Compare commits
3
Commits
| Author | SHA1 | Date | |
|---|---|---|---|
 Aryan Kumar
|
5754bce54b | ||
|
Aryan Kumar
|
805d2f7c2d | ||
|
Aryan Kumar
|
975e172ba6 |
@@ -8,24 +8,18 @@
|
||||
|
||||
**FastVideo is a unified post-training and real-time inference framework for accelerated video generation.**
|
||||
|
||||
## NEWS
|
||||
- `2026/10/06`: FastH3 V2 now runs on a single consumer machine: NVIDIA RTX 5090, RTX 4090 and RTX PRO 6000 GPUs, DGX Spark and Apple Silicon. We also release [FastH3 Trim](https://huggingface.co/FastVideo/FastVideo-FastH3-Trim-8-Step-NVFP4), an experimental pruned model that is 4.2× smaller than base H3 and runs in as little as 8 GB of GPU memory. Get the [models](https://huggingface.co/collections/FastVideo/fastvideo-fasth3) and read the [Blog](https://haoailab.com/blogs/fasth3-rtx/).
|
||||
## Recent news
|
||||
- `2026/10/21`: Release [FastH3 Omni](https://huggingface.co/FastVideo/FastVideo-FastH3-Omni-8-Step-V1) (Ref2VA): reference-conditioned video-and-audio generation from a prompt plus ordered image, video and audio references, in eight fused forwards. It runs on consumer Blackwell (RTX 5090, RTX PRO 6000, DGX Spark — [NVFP4 QAD](https://huggingface.co/FastVideo/FastH3-OmniRef-NVFP4-QAD)), on RTX 4090 (FP8) and on Apple Silicon (MLX). See the [cookbook](https://haoailab.com/FastVideo/cookbook/minimax-h3/) and the [inference guide](https://hao-ai-lab.github.io/FastVideo/inference/fasth3_omniref/).
|
||||
- `2026/10/06`: FastVideo now supports [Kandinsky 6](https://x.com/kandinskylab_ai/status/2107374635218055345) from Kandinsky Lab: text- and image-to-video with synchronized audio (base and 10-step distilled pi-Flow checkpoints) plus video super-resolution up to 4x. See the [Kandinsky 6 recipes](https://haoailab.com/FastVideo/cookbook/kandinsky6/).
|
||||
- `2026/10/06`: FastH3 V2 now runs on a single consumer machine: NVIDIA RTX 5090, RTX 4090 and RTX PRO 6000 GPUs, DGX Spark and Apple Silicon. We also release [FastH3 Trim](https://huggingface.co/FastVideo/FastVideo-FastH3-Trim-8-Step-NVFP4), an experimental pruned model that is 4.2× smaller than base H3 and runs in as little as 8 GB of GPU memory. Get the [models](https://huggingface.co/collections/FastVideo/fastvideo-fasth3) and read the [Blog](https://haoailab.com/blogs/fasth3-rtx/).
|
||||
- `2026/09/15`: Release [FastH3 8-Step V2](https://huggingface.co/FastVideo/FastVideo-FastH3-8-Step-V2), an eight-forward data-free DMD2 checkpoint distilled from MiniMax-H3 with 80% Video Sparse Attention. Run it with `examples/inference/basic/basic_fasth3_8step.py` or the [FastH3 8-Step V2 recipe](https://haoailab.com/FastVideo/cookbook/minimax-h3/).
|
||||
- `2026/09/01`: FastH3 now runs locally on Apple Silicon through MLX and on NVIDIA DGX Spark through CUDA 13, including two-Spark inference. Follow the [FastH3 recipes](https://haoailab.com/FastVideo/cookbook/minimax-h3/) and read the [Blog](https://haoailab.com/blogs/fasth3-local/).
|
||||
- `2026/08/27`: [FastH3 Preview v1](https://haoailab.com/blogs/fasth3-preview/) is an open-weight 4-step sparse-distilled MiniMax-H3 model for synchronized video-and-audio generation, developed in collaboration with [Nuva Lab](https://nuvalab.ai/) and the [NVIDIA FastGen team](https://github.com/NVlabs/FastGen). Download the recommended [VSA / Data-Free weights](https://huggingface.co/FastVideo/FastVideo-FastH3-4-step-Preview-v1-VSA-DataFree), or see the [full FastH3 collection](https://huggingface.co/collections/FastVideo/fastvideo-fasth3).
|
||||
- `2026/08/19`: FastVideo now supports MLX on Apple Silicon with [FastMetal-QAD](https://huggingface.co/collections/FastVideo/fastmetal), a family of 1.3B, 5B, and 14B models optimized for Mac. Follow the [MLX install guide](https://hao-ai-lab.github.io/FastVideo/getting_started/installation/mlx/) and read the [Blog](https://haoailab.com/blogs/fastmetal/).
|
||||
- `2026/06/23`: Release FastWan-QAD: 5s of Video generated in 1.8s E2E. See the [FastWan-QAD models](https://huggingface.co/FastVideo/FastWan-QAD-FP8-1.3B), [Attn-QAT training guide](https://haoailab.com/FastVideo/training/attn_qat/), and [blog](https://haoailab.com/blogs/fastwan-qad/).
|
||||
- `2026/03/17`: Release demo: Into the Dreamverse: Vibe Directing in FastVideo, check out the [Blog](https://haoailab.com/blogs/dreamverse/).
|
||||
- `2026/03/13`: Release demo: Create a 5s 1080p Video in 4.5s with FastVideo on a Single GPU, check out the [Blog](https://haoailab.com/blogs/fastvideo_realtime_1080p/).
|
||||
- `2025/11/19`: Release [CausalWan2.2 I2V A14B Preview](https://huggingface.co/FastVideo/CausalWan2.2-I2V-A14B-Preview-Diffusers) models, [Blog](https://hao-ai-lab.github.io/blogs/fastvideo_causalwan_preview/) and [Inference Code!](https://github.com/hao-ai-lab/FastVideo/blob/main/examples/inference/basic/basic_self_forcing_causal_wan2_2_i2v.py).
|
||||
- `2025/08/04`: Release [FastWan](https://hao-ai-lab.github.io/FastVideo/distillation/dmd) models and [Sparse-Distillation](https://hao-ai-lab.github.io/blogs/fastvideo_post_training/).
|
||||
|
||||
### More News
|
||||
|
||||
- `2025/06/14`: Release finetuning and inference code for [VSA](https://arxiv.org/pdf/2505.13389).
|
||||
- `2025/04/24`: [FastVideo V1](https://hao-ai-lab.github.io/blogs/fastvideo/) is released!
|
||||
- `2025/02/18`: Release the inference code for [Sliding Tile Attention](https://hao-ai-lab.github.io/blogs/sta/).
|
||||
- [All announcements and earlier milestones](https://hao-ai-lab.github.io/FastVideo/news/).
|
||||
|
||||
## Key Features
|
||||
|
||||
@@ -54,59 +48,23 @@ FastVideo has the following features:
|
||||
|
||||
## Getting Started
|
||||
|
||||
We recommend using [uv](https://docs.astral.sh/uv/) to create a clean environment. If you previously used Conda, switching to uv generally gives faster and more stable installs.
|
||||
Install FastVideo with [uv](https://docs.astral.sh/uv/):
|
||||
|
||||
```bash
|
||||
# Create and activate a new uv environment
|
||||
uv venv --python 3.12 --seed
|
||||
source .venv/bin/activate
|
||||
|
||||
# Install FastVideo on NVIDIA CUDA 12
|
||||
# NVIDIA CUDA 12; use cu130 on CUDA 13
|
||||
UV_TORCH_BACKEND=cu126 uv pip install fastvideo
|
||||
```
|
||||
|
||||
Use `UV_TORCH_BACKEND=cu130` on CUDA 13. Apple silicon users should follow the
|
||||
[MLX install guide](https://hao-ai-lab.github.io/FastVideo/getting_started/installation/mlx/).
|
||||
Platform guides: [NVIDIA GPUs](https://hao-ai-lab.github.io/FastVideo/getting_started/installation/gpu/), [DGX Spark / GB10](https://hao-ai-lab.github.io/FastVideo/getting_started/installation/spark/), [Apple Silicon MLX](https://hao-ai-lab.github.io/FastVideo/getting_started/installation/mlx/), [AMD / ROCm](https://hao-ai-lab.github.io/FastVideo/getting_started/installation/amd/) — or the [full installation guide](https://hao-ai-lab.github.io/FastVideo/getting_started/installation/).
|
||||
|
||||
> **On an Apple Silicon Mac?** Install with `uv pip install -e '.[mlx]'` from
|
||||
> a clone, then pick a recipe in the
|
||||
> [cookbook](https://haoailab.com/FastVideo/cookbook/). See the
|
||||
> [MLX install guide](https://hao-ai-lab.github.io/FastVideo/getting_started/installation/mlx/).
|
||||
|
||||
Please see our [docs](https://hao-ai-lab.github.io/FastVideo/getting_started/installation/) for more detailed installation instructions.
|
||||
|
||||
> **On an NVIDIA DGX Spark (GB10 / ARM64 + CUDA 13)?** There's no prebuilt ARM wheel for the FastVideo CUDA kernel, so it's an editable from-source install (`UV_TORCH_BACKEND=cu130 uv pip install -e .`, which compiles that kernel for you) rather than `UV_TORCH_BACKEND=cu130 uv pip install fastvideo`. A compatible prebuilt ARM64 FlashAttention wheel is available separately. Follow the [DGX Spark install guide](https://hao-ai-lab.github.io/FastVideo/getting_started/installation/spark/).
|
||||
|
||||
### Install with an AI coding agent
|
||||
|
||||
FastVideo is a monorepo with rich agent guidance (see [`AGENTS.md`](AGENTS.md)). If you use Claude Code, Cursor, or another coding agent, paste the prompt below — it detects your platform and follows the matching guide:
|
||||
|
||||
```text
|
||||
Install FastVideo (https://github.com/hao-ai-lab/FastVideo) into a fresh uv virtual environment.
|
||||
|
||||
1. Detect the platform: run `uname -m`, `nvidia-smi`, and `nvcc --version`.
|
||||
2. Read and follow the matching install guide exactly (in this repo, or at
|
||||
https://hao-ai-lab.github.io/FastVideo/getting_started/installation/):
|
||||
- NVIDIA GPU, x86_64 -> docs/getting_started/installation/gpu.md
|
||||
- NVIDIA DGX Spark / GB10, aarch64, CUDA 13 -> docs/getting_started/installation/spark.md
|
||||
- Apple Silicon, macOS -> docs/getting_started/installation/mlx.md
|
||||
3. Use uv for every step. If a command fails, debug it and tell me what you changed.
|
||||
4. Verify the result:
|
||||
python -c "import fastvideo, torch; print('cuda', torch.cuda.is_available())"
|
||||
fastvideo --help
|
||||
5. Report which platform you detected and any deviations you had to make.
|
||||
```
|
||||
To have a coding agent do the install, give it the [agent install prompt](https://hao-ai-lab.github.io/FastVideo/getting_started/installation/#install-with-an-ai-coding-agent).
|
||||
|
||||
## Sparse Distillation
|
||||
|
||||
For our sparse distillation techniques, please see our [distillation docs](https://hao-ai-lab.github.io/FastVideo/distillation/dmd/) and check out our [blog](https://hao-ai-lab.github.io/blogs/fastvideo_post_training/).
|
||||
|
||||
See below for recipes and datasets:
|
||||
|
||||
| Model | Sparse Distillation | Dataset |
|
||||
| ------------------------------------------------------------------------------------- | --------------------------------------------------------------------------------------------------------------- | -------------------------------------------------------------------------------------------------------- |
|
||||
| [FastWan2.1-T2V-1.3B](https://huggingface.co/FastVideo/FastWan2.1-T2V-1.3B-Diffusers) | [Recipe](https://github.com/hao-ai-lab/FastVideo/tree/main/examples/distill/Wan2.1-T2V/Wan-Syn-Data-480P) | [FastVideo Synthetic Wan2.1 480P](https://huggingface.co/datasets/FastVideo/Wan-Syn_77x448x832_600k) |
|
||||
| [FastWan2.2-TI2V-5B](https://huggingface.co/FastVideo/FastWan2.2-TI2V-5B-Diffusers) | [Recipe](https://github.com/hao-ai-lab/FastVideo/tree/main/examples/distill/Wan2.2-TI2V-5B-Diffusers/Data-free) | [FastVideo Synthetic Wan2.2 720P](https://huggingface.co/datasets/FastVideo/Wan2.2-Syn-121x704x1280_32k) |
|
||||
Sparse distillation reaches >50× denoising speedup. See the [distillation guide](https://hao-ai-lab.github.io/FastVideo/distillation/dmd/) and the [blog](https://hao-ai-lab.github.io/blogs/fastvideo_post_training/); the training recipes live under [`examples/distill/`](examples/distill/) with the [FastVideo Synthetic datasets](https://huggingface.co/FastVideo).
|
||||
|
||||
## Dreamverse — Realtime Video Generation & Editing
|
||||
|
||||
|
||||
@@ -1,4 +1,4 @@
|
||||
const NEWS_HEADINGS = ["news", "recent updates", "what's new", "latest news", "announcements"];
|
||||
const NEWS_HEADINGS = ["news", "recent news", "recent updates", "what's new", "latest news", "announcements"];
|
||||
|
||||
const PROJECT_HEADINGS = [
|
||||
"awesome work using fastvideo or our research projects",
|
||||
|
||||
@@ -8,6 +8,10 @@
|
||||
"href": "/fast-models",
|
||||
"icon": "boxes"
|
||||
},
|
||||
{
|
||||
"label": "News",
|
||||
"href": "/news"
|
||||
},
|
||||
{
|
||||
"label": "Get started",
|
||||
"icon": "rocket",
|
||||
|
||||
@@ -265,7 +265,7 @@ export async function transferDocs() {
|
||||
.replace(/\[([^\]]+)\]\([^)]+\)/g, "$1")
|
||||
.split(/:\s|\.\s/)[0]
|
||||
.replace(/\.$/, "");
|
||||
const href = links.at(-1)?.[2] ?? "https://github.com/hao-ai-lab/FastVideo#news";
|
||||
const href = links.at(-1)?.[2] ?? "https://hao-ai-lab.github.io/FastVideo/news/";
|
||||
const iso = date.replaceAll("/", "-");
|
||||
const label = new Intl.DateTimeFormat("en", {
|
||||
month: "short",
|
||||
@@ -277,7 +277,7 @@ export async function transferDocs() {
|
||||
});
|
||||
await writeFile(
|
||||
join(content, "_snippets/home-news.mdx"),
|
||||
`<section className="home-news not-prose" aria-labelledby="recent-news">\n<div className="home-news-heading"><h2 id="recent-news">Recent news</h2><a className="home-news-more" href="https://github.com/hao-ai-lab/FastVideo#news">All announcements →</a></div>\n<ul>\n${recent.join("\n")}\n</ul>\n</section>\n`,
|
||||
`<section className="home-news not-prose" aria-labelledby="recent-news">\n<div className="home-news-heading"><h2 id="recent-news">Recent news</h2><a className="home-news-more" href="https://hao-ai-lab.github.io/FastVideo/news/">All announcements →</a></div>\n<ul>\n${recent.join("\n")}\n</ul>\n</section>\n`,
|
||||
);
|
||||
await writeFile(join(content, "_snippets/home-projects.mdx"), asMdx(projects.trim()));
|
||||
await rm(join(content, "index.md"));
|
||||
|
||||
@@ -59,6 +59,15 @@ test("stops at the next heading and ignores other sections", () => {
|
||||
assert.doesNotMatch(news[1].text, /Awesome work/);
|
||||
});
|
||||
|
||||
test("accepts the Recent news heading used by the README", () => {
|
||||
const { news } = readReadmeSections(legacy.replace("## NEWS", "## Recent news"));
|
||||
assert.equal(news.length, 2);
|
||||
assert.deepEqual(
|
||||
news.map((item) => item.date),
|
||||
["2026/10/06", "2026/09/15"],
|
||||
);
|
||||
});
|
||||
|
||||
test("returns null when a section is missing", () => {
|
||||
assert.deepEqual(readReadmeSections("# FastVideo\n\n## Overview\n"), { news: null, projects: null });
|
||||
});
|
||||
|
||||
@@ -60,6 +60,29 @@ UV_TORCH_BACKEND=cu126 uv pip install -e .
|
||||
uv pip install flash-attn --no-build-isolation -v
|
||||
```
|
||||
|
||||
## Install with an AI coding agent
|
||||
|
||||
FastVideo is a monorepo with rich agent guidance (see
|
||||
[`AGENTS.md`](https://github.com/hao-ai-lab/FastVideo/blob/main/AGENTS.md)).
|
||||
If you use Claude Code, Cursor, or another coding agent, paste the prompt
|
||||
below; it detects your platform and follows the matching guide:
|
||||
|
||||
```text
|
||||
Install FastVideo (https://github.com/hao-ai-lab/FastVideo) into a fresh uv virtual environment.
|
||||
|
||||
1. Detect the platform: run `uname -m`, `nvidia-smi`, and `nvcc --version`.
|
||||
2. Read and follow the matching install guide exactly (in this repo, or at
|
||||
https://hao-ai-lab.github.io/FastVideo/getting_started/installation/):
|
||||
- NVIDIA GPU, x86_64 -> docs/getting_started/installation/gpu.md
|
||||
- NVIDIA DGX Spark / GB10, aarch64, CUDA 13 -> docs/getting_started/installation/spark.md
|
||||
- Apple Silicon, macOS -> docs/getting_started/installation/mlx.md
|
||||
3. Use uv for every step. If a command fails, debug it and tell me what you changed.
|
||||
4. Verify the result:
|
||||
python -c "import fastvideo, torch; print('cuda', torch.cuda.is_available())"
|
||||
fastvideo --help
|
||||
5. Report which platform you detected and any deviations you had to make.
|
||||
```
|
||||
|
||||
## Requirements
|
||||
|
||||
- **Python**: 3.10-3.12 is the tested and recommended range (the commands
|
||||
|
||||
@@ -0,0 +1,27 @@
|
||||
# News
|
||||
|
||||
All FastVideo announcements, newest first. The README carries the latest items; this page keeps the full history.
|
||||
|
||||
## Recent news
|
||||
|
||||
- `2026/10/21`: Release [FastH3 Omni](https://huggingface.co/FastVideo/FastVideo-FastH3-Omni-8-Step-V1) (Ref2VA): reference-conditioned video-and-audio generation from a prompt plus ordered image, video and audio references, in eight fused forwards. It runs on consumer Blackwell (RTX 5090, RTX PRO 6000, DGX Spark — [NVFP4 QAD](https://huggingface.co/FastVideo/FastH3-OmniRef-NVFP4-QAD)), on RTX 4090 (FP8) and on Apple Silicon (MLX). See the [cookbook](https://haoailab.com/FastVideo/cookbook/minimax-h3/) and the [inference guide](https://hao-ai-lab.github.io/FastVideo/inference/fasth3_omniref/).
|
||||
- `2026/10/06`: FastVideo now supports [Kandinsky 6](https://x.com/kandinskylab_ai/status/2107374635218055345) from Kandinsky Lab: text- and image-to-video with synchronized audio (base and 10-step distilled pi-Flow checkpoints) plus video super-resolution up to 4x. See the [Kandinsky 6 recipes](https://haoailab.com/FastVideo/cookbook/kandinsky6/).
|
||||
- `2026/10/06`: FastH3 V2 now runs on a single consumer machine: NVIDIA RTX 5090, RTX 4090 and RTX PRO 6000 GPUs, DGX Spark and Apple Silicon. We also release [FastH3 Trim](https://huggingface.co/FastVideo/FastVideo-FastH3-Trim-8-Step-NVFP4), an experimental pruned model that is 4.2× smaller than base H3 and runs in as little as 8 GB of GPU memory. Get the [models](https://huggingface.co/collections/FastVideo/fastvideo-fasth3) and read the [Blog](https://haoailab.com/blogs/fasth3-rtx/).
|
||||
- `2026/09/15`: Release [FastH3 8-Step V2](https://huggingface.co/FastVideo/FastVideo-FastH3-8-Step-V2), an eight-forward data-free DMD2 checkpoint distilled from MiniMax-H3 with 80% Video Sparse Attention. Run it with `examples/inference/basic/basic_fasth3_8step.py` or the [FastH3 8-Step V2 recipe](https://haoailab.com/FastVideo/cookbook/minimax-h3/).
|
||||
- `2026/09/01`: FastH3 now runs locally on Apple Silicon through MLX and on NVIDIA DGX Spark through CUDA 13, including two-Spark inference. Follow the [FastH3 recipes](https://haoailab.com/FastVideo/cookbook/minimax-h3/) and read the [Blog](https://haoailab.com/blogs/fasth3-local/).
|
||||
- `2026/08/27`: [FastH3 Preview v1](https://haoailab.com/blogs/fasth3-preview/) is an open-weight 4-step sparse-distilled MiniMax-H3 model for synchronized video-and-audio generation, developed in collaboration with [Nuva Lab](https://nuvalab.ai/) and the [NVIDIA FastGen team](https://github.com/NVlabs/FastGen). Download the recommended [VSA / Data-Free weights](https://huggingface.co/FastVideo/FastVideo-FastH3-4-step-Preview-v1-VSA-DataFree), or see the [full FastH3 collection](https://huggingface.co/collections/FastVideo/fastvideo-fasth3).
|
||||
- `2026/08/19`: FastVideo now supports MLX on Apple Silicon with [FastMetal-QAD](https://huggingface.co/collections/FastVideo/fastmetal), a family of 1.3B, 5B, and 14B models optimized for Mac. Follow the [MLX install guide](https://hao-ai-lab.github.io/FastVideo/getting_started/installation/mlx/) and read the [Blog](https://haoailab.com/blogs/fastmetal/).
|
||||
|
||||
## Earlier news
|
||||
|
||||
- `2026/06/23`: Release FastWan-QAD: 5s of Video generated in 1.8s E2E. See the [FastWan-QAD models](https://huggingface.co/FastVideo/FastWan-QAD-FP8-1.3B), [Attn-QAT training guide](https://haoailab.com/FastVideo/training/attn_qat/), and [blog](https://haoailab.com/blogs/fastwan-qad/).
|
||||
- `2026/03/17`: Release demo: Into the Dreamverse: Vibe Directing in FastVideo, check out the [Blog](https://haoailab.com/blogs/dreamverse/).
|
||||
- `2026/03/13`: Release demo: Create a 5s 1080p Video in 4.5s with FastVideo on a Single GPU, check out the [Blog](https://haoailab.com/blogs/fastvideo_realtime_1080p/).
|
||||
- `2025/11/19`: Release [CausalWan2.2 I2V A14B Preview](https://huggingface.co/FastVideo/CausalWan2.2-I2V-A14B-Preview-Diffusers) models, [Blog](https://hao-ai-lab.github.io/blogs/fastvideo_causalwan_preview/) and [Inference Code!](https://github.com/hao-ai-lab/FastVideo/blob/main/examples/inference/basic/basic_self_forcing_causal_wan2_2_i2v.py).
|
||||
- `2025/08/04`: Release [FastWan](https://hao-ai-lab.github.io/FastVideo/distillation/dmd) models and [Sparse-Distillation](https://hao-ai-lab.github.io/blogs/fastvideo_post_training/).
|
||||
|
||||
## Earlier milestones
|
||||
|
||||
- `2025/06/14`: Release finetuning and inference code for [VSA](https://arxiv.org/pdf/2505.13389).
|
||||
- `2025/04/24`: [FastVideo V1](https://hao-ai-lab.github.io/blogs/fastvideo/) is released!
|
||||
- `2025/02/18`: Release the inference code for [Sliding Tile Attention](https://hao-ai-lab.github.io/blogs/sta/).
|
||||
Reference in New Issue
Block a user