Temporal Prompt Engine: Local, Open-Source, Intuitive, Cinematic Prompt Engine + Video and Audio Generation Suite for Nvidia GPUs
##NOW FEATURING custom 12b HunYuanVideo script with Incorporated MMAudio for 80gb cards.
##MASSIVE UPDATE TO INSTRUCTIONS BELOW COMING VERY SOON (12/11/2024)
I am looking for a volunteer assistant if you're interested reach out at [email protected] - This is going to a webapp version VERY soon.
- Introduction
- Features Overview
- Installation
- Quick Start Guide
- API Key Setup
- Story Mode: Unleash Epic Narratives
- Inspirational Use Cases
- Harnessing the Power of ComfyUI
- Local Video Generation Using CogVideo
- Join the Temporal Labs Journey
- Donations and Support
- Additional Services Offered
- Attribution and Courtesy Request
- Contact
- Acknowledgments
Welcome to the Temporal Prompt Engine, your ultimate tool for crafting immersive video and audio experiences. This engine empowers you to generate high-quality prompts with unparalleled control over cinematic elements, all while being intuitive and accessible for users. I'm still currently experimenting with options a lot and will be honing the variety down a bit as I go.
Imagine capturing the world through the eyes of an ancient philosopher contemplating the cosmos, visualizing crypto-animals roaming digital landscapes, or peering into potential futures shaped by advanced technologies and societal shifts. With the Temporal Prompt Engine, these narratives are not just possible—they're straightforward and exhilarating to create.
This engine allows you to select standard film terms through user-friendly dropdown menus combined with a basic input concept. These selections are transformed into detailed lists of visual and audio prompts, ensuring that every aspect of your scene aligns perfectly with your creative vision. Additionally, the engine is capable of generating custom sound effects using the powerful AudioLDM2, giving you the tools to create truly unique and immersive audio landscapes.
- Cinematic Video Prompts: Tailor every aspect of your scene—from camera type and lens to lighting and framing.
- Adaptive Audio Prompts: Generate immersive soundscapes that perfectly match your visuals.
- Historical and Futuristic Perspectives: Choose from ancient viewpoints to futuristic visions, adding depth and context to your scenes.
- Dynamic Variables: Adjust settings like resolution, frame rate, and aspect ratio to guide the AI in generating content that matches your vision.
- Special Modes:
- Story Mode: Seamlessly blend prompts across frames to create cohesive narratives, enabling you to craft epic stories or intimate tales.
- Holiday Mode: Generate seasonal content tailored to holidays, perfect for festive branding and marketing.
- Chaos Mode: Add unpredictability with Chaos Mode.
- Interconnected Settings: Experience how choices like selecting an ancient art style influence other variables like color palette and texture, creating a cohesive and authentic output.
- Cross-Platform Compatibility: Available for Windows, with Linux support coming soon.
Setting up the Temporal Prompt Engine is simple and hassle-free, allowing you to focus on creation rather than configuration. This section provides installation steps for Windows users.
- Operating System: Windows 10 or later
- Python: Version 3.10.9
- Download Python Installer:
- Visit the Python Downloads page.
- Download the Python 3.10.9 installer for Windows.
- Install Python:
- Run the Installer:
- Double-click the downloaded EXE installer file.
- Important:
- If you already have Python installed:
- Be cautious when adding Python to PATH, as it may overwrite your existing Python version in the system PATH.
- To avoid conflicts, you can install Python 3.10.9 without adding it to the PATH. Instead, use the Python Launcher for Windows (
py
) to specify the version when running commands.
- If you don't have Python installed:
- Check the box "Add Python to PATH" at the bottom of the installer window.
- If you already have Python installed:
- Click "Install Now".
- Run the Installer:
- Verify Installation:
- Open a new Command Prompt.
- Run:
py -3.10 --version
- You should see
Python 3.10.9
.
- Download Python Installer:
##If you want the Video Generation capabilities to work you will also need a local Python 3.12.4 Installed as well. If this is in place the setup.py will handle everything. ##If you only want to use the prompt engine and sound effects generator then you are fine with only 3.10.9 even if the setup.py complains a little.
-
Operating System: Windows 10 or later
-
Python: Version 3.12.4
- Download Python Installer:
- Visit the Python Downloads page.
- Download the Python 3.12.4 installer for Windows.
- Install Python:
- Run the Installer:
- Double-click the downloaded EXE installer file.
- Important:
- If you already have Python installed:
- Be cautious when adding Python to PATH, as it may overwrite your existing Python version in the system PATH.
- To avoid conflicts, you can install Python 3.10.9 without adding it to the PATH. Instead, use the Python Launcher for Windows (
py
) to specify the version when running commands.
- If you don't have Python installed:
- Check the box "Add Python to PATH" at the bottom of the installer window.
- If you already have Python installed:
- Click "Install Now".
- Run the Installer:
- Verify Installation:
- Open a new Command Prompt.
- Run:
py -3.12 --version
- You should see
Python 3.12.4
. - Operating System: Windows 10 or later
- Py -3.12 is used by important video generation environment back-end processes.
- Download Python Installer:
-
Git: Installed and configured
- Install Git:
- Download Git from Git for Windows.
- Run the installer with default settings.
- Verify Installation:
- Open a new Command Prompt.
- Run:
git --version
- You should see the installed Git version.
- Install Git:
-
NVIDIA GPU: CUDA-enabled NVIDIA GPU for optimal performance
- Ensure you have an NVIDIA GPU with the latest drivers.
-
CUDA Toolkit: Version 11.8 compatible with your GPU and installed
- Install CUDA Toolkit:
- Download the CUDA Toolkit from CUDA Toolkit Download.
- Run the installer and follow the on-screen instructions.
- Verify Installation:
- Open a new Command Prompt and run:
nvcc --version
- You should see the CUDA compilation tools version information.
- Open a new Command Prompt and run:
- Install CUDA Toolkit:
-
FFmpeg: Installed and added to system
PATH
- Step 1: Click here to download the zip file of the latest version.
- Step 2: Unzip this file by using any file archiver such as Winrar or 7z.
- Step 3: Rename the extracted folder to ffmpeg and move it into the root of C: drive.
setx /m PATH "C:\ffmpeg\bin;%PATH%"
- Verify Installation:
- Open a new Command Prompt and run:
ffmpeg -version
- You should see FFmpeg version information.
- Open a new Command Prompt and run:
-
Ollama: Download from Ollama Setup and follow the on-screen instructions to install.
##[Setup Video Guide Part 2)[https://youtu.be/_8TXLNwA9ak?si=EO73EDqXafBLqA_O]
Clone the Repo
git clone https://github.com/TemporalLabsLLC-SOL/TemporalPromptEngine.git
Manually Download the Repository:
- Visit the TemporalPromptGenerator GitHub Repository.
- Click on the "Code" button and select "Download ZIP".
- Extract the downloaded ZIP file to your desired location (e.g.,
C:\TemporalPromptEngine
).
EASY-ONE-CLICK-INSTALLER
Open the Extracted Archive (or cloned repo) and click
RUN-FIRST-PRE-FLIGHT-CHECK.bat
- Follow the on-screen prompts. The script will automatically set up the python environment(s), install necessary packages, and configure settings.
The application will launch, guiding you through the initial setup. DURING THIS SETUP, either after you close the app OR denying to open the open during setup, you will have the option to add a work-around shortcut to your desktop. It currently does not have an icon. That will load the env and scripts reliably going forward.
OR
Manually Run the Setup Script:
- Open Command Prompt and navigate to the extracted
TemporalPromptEngine
directory:cd C:\TemporalPromptEngine-main
- Navigate to the
WINDOWS
folder:cd WINDOWS
- Run the setup script:
py -3.10 SETUP.py
IF YOU ADDED THE SHORTCUT TO DESKTOP DURING SETUP
Click the Temporal Prompt Engine Shortcut on your Windows Desktop
IF YOU DID NOT ADD SHORTCUT TO DESKTOP DURING SETUP
Activate the Virtual Environment:
cd C:\TemporalPromptEngine-main
TemporalPromptEngineEnv\Scripts\activate
Launch the Application:
python TemporalPromptEngine.py
-
Enter API Key in the Application:
- Paste your HuggingFace API key when prompted during the setup of the Temporal Prompt Engine.
- You only will need to enter this one time within a setup engine environment. You can simply close this popup in subsequent uses.
-
Enter Your Scene Concept:
Input your creative idea or scene description (up to 400 characters).
Examples:
- Ancient Perspectives:
"View the world through the eyes of an ancient astronomer, mapping the stars with rudimentary tools under a vast, unpolluted night sky."
- Crypto-Animal Footage:
"Documentary-style footage capturing the elusive 'cryptolion,' a mythical creature that embodies digital patterns and luminescent fur."
-
Configure Video and Audio Prompt Options: Tailor your video and audio prompts using dropdowns and input fields.
-
Generate Video and Audio Prompts:
- Click Generate Video Prompts.
- Click Generate Audio Prompts.
-
Save and Export: All media and prompts are saved in your designated output directory.
Story Mode allows you to create cohesive stories by blending prompts across frames. Adjust variables like pacing, lighting, and camera movement to reflect the progression of your narrative.
Explore the full potential of the Temporal Prompt Engine with these unique configurations:
-
Ancient Philosophical Journey:
"Through the eyes of an ancient philosopher, journey from the bustling streets of Athens to the serene temples of the Far East."
-
Crypto-Animal Documentary:
"Track the elusive 'cryptowolf' through digital forests where code and nature intertwine."
-
Visions of the Future:
"A day in the life of a future city powered by renewable energy and AI."
You can generate videos locally using CogVideo directly through the Temporal Prompt Engine. The engine now simplifies the entire CogVideo process for video generation.
-
Select CogVideo Version:
-
Click Local Video Generation from the Temporal Prompt Engine interface.
-
Enter the desired version of CogVideo (e.g., CogVideoX or another available version).
-
On Lower Vram (20gb-ish) you may benefit from running the script individually outside of the python app to do that follow the instructions below: Navigate to VideoGeneratorUtilities Folder
Open VideoGeneratorUtilities folder within the cloned repo directory
Run 5b Video Generation Script
One-Click-Utility-TemporalCog-5b
-
-
Prompt List Selection:
- Once the version is entered, the engine will initialize.
- Simply select the video prompt list to generate the videos using the settings you provide.
-
Automated Processing:
- The Temporal Prompt Engine manages the entire video generation process, allowing you to focus on creativity rather than configuration.
- Videos will be generated locally using your NVIDIA GPU for optimal performance. You will have the opportunity to entered your desired Guidance Scale and Steps during this intial setup.
Support the mission of pushing the boundaries of AI and technology. Join as an investor, developer, or client.
If you'd like to support the project, you can donate via the following crypto addresses:
- Ethereum (ETH):
0x5616b3415ED6Ea7005595eF144A2054d4cD5767B
- Bitcoin (BTC):
bc1qpsfn8a7cs75fxxwv3ax7gtnurm44n5x2fmh59c
- Solana (SOL):
FVPGxfGT7QWfQEWvXpFkwdgiiKFM3VdvzNG6mEmX8pgi
- Litecoin (LTC):
ltc1qwlyjz8aymy9uagqhht5a4kaq06kmv58dxlzyww
- Dogecoin (DOGE):
DAeWAroHCy8nXCoUsobderPRSNXNu1WY34
- Venmo:
@Utah-DM
- Tutoring, Development, Design, Consulting, and Workshops are available to meet your AI and technology needs.
If you create videos or content using prompts generated by the Temporal Prompt Engine, we kindly request that you include the phrase "using Temporal Prompt Engine" in your final release.
For questions, support, collaboration opportunities, or to discuss how we can work together:
- Email: [email protected]
- Phone: +1-385-222-9920
Thanks to the developers and communities behind Git, Python, FFmpeg, Ollama, AudioLDM2, CogVideo, ComfyUI, and HuggingFace for making this project possible.