townwish-mindone-testing 0.2.0

ONE for all, Optimal generator with No Exception.

MindONE

This repository contains SoTA algorithms, models, and interesting projects in the area of multimodal understanding and content generation

ONE is short for "ONE for all"

News

Hello MindSpore from Stable Diffusion 3!

• mindone/diffusers now supports Stable Diffusion 3. Give it a try yourself!

  import mindspore
  from mindone.diffusers import StableDiffusion3Pipeline
  
  pipe = StableDiffusion3Pipeline.from_pretrained(
      "stabilityai/stable-diffusion-3-medium-diffusers",
      mindspore_dtype=mindspore.float16,
  )
  prompt = "A cat holding a sign that says 'Hello MindSpore'"
  image = pipe(prompt)[0][0]
  image.save("sd3.png")
  

supported models under mindone/examples

model	features
cambrian	working on it
minicpm-v	working on v2.6
internvl	working on v1.0 v1.5 v2.0
llava	working on llava 1.5 & 1.6
vila	working on it
pllava	working on it
hpcai open sora	support v1.0/1.1/1.2 large scale training with dp/sp/zero
open sora plan	support v1.0/1.1/1.2 large scale training with dp/sp/zero
stable diffusion	support sd 1.5/2.0/2.1, vanilla fine tune, lora, dreambooth, text inversion
stable diffusion xl	support sai style(stability AI) vanilla fine tune, lora, dreambooth
dit	support text to image fine tune
latte	support uncondition text to image fine tune
animate diff	support motion module and lora training
video composer	support conditional video generation with motion transfer and etc.
ip adapter	refactoring
t2i-adapter	refactoring

run hf diffusers on mindspore

mindone diffusers is under active development, most tasks were tested with mindspore 2.2.10 and ascend 910 hardware.

component	features
pipeline	support text2image,text2video,text2audio tasks 30+
models	support audoencoder & transformers base models same as hf diffusers
schedulers	support ddpm & dpm solver 10+ schedulers same as hf diffusers

TODO

• mindspore 2.3.0 version adaption
• hf diffusers 0.30.0 version adaption