diffusers
diffusers copied to clipboard
🤗 Diffusers: State-of-the-art diffusion models for image, video, and audio generation in PyTorch.
### Describe the bug When loading a model using `from_single_file()`, the RAM usage is really high possibly because the weights are loaded in FP32 before conversion. ### Reproduction ```python import...
# What does this PR do? In the `scale_noise` functions of all flow matching schedulers, the definitions and docs are wrong: * `timestep` should be a `torch.FloatTensor`, not a float...
## Changes - Fixed the previously skipped [test_solver_order_and_type](cci:1://file:///d:/Github/diffusers/tests/schedulers/test_scheduler_edm_dpmsolver_multistep.py:167:4-210:97) in [test_scheduler_edm_dpmsolver_multistep.py](cci:7://file:///d:/Github/diffusers/tests/schedulers/test_scheduler_edm_dpmsolver_multistep.py:0:0-0:0) - Properly separated testing of `dpmsolver++` and `sde-dpmsolver++` algorithms - Added comprehensive validation for all supported solver orders and types...
### Model/Pipeline/Scheduler description MAGI-1 is a video generation model that has achieved stunning visual effects. ### Open source status - [x] The model implementation is available. - [x] The model...
### Describe the bug I am trying to load lora model trained with flux-fill pipeline using `FluxControlInpaintPipeline`. But it is not able to load the lora model into transformers. Any...
### Sys env: OS Ubuntu 22.04 PyTorch 2.4.0+cu121 sana == 0.0.1 Diffusers == 0.34.0.dev0 ### Reproduce: Try the demo test code: ``` import torch from diffusers import SanaPAGPipeline pipe =...
# What does this PR do? Fixes # (issue) ## Before submitting - [x] This PR fixes a typo or improves the docs (you can dismiss the other checks if...
### Describe the bug `RuntimeError: Invalid buffer size: 635.13 GB` on Apple Silicon with mps, when the readme claims 60GB. Smaller res possible but still with 10x the memory claimed....
# What does this PR do? Adds a very mechanical conversion of other img2img pipelines (mostly SD3/Flux) to support AuraFlow. Seems to require a bit more strength (0.75+) compared to...
# What does this PR do? This PR adds support for Perturbed Attention Guidance (PAG) to the FluxPipeline Fixes #11488 ## Before submitting - [ ] This PR fixes a...