Skip to content

GitLab

Switch branch/tag
  1. 10 Oct, 2022 1 commit
    • Patrick von Platen's avatar
      [Low CPU memory] + device map (#772) · fab17528
      Patrick von Platen authored 
      

* add accelerate to load models with smaller memory footprint

* remove low_cpu_mem_usage as it is reduntant

* move accelerate init weights context to modelling utils

* add test to ensure results are the same when loading with accelerate

* add tests to ensure ram usage gets lower when using accelerate

* move accelerate logic to single snippet under modelling utils and remove it from configuration utils

* format code using to pass quality check

* fix imports with isor

* add accelerate to test extra deps

* only import accelerate if device_map is set to auto

* move accelerate availability check to diffusers import utils

* format code

* add device map to pipeline abstraction

* lint it to pass PR quality check

* fix class check to use accelerate when using diffusers ModelMixin subclasses

* use low_cpu_mem_usage in transformers if device_map is not available

* NoModuleLayer

* comment out tests

* up

* uP

* finish

* Update src/diffusers/pipelines/stable_diffusion/safety_checker.py

* finish

* uP

* make style
Co-authored-by: default avatarPi Esposito <piero.skywalker@gmail.com>
      fab17528
  3. 07 Oct, 2022 1 commit
    • Suraj Patil's avatar
      [img2img, inpainting] fix fp16 inference (#769) · 92d70863
      Suraj Patil authored 
      * handle dtype in vae and image2image pipeline

* fix inpaint in fp16

* dtype should be handled in add_noise

* style

* address review comments

* add simple fast tests to check fp16

* fix test name

* put mask in fp16
      92d70863
  5. 06 Oct, 2022 3 commits
  7. 05 Oct, 2022 6 commits
  9. 04 Oct, 2022 1 commit
  11. 03 Oct, 2022 4 commits
  13. 02 Oct, 2022 1 commit
  15. 30 Sep, 2022 1 commit
    • Nouamane Tazi's avatar
      Optimize Stable Diffusion (#371) · 9ebaea54
      Nouamane Tazi authored 
      * initial commit

* make UNet stream capturable

* try to fix noise_pred value

* remove cuda graph and keep NB

* non blocking unet with PNDMScheduler

* make timesteps np arrays for pndm scheduler
because lists don't get formatted to tensors in `self.set_format`

* make max async in pndm

* use channel last format in unet

* avoid moving timesteps device in each unet call

* avoid memcpy op in `get_timestep_embedding`

* add `channels_last` kwarg to `DiffusionPipeline.from_pretrained`

* update TODO

* replace `channels_last` kwarg with `memory_format` for more generality

* revert the channels_last changes to leave it for another PR

* remove non_blocking when moving input ids to device

* remove blocking from all .to() operations at beginning of pipeline

* fix merging

* fix merging

* model can run in other precisions without autocast

* attn refactoring

* Revert "attn refactoring"

This reverts commit 0c70c0e189cd2c4d8768274c9fcf5b940ee310fb.

* remove restriction to run conv_norm in fp32

* use `baddbmm` instead of `matmul`for better in attention for better perf

* removing all reshapes to test perf

* Revert "removing all reshapes to test perf"

This reverts commit 006ccb8a8c6bc7eb7e512392e692a29d9b1553cd.

* add shapes comments

* hardcore whats needed for jitting

* Revert "hardcore whats needed for jitting"

This reverts commit 2fa9c698eae2890ac5f8e367ca80532ecf94df9a.

* Revert "remove restriction to run conv_norm in fp32"

This reverts commit cec592890c32da3d1b78d38b49e4307aedf459b9.

* revert using baddmm in attention's forward

* cleanup comment

* remove restriction to run conv_norm in fp32. no quality loss was noticed

This reverts commit cc9bc1339c998ebe9e7d733f910c6d72d9792213.

* add more optimizations techniques to docs

* Revert "add shapes comments"

This reverts commit 31c58eadb8892f95478cdf05229adf678678c5f4.

* apply suggestions

* make quality

* apply suggestions

* styling

* `scheduler.timesteps` are now arrays so we dont need .to()

* remove useless .type()

* use mean instead of max in `test_stable_diffusion_inpaint_pipeline_k_lms`

* move scheduler timestamps to correct device if tensors

* add device to `set_timesteps` in LMSD scheduler

* `self.scheduler.set_timesteps` now uses device arg for schedulers that accept it

* quick fix

* styling

* remove kwargs from schedulers `set_timesteps`

* revert to using max in K-LMS inpaint pipeline test

* Revert "`self.scheduler.set_timesteps` now uses device arg for schedulers that accept it"

This reverts commit 00d5a51e5c20d8d445c8664407ef29608106d899.

* move timesteps to correct device before loop in SD pipeline

* apply previous fix to other SD pipelines

* UNet now accepts tensor timesteps even on wrong device, to avoid errors
- it shouldnt affect performance if timesteps are alrdy on correct device
- it does slow down performance if they're on the wrong device

* fix pipeline when timesteps are arrays with strides
      9ebaea54
  17. 27 Sep, 2022 5 commits
    • Anton Lozhkov's avatar
      Fix onnx tensor format (#654) · d8572f20
      Anton Lozhkov authored 
      fix np onnx
      d8572f20
    • Kashif Rasul's avatar
      [Pytorch] Pytorch only schedulers (#534) · bd8df2da
      Kashif Rasul authored 
      

* pytorch only schedulers

* fix style

* remove match_shape

* pytorch only ddpm

* remove SchedulerMixin

* remove numpy from karras_ve

* fix types

* remove numpy from lms_discrete

* remove numpy from pndm

* fix typo

* remove mixin and numpy from sde_vp and ve

* remove remaining tensor_format

* fix style

* sigmas has to be torch tensor

* removed set_format in readme

* remove set format from docs

* remove set_format from pipelines

* update tests

* fix typo

* continue to use mixin

* fix imports

* removed unsed imports

* match shape instead of assuming image shapes

* remove import typo

* update call to add_noise

* use math instead of numpy

* fix t_index

* removed commented out numpy tests

* timesteps needs to be discrete

* cast timesteps to int in flax scheduler too

* fix device mismatch issue

* small fix

* Update src/diffusers/schedulers/scheduling_pndm.py
Co-authored-by: default avatarPatrick von Platen <patrick.v.platen@gmail.com>
      bd8df2da
    • Pedro Cuenca's avatar
      Flax pipeline pndm (#583) · ab3fd671
      Pedro Cuenca authored 
      

* WIP: flax FlaxDiffusionPipeline & FlaxStableDiffusionPipeline

* todo comment

* Fix imports

* Fix imports

* add dummies

* Fix empty init

* make pipeline work

* up

* Allow dtype to be overridden on model load.

This may be a temporary solution until #567 is addressed.

* Convert params to bfloat16 or fp16 after loading.

This deals with the weights, not the model.

* Use Flax schedulers (typing, docstring)

* PNDM: replace control flow with jax functions.

Otherwise jitting/parallelization don't work properly as they don't know
how to deal with traced objects.

I temporarily removed `step_prk`.

* Pass latents shape to scheduler set_timesteps()

PNDMScheduler uses it to reserve space, other schedulers will just
ignore it.

* Wrap model imports inside availability checks.

* Optionally return state in from_config.

Useful for Flax schedulers.

* Do not convert model weights to dtype.

* Re-enable PRK steps with functional implementation.

Values returned still not verified for correctness.

* Remove left over has_state var.

* make style

* Apply suggestion list -> tuple
Co-authored-by: default avatarSuraj Patil <surajp815@gmail.com>

* Apply suggestion list -> tuple
Co-authored-by: default avatarSuraj Patil <surajp815@gmail.com>

* Remove unused comments.

* Use zeros instead of empty.
Co-authored-by: default avatarMishig Davaadorj <dmishig@gmail.com>
Co-authored-by: default avatarMishig Davaadorj <mishig.davaadorj@coloradocollege.edu>
Co-authored-by: default avatarPatrick von Platen <patrick.v.platen@gmail.com>
Co-authored-by: default avatarSuraj Patil <surajp815@gmail.com>
      ab3fd671
    • Pedro Cuenca's avatar
      Remove deprecated `torch_device` kwarg (#623) · b671cb09
      Pedro Cuenca authored 
      * Remove deprecated `torch_device` kwarg.

* Remove unused imports.
      b671cb09
    • Yuta Hayashibe's avatar
      Warning for too long prompts in DiffusionPipelines (Resolve #447) (#472) · f7ebe569
      Yuta Hayashibe authored 
      * Return encoded texts by DiffusionPipelines

* Updated README to show hot to use enoded_text_input

* Reverted examples in README.md

* Reverted all

* Warning for long prompts

* Fix bugs

* Formatted
      f7ebe569
  19. 23 Sep, 2022 2 commits
  21. 21 Sep, 2022 1 commit
    • Pedro Cuenca's avatar
      Allow dtype to be specified in Flax pipeline (#600) · fb2fbab1
      Pedro Cuenca authored 
      * Fix typo in docstring.

* Allow dtype to be overridden on model load.

This may be a temporary solution until #567 is addressed.

* Create latents in float32

The denoising loop always computes the next step in float32, so this
would fail when using `bfloat16`.
      fb2fbab1
  23. 20 Sep, 2022 4 commits
  25. 19 Sep, 2022 2 commits
  27. 17 Sep, 2022 1 commit
  29. 16 Sep, 2022 2 commits
  31. 13 Sep, 2022 1 commit
  33. 12 Sep, 2022 1 commit
    • Kashif Rasul's avatar
      update expected results of slow tests (#268) · f4781a0b
      Kashif Rasul authored 
      

* update expected results of slow tests

* relax sum and mean tests

* Print shapes when reporting exception

* formatting

* fix sentence

* relax test_stable_diffusion_fast_ddim for gpu fp16

* relax flakey tests on GPU

* added comment on large tolerences

* black

* format

* set scheduler seed

* added generator

* use np.isclose

* set num_inference_steps to 50

* fix dep. warning

* update expected_slice

* preprocess if image

* updated expected results

* updated expected from CI

* pass generator to VAE

* undo change back to orig

* use orignal

* revert back the expected on cpu

* revert back values for CPU

* more undo

* update result after using gen

* update mean

* set generator for mps

* update expected on CI server

* undo

* use new seed every time

* cpu manual seed

* reduce num_inference_steps

* style

* use generator for randn
Co-authored-by: default avatarPatrick von Platen <patrick.v.platen@gmail.com>
      f4781a0b
  35. 08 Sep, 2022 3 commits