LTX-2 Text to Video fp4 quantized

Note: All model files and custom nodes below must be downloaded via the client. If you do not have the client, Download >>>

LTX-2 is an AI video generation model developed by Lightricks, which is very user-friendly for beginners. It can generate videos with audio in one go—for example, the character's lip movements match the speech, and the sound effect synchronizes with the visuals when an explosion happens, without the need for manual dubbing later. It can produce videos up to 4K high definition, with images clear enough for advertisements or short films. It supports generating videos using text, images, even sketches as materials, and allows adjusting details like camera angles and picture styles.

LTX-2 >5 files

Name	Size	Target Dir
ltx-2-19b-dev-fp4.safetensors	19.99 GB	checkpoints
ltx-2-19b-dev-fp8.safetensors	27.08 GB	checkpoints
ltx-2-19b-distilled-fp8.safetensors	27.13 GB	checkpoints
ltx-2-19b-dev.safetensors	43.29 GB	checkpoints
ltx-2-19b-distilled.safetensors	43.33 GB	checkpoints

ltx-2-19b-distilled-lora-384 >1 files

Name	Size	Target Dir
ltx-2-19b-distilled-lora-384.safetensors	7.67 GB	loras

LTX-2 19B LoRA - Camera Control (Dolly Left) >1 files

Name	Size	Target Dir
ltx-2-19b-lora-camera-control-dolly-left.safetensors	327.31 MB	loras

ltx-2-upscaler >2 files

Name	Size	Target Dir
ltx-2-temporal-upscaler-x2-1.0.safetensors	261.97 MB	latent_upscale_models
ltx-2-spatial-upscaler-x2-1.0.safetensors	995.77 MB	latent_upscale_models

gemma_3_12B_it >2 files

Name	Size	Target Dir
gemma_3_12B_it.safetensors	24.38 GB	text_encoders
gemma_3_12B_it_fp8_e4m3fn.safetensors	13.21 GB	text_encoders