diffusers

diffusers copied to clipboard

Add attentionless VAE support

1

# What does this PR do? Exposes the ```mid_block_add_attention``` parameter in Encoder and Decoder to enable attentionless VAEs using AutoencoderKL config . ## Before submitting - [x] Did you read...

Gothos

Fix Colab and Notebook checks for `diffusers-cli env`

16

This pull request updates the `is_google_colab` check to use an environment variable instead of checking for the presence of the `google.colab` module. This change ensures that the check is more...

tolgacangoz

# What does this PR do? https://github.com/InstantID/InstantID has more than 10k stars. So, makes sense to include it IMO.

sayakpaul

Tensor size mismatch for non pow of 2 sized image, SD3ControlNetModel

11

### Describe the bug There seems to be an issue with certain non power of 2 sized control net guidance images when using SD3ControlNetModel ### Reproduction ``` import diffusers import...

Teriks

Removing LoRAAttnProcessor causes many dependencies to fail

1

### Describe the bug https://github.com/huggingface/diffusers/pull/8623 removed obsolete `LoRAAttnProcessor` which in principle is a good thing, but it was done without considerations where is that feature currently in-use so it breaks...

vladmandic

[Core] introduce `controlnet` module

3

# What does this PR do? We have several ControlNet models now and I think it makes sense to now have a separate `controlnets` module under `src/diffisers/models`.

sayakpaul

Advanced training SD1.5 has an issue when saving checkpoints

5

### Describe the bug Today I trained using examples/dreambooth/train_dreambooth_lora.py in google colab, everything was OK I wanted to try examples/advanced_diffusion_training/train_dreambooth_lora_sd15_advanced.py I use the stable diffusion 1.5 model original (which I...

josemerinom

[advanced dreambooth lora] add clip_skip arg

1

adds `--clip_skip` to the advanced sdxl dreambooth lora trainer

linoytsaban

[Chore] move `vae_flax` to `autoencoders`

1

Following other modules.

sayakpaul

[Tests] fix sharding tests

1

# What does this PR do? So the following tests pass: ``` tests/models/unets/test_models_unet_2d.py::NCSNppModelTests::test_sharded_checkpoints tests/models/autoencoders/test_models_vae.py::ConsistencyDecoderVAETests::test_sharded_checkpoints ``` Some comments in line.

sayakpaul