EVA3D

Visit Tool

EVA3D is an AI tool that enables compositional 3D human generation from 2D image collections. It allows users to create high-quality, unconditional 3D human models, requiring only 2D images for training.

Claim this tool

1View

At a glance

Pricing

Open Source

Free tier

Yes

API

Skill level

Technical

About

What is EVA3D?

EVA3D is an open-source AI tool designed for compositional 3D human generation, leveraging 2D image collections for training. This innovative model, recognized as an ICLR 2023 Spotlight, allows users to create high-quality, unconditional 3D human models without the need for extensive 3D data. The project provides both inference and training codes, making it accessible for researchers and developers. It supports generation on various datasets like DeepFashion, SHHQ, UBCFashion, and AIST, with pre-trained models available for download. The tool is built on Python and requires NVIDIA GPUs, with compatibility tested across several NVIDIA GPU models. It offers a robust solution for AI-driven 3D content creation, particularly for human models.

Best used for

Ideal for developers and data scientists who need to generate realistic 3D human models from 2D image data, train custom generative AI models, and explore novel pose generation. Especially valuable for researchers and content creators working on virtual reality, gaming, or animation projects requiring high-quality 3D human assets.

Common actions

generate 3D models

create human avatars

train generative AI

research 3D generation

workflowsopen-sourcedeepfakelow-code/no-code"AI Agents"github copilotface swappingcollaborationautomated workflow

Capabilities

Key features

Compositional 3D human generation
2D image collection training
Pre-trained model inference
DeepFashion training code
NVIDIA GPU support

Target Audience

developerdata scientistcontent creator

Integrations

Not yet documented

Pricing & Plans

Open Source

Free

FAQs

What kind of hardware is required to run EVA3D?

EVA3D requires NVIDIA GPUs for both inference and training. It has been tested on NVIDIA RTX2080Ti, V100, A100, and T4 for inference, and V100 and A100 for training. Anaconda is recommended for managing Python environments.

What datasets does EVA3D support for generation and training?

EVA3D supports inference on SHHQ, UBCFashion, and AIST datasets. For training, it provides codes for DeepFashion using a processed dataset. Support for additional datasets is planned for future releases.

How do I get started with EVA3D for inference?

To start inference, you need to download pre-trained models using the provided script and register/download SMPL models, placing them in the specified 'smpl_models' folder. Then, you can use the provided bash scripts for different datasets.

Trending

Subcategories trending in Content & Design

Image Generation AI Writing Assistants Audio & Music Video Generation Photo Editing Graphic Design

Trending

Explore

Browse AI tools by category

Content & Design Productivity & Business Coding & Development AI Agents & Automation Research & Education Wellness & Lifestyle Career Development Marketing & Growth Data & Analytics Customer Support & CX Finance E-commerce