You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Thanks for your work, the model is impressive.
I’ve created a tool for prompt scheduling using LTXV, which allows for easy interpolation between two tensors to control the context of animation and create smooth transitions between two concepts. I’m using the CLIP model (t5_xxl_fp8) to tokenize each prompt and apply the results throughout the entire inference process.
However, the video output is quite strange and lacks coherence. When I use the same tensor for the entire process, everything works perfectly. But when I change the tensor every 3-5 frames, the rendered video becomes unusable.
Do you have another document or resource to find my solution
The text was updated successfully, but these errors were encountered:
Hey,
Thanks for your work, the model is impressive.
I’ve created a tool for prompt scheduling using LTXV, which allows for easy interpolation between two tensors to control the context of animation and create smooth transitions between two concepts. I’m using the CLIP model (t5_xxl_fp8) to tokenize each prompt and apply the results throughout the entire inference process.
However, the video output is quite strange and lacks coherence. When I use the same tensor for the entire process, everything works perfectly. But when I change the tensor every 3-5 frames, the rendered video becomes unusable.
Do you have another document or resource to find my solution
The text was updated successfully, but these errors were encountered: