Skip to content

GitLab

Switch branch/tag
  1. 15 Feb, 2025 1 commit
  3. 14 Feb, 2025 1 commit
    • Aryan's avatar
      Module Group Offloading (#10503) · 9a147b82
      Aryan authored 
      

* update

* fix

* non_blocking; handle parameters and buffers

* update

* Group offloading with cuda stream prefetching (#10516)

* cuda stream prefetch

* remove breakpoints

* update

* copy model hook implementation from pab

* update; ~very workaround based implementation but it seems to work as expected; needs cleanup and rewrite

* more workarounds to make it actually work

* cleanup

* rewrite

* update

* make sure to sync current stream before overwriting with pinned params

not doing so will lead to erroneous computations on the GPU and cause bad results

* better check

* update

* remove hook implementation to not deal with merge conflict

* re-add hook changes

* why use more memory when less memory do trick

* why still use slightly more memory when less memory do trick

* optimise

* add model tests

* add pipeline tests

* update docs

* add layernorm and groupnorm

* address review comments

* improve tests; add docs

* improve docs

* Apply suggestions from code review
Co-authored-by: default avatarSteven Liu <59462357+stevhliu@users.noreply.github.com>

* apply suggestions from code review

* update tests

* apply suggestions from review

* enable_group_offloading -> enable_group_offload for naming consistency

* raise errors if multiple offloading strategies used; add relevant tests

* handle .to() when group offload applied

* refactor some repeated code

* remove unintentional change from merge conflict

* handle .cuda()

---------
Co-authored-by: default avatarSteven Liu <59462357+stevhliu@users.noreply.github.com>
      9a147b82
  5. 13 Feb, 2025 2 commits
  7. 12 Feb, 2025 3 commits
  9. 11 Feb, 2025 3 commits
  11. 10 Feb, 2025 1 commit
  13. 31 Jan, 2025 1 commit
  15. 28 Jan, 2025 2 commits
  17. 27 Jan, 2025 1 commit
  19. 24 Jan, 2025 1 commit
  21. 22 Jan, 2025 1 commit
    • Aryan's avatar
      [core] Layerwise Upcasting (#10347) · beacaa55
      Aryan authored 
      

* update

* update

* make style

* remove dynamo disable

* add coauthor
Co-Authored-By: default avatarDhruv Nair <dhruv.nair@gmail.com>

* update

* update

* update

* update mixin

* add some basic tests

* update

* update

* non_blocking

* improvements

* update

* norm.* -> norm

* apply suggestions from review

* add example

* update hook implementation to the latest changes from pyramid attention broadcast

* deinitialize should raise an error

* update doc page

* Apply suggestions from code review
Co-authored-by: default avatarSteven Liu <59462357+stevhliu@users.noreply.github.com>

* update docs

* update

* refactor

* fix _always_upcast_modules for asym ae and vq_model

* fix lumina embedding forward to not depend on weight dtype

* refactor tests

* add simple lora inference tests

* _always_upcast_modules -> _precision_sensitive_module_patterns

* remove todo comments about review; revert changes to self.dtype in unets because .dtype on ModelMixin should be able to handle fp8 weight case

* check layer dtypes in lora test

* fix UNet1DModelTests::test_layerwise_upcasting_inference

* _precision_sensitive_module_patterns -> _skip_layerwise_casting_patterns based on feedback

* skip test in NCSNppModelTests

* skip tests for AutoencoderTinyTests

* skip tests for AutoencoderOobleckTests

* skip tests for UNet1DModelTests - unsupported pytorch operations

* layerwise_upcasting -> layerwise_casting

* skip tests for UNetRLModelTests; needs next pytorch release for currently unimplemented operation support

* add layerwise fp8 pipeline test

* use xfail

* Apply suggestions from code review
Co-authored-by: default avatarDhruv Nair <dhruv.nair@gmail.com>

* add assertion with fp32 comparison; add tolerance to fp8-fp32 vs fp32-fp32 comparison (required for a few models' test to pass)

* add note about memory consumption on tesla CI runner for failing test

---------
Co-authored-by: default avatarDhruv Nair <dhruv.nair@gmail.com>
Co-authored-by: default avatarSteven Liu <59462357+stevhliu@users.noreply.github.com>
      beacaa55
  23. 21 Jan, 2025 2 commits
  25. 20 Jan, 2025 1 commit
  27. 19 Jan, 2025 1 commit
  29. 16 Jan, 2025 4 commits
  31. 14 Jan, 2025 2 commits
  33. 13 Jan, 2025 1 commit
  35. 11 Jan, 2025 1 commit
    • Junyu Chen's avatar
      [DC-AE] support tiling for DC-AE (#10510) · e7db062e
      Junyu Chen authored 
      

* autoencoder_dc tiling

* add tiling and slicing support in SANA pipelines

* create variables for padding length because the line becomes too long

* add tiling and slicing support in pag SANA pipelines

* revert changes to tile size

* make style

* add vae tiling test

---------
Co-authored-by: default avatarAryan <aryan@huggingface.co>
      e7db062e
  37. 10 Jan, 2025 1 commit
    • Daniel Hipke's avatar
      Add a `disable_mmap` option to the `from_single_file` loader to improve load... · 52c05bd4
      Daniel Hipke authored 
      
Add a `disable_mmap` option to the `from_single_file` loader to improve load performance on network mounts (#10305)

* Add no_mmap arg.

* Fix arg parsing.

* Update another method to force no mmap.

* logging

* logging2

* propagate no_mmap

* logging3

* propagate no_mmap

* logging4

* fix open call

* clean up logging

* cleanup

* fix missing arg

* update logging and comments

* Rename to disable_mmap and update other references.

* [Docs] Update ltx_video.md to remove generator from `from_pretrained()` (#10316)

Update ltx_video.md to remove generator from `from_pretrained()`

* docs: fix a mistake in docstring (#10319)

Update pipeline_hunyuan_video.py

docs: fix a mistake

* [BUG FIX] [Stable Audio Pipeline] Resolve torch.Tensor.new_zeros() TypeError in function prepare_latents caused by audio_vae_length (#10306)

[BUG FIX] [Stable Audio Pipeline] TypeError: new_zeros(): argument 'size' failed to unpack the object at pos 3 with error "type must be tuple of ints,but got float"

torch.Tensor.new_zeros() takes a single argument size (int...) – a list, tuple, or torch.Size of integers defining the shape of the output tensor.

in function prepare_latents:
audio_vae_length = self.transformer.config.sample_size * self.vae.hop_length
audio_shape = (batch_size // num_waveforms_per_prompt, audio_channels, audio_vae_length)
...
audio = initial_audio_waveforms.new_zeros(audio_shape)

audio_vae_length evaluates to float because self.transformer.config.sample_size returns a float
Co-authored-by: default avatarhlky <hlky@hlky.ac>

* [docs] Fix quantization links (#10323)

Update overview.md

* [Sana]add 2K related model for Sana (#10322)

add 2K related model for Sana

* Update src/diffusers/loaders/single_file_model.py
Co-authored-by: default avatarDhruv Nair <dhruv.nair@gmail.com>

* Update src/diffusers/loaders/single_file.py
Co-authored-by: default avatarDhruv Nair <dhruv.nair@gmail.com>

* make style

---------
Co-authored-by: default avatarhlky <hlky@hlky.ac>
Co-authored-by: default avatarSayak Paul <spsayakpaul@gmail.com>
Co-authored-by: default avatarLeojc <liao_junchao@outlook.com>
Co-authored-by: default avatarAditya Raj <syntaxticsugr@gmail.com>
Co-authored-by: default avatarSteven Liu <59462357+stevhliu@users.noreply.github.com>
Co-authored-by: default avatarJunsong Chen <cjs1020440147@icloud.com>
Co-authored-by: default avatarDhruv Nair <dhruv.nair@gmail.com>
      52c05bd4
  39. 09 Jan, 2025 1 commit
  41. 08 Jan, 2025 4 commits
  43. 07 Jan, 2025 1 commit
  45. 06 Jan, 2025 2 commits
    • Ameer Azam's avatar
      Regarding the RunwayML path for V1.5 did change to... · 4f5e3e35
      Ameer Azam authored 
      Regarding the RunwayML path for V1.5 did change to stable-diffusion-v1-5/[stable-diffusion-v1-5/ stable-diffusion-inpainting] (#10476)

* Update pipeline_controlnet.py

* Update pipeline_controlnet_img2img.py

runwayml Take-down so change all from to this
stable-diffusion-v1-5/stable-diffusion-v1-5

* Update pipeline_controlnet_inpaint.py

* runwayml take-down make change to sd-legacy

* runwayml take-down make change to sd-legacy

* runwayml take-down make change to sd-legacy

* runwayml take-down make change to sd-legacy

* Update convert_blipdiffusion_to_diffusers.py

style change
      4f5e3e35
    • Aryan's avatar
      Fix hunyuan video attention mask dim (#10454) · 7747b588
      Aryan authored 
      

* fix

* add coauthor
Co-Authored-By: default avatarNerogar <nerogar@arcor.de>

---------
Co-authored-by: default avatarNerogar <nerogar@arcor.de>
      7747b588
  47. 02 Jan, 2025 2 commits