Upload README.md

README.md

@@ -1,7 +1,16 @@
+---
+license: apache-2.0
+
+---
+
 <h1 align="center">SwiftVR: Real-Time One-Step Generative Video Restoration</h1>
 
+
+
 <p align="center"><img src="assets/teaser.avif" width="100%" alt="SwiftVR teaser"></p>
 
+
+
 > **SwiftVR** is the first generative video restoration model to reach **real-time 1080p streaming on a consumer-grade GPU** (≈26 FPS on a single RTX 5090), sustains **31 FPS at QHD (2560×1440)** and **14 FPS at 4K (3840×2160)** on a single H100, and streams at resolutions where every compared diffusion-based VR baseline runs out of memory.
 
 <p>
@@ -9,24 +18,31 @@
   <a href="https://github.com/H-oliday/SwiftVR/"><img src="https://img.shields.io/badge/Project-Page-1f8acb.svg?style=flat-square" alt="Project Page"></a>
   <a href="https://huggingface.co/H-oliday/SwiftVR"><img src="https://img.shields.io/badge/%F0%9F%A4%97%20HuggingFace-Model-ffce00.svg?style=flat-square" alt="HuggingFace"></a>
   <a href="https://github.com/H-oliday/SwiftVR/blob/main/LICENSE"><img src="https://img.shields.io/badge/License-Apache%202.0-green.svg?style=flat-square" alt="License"></a>
+
 </p>
 
+
+
 ## Updates
+---
+- [2026/06] Release the inference code and pretrained weights 🎉
+
+
+
 
-* [2026/06] Release the inference code and pretrained weights 🎉
 
 ## ✨ Highlights
+---
 
-* **Mask-free shifted-window self-attention (MFSWA).** Each spatial window is **pre-gathered into a dense tensor**, so every attention call reduces to a single standard scaled-dot-product (SDPA) call — *no attention mask, cyclic shift, or padding ever enters the graph*. This gives a **1.62× throughput gain over its full-attention teacher** at essentially identical quality, with **no dedicated sparse kernel**.
-* **Restoration-aware Autoencoder (ReAE).** A lightweight encoder–decoder jointly fine-tuned with the DiT in pixel space removes the heavy-3D-VAE / tiled-decoding bottleneck.
-* **Causal chunk-wise streaming.** A minimal causal protocol (no rolling KV cache, no overlapped DiT inference) bounds the temporal axis, confining the residual (\mathcal{O}(N^2)) cost to the spatial axes.
+- **Mask-free shifted-window self-attention (MFSWA).** Each spatial window is **pre-gathered into a dense tensor**, so every attention call reduces to a single standard scaled-dot-product (SDPA) call — *no attention mask, cyclic shift, or padding ever enters the graph*. This gives a **1.62× throughput gain over its full-attention teacher** at essentially identical quality, with **no dedicated sparse kernel**.
+- **Restoration-aware Autoencoder (ReAE).** A lightweight encoder–decoder jointly fine-tuned with the DiT in pixel space removes the heavy-3D-VAE / tiled-decoding bottleneck.
+- **Causal chunk-wise streaming.** A minimal causal protocol (no rolling KV cache, no overlapped DiT inference) bounds the temporal axis, confining the residual \(\mathcal{O}(N^2)\) cost to the spatial axes.
 
 
 ## 📊 Results
+---
 
-### Efficiency at 2560×1440
-
-Single H100, causal streaming, 24 frames.
+### Efficiency at 2560×1440 (single H100, causal streaming, 24 frames)
 
 | Metric | DOVE (tile) | SeedVR2-3B (tile)| FlashVSR-Tiny | **SwiftVR (Ours)** |
 |---|:---:|:---:|:---:|:---:|
@@ -38,12 +54,15 @@ Single H100, causal streaming, 24 frames.
 
 ### Qualitative comparison
 
-<img src="assets/qualitative.png" width="100%" alt="SwiftVR qualitative comparison">
+<img src="assets/qualitative.png" width="100%" alt="SwiftVR teaser">
+
+
 
 ## 🛠 Installation
+---
 
 ```bash
-git clone https://github.com/H-oliday/SwiftVR.git
+git clone https://github.com/Holiday/SwiftVR.git
 cd SwiftVR
 
 conda create -n swiftvr python=3.10 -y
@@ -59,35 +78,37 @@ pip install -e .
 <details>
 <summary><b>Hardware notes</b></summary>
 
-* **Server:** single H100-80G reproduces the QHD/4K numbers above.
-* **Consumer:** single RTX 5090 reaches ≈26 FPS at 1080p with the *same checkpoint* (default PyTorch SDPA path, bfloat16, causal chunk protocol).
-* No hardware-specific retraining or kernel rewrite is required on any platform.
-
+- **Server:** single H100-80G reproduces the QHD/4K numbers above.
+- **Consumer:** single RTX 5090 reaches ≈26 FPS at 1080p with the *same checkpoint* (default PyTorch SDPA path, bfloat16, causal chunk protocol).
+- No hardware-specific retraining or kernel rewrite is required on any platform.
 </details>
 
-## 🗂 Model Zoo
 
-| Model Name | Date    | Backbone       | Link                                                  |
-| ---------- | ------- | -------------- | ----------------------------------------------------- |
-| SwiftVR        | 2026.06 | Wan2.2-TI2V-5B | [🤗 HuggingFace](https://huggingface.co/H-oliday/SwiftVR) |
+
+## 🗂 Model Zoo
+---
+| Model Name | Date | Backbone | Link |
+|---|---|---|---|
+| SwiftVR | 2026.06 | Wan2.2-TI2V-5B | [🤗 HuggingFace](https://huggingface.co/H-oliday/SwiftVR) |
 
 ```bash
 huggingface-cli download H-oliday/SwiftVR --local-dir checkpoints/
 ```
 
-Expected checkpoint layout, where `checkpoints/` is the directory passed to `from_pretrained`:
+Expected checkpoint layout (the directory passed to `from_pretrained`):
 
-```text
+```
 checkpoints/
-├── reae.safetensors             # Restoration-aware Autoencoder weights
-├── prompt_embedding.safetensors # precomputed empty-prompt text embedding, key: "prompt_emb"
-└── transformer/                 # diffusers-format DiT
+├── reae.safetensors            # Restoration-aware Autoencoder weights
+├── prompt_embedding.safetensors# precomputed empty-prompt text embedding (key: "prompt_emb")
+└── transformer/                # diffusers-format DiT
     ├── config.json
     └── diffusion_pytorch_model.safetensors
 ```
 
-## 🚀 Quick Start
 
+## 🚀 Quick Start
+---
 ### Python API
 
 ```python
@@ -108,7 +129,8 @@ Tunable knobs include:
 * `quality`: 0–100, mapped to x265 CRF
 * `queue_size`: pipeline queue size
 
-### Streaming
+
+### Streaming (causal, chunk by chunk, no future frames)
 
 Causal, chunk-by-chunk restoration without future frames.
 
@@ -116,11 +138,11 @@ Causal, chunk-by-chunk restoration without future frames.
 session = pipe.stream(clip_len=24, resolution=(1920, 1080))
 
 for lq_chunk in read_chunks("low_quality.mp4", n=24):   # lq_chunk: [T, H, W, 3] uint8
-    hq = session.step(lq_chunk)                         # [1, T', 3, H', W'] in [0, 1], or None if buffered
+    hq = session.step(lq_chunk)        # [1, T', 3, H', W'] in [0, 1], or None if buffered
     if hq is not None:
         write(hq)
 
-tail = session.flush()                                  # flush the final buffered frames
+tail = session.flush()                 # flush the final buffered frames
 ```
 
 ### Command line
@@ -132,40 +154,31 @@ python scripts/inference.py \
   --checkpoint checkpoints/ \
   --upscale 4 \
   --clip-len 24 \
-  --dtype bfloat16
+  --dtype bfloat16 \
 ```
 
 Use `--png` to write a PNG sequence.
 
-## 📁 Repository Structure
-
-```text
-SwiftVR/
-├── README.md
-├── LICENSE
-├── requirements.txt
-├── setup.py
-├── scripts/
-│   └── inference.py              # CLI entry point, thin wrapper over SwiftVRPipeline
-└── swiftvr/
-    ├── __init__.py               # exports SwiftVRPipeline
-    ├── pipeline.py               # SwiftVRPipeline: from_pretrained / to / restore_video / stream
-    ├── runner.py                 # four-stage pipelined runner: reader → H2D → GPU → writer
-    ├── io.py                     # frame reading, GPU preprocessing, mp4 / PNG writing
-    ├── models/
-    │   ├── reae.py               # Restoration-aware Autoencoder
-    │   └── transformer.py        # DiT + mask-free shifted-window self-attention
-    └── streaming/
-        ├── chunk.py              # fixed-size causal chunk protocol
-        ├── tae.py                # streaming autoencoder with causal boundary state
-        └── dit.py                # one-step streaming DiT with fixed timestep and RoPE offsets
-```
+
+## 🎬 More Visual Results
+---
+> Full-length restored clips (low-quality input → SwiftVR, played back to back).
+
+
+<video src="https://huggingface.co/H-oliday/SwiftVR/resolve/main/assets/demo_1.mp4" controls width="100%"></video>
+
+<video src="https://huggingface.co/H-oliday/SwiftVR/resolve/main/assets/demo_2.mp4" controls width="100%"></video>
+
+<video src="https://huggingface.co/H-oliday/SwiftVR/resolve/main/assets/demo_3.mp4" controls width="100%"></video>
+
+
+
 
 ## 🙏 Acknowledgements
+---
+SwiftVR builds on [Wan2.2-TI2V-5B](https://github.com/Wan-Video), the lightweight autoencoder [TAEHV](https://github.com/madebyollin/taehv), and the [RealBasicVSR](https://github.com/ckkelvinchan/RealBasicVSR) degradation pipeline. We thank the authors of [DOVE](https://github.com/zhengchen1999/DOVE), [SeedVR2](https://github.com/ByteDance-Seed/SeedVR), and [FlashVSR](https://github.com/OpenImagingLab/FlashVSR) for releasing strong baselines, and the [UltraVideo](https://github.com/Tele-AI/UltraVideo) team for the training corpus.
 
-SwiftVR builds on [Wan2.2-TI2V-5B](https://github.com/Wan-Video), the lightweight autoencoder [TAEHV](https://github.com/madebyollin/taehv), and the [RealBasicVSR](https://github.com/ckkelvinchan/RealBasicVSR) degradation pipeline.
 
-We thank the authors of [DOVE](https://github.com/zhengchen1999/DOVE), [SeedVR2](https://github.com/ByteDance-Seed/SeedVR), and [FlashVSR](https://github.com/OpenImagingLab/FlashVSR) for releasing strong baselines, and the [UltraVideo](https://huggingface.co/datasets/APRIL-AIGC/UltraVideo) team for the training corpus.
 
 ## 📜 License
 
@@ -187,6 +200,3 @@ If you have any questions, feel free to reach out:
 
 * Email: [kakibluee@gmail.com](mailto:kakibluee@gmail.com)
 
-<div align="center">
-<sub>If SwiftVR is useful to your research or product, please consider giving it a ⭐.</sub>
-</div>