| domain | framepack-ai.com |
| summary | FramePack is a tool for practical video diffusion using consumer GPUs. It’s built around a research paper titled “Packing Input Frame Contexts in Next-Frame Prediction Models for Video Generation” by Lvmin Zhang and Maneesh Agrawala, published on Arxiv in 2025.
Key Features & Requirements:
* Supported GPUs: Supported by unspecified GPUs (details not provided). * VRAM Requirements: VRAM needs are not specified. * Inference Speed: Inference speed isn't detailed. * Operating System Support: Supports Windows and Linux. * Installation: Requires Python 3.10 and installation of PyTorch, torchvision, torchaudio, and the `requirements.txt` file via `pip`. The GUI is launched with `python demo_gradio.py`, supporting options like `--share`, `--port`, etc.
Additional Information:
* The project provides useful resources related to video generation and AI, including links to Kyncept, Vidione, Maxrankr, etc. |
| title | FramePack - Practical Video Diffusion on Consumer GPUs |
| description | Generate full-length 30 fps videos on laptops and desktops with only 6 GB VRAM. Open-source AI video platform for creators, researchers, and studios. |
| keywords | video, videos, motion, frame, prompts, image, attention, diffusion, full, results, quality, movements, more, python, images, dynamic, prompt |
| upstreams |
|
| downstreams |
|
| nslookup | A 172.67.203.92, A 104.21.90.180 |
| created | 2025-11-05 |
| updated | 2026-01-27 |
| summarized | 2026-01-27 |
|
|