diffusion
diffusion copied to clipboard
mosaicml
This PR includes several changes for training a transformer based model. Changes include: - Refactored model creation - Callback for logging activation norms - FP32 layernorms and attention - mu-Parameterization
Bumps [gradio](https://github.com/gradio-app/gradio) from 4.44.0 to 5.5.0. Release notes Sourced from gradio's releases. [email protected] Features #9875 8305ff8 - Adds .expand() and .collapse() events to gr.Accordion. Thanks @abidlabs! #9424 a1582a6 - Lite...
![dependabot[bot] avatar](https://avatars.githubusercontent.com/in/29110?v=4 "dependabot[bot] avatar"){kind=avatar}
This PR adds a Load Planner that renames target modules of LoRA so that we can take a Composer checkpoint without LoRA and then use it to load and train...
Adds ControlNet to the diffusion repo for both SDXL and SD2 style models. Some of the highlights of the work here: 1. Custom callback to handle initializing a ControlNet from...
Hi, @coryMosaicML Thank you for merging the recent PR https://github.com/mosaicml/diffusion/pull/155 related to Diffusion Transformer (DiT). I have a couple of questions regarding the use of DiT in this repository: 1....
Hi, I am looking for sample code on how to train models using Huggingface datasets. Does anyone have any examples or tips for training with Huggingface datasets that they could...
[Patch Diffusion](https://github.com/Zhendong-Wang/Patch-Diffusion/tree/main) can x2 training speed even on 256x256 ImageNet. If this works out between Mosaic Diffusion and Patch-Diffusion, that is potentially x10 cumulative boost. The issue is both have...
Could you help me with the Azure Machine configuration which I can use to train stable diffusion configuration. I am unable to signin to MosaicML as it says I have...
It seems like the repo is missing the yaml for training the autoencoder though the encoder training code is provided?
Hi, for example I am training a job using this [yaml](https://github.com/mosaicml/diffusion/blob/main/yamls/hydra-yamls/SD-2-base-512.yaml), how to do continue training if this job failed? Thanks.
