Image to video LTX2.0 [Accelerate]

Click the avatar in the upper right corner - Invitation Code - Enter invitation code: rh-v1408 to get 1000 RH Coins, and log in daily to get 100 coins. If you are interested or have modification needs, you can join the discussion QQ group: 149735439.Major update!! LTX2.3 Image-to-Video, click the link to go and fix subtitle and face issues: https://www.RunningHub.cn/ai-detail/2029954327451803650/?inviteCode=rh-v1408

The model has been fixed, but the running speed feels slightly slower than initially. If creating characters, Sigama must be switched. For LTX2.0 Image-to-Video [Accelerated], 720P generation time may take between 2-4 minutes, 1080P generation speed for 10s is estimated at 5-6 minutes, and 15s is estimated at around 7 minutes. 3K resolution is available for 5s, 10s not tested.

Auto Prompts: Currently requires simply inputting action + camera movement, but it is more controllable than before and supports speaking.

New ways to play with LTX2.0, below are applications of other LTX2. series:

LTX2.0 Start/End Frames + Lip Sync: https://www.RunningHub.cn/post/2016871878190702594/?inviteCode=rh-v1408
LTX2.0 + QwenTTS + Character Speaking: https://www.RunningHub.cn/post/2017601157362884610/?inviteCode=rh-v1408
LTX2.0 Start/End Frames: https://www.RunningHub.cn/post/2009910509788733441/?inviteCode=rh-v1408
LTX2.0 Audio Lip Sync [Digital Human / Singing / Film & TV]: https://www.RunningHub.cn/post/2011770027745222658/?inviteCode=rh-v1408
LTX2.0 Image-to-Video: https://www.RunningHub.cn/post/2008764920220880897/?inviteCode=rh-v1408

Friends who are interested are welcome to follow me
Youtube: https://www.youtube.com/@MoteGlow-%E5%B0%98%E5%85%89AI
Bilibili: https://space.bilibili.com/1285669374?spm_id_from=333.1387.0.0

Since Qwen reverse-prompting was still a bit slow, it has been improved, and the speed seems okay now. Writing prompts is still faster when connected to the API.

Camera LoRA selection has been updated and now supports all directions. Normal left and right movements are more natural; in close-ups or related situations, you need to select the corresponding camera, otherwise left and right movements will forcefully pull the frame away.

Although the prompt is built-in, friends who write well can actually write their own, which roughly includes: 1. Detailed Chinese scene description (including subject action, micro-environment dynamics, dialogue content and lip sync, specific camera movement trajectory) + English Quality Tags. 2. Environmental sound effects + action sound effects (describing sounds corresponding to the scene). 3. Tone: emotional words (omit this field if there is no dialogue). 4. Music style / instruments.