Code

Add Stable Diffusion demo#100

Code

Draft

akhilg-nv wants to merge 16 commits intomainNVIDIA/TensorRT-Incubator:mainfrom dev-akhilg-demo-diffusionNVIDIA/TensorRT-Incubator:dev-akhilg-demo-diffusionCopy head branch name to clipboard

+1,254

Commits

Commits on Nov 16, 2024

• Initial DemoDiffusion

  Show description for 5f4fac5
  akhilg-nv

  committed
  5f4fac5

• Clean up example code

  Show description for 730b956
  akhilg-nv

  committed
  730b956

• Fix accuracy regression, resume clean-up

  Show description for 8829f68
  akhilg-nv

  committed
  8829f68

• Remove inlined CLIP model, use K-LMS scheduler for HF

  Show description for c54431d
  akhilg-nv

  committed
  c54431d

• Add accuracy test

  Show description for 49f0e19
  akhilg-nv

  committed
  49f0e19

• Split and simplify model code

  Show description for 43249e3
  akhilg-nv

  committed
  43249e3

• Add fp16 support

  Show description for fc953d0
  akhilg-nv

  committed
  fc953d0

• Fix NaN bug in normalization for fp16

  akhilg-nv

  committed
  e361f94

• Update image saving and test

  akhilg-nv

  committed
  368a56b

• Update for latest shape changes, improve perf.

  Show description for 662e631
  akhilg-nv

  committed
  662e631

• Remove Tripy casts in critical loop

  akhilg-nv

  committed
  e7a02f6

Commits on Dec 13, 2024

• Additional performance profiling

  akhilg-nv

  committed
  75efd7d

• Remove unnecessary stream sync

  yizhuoz004

  committed
  45e048e

• Merge branch 'main' of github.com:NVIDIA/TensorRT-Incubator into dev-akhilg-demo-diffusion

  yizhuoz004

  committed
  2646d24

• Rebase to TOT, small fixes

  yizhuoz004

  committed
  8c8e5c4

Commits on Dec 18, 2024

• Small change to scale dot attn

  yizhuoz004

  committed
  b5a9c9c