---
license: creativeml-openrail-m
base_model: ptx0/pixart-900m-1024-ft-large
tags:
- stable-diffusion
- stable-diffusion-diffusers
- text-to-image
- diffusers
- full
- pixart
- pixart sigma
inference: true
library_name: diffusers
pipeline_tag: text-to-image
---
# realMODE PixArt Sigma 900M v1

## Inference

### ComfyUI
- Download model file `transformer/diffusion_pytorch_model.safetensors` and put into `ComfyUI/models/checkpoints`
- Use ExtraModels node: https://github.com/city96/ComfyUI_ExtraModels?tab=readme-ov-file#pixart

![image/png](https://cdn-uploads.huggingface.co/production/uploads/643c7e91b409fef15e0bd11b/MJfTShin1fYOOCo4mTv2-.png)

```python
import torch
from diffusers import DiffusionPipeline, EulerAncestralDiscreteScheduler
from diffusers.models import PixArtTransformer2DModel


model_id = "anyMODE/realMODE_900M"
negative_prompt = "malformed, disgusting, overexposed, washed-out"

pipeline = DiffusionPipeline.from_pretrained(model_id, torch_dtype=torch.float16)
pipeline.transformer = PixArtTransformer2DModel.from_pretrained(model_id, subfolder="transformer", torch_dtype=torch.float16)
pipeline.scheduler = EulerAncestralDiscreteScheduler.from_config(pipeline.scheduler.config)
pipeline.to('cuda' if torch.cuda.is_available() else 'cpu')

prompt = "On the left, there is a red cube. On the right, there is a blue sphere. On top of the red cube is a dog. On top of the blue sphere is a cat"
image = pipeline(
    prompt=prompt,
    negative_prompt='blurry, cropped, ugly',
    num_inference_steps=30,
    generator=torch.Generator(device='cuda' if torch.cuda.is_available() else 'mps' if torch.backends.mps.is_available() else 'cpu').manual_seed(1641421826),
    width=1024,
    height=1024,
    guidance_scale=5.5,
).images[0]
image.save("output.png", format="JPEG")
```