LTX2.3 GGUF Simple i2v + Last Frame 16GB/low vram — v3 — v5

This LTX 2.3 ComfyUI workflow is for image-to-video creation: it takes an image, supports last-frame guidance and reference audio, and includes latent upscaling and a video-combine output.

Its graph includes LTXVLatentUpsampler, RIFE VFI, VHS_VideoCombine, and SaveImage nodes.

Named input nodes include LoadImage and LoadAudio; LTXVReferenceAudio, LTXVAddGuideMulti, and CLIPTextEncode are also present, with VHS_VideoCombine and SaveImage as output nodes.

The description offers two diffuser choices—10Eros_v1.4-Q4_K_S.gguf or ltx-2.3-22b-dev-Q4_K_S.gguf—and calls for both text encoders: gemma_3_12B_it_fp8_scaled.safetensors and ltx-2.3-22b-dev_embeddings_connectors.safetensors.

Listed model files include 10Eros_v1.4-Q4_K_S.gguf, gemma_3_12B_it_fp8_scaled.safetensors, lora_weights.safetensors, ltx-2-spatial-upscaler-x2-1.0.safetensors, and ltx-2.3-22b-dev_embeddings_connectors.safetensors; also ltx-2.3-22b-distilled-lora-dynamic_fro09_avg_rank_105_bf16.s

Other listed model files are LTX23_audio_vae_bf16.safetensors, LTX23_video_vae_bf16.safetensors, rife49.pth, and taeltx2_3.safetensors.

Named node requirements include Bookmark (rgthree), comfyui-kjnodes, comfyui-videohelpersuite, easy showAnything, ImageBatchMulti, ImageResizeKJv2, LTX2SamplingPreviewOverride, LTXVChunkFeedForward, rgthree-comfy, RIFE VFI, Ryuu_TokenCountTextBox, StringConstantMultiline, Unetứ

In the listed files, lora_weights.safetensors is optional; the description says the distilled LoRA is needed and should not be disabled.

The description names ltx-2.3-spatial-upscaler-x2-1.1.safetensors, while the workflow file list names ltx-2-spatial-upscaler-x2-1.0.safetensors; verify which matches the v5 graph before setup.

Suggestion · not verified

The description warns that comfyui_smznodes breaks LTXV workflows; if you use it, check compatibility before running this workflow.

Suggestion · not verified