Instructions to use SG161222/SPARK.Chroma_v1 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Diffusers
How to use SG161222/SPARK.Chroma_v1 with Diffusers:
pip install -U diffusers transformers accelerate
import torch from diffusers import DiffusionPipeline # switch to "mps" for apple devices pipe = DiffusionPipeline.from_pretrained("SG161222/SPARK.Chroma_v1", dtype=torch.bfloat16, device_map="cuda") prompt = "Astronaut in a jungle, cold color palette, muted colors, detailed, 8k" image = pipe(prompt).images[0] - Notebooks
- Google Colab
- Kaggle
- Local Apps
- Draw Things
- DiffusionBee
How the training going for the new chroma sparks?
#3
by sasori7713 - opened
How the training going for the new chroma sparks? It's been on 80% for weeks?
Hi!
I am currently testing new settings for two-stage training. This might resolve artifacts on the edges of the images.
I have almost finished refining the captions for the images. I still need to manually check and correct about 350 descriptions.
I think I will manage to sort everything out before February 25th and start training SPARK v1. However, the training might take almost twice as long because it will be performed first at 512 resolution, and then at 1024. The preview version was trained only at 768 and had an error in the training configuration.
awesome. keep going strong mate.