1
0
mirror of https://github.com/huggingface/diffusers.git synced 2026-01-29 07:22:12 +03:00
Commit Graph

2135 Commits

Author SHA1 Message Date
Pedro Cuenca
a26f35ff2e Initial draft of Core ML docs (#2987)
* Initial draft of Core ML docs.

* Apply suggestions from code review

Co-authored-by: Steven Liu <59462357+stevhliu@users.noreply.github.com>

* Fix Core ML spelling

* Apply the rest of suggestions.

* Attempt to fix hyperlink inside Tip.

* Apply suggestions from code review

Co-authored-by: Steven Liu <59462357+stevhliu@users.noreply.github.com>

* Apply suggestions from code review

---------

Co-authored-by: Steven Liu <59462357+stevhliu@users.noreply.github.com>
2023-05-05 07:22:12 -07:00
William Berman
93ee47b255 add encoder_hid_dim to unet
`encoder_hid_dim` provides an additional projection for the input `encoder_hidden_states` from `encoder_hidden_dim` to `cross_attention_dim`
2023-05-05 07:22:12 -07:00
William Berman
a87520f469 allow unet varying number of layers per block 2023-05-05 07:22:12 -07:00
William Berman
c51b7c94d0 resnet skip time activation and output scale factor 2023-05-05 07:22:12 -07:00
William Berman
2414db30e0 do not overwrite scheduler instance variables with type casted versions 2023-05-05 07:22:12 -07:00
William Berman
b1d42e7df6 add missing AttnProcessor2_0 to AttentionProcessor union 2023-05-05 07:22:12 -07:00
William Berman
51c7041e0e fix simple attention processor encoder hidden states ordering 2023-05-05 07:22:12 -07:00
William Berman
bd51557816 clamp comment @sayakpaul 2023-05-05 07:22:12 -07:00
William Berman
edd9ad88d8 remove extra min arg @sayakpaul 2023-05-05 07:22:12 -07:00
William Berman
9f85701567 ddpm scheduler variance fixes 2023-05-05 07:22:12 -07:00
Will Berman
025609e5f3 dynamic threshold sampling bug fixes and docs (#3003)
dynamic threshold sampling bug fix and docs
2023-05-05 07:22:12 -07:00
YiYi Xu
7ee2817ae2 add Min-SNR loss to Controlnet flax train script (#3016)
* add wandb team and min-snr loss

* make style

* apply feedbacks
2023-05-05 07:22:12 -07:00
Patrick von Platen
5326243241 Update one_step_unet.py
Fix dummy community pipeline
2023-05-05 07:22:12 -07:00
Guspan Tanadi
229f9a22c9 docs: Link Navigation Path API Pipelines (#2976)
* docs: link navigation Safe Stable Diffusion

Link navigation API pipelines text2img and using diffusers Conditional Image Generation.

* docs: link navigation Versatile Diffusion

Removing exceeding path Stable Diffusion Overview.

* docs: Python extension Spectrogram Diffusion

Link navigation Spectrogram Diffusion Pipeline source code

* docs: Link navigation AltDiffusion Pipelines

Stable Diffusion Overview and Using Diffusers path.
2023-05-05 07:22:12 -07:00
Pedro Cuenca
3dc4a2a6f7 Explain how to install test dependencies (#2983)
As pointed out by @Birch-san: https://github.com/huggingface/diffusers/pull/2634#issuecomment-1496517210
2023-05-05 07:22:12 -07:00
YiYi Xu
2f6b212c78 minor fix in controlnet flax example (#2986)
* fix the error when push_to_hub but not log validation

* contronet_from_pt & controlnet_revision

* add intermediate checkpointing to the guide
2023-05-05 07:22:12 -07:00
FurryPotato
f7f096f382 [scheduler] fix some scheduler dtype error (#2992)
Co-authored-by: wangguan <dizhipeng.dzp@alibaba-inc.com>
Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com>
2023-05-05 07:22:12 -07:00
Sayak Paul
a7da9c1a8f [Examples] Add support for Min-SNR weighting strategy for better convergence (#2899)
* improve stable unclip doc.

* feat: support for applying min-snr weighting for faster convergence.

* add: support for validation logging with wandb

* make  not a required arg.

* fix: arg name.

* fix: cli args.

* fix: tracker config.

* fix: loss calculation.

* fix: validation logging.

* fix: unwrap call.

* fix: validation logging.

* fix: internval.

* fix: checkpointing push to hub.

* fix: c8a2856c6d\#commitcomment-106913193

* fix: norm group test for UNet3D.

* address PR comments.

* remove unneeded code.

* add: entry in the readme and docs.

* Apply suggestions from code review

Co-authored-by: Suraj Patil <surajp815@gmail.com>

---------

Co-authored-by: Suraj Patil <surajp815@gmail.com>
2023-05-05 07:22:12 -07:00
cmdr2
30e1d9aa13 Update the K-Diffusion SD pipeline, to allow calling it with only prompt_embeds (instead of always requiring a prompt) (#2962) 2023-05-05 07:22:12 -07:00
Nipun Jindal
f704da537d [2905]: Add Karras pattern to discrete euler (#2956)
* [2905]: Add Karras pattern to discrete euler

* [2905]: Add Karras pattern to discrete euler

* Review comments

* Review comments

* Review comments

* Review comments

---------

Co-authored-by: njindal <njindal@adobe.com>
2023-05-05 07:22:12 -07:00
Kadir Nar
912b8391b0 The variable name has been updated. (#2970) 2023-05-05 07:22:12 -07:00
Patrick von Platen
0aa92a53ad [Pipeline download] Improve pipeline download for index and passed co… (#2980)
* [Pipeline download] Improve pipeline download for index and passed components

* correct

* add more tests

* up
2023-05-05 07:22:12 -07:00
YiYi Xu
fe03d5bce4 update flax controlnet training script (#2951)
* load_from_disk + checkpointing_steps

* apply feedback
2023-05-05 07:22:11 -07:00
Steven Liu
72f37316ee [docs] Simplify loading guide (#2694)
* simplify loading guide

* apply feedbacks

* clarify variants

* clarify torch_dtype and variant

* remove conceptual pipeline doc
2023-05-05 07:22:11 -07:00
YiYi Xu
39ddd78c58 fix post-processing (#2968)
Co-authored-by: yiyixuxu <yixu310@gmail,com>
2023-05-05 07:22:11 -07:00
YiYi Xu
a6f2d7224e allow use custom local dataset for controlnet training scripts (#2928)
use custom local datset

Co-authored-by: yiyixuxu <yixu310@gmail,com>
Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com>
2023-05-05 07:22:11 -07:00
Lucain
1d3758f0a4 Use upload_folder in training scripts (#2934)
use upload folder in training scripts

Co-authored-by: testbot <lucainp@hf.co>
2023-05-05 07:22:11 -07:00
Patrick von Platen
72278553a3 make style 2023-05-05 07:22:11 -07:00
Ernie Chu
aee06e7cbb Ensure validation image RGB not RGBA (#2945)
* ensure validation image RGB not RGBA

* ensure validation image RGB not RGBA

---------

Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com>
2023-05-05 07:22:11 -07:00
Guspan Tanadi
11bcc6e895 Removing explicit markdown extension (#2944)
Trigger from previous PR. Build the page once again.
2023-05-05 07:22:11 -07:00
M. Tolga Cangöz
9ddd0ebaed Update ddpm.mdx (#2929) 2023-05-05 07:22:11 -07:00
M. Tolga Cangöz
353e6b5fb3 Update ddim.mdx (#2926) 2023-05-05 07:22:11 -07:00
M. Tolga Cangöz
32039273b7 Update score_sde_vp.mdx (#2938) 2023-05-05 07:22:11 -07:00
M. Tolga Cangöz
909a0d86a4 Update score_sde_ve.mdx (#2937) 2023-05-05 07:22:11 -07:00
M. Tolga Cangöz
8ab78d3a23 Update unipc.mdx (#2936) 2023-05-05 07:22:11 -07:00
M. Tolga Cangöz
98e9d4d337 Update euler_ancestral.mdx (#2932) 2023-05-05 07:22:11 -07:00
Daniel Gu
10a174a12c Fix code style with make style. 2023-05-05 05:21:01 -07:00
Daniel Gu
f2f336c871 Fix more bugs and make PatchEmbed position embedding configurable. 2023-05-05 05:19:00 -07:00
Daniel Gu
240940a1ea Fix small bugs in pipeline_unidiffuser.py and convert_from_ckpt.py. 2023-04-24 21:14:06 -07:00
Daniel Gu
48f2c25f36 Fix bugs in convert_from_ckpt.py and UniDiffuserTextDecoder encode_prefix() 2023-04-24 21:02:05 -07:00
Daniel Gu
2fdb7b871b Add convert_from_ckpt.py script to convert UniDiffuser checkpoints to their diffusers equivalents. 2023-04-20 22:02:20 -07:00
Daniel Gu
3c9b20e066 Alter the U-ViT implementation (UniDiffuserModel and associated building blocks) to more closely match the original UniDiffusers implementation. 2023-04-19 22:37:00 -07:00
Daniel Gu
84781fbd67 Fix noise pred timestep, clip_tokenizer, CLIP image encoding, and other bugs. 2023-04-17 19:06:19 -07:00
Daniel Gu
0300563861 Simplify the UTransformer2DModel / UniDiffuserModel implementation and fix some more bugs. 2023-04-16 22:25:15 -07:00
Daniel Gu
a492e0c86f Add UniDiffuser classes to __init__ files, modify transformer block to support pre- and post-LN, add fast default tests, fix some bugs. 2023-04-14 00:54:02 -07:00
Daniel Gu
0140e33564 Refactor pipeline based on review (change __init__ design for UniDiffuserTextDecoder, add text_tokenizer to __init__ for UniDiffuserPipeline) and add mode inference and mode setting functions to UniDiffuserPipeline. 2023-04-06 01:45:16 -07:00
Daniel Gu
afe5ba0f20 Initial commit for a image-text UniDiffuser pipeline. 2023-04-03 23:40:20 -07:00
Sayak Paul
7139f0e874 fix: norm group test for UNet3D. (#2959) 2023-04-04 09:01:15 +01:00
Patrick von Platen
8c530fc2f6 make style 2023-03-31 23:46:28 +02:00
Patrick von Platen
723933f5f1 add another import 2023-03-31 23:45:05 +02:00