MagicDrive: Street View Generation with Diverse 3D Geometry Control
Paper • 2310.02601 • Published • 1
pip install -U diffusers transformers accelerate
import torch
from diffusers import DiffusionPipeline
# switch to "mps" for apple devices
pipe = DiffusionPipeline.from_pretrained("flymin/MagicDrive-424x800-450ep", dtype=torch.bfloat16, device_map="cuda")
prompt = "Astronaut in a jungle, cold color palette, muted colors, detailed, 8k"
image = pipe(prompt).images[0]MagicDrive
ArXiv | Web | License | GitHub
This repository contains high-resolution (424x800) driving-view image generation model checkpoint.
- We train this model by resuming from the 224x400 checkpoint at 200 epochs (actual training from 200 to 450 epochs).
- Please check
hydra/overrides.yamlfor command overrides in the training config.
MagicDrive: Street View Generation with Diverse 3D Geometry Control
Ruiyuan Gao1*, Kai Chen2*, Enze Xie3^, Lanqing Hong3, Zhenguo Li3, Dit-Yan Yeung2, Qiang Xu1^
1CUHK 2HKUST 3Huawei Noah's Ark Lab
*Equal Contribution ^Corresponding Authors
For more information, please refer to our github: https://github.com/cure-lab/MagicDrive
- Downloads last month
- -
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support
# Gated model: Login with a HF token with gated access permission hf auth login