WorkflowListedExternal

LTX2.3 GGUF Simple i2v + Last Frame 16GB/low vram — v3 — v4

Swapped to a finetuned diffusion model, cleaned up nodes.

VersionRelease 3187934 — ltxv23GGUFSimpleI2v_v4.jsonEstimated VRAMNot yet characterizedUnratedgrade

Runtime profile

Remaining hardware or download-size estimates are not available yet.

Requirements

10Eros_v1.4-Q4_K_S.gguf
Repository matched
Diffusion model · Hugging Face · vantagewithai/LTX2.3-10Eros-1.4-GGUF
gemma_3_12B_it_fp8_scaled.safetensors
Repository matched
Text encoder · Hugging Face · Comfy-Org/ltx-2
ltx-2-spatial-upscaler-x2-1.0.safetensors
Found
Upscaler · 950 MB · SAFETENSORS · FP8 (unverified) · Hugging Face · Lightricks/LTX-2
ltx-2.3-22b-dev_embeddings_connectors.safetensors
Found
Text encoder · 2.15 GB · SAFETENSORS · Hugging Face · unsloth/LTX-2.3-GGUF
ltx-2.3-22b-distilled-lora-dynamic_fro09_avg_rank_105_bf16.safetensors
File specified
LoRA · 2.41 GB · Hugging Face · Kijai/LTX2.3_comfy
LTX23_audio_vae_bf16.safetensors
Found
VAE · 348 MB · SAFETENSORS · BF16 (unverified) · Unknown
LTX23_video_vae_bf16.safetensors
Repository matched
VAE · Unknown
rife49.pth
Found
Checkpoint · 20.4 MB · PTH · Unknown

AI-generated commentary

AI-generated explanation based on source and configuration details. Suggestions are clearly labeled.

This LTX 2.3 ComfyUI workflow is for turning an input image and text prompt into an image-to-video clip, with a last-frame option and an upscale path.

Its 105-node graph includes image loading and text encoding, LTXV video processing, video combining, and image saving.

Use an image and text prompt as inputs; the graph includes nodes for combining video and saving images.

The workflow file is named ltxv23GGUFSimpleI2v_v4.json, and the listed version is v4.

Model files listed for this workflow: 10Eros_v1.4-Q4_K_S.gguf, gemma_3_12B_it_fp8_scaled.safetensors, ltx-2-spatial-upscaler-x2-1.0.safetensors, ltx-2.3-22b-dev_embeddings_connectors.safetensors, LTX23_audio_vae_bf16.safetensors, LTX23_video_vae_bf16.safetensors, and rife49.pth.
Node packs named in the workflow: comfyui-kjnodes, comfyui-videohelpersuite, rgthree-comfy, and was-ns.

The author says to choose one diffuser model, use both listed text encoders, and keep the distilled LoRA enabled.

The author warns that the comfyui_smznodes extension breaks LTXV workflows and says to disable it if installed.
Check the spatial upscaler file before setup: the workflow lists ltx-2-spatial-upscaler-x2-1.0.safetensors, while the author’s notes list ltx-2.3-spatial-upscaler-x2-1.1.safetensors.
Suggestion

Does this need to be edited?

Sources

1 source

CivitaiPrimaryVersion 3187934

Verified Nov 3