Skip to content

Official implementation of MAGREF: Masked Guidance for Any-Reference Video Generation

active 2025-05-292026-04-19 (UTC)

Partial coverage11,924 / 13,348 hourly files (89%) · 2 absent upstream · 1,420 failed, retryable2025-01-312026-08-10 (UTC)— sampled evenly across the window, so rankings and trends hold; absolute counts scale up.
Events
236
Pushes
27
Pull requests
0
Issues
14
Stars
169
Forks
7

Activity over time

Daily event counts in the loaded window

Line chart, 326 days from 2025-05-29 to 2026-04-19. Pushes: 27 total, peak 7 in a day. Pull requests: 0 total, peak 0 in a day. Issues: 14 total, peak 4 in a day. Comments: 12 total, peak 7 in a day. Stars: 169 total, peak 17 in a day.

  • Pushes
  • Pull requests
  • Issues
  • Comments
  • Stars

Stars, PRs, issues and forks are under-captured in the later part of this window. GH Archive progressively stopped capturing non-push events during 2026 — −95% or worse by the end of the window. Every series here except Pushes fades for that reason, so a decline above reflects the archive, not this repository. Pushes stay reliable throughout, so read them, and the contributor counts derived from them, as the real signal. Data health has the measurements.

Top contributors

Pushes, PRs, issues, reviews and comments — stars and forks excluded, so this is contribution rather than popularity

ContributorContributionsPushesPRsComments
MAGREF-Video382509
keetsky2001
sunjian20152000
mingtiannihao2002
gangxu8222000
Idmon1000
libai-lab1000
RobertAckleyLin1000
lionalyf101100
younestft1000
NielsRogge1000

Recent activity

Latest issues, pull requests and releases

  • Issue#21RobertAckleyLin2025-12-05 07:53
    Subject Disentanglement 只在训练做了吗,这版本的推理代码好像没看到这个设计
  • Issue#17libai-lab2025-08-04 15:01
    is support wan2.2?
  • Issue#16Idmon2025-07-15 14:22
    MAGREF-14B Pro release?
  • Issue#13younestft2025-06-29 03:07
    What FPS to use?
  • Issue#12MAGREF-Video2025-06-21 14:42
    看下我跑出的demo感觉很闪,比作者公布的效果差很多,这是为什么
  • Issue comment#12keetsky2025-06-20 08:14
    看下我跑出的demo感觉很闪,比作者公布的效果差很多,这是为什么
  • Issue comment#9mingtiannihao2025-06-20 07:52
    training problem!!!
  • Issue comment#9MAGREF-Video2025-06-19 15:35
    training problem!!!
  • Issue comment#1MAGREF-Video2025-06-19 15:30
    Release MAGREF model and benchmark dataset on Hugging Face
  • Issue comment#12MAGREF-Video2025-06-19 13:10
    看下我跑出的demo感觉很闪,比作者公布的效果差很多,这是为什么
  • Issue#2MAGREF-Video2025-06-19 12:46
    这个是个了不起的工作,请问什么时候开源训练和推理的代码啊?
  • Issue comment#12MAGREF-Video2025-06-19 12:45
    看下我跑出的demo感觉很闪,比作者公布的效果差很多,这是为什么
  • Issue#12keetsky2025-06-19 11:42
    看下我跑出的demo感觉很闪,比作者公布的效果差很多,这是为什么
  • Issue comment#10MAGREF-Video2025-06-19 06:37
    ckpt upload to modelscope?
  • Issue comment#9mingtiannihao2025-06-19 06:02
    training problem!!!
  • Issue#8sunjian20152025-06-19 05:45
    mask channel 为什么是 4
  • Issue comment#8MAGREF-Video2025-06-19 04:40
    mask channel 为什么是 4
  • Issue#8sunjian20152025-06-18 09:41
    mask channel 为什么是 4
  • Issue comment#7MAGREF-Video2025-06-17 13:54
    这样运行也报错:CUDA_VISIBLE_DEVICES=1,2,3,7 torchrun --nproc_per_node=4 generate.py \ --dit_fsdp --t5_fsdp --ulysses_size 4 \ --ckpt_dir ./ckpts/magref \ --save_dir ./samples \ --prompt_path ./assets/multi_id_v2.txt
  • Issue#4MAGREF-Video2025-06-17 13:28
    有报错哎,怎么解决?
  • Issue#5MAGREF-Video2025-06-17 13:27
    好像少了这个模型
  • Issue#7gangxu8222025-06-17 13:01
    这样运行也报错:CUDA_VISIBLE_DEVICES=1,2,3,7 torchrun --nproc_per_node=4 generate.py \ --dit_fsdp --t5_fsdp --ulysses_size 4 \ --ckpt_dir ./ckpts/magref \ --save_dir ./samples \ --prompt_path ./assets/multi_id_v2.txt
  • Issue#6gangxu8222025-06-17 12:52
    这个需要多大的显存卡啊?46g显存跑不起来
  • Issue comment#4MAGREF-Video2025-06-17 12:37
    有报错哎,怎么解决?
  • Issue comment#2MAGREF-Video2025-06-16 13:58
    这个是个了不起的工作,请问什么时候开源训练和推理的代码啊?

Totals cover only the window loaded into ClickHouse and count events, not GitHub's lifetime totals — 169 stars here means stars gained during the window, not the repo's star count.