RunningHub
Search
Invite
Consistent Pose Change
Back
Consistent Pose Change
Consistent Pose Change
Consistent Pose Change
Consistent Pose Change
Consistent Pose Change
Consistent Pose Change

Consistent Pose Change

1607 Views1 Likes20 Favorites
CharacterPosture

我不吃番茄

我不吃番茄

CharacterPosture

Model Information

Active
Model Type：
LoRA
Basic Model：
F1基础-Kontext
Trigger Words：
Change the pose of the person in the left picture to the one in the right picture,
Resource Name：
models/loras/一致性稳定换姿势1.safetensors
MD5：
b372fbef462cbdef4d52938ee9ed94bf

Workflow Demo URL:https://www.RunningHub.cn/workflow/2033402926592434178

🌟 Key Highlights:

1️⃣ Background Locked, Pose Switched Only
No matter how complex the target pose is, the original background is steadily preserved, saving you the hassle of post-processing cutout and compositing.

2️⃣ Perfect Adaptation for Half-Body Photos
Even if your uploaded original image is only a half-body or face close-up shot, it can still precisely convert to the specified pose based on the reference image, offering exceptional versatility.

3️⃣ Minimalist Operation, High ID Consistency
Fully automated process, you only need to upload the [Original Image] and [Target Pose Image]. This feature mainly benefits from the built-in exclusive LoRA model, which can not only handle pose changes in various complex scenes, but also well maintain the consistency of the character's facial ID.

Workflow Demo URL:https://www.RunningHub.cn/workflow/2033402926592434178

🌟 Key Highlights:

1️⃣ Background Locked, Pose Switched Only
No matter how complex the target pose is, the original background is steadily preserved, saving you the hassle of post-processing cutout and compositing.

2️⃣ Perfect Adaptation for Half-Body Photos
Even if your uploaded original image is only a half-body or face close-up shot, it can still precisely convert to the specified pose based on the reference image, offering exceptional versatility.

3️⃣ Minimalist Operation, High ID Consistency
Fully automated process, you only need to upload the [Original Image] and [Target Pose Image]. This feature mainly benefits from the built-in exclusive LoRA model, which can not only handle pose changes in various complex scenes, but also well maintain the consistency of the character's facial ID.