Description:
This workflow is based on the LTX2.3 Distilled v1.1 large model and incorporates the VBVR Spatial Reasoning LoRA to generate videos from images, with automatic audio generation included.
The workflow execution is divided into two sampling stages. In the second stage, the workflow uses the official upscaling model to perform a secondary video enhancement based on the original input first frame, producing high-definition video results. During this second sampling stage, the workflow also integrates the IC Detail LoRA and Refocus LoRA to further improve video details.
Model Requirements
- ltx-2.3-22b-distilled-1.1_transformer_only_fp8_scaled.safetensors Location:
models\diffusion_models - Ltx2.3-Licon-VBVR-I2V-240K-R32.safetensors Location:models\loras
- ltx-2.3-22b-ic-lora-refocus.safetensors Location:models\loras
- ltx-2-19b-ic-lora-detailer.safetensors Location:models\loras
- gemma_3_12B_it_fp8_scaled.safetensors Location：models\clip
- ltx-2.3_text_projection_bf16.safetensors Location：models\text_encoders
- LTX23_audio_vae_bf16.safetensors Location：models\vae
- LTX23_video_vae_bf16.safetensors Location：models\vae
- ltx-2.3-spatial-upscaler-x2-1.1.safetensors Location：models\latent_upscale_models
