LIVE: Long-horizon Interactive Video World Modeling
Copyright 2026 the LIVE authors

This distribution includes or adapts components from Stability AI Generative
Models, CompVis Latent Diffusion, NVIDIA diffusion implementations, OpenAI
diffusion implementations, LPIPS/PerceptualSimilarity, pytorch-fid, Karpathy's
deep-vector-quantization, and Black Forest Labs FLUX. Their original copyright
notices and license terms remain applicable.

Stability AI Generative Models components are distributed under the MIT
license in licenses/Stability-AI-MIT.txt. CompVis Latent Diffusion components
are distributed under the MIT license in licenses/CompVis-LDM-MIT.txt.

The dataset loader was adapted from Boyuan Chen's research template and is
distributed under the MIT license in licenses/research-template-MIT.txt.

The diffusion files that carry NVIDIA Apache-2.0 headers retain those headers.
The remaining OpenAI diffusion-derived files are distributed under the MIT
license in licenses/OpenAI-MIT.txt.

The quantizer code credited to Andrej Karpathy is distributed under the MIT
license in licenses/Karpathy-DVQ-MIT.txt.

The MLP embedder adapted from Black Forest Labs FLUX and Martin Seitzer's
pytorch-fid inception implementation are distributed under Apache License 2.0;
see the root LICENSE file.

The LPIPS implementation in tvae/taming/lpips.py is derived from
https://github.com/richzhang/PerceptualSimilarity and is distributed under its
BSD license. See licenses/LPIPS-BSD-2-Clause.txt.