From 76b1ee3221d356096330acd4d43d41a9523b733e Mon Sep 17 00:00:00 2001 From: Harahan Date: Wed, 11 Feb 2026 09:39:00 +0800 Subject: [PATCH 1/2] docs: announce GenRL framework release - Add GenRL release announcement to Latest News section - Add prominent GenRL highlight after online service notice - Include links to GitHub repo and HuggingFace model collection - Emphasize multi-node multi-GPU training and call for stars - Update both English and Chinese README versions --- README.md | 4 ++++ README_zh.md | 4 ++++ 2 files changed, 8 insertions(+) diff --git a/README.md b/README.md index 11187885..9d040fc0 100755 --- a/README.md +++ b/README.md @@ -20,6 +20,8 @@ > 🌐 **Try it online now!** Experience LightX2V without installation: **[LightX2V Online Service](https://x2v.light-ai.top/login)** - Free, lightweight, and fast AI digital human video generation platform. +> 🎉 **NEW: GenRL is here!** Check out our new **[GenRL Framework](https://github.com/ModelTC/GenRL)** for training visual generation models with reinforcement learning! High-performance RL-trained checkpoints now available on **[HuggingFace](https://huggingface.co/collections/lightx2v/genrl)**. + > 👋 **Join our WeChat group! LightX2V Rotbot WeChat ID: random42seed** ## 🧾 Community Code Contribution Guidelines @@ -40,6 +42,8 @@ Besides the contributions from the LightX2V team, we have received contributions ## :fire: Latest News +- **February 11, 2026:** 🎉 We are excited to announce **[GenRL](https://github.com/ModelTC/GenRL)** - a scalable reinforcement learning framework for visual generation! GenRL enables training diffusion/flow models with multi-reward optimization (HPSv3, VideoAlign, etc.) using GRPO algorithm. We've released high-performance LoRA checkpoints trained with multi-node multi-GPU setup, demonstrating significant improvements in aesthetic quality, motion coherence, and text-video alignment. Check out our [model collection](https://huggingface.co/collections/lightx2v/genrl) on HuggingFace! ⭐ Give us a star if you find it useful! + - **January 20, 2026:** 🚀 We support the [LTX-2](https://huggingface.co/Lightricks/LTX-2) audio-video generation model, featuring CFG parallelism, block-level offload, and FP8 per-tensor quantization. Usage examples can be found in [examples/ltx2](https://github.com/ModelTC/LightX2V/tree/main/examples/ltx2) and [scripts/ltx2](https://github.com/ModelTC/LightX2V/tree/main/scripts/ltx2). - **January 6, 2026:** 🚀 We updated the 8-step CFG/step-distilled models for [Qwen-Image-2512](https://huggingface.co/Qwen/Qwen-Image-2512) and [Qwen/Qwen-Image-Edit-2511](https://huggingface.co/Qwen/Qwen-Image-Edit-2511). You can download the corresponding weights from [Qwen-Image-Edit-2511-Lightning](https://huggingface.co/lightx2v/Qwen-Image-Edit-2511-Lightning) and [Qwen-Image-2512-Lightning](https://huggingface.co/lightx2v/Qwen-Image-2512-Lightning) for use. Usage tutorials can be found [here](https://github.com/ModelTC/LightX2V/tree/main/examples/qwen_image). diff --git a/README_zh.md b/README_zh.md index 371f96e4..2d048a91 100755 --- a/README_zh.md +++ b/README_zh.md @@ -20,6 +20,8 @@ > 🌐 **立即在线体验！** 无需安装即可体验 LightX2V：**[LightX2V 在线服务](https://x2v.light-ai.top/login)** - 免费、轻量、快速的AI数字人视频生成平台。 +> 🎉 **新品发布：GenRL 来了！** 查看我们全新的 **[GenRL 框架](https://github.com/ModelTC/GenRL)**，使用强化学习训练视觉生成模型！高性能 RL 训练的 checkpoint 现已在 **[HuggingFace](https://huggingface.co/collections/lightx2v/genrl)** 开放下载。 + > 👋 **加入微信交流群，LightX2V加群机器人微信号: random42seed** ## 🧾 社区代码贡献指南 @@ -40,6 +42,8 @@ pre-commit run --all-files ## :fire: 最新动态 +- **2026年2月11日:** 🎉 我们很高兴宣布推出 **[GenRL](https://github.com/ModelTC/GenRL)** —— 一个用于视觉生成的可扩展强化学习训练框架！GenRL 支持使用 GRPO 算法对 diffusion/flow 模型进行多奖励优化训练（HPSv3、VideoAlign等）。我们已经发布了在多机多卡上训练的高性能 LoRA checkpoints，在美学质量、运动连贯性和文本-视频对齐等方面都有显著提升。欢迎查看我们在 HuggingFace 上的[模型合集](https://huggingface.co/collections/lightx2v/genrl)！⭐ 觉得有用的话欢迎给个 star！ + - **2026年1月20日:** 🚀 我们支持了[LTX-2](https://huggingface.co/Lightricks/LTX-2)音频-视频生成模型，包含CFG并行、block级别offload、FP8 per-tensor量化等先进特性。使用示例可参考[examples/ltx2](https://github.com/ModelTC/LightX2V/tree/main/examples/ltx2)和[scripts/ltx2](https://github.com/ModelTC/LightX2V/tree/main/scripts/ltx2)。 - **2026年1月6日:** 🚀 我们更新了[Qwen-Image-2512](https://huggingface.co/Qwen/Qwen-Image-2512)和[Qwen/Qwen-Image-Edit-2511](https://huggingface.co/Qwen/Qwen-Image-Edit-2511)的8步的CFG/步数蒸馏模型。可以在[Qwen-Image-Edit-2511-Lightning](https://huggingface.co/lightx2v/Qwen-Image-Edit-2511-Lightning)和[Qwen-Image-2512-Lightning](https://huggingface.co/lightx2v/Qwen-Image-2512-Lightning)下载对应的权重进行使用。使用教程参考[这里](https://github.com/ModelTC/LightX2V/tree/main/examples/qwen_image)。 From ae5101664b3c35921f3e21ced29c7eb3e78e8be0 Mon Sep 17 00:00:00 2001 From: Harahan Date: Wed, 11 Feb 2026 09:46:55 +0800 Subject: [PATCH 2/2] docs: announce GenRL framework release - Add GenRL release announcement to Latest News section - Add prominent GenRL highlight after online service notice - Include links to GitHub repo and HuggingFace model collection - Emphasize multi-node multi-GPU training and call for stars - Update both English and Chinese README versions --- README.md | 2 +- README_zh.md | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-) diff --git a/README.md b/README.md index 9d040fc0..886b75bc 100755 --- a/README.md +++ b/README.md @@ -42,7 +42,7 @@ Besides the contributions from the LightX2V team, we have received contributions ## :fire: Latest News -- **February 11, 2026:** 🎉 We are excited to announce **[GenRL](https://github.com/ModelTC/GenRL)** - a scalable reinforcement learning framework for visual generation! GenRL enables training diffusion/flow models with multi-reward optimization (HPSv3, VideoAlign, etc.) using GRPO algorithm. We've released high-performance LoRA checkpoints trained with multi-node multi-GPU setup, demonstrating significant improvements in aesthetic quality, motion coherence, and text-video alignment. Check out our [model collection](https://huggingface.co/collections/lightx2v/genrl) on HuggingFace! ⭐ Give us a star if you find it useful! +- **February 11, 2026:** 🎉 We are excited to announce **[GenRL](https://github.com/ModelTC/GenRL)** - a scalable reinforcement learning framework for visual generation! GenRL enables training diffusion/flow models with multi-reward optimization (HPSv3, VideoAlign, etc.) using GRPO algorithm. We've released high-performance LoRA checkpoints trained with multi-node multi-GPU setup, demonstrating significant improvements in aesthetic quality, motion coherence, and text-video alignment. Check out our [model collection](https://huggingface.co/collections/lightx2v/genrl) on HuggingFace! Give us a ⭐ if you find it useful! - **January 20, 2026:** 🚀 We support the [LTX-2](https://huggingface.co/Lightricks/LTX-2) audio-video generation model, featuring CFG parallelism, block-level offload, and FP8 per-tensor quantization. Usage examples can be found in [examples/ltx2](https://github.com/ModelTC/LightX2V/tree/main/examples/ltx2) and [scripts/ltx2](https://github.com/ModelTC/LightX2V/tree/main/scripts/ltx2). diff --git a/README_zh.md b/README_zh.md index 2d048a91..3ec55e1a 100755 --- a/README_zh.md +++ b/README_zh.md @@ -42,7 +42,7 @@ pre-commit run --all-files ## :fire: 最新动态 -- **2026年2月11日:** 🎉 我们很高兴宣布推出 **[GenRL](https://github.com/ModelTC/GenRL)** —— 一个用于视觉生成的可扩展强化学习训练框架！GenRL 支持使用 GRPO 算法对 diffusion/flow 模型进行多奖励优化训练（HPSv3、VideoAlign等）。我们已经发布了在多机多卡上训练的高性能 LoRA checkpoints，在美学质量、运动连贯性和文本-视频对齐等方面都有显著提升。欢迎查看我们在 HuggingFace 上的[模型合集](https://huggingface.co/collections/lightx2v/genrl)！⭐ 觉得有用的话欢迎给个 star！ +- **2026年2月11日:** 🎉 我们很高兴宣布推出 **[GenRL](https://github.com/ModelTC/GenRL)** —— 一个用于视觉生成的可扩展强化学习训练框架！GenRL 支持使用 GRPO 算法对 diffusion/flow 模型进行多奖励优化训练（HPSv3、VideoAlign等）。我们已经发布了在多机多卡上训练的高性能 LoRA checkpoints，在美学质量、运动连贯性和文本-视频对齐等方面都有显著提升。欢迎查看我们在 HuggingFace 上的[模型合集](https://huggingface.co/collections/lightx2v/genrl)！觉得有用的话欢迎给个 ⭐！ - **2026年1月20日:** 🚀 我们支持了[LTX-2](https://huggingface.co/Lightricks/LTX-2)音频-视频生成模型，包含CFG并行、block级别offload、FP8 per-tensor量化等先进特性。使用示例可参考[examples/ltx2](https://github.com/ModelTC/LightX2V/tree/main/examples/ltx2)和[scripts/ltx2](https://github.com/ModelTC/LightX2V/tree/main/scripts/ltx2)。