LTX 2.3 Finetuned I2V

LTX 2.3 Finetuned Advanced I2V with identity & voice conditioning. Upload an image, write a prompt, generate.

Please beware this space may not represent LTX 2.3 properly, and it may be slow. This space stands as an opinionated feature test not an official LTX 2.3 demo (consider using Lightricks' official space for that use case). Also, all generated content must follow Huggingface policy.
preset (sets all lora, targeting, and sigma defaults)
1 41
0 1
snap to multiple
face mode
0 1
0 1
0 0.5
0 1
0 2
2 4
0 24
0.25 0.5
0 10
0 1
0 12
10 2000
0 1
20 540
0 540

export all current settings to a json file, or import a previously saved profile. images, prompts, and audio files are not included.