Train real-time, playable video world models. ForgeWM is fully open and reproducible, with keyboard, mouse, and gamepad control. 🔨</p>\n","updatedAt":"2026-08-21T02:03:19.654Z","author":{"_id":"65df4f5af3efe60b06ea1409","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/65df4f5af3efe60b06ea1409/u-9mnMU8bNXvyQ8z1CG5m.png","fullname":"Xinye Li","name":"asdfo123","type":"user","isPro":false,"isHf":false,"isHfAdmin":false,"isMod":false,"followerCount":3,"isUserFollowing":false}},"numEdits":0,"identifiedLanguage":{"language":"en","probability":0.871147632598877},"editors":["asdfo123"],"editorAvatarUrls":["https://cdn-avatars.huggingface.co/v1/production/uploads/65df4f5af3efe60b06ea1409/u-9mnMU8bNXvyQ8z1CG5m.png"],"reactions":[],"isReport":false}}],"primaryEmailConfirmed":false,"paper":{"id":"2608.14022","authors":[{"_id":"6a829cdfb601d59c652814d7","user":{"_id":"65df4f5af3efe60b06ea1409","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/65df4f5af3efe60b06ea1409/u-9mnMU8bNXvyQ8z1CG5m.png","isPro":false,"fullname":"Xinye Li","user":"asdfo123","type":"user","name":"asdfo123"},"name":"Xinye Li","status":"claimed_verified","statusLastChangedAt":"2026-08-17T08:45:04.805Z","hidden":false},{"_id":"6a829cdfb601d59c652814d8","name":"Lingshuai Lin","hidden":false},{"_id":"6a829cdfb601d59c652814d9","name":"Lei Wang","hidden":false},{"_id":"6a829cdfb601d59c652814da","name":"Liuzhou Zhang","hidden":false},{"_id":"6a829cdfb601d59c652814db","name":"Jialin Cui","hidden":false},{"_id":"6a829cdfb601d59c652814dc","name":"Qingshan Li","hidden":false},{"_id":"6a829cdfb601d59c652814dd","name":"Guanchu Wang","hidden":false},{"_id":"6a829cdfb601d59c652814de","name":"Qingbin Liu","hidden":false},{"_id":"6a829cdfb601d59c652814df","name":"Xi Chen","hidden":false},{"_id":"6a829cdfb601d59c652814e0","name":"Jiang Bian","hidden":false},{"_id":"6a829cdfb601d59c652814e1","name":"Wai Lam","hidden":false}],"mediaUrls":["https://cdn-uploads.huggingface.co/production/uploads/65df4f5af3efe60b06ea1409/24zQVmFj2pY1wQeypzqqf.png"],"publishedAt":"2026-08-14T07:13:34.000Z","submittedOnDailyAt":"2026-08-21T00:00:00.000Z","title":"ForgeWM: Progressive Causal Training for Few-Step Action-Conditioned Video World Models","submittedOnDailyBy":{"_id":"65df4f5af3efe60b06ea1409","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/65df4f5af3efe60b06ea1409/u-9mnMU8bNXvyQ8z1CG5m.png","isPro":false,"fullname":"Xinye Li","user":"asdfo123","type":"user","name":"asdfo123"},"summary":"Action-conditioned video world models require low-latency causal generation and reliable responses to game-native controls. Although causal distillation enables one- or few-step video synthesis, extending it to interactive world models remains challenging, as discrete keyboard states and continuous mouse motion must remain aligned with temporally compressed latent chunks during causal training and autoregressive rollout. We introduce ForgeWM, a progressive framework that transforms a bidirectional action-conditioned video generator into efficient few-step world models through domain adaptation, teacher-forced causal training, causal consistency distillation, and on-policy distribution matching with a bidirectional teacher. The resulting budget-specialized students operate at steady-state denoising budgets of 1, 2, and 4 steps. ForgeWM further supports a dual-path deployment protocol combining latency-critical interaction with optional replay-time refinement, where the one-step student re-noises and refines its saved draft. On paired Minecraft trajectories, ForgeWM leads the evaluated systems in Imaging Quality, reference-aligned motion-profile agreement, action-sign accuracy, and mouse-control accuracy, while achieving the lowest reference LPIPS; the same four-stage recipe transfers to gamepad-controlled FPS gameplay. Replay-time refinement matches four-step reference quality while remaining roughly three times closer to the experienced trajectory than regeneration from noise. These results demonstrate ForgeWM's effectiveness for controllable few-step video generation.","upvotes":12,"discussionId":"6a829cdfb601d59c652814e2","projectPage":"https://asdfo123.github.io/ForgeWM/","githubRepo":"https://github.com/asdfo123/ForgeWM","githubRepoAddedBy":"user","ai_summary":"ForgeWM progressively distills bidirectional video generators into efficient few-step interactive world models with aligned discrete and continuous controls, supporting low-latency interaction and replay-time refinement.","ai_keywords":["action-conditioned video world models","causal distillation","causal consistency distillation","on-policy distribution matching","bidirectional teacher","few-step video synthesis","latent chunks","autoregressive rollout","replay-time refinement"],"ai_summary_model":"thinkingmachines/Inkling-Small","githubStars":76,"organization":{"_id":"6a47c76878735a616d396f98","name":"ForgeWM","fullname":"ForgeWM","avatar":"https://cdn-avatars.huggingface.co/v1/production/uploads/65df4f5af3efe60b06ea1409/Y0AXxJdaoTELKA0S9Lch7.png"}},"canReadDatabase":false,"canManagePapers":false,"canSubmit":false,"hasHfLevelAccess":false,"upvoted":false,"upvoters":[{"_id":"65df4f5af3efe60b06ea1409","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/65df4f5af3efe60b06ea1409/u-9mnMU8bNXvyQ8z1CG5m.png","isPro":false,"fullname":"Xinye Li","user":"asdfo123","type":"user"},{"_id":"668fcd97495674543600db0b","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/noauth/JfBeJ6gSXe379G9ejjrLR.jpeg","isPro":false,"fullname":"Gao Mingqi","user":"EchoMinkki","type":"user"},{"_id":"665abce1d599adace6d66674","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/665abce1d599adace6d66674/pd6IiwPkbT4ocOySleXga.jpeg","isPro":false,"fullname":"Selen Su","user":"selensu","type":"user"},{"_id":"687f853bb39262ba84f3eeff","avatarUrl":"/avatars/cdfc44fde8237f08f10192553fe5a075.svg","isPro":false,"fullname":"Junhao Shen","user":"shenjunhao","type":"user"},{"_id":"66d05337e62d6bbf50186c2f","avatarUrl":"/avatars/f5be15e754f0fbbb37d2cc5ea417f729.svg","isPro":false,"fullname":"Yijie Jin","user":"DrewJin0827","type":"user"},{"_id":"665c91e15b11dca02f0c5891","avatarUrl":"/avatars/49a4ee76c3edfe5b0916051a5ac4acfd.svg","isPro":false,"fullname":"Ye Huang","user":"henry-y1","type":"user"},{"_id":"6712510a164bbd72c709c428","avatarUrl":"/avatars/63e44aa1e9f7f6b1f3a53455f53e85a1.svg","isPro":false,"fullname":"Lingshuai Lin","user":"Lynn031108","type":"user"},{"_id":"620783f24e28382272337ba4","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/620783f24e28382272337ba4/zkUveQPNiDfYjgGhuFErj.jpeg","isPro":false,"fullname":"GuoLiangTang","user":"Tommy930","type":"user"},{"_id":"6858f7393794a7a08a254364","avatarUrl":"/avatars/aa2870bed47f7ef25356a4e9ff54ce38.svg","isPro":false,"fullname":"kexu Cheng*","user":"sediment1024","type":"user"},{"_id":"67efc3ebb34bbe2671b24c79","avatarUrl":"/avatars/f32619856d7f5f75cbeb1ea9556560fd.svg","isPro":false,"fullname":"lhz","user":"lhz191","type":"user"},{"_id":"65d6ac995e971572dafe2576","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/65d6ac995e971572dafe2576/4lDwi-j2M9V-myJF_PKyz.png","isPro":false,"fullname":"Shuaiyu Zhang","user":"ChefCurryGoat","type":"user"},{"_id":"66f4b264ee391e4c1f71f75f","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/noauth/vaky0_pbEpwupDrsjx_y5.jpeg","isPro":false,"fullname":"Wencan Liu","user":"HippoH","type":"user"}],"acceptLanguages":["en"],"dailyPaperRank":0,"organization":{"_id":"6a47c76878735a616d396f98","name":"ForgeWM","fullname":"ForgeWM","avatar":"https://cdn-avatars.huggingface.co/v1/production/uploads/65df4f5af3efe60b06ea1409/Y0AXxJdaoTELKA0S9Lch7.png"},"markdownContentUrl":"https://huggingface.co/buckets/huggingchat/papers-content/resolve/2608/2608.14022.md","query":{}}">

Get this paper in your agent:

Don't have the latest CLI?