NVIGI-For-Unity

KO | EN

gitlite — search

#python #react #typescript #llm #go #deep-learning #linux #docker #macos #nodejs #java #kubernetes #ios #openai

NVIGI-For-Unity

(★ 33)

Nvidia In-Game Inference Framework Adapt to Unity

NVIGI-For-Unity 최신버젼 다운로드

최종 버전 다운로드 (.zip)

// repository documentation

Was this content helpful?

★ 0 (0 ratings)

## Introduction
This project adapts the NVIDIA In-Game Inferencing (NVIGI) SDK for use in Unity, enabling the following features on your local machine:

- Automatic Speech Recognition (ASR)
- Text-to-Speech (TTS)
- Large Language Model (LLM / GPT)

## Installation

### 1. Download NVIGI SDK 1.4.0

1. Visit https://developer.nvidia.com/rtx/in-game-inferencing
2. Download “NVIDIA In-Game Inferencing (NVIGI) SDK 1.4.0”
3. After extracting, copy the official NVIGI Developer Pack to your project directory:
	 - Recommended: `Assets/Plugins/x86_64/nvigi/nvigi_pack_1_4_0`
	 - Or adjust to your own path, but keep the internal structure unchanged

(If you use this Unity project template directly, simply place the entire `nvigi` folder under `Assets/Plugins/x86_64/`.)

### 2. Unity DLL Setup

On Windows:

- For initial setup: double-click `set_unity_dll.bat`
- To reset/restore: double-click `unset_unity_dll.bat`

These scripts will set up or clean the required NVIGI DLLs for the Unity Editor/Player.

### 3. Check DLL Dependencies (Optional)

- Recommended tool: Dependencies  
	https://github.com/lucasg/Dependencies

If you encounter missing DLL issues, use this tool to check the dependency chain.

### 4. Multiple CUDA Versions (Optional)

If your machine has multiple CUDA versions, refer to:  
https://blog.csdn.net/qq_50677040/article/details/132131346

## Model Support

Currently validated models are shown below:

![Model Supports](https://raw.githubusercontent.com/doraemonaaaa/NVIGI-For-Unity/main/docs/model_supports.png)

You can download all models at once using the provided script:

- Run: `nvigi_pack_1_4_0/download_all_data.bat`

### Download a Single Model Manually (Optional)

You can also manually download a specific model, for example TTS:

- Example path:  
	`nvigi_pack_1_4_0/data/nvigi.models/nvigi.plugin.tts.asqflow-trt/{81320D1D-DF3C-4CFC-B9FA-4D3FF95FC35F}/download.bat`
- Simply double-click `download.bat` to download the corresponding model

## Examples & Results

Simple Unity scene examples are provided (see the `Examples` directory):

- `ASR Test.unity`: Speech Recognition
- `LLM Test.unity`: Large Language Model
- `TTS Test.unity`: Text-to-Speech

Screenshots of actual results:

### ASR Recognition Result

![ASR Results](https://raw.githubusercontent.com/doraemonaaaa/NVIGI-For-Unity/main/docs/ASRResults.png)

### LLM Dialogue Result

![LLM Results](https://raw.githubusercontent.com/doraemonaaaa/NVIGI-For-Unity/main/docs/LLMResults.png)

## Future Plans
- Integrate directly by dll, not exe
- RAG is not implemented, but I suggest implement it upon on LLM, not original RAG
- Audio2MetaHuman integration (Audio2Face / Audio2Gesture)

Feel free to modify scripts and scenes as needed. If you have questions, you can add further notes in the project.

NVIGI-For-Unity

NVIGI-For-Unity 최신버젼 다운로드

오늘의 추천 프로젝트

Was this content helpful?

Recent Feedback

NVIGI-For-Unity

NVIGI-For-Unity 최신버젼 다운로드

오늘의 추천 프로젝트

Was this content helpful?

Recent Feedback

Download README