AI/diffusers
1
0
Fork 0
mirror of https://github.com/huggingface/diffusers.git synced 2026-01-29 07:22:12 +03:00
Code Activity
2,518 Commits 742 Branches 94 Tags
10e3774b8eef40b904bd645cf84aab4c9abc2ffe
Commit Graph

2518 Commits

This Branch
This Branch
All Branches
Author SHA1 Message Date
Daniel Gu
10e3774b8e Make style and remove some testing code. 2023-05-11 16:32:20 -07:00
Daniel Gu
54cfa3b370 Use a CLIPVisionModelWithProjection instead of CLIPVisionModel for image_encoder to better match the original UniDiffuser implementation. 2023-05-11 16:19:49 -07:00
Daniel Gu
b1a6f228cb make style 2023-05-11 10:54:31 -07:00
Daniel Gu
5341450e14 Fix the mixed precision issue and add additional tests of the pipeline cuda/fp16 functionality. 2023-05-11 10:53:45 -07:00
Daniel Gu
1bc2b91dfc Add fast test for CUDA/fp16 model behavior (currently failing). 2023-05-11 10:05:04 -07:00
Daniel Gu
f36df41999 Revert "Fix mixed precision issue by wrapping the offending code with the torch.autocast context manager." 
This reverts commit 1a58958ab4.
 2023-05-11 08:07:54 -07:00
Daniel Gu
1a58958ab4 Fix mixed precision issue by wrapping the offending code with the torch.autocast context manager. 2023-05-11 07:44:15 -07:00
Daniel Gu
7242f1bba9 make style 2023-05-11 07:22:58 -07:00
Daniel Gu
de8794ce50 Add slow test on full checkpoint for joint mode and correct expected image slices/text prefixes. 2023-05-11 07:22:21 -07:00
Daniel Gu
28dda62f53 make style 2023-05-11 03:28:37 -07:00
Daniel Gu
19a20a55b9 Remove dependency on einops by refactoring einops operations to pure torch operations. 2023-05-11 03:27:33 -07:00
Daniel Gu
fa9e3879e8 Remove outdated logic based on transformers version in UniDiffuser pipeline __init__.py 2023-05-11 02:05:27 -07:00
Daniel Gu
a54d6318df Fix some typos in the UniDiffuser docs. 2023-05-11 00:06:14 -07:00
Daniel Gu
8f2d325c2d make style 2023-05-10 23:53:40 -07:00
Daniel Gu
006ab4997d Improve UniDiffuser slow tests. 2023-05-10 23:53:01 -07:00
Daniel Gu
8a57342d12 Fix code quality via make style. 2023-05-10 23:00:46 -07:00
Daniel Gu
a7f50f4c9c Move the UniDiffusers convert_from_ckpy.py script to diffusers/scripts/convert_unidiffuser_to_diffusers.py 2023-05-10 22:56:36 -07:00
Daniel Gu
a46e1ecb3c Refactor and improve the UniDiffuser convert_from_ckpt.py script. 2023-05-10 22:49:25 -07:00
Daniel Gu
2b92111c4f Fix code with make style. 2023-05-10 20:14:49 -07:00
Daniel Gu
ae7d549e0b Add documentation for UniDiffuser and fix some typos/formatting in docstrings. 2023-05-10 20:13:28 -07:00
Daniel Gu
abd6fca81e Add/edit docstrings for added classes and public pipeline methods. Also do some light refactoring. 2023-05-10 19:14:20 -07:00
Daniel Gu
5728328545 Fix code style with make style. 2023-05-09 09:02:05 -07:00
Daniel Gu
16fd515d3c Fix bugs and make improvements to the UniDiffuser pipeline: 
- Improve batch size inference and fix bugs when num_images_per_prompt or num_prompts_per_image > 1
	- Add tests for num_images_per_prompt, num_prompts_per_image > 1
	- Improve check_inputs, especially regarding checking supplied latents
	- Add reset_mode method so that mode inference can be re-enabled after mode is set manually
	- Fix some warnings related to accessing class members directly instead of through their config
	- Small amount of refactoring in pipeline_unidiffuser.py
 2023-05-09 08:54:43 -07:00
Daniel Gu
0cddc3cadb Add UniDiffuser dummy objects via make fix-copies. 2023-05-08 18:38:46 -07:00
Daniel Gu
34a40ad913 Make changes so that the repository consistency tests pass. 2023-05-08 18:30:39 -07:00
Daniel Gu
8dd7b0be2c Add fast test for checking UniDiffuser-v1 sampling. 2023-05-08 17:50:46 -07:00
Daniel Gu
54c495f175 Support using a data type embedding for UniDiffuser-v1. 2023-05-06 00:06:14 -07:00
Daniel Gu
fc540b5511 Fix code quality with make style. 2023-05-05 12:28:12 -07:00
Daniel Gu
e62b32aadc Add self.image_encoder, self.text_decoder to list of models to offload to CPU in the enable_sequential_cpu_offload(...)/enable_model_cpu_offload(...) methods to make test_cpu_offload_forward_pass pass. 2023-05-05 12:12:42 -07:00
Daniel Gu
1cb726a005 Merge branch 'main' into unidiffuser-pipeline 2023-05-05 09:23:41 -07:00
Isotr0py
0ffac97933 Add use_Karras_sigmas to LMSDiscreteScheduler (#3351) 
* add karras sigma to lms discrete scheduler

* add test for lms_scheduler karras

* reformat test lms
 2023-05-06 12:19:27 +01:00
Lysandre Debut
b0966f5801 Inpainting: typo in docs (#3331) 
Typo in docs

Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com>
 2023-05-06 12:13:33 +01:00
Lucca Zenóbio
0407c3e7d0 Fix pipeline class on README (#3345) 
Update README.md
 2023-05-06 12:06:52 +01:00
At-sushi
7ce3fa010a Fix TypeError when using prompt_embeds and negative_prompt (#2982) 
* test: Added test case

* fix: fixed type checking issue on _encode_prompt

* fix: fixed copies consistency

* fix: one copy was not sufficient
 2023-05-06 12:04:07 +01:00
Sanchit Gandhi
abd86d1c17 [AudioLDM] Generalise conversion script (#3328) 
Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com>
 2023-05-06 12:00:42 +01:00
Adrià Arrufat
e9aa0925a8 Rename --only_save_embeds to --save_as_full_pipeline (#3206) 
* Set --only_save_embeds to False by default

Due to how the option is named, it makes more sense to behave like this.

* Refactor only_save_embeds to save_as_full_pipeline
 2023-05-06 12:00:30 +01:00
Will Rice
36f43ea75a Add upsample_size to AttnUpBlock2D, AttnDownBlock2D (#3275) 
The argument `upsample_size` needs to be added to these modules to allow compatibility with other blocks that require this argument.
 2023-05-05 19:50:41 +01:00
Daniel Gu
9d39bef45f Revert "Fix code style with make style." 
This reverts commit 10a174a12c.
 2023-05-05 09:08:07 -07:00
Cheng Lu
27522b585b Add the SDE variant of DPM-Solver and DPM-Solver++ (#3344) 
* add SDE variant of DPM-Solver and DPM-Solver++

* add test

* fix typo

* fix typo
 2023-05-05 16:03:47 +01:00
Daniel Gu
fc8526354f Fix code with make style. 2023-05-05 07:55:15 -07:00
Daniel Gu
7097dd74f5 Update fast tests to use test checkpoints stored on the hub and to better match the reference UniDiffuser implementation. 2023-05-05 07:49:46 -07:00
Daniel Gu
434b255060 Add UniDiffuser classes to __init__ files, modify transformer block to support pre- and post-LN, add fast default tests, fix some bugs. 2023-05-05 07:49:46 -07:00
Will Rice
1d213def63 Add upsample_size to AttnUpBlock2D, AttnDownBlock2D (#3275) 
The argument `upsample_size` needs to be added to these modules to allow compatibility with other blocks that require this argument.
 2023-05-05 07:23:51 -07:00
Cheng Lu
652dbaad8d Add the SDE variant of DPM-Solver and DPM-Solver++ (#3344) 
* add SDE variant of DPM-Solver and DPM-Solver++

* add test

* fix typo

* fix typo
 2023-05-05 07:23:51 -07:00
Patrick von Platen
140ab748e8 Fix config dpm (#3343) 2023-05-05 07:23:51 -07:00
Patrick von Platen
458847e25b [Quality] Make style (#3341) 2023-05-05 07:23:51 -07:00
Sayak Paul
b4aa419f9b update controlling generation doc with latest goodies. (#3321) 2023-05-05 07:23:51 -07:00
Cesar Aybar
3e8d3d86ac Update write_own_pipeline.mdx (#3323) 2023-05-05 07:23:51 -07:00
Isamu Isozaki
79295870f7 Added input pretubation (#3292) 
* Added input pretubation

* Fixed spelling
 2023-05-05 07:23:51 -07:00
Steven Liu
0e8f4f0f49 [docs] Improve LoRA docs (#3311) 
* update docs

* add to toctree

* apply feedback
 2023-05-05 07:23:51 -07:00