced7c96
  patil-suraj

  authored
  ced7c96

  Copy full SHA for ced7c96

• [UNet2DConditionModel] add an option to upcast attention to fp32 (#1590)

  Show description for 170ebd2
  patil-suraj

  authored
  170ebd2

  Copy full SHA for 170ebd2

• Add paint by example (#1533)

  Show description for 896c98a
  

  patrickvonplaten

  and

  pcuenca

  authored
  896c98a

  Copy full SHA for 896c98a

Commits on Dec 5, 2022

• [Docs] Correct docs (#1554)

  patrickvonplaten

  authored
  62b497c

  Copy full SHA for 62b497c

• [refactor] make set_attention_slice recursive (#1532)

  Show description for bce65cd
  

  patil-suraj

  and

  patrickvonplaten

  authored
  bce65cd

  Copy full SHA for bce65cd

• Compute embedding distances with torch.cdist (#1459)

  Show description for 720dbfc
  blefaudeux

  authored
  720dbfc

  Copy full SHA for 720dbfc

Commits on Dec 3, 2022

• Add xformers attention to VAE (#1507)

  Show description for daebee0
  

  3 peopleauthored
  daebee0

  Copy full SHA for daebee0

Commits on Dec 2, 2022

• fix tests

  patrickvonplaten

  committed
  cf4664e

  Copy full SHA for cf4664e

• Do not use torch.long in mps (#1488)

  Show description for 3ceaa28
  

  pcuenca

  and

  anton-l

  authored
  3ceaa28

  Copy full SHA for 3ceaa28

• [refactor] Making the xformers mem-efficient attention activation recursive (#1493)

  Show description for a816a87
  

  blefaudeux

  and

  Benjamin Lefaudeux

  authored
  a816a87

  Copy full SHA for a816a87

Commits on Dec 1, 2022

• Fix Flax flip_sin_to_cos (#1369)

  Show description for a6a25ce
  

  akashgokul

  and

  kashif

  authored
  a6a25ce

  Copy full SHA for a6a25ce

• simplyfy AttentionBlock (#1492)

  patil-suraj

  authored
  2bbf8b6

  Copy full SHA for 2bbf8b6

Commits on Nov 29, 2022

• StableDiffusion: Decode latents separately to run larger batches (#1150)

  Show description for c28d3c8
  

  kig

  and

  Ilmari Heikkinen

  authored
  c28d3c8

  Copy full SHA for c28d3c8

• Flax support for Stable Diffusion 2 (#1423)

  Show description for 4d1e4e2
  

  pcuenca

  and

  patrickvonplaten

  authored
  4d1e4e2

  Copy full SHA for 4d1e4e2

Commits on Nov 25, 2022

• Allow to set config params directly in init (#1419)

  Show description for 8faa822
  patrickvonplaten

  authored
  8faa822

  Copy full SHA for 8faa822

• [MPS] call contiguous after permute (#1411)

  Show description for babfb8a
  kashif

  authored
  babfb8a

  Copy full SHA for babfb8a

Commits on Nov 24, 2022

• Support SD2 attention slicing (#1397)

  Show description for d50e321
  anton-l

  authored
  d50e321

  Copy full SHA for d50e321

• Adapt UNet2D for supre-resolution (#1385)

  Show description for cecdd8b
  patil-suraj

  authored
  cecdd8b

  Copy full SHA for cecdd8b

Commits on Nov 23, 2022

• [Transformer2DModel] don't norm twice (#1381)

  Show description for 1524122
  patil-suraj

  authored
  1524122

  Copy full SHA for 1524122

• update unet2d (#1376)

  Show description for f07a16e
  patil-suraj

  authored
  f07a16e

  Copy full SHA for f07a16e

• [Versatile Diffusion] Add versatile diffusion model (#1283)

  Show description for 2625fb5
  

  patrickvonplaten

  and

  anton-l

  authored
  2625fb5

  Copy full SHA for 2625fb5

• handle fp16 in `UNet2DModel` (#1216)

  Show description for 9e234d8
  

  patil-suraj

  and

  pcuenca

  authored
  9e234d8

  Copy full SHA for 9e234d8

• Fix using non-square images with UNet2DModel and DDIM/DDPM pipelines (#1289)

  Show description for 8fd3a74
  

  jenkspt

  and

  patrickvonplaten

  authored
  8fd3a74

  Copy full SHA for 8fd3a74

Commits on Nov 22, 2022

• use memory_efficient_attention by default (#1354)

  Show description for 2d6d4ed
  

  patil-suraj

  and

  pcuenca

  authored
  2d6d4ed

  Copy full SHA for 2d6d4ed

Commits on Nov 21, 2022

• perf: prefer batched matmuls for attention (#1203)

  Show description for ad93593
  Birch-san

  authored
  ad93593

  Copy full SHA for ad93593

Commits on Nov 16, 2022

• doc string args shape fix (#1243)

  Show description for aa5c4c2
  kamalkraj

  authored
  aa5c4c2

  Copy full SHA for aa5c4c2

Commits on Nov 14, 2022

• Fix documentation typo for `UNet2DModel` and `UNet2DConditionModel` (#1275)

  Show description for 57525bb
  xenova

  authored
  57525bb

  Copy full SHA for 57525bb

• Add UNet 1d for RL model for planning + colab (#105)

  Show description for 7c5fef8
  

  Nathan Lambert

  and

  bglick13

  authored
  7c5fef8

  Copy full SHA for 7c5fef8

• Edited attention.py for older xformers (#1270)

  Show description for 33d7e89
  Lime-Cakes

  authored
  33d7e89

  Copy full SHA for 33d7e89

Commits on Nov 8, 2022

• handle dtype xformers attention (#1196)

  Show description for 5786b0e
  patil-suraj

  authored
  5786b0e

  Copy full SHA for 5786b0e

Commits on Nov 5, 2022

• Flax: Flip sin to cos in time embeddings (#1149)

  Show description for 08a6dc8
  pcuenca

  authored
  08a6dc8

  Copy full SHA for 08a6dc8