LTX-2 Full GGUF Workflow Setup Tutorial (Text, Image First/Last Frame, Custom Audio)
Автор: Atelier Darren
Загружено: 2026-01-19
Просмотров: 1813
As long as your RAM is large enough, you can run the workflow even on a 6GB GPU. The Reserved VRAM node helps prevent OOM (out-of-memory) errors.
The workflow is png file. If you cant load my workflow, update your comfyui...
Artifact noise can be fixed by swapping the VAE Loader to the KJ VAE Loader instead.
diffusion model (need distill lora):
ltx-2-19b-dev-fp8.safetensors
https://huggingface.co/Kijai/LTXV2_co...
MelBandRoformer_fp16.safetensors
https://huggingface.co/Kijai/MelBandR...
Unet (recommended):
ltx-2-19b-distilled_Q4_K_M.gguf
https://huggingface.co/Kijai/LTXV2_co...
text encoder:
gemma_3_12B_it_fp8_e4m3fn.safetensors
https://huggingface.co/GitMylo/LTX-2-...
ltx-2-19b-embeddings_connector_distill_bf16.safetensors
https://huggingface.co/Kijai/LTXV2_co...
VAE:
LTX2_audio_vae_bf16.safetensors
LTX2_video_vae_bf16.safetensors
https://huggingface.co/Kijai/LTXV2_co...
lora:
ltx-2-19b-distilled-lora-384.safetensors (if using dev gguf)
ltx-2-spatial-upscaler-x2-1.0.safetensors (if necessary)
https://huggingface.co/Lightricks/LTX...
All 4 workflow here:
https://huggingface.co/AtelierDarren/...
Доступные форматы для скачивания:
Скачать видео mp4
-
Информация по загрузке:
