<h3 id="ltx2.3_flf2v_mod-simple-workflow-v2">LTX2.3_FLF2V_Mod simple workflow v2</h3><img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/b2ff95d2-aff4-404f-8852-0c9278deb336/original=true/b2ff95d2-aff4-404f-8852-0c9278deb336.jpeg" />Added gemma-3-12b-it-abliterated_heretic_lora_rank64_bf16.safetensors<img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/ec605631-2f47-418d-86ca-424c36c555c7/original=true/ec605631-2f47-418d-86ca-424c36c555c7.jpeg" /><a target="_blank" rel="ugc" href="https://huggingface.co/Comfy-Org/ltx-2/resolve/main/split_files/loras/gemma-3-12b-it-abliterated_heretic_lora_rank64_bf16.safetensors">gemma-3-12b-it-abliterated_heretic_lora_rank64_bf16.safetensors</a><hr /><h3 id="ltx2.3_flf2v_mod-simple-workflow-v1">LTX2.3_FLF2V_Mod simple workflow v1</h3>FLF2V (First-Last Frame Video Generation) suitable for short clips.<img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/be009b78-4101-4ca4-ab3d-76349a917a35/original=true/be009b78-4101-4ca4-ab3d-76349a917a35.jpeg" />This is a ComfyUI template workflow-based mod.Includes subgraphs and standard workflows.This workflow supportedhigh-speed upscaling of RTX Video Super ResolutionThe initial scale value is "2"(Within a subgraph node)<img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/316214ad-0931-4ebb-ac3d-39a262ccb84d/original=true/316214ad-0931-4ebb-ac3d-39a262ccb84d.jpeg" /><img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/472e3c6d-b0e3-4843-91e9-ca2ce3060088/original=true/472e3c6d-b0e3-4843-91e9-ca2ce3060088.jpeg" />The workflow incorporates ideas from matros99, thank you!<hr />Custom Nodes<a target="_blank" rel="ugc" href="https://github.com/Lightricks/ComfyUI-LTXVideo">https://github.com/Lightricks/ComfyUI-LTXVideo</a><a target="_blank" rel="ugc" href="https://github.com/rgthree/rgthree-comfy">https://github.com/rgthree/rgthree-comfy</a><a target="_blank" rel="ugc" href="https://github.com/kijai/ComfyUI-KJNodes">https://github.com/kijai/ComfyUI-KJNodes</a><a target="_blank" rel="ugc" href="https://github.com/yolain/ComfyUI-Easy-Use">https://github.com/yolain/ComfyUI-Easy-Use</a><a target="_blank" rel="ugc" href="https://github.com/Comfy-Org/Nvidia_RTX_Nodes_ComfyUI">https://github.com/Comfy-Org/Nvidia_RTX_Nodes_ComfyUI</a><hr />checkpoints<ul><li><a target="_blank" rel="ugc" href="https://huggingface.co/Lightricks/LTX-2.3-fp8/resolve/main/ltx-2.3-22b-distilled-fp8.safetensors">ltx-2.3-22b-distilled-fp8.safetensors</a></li></ul>text_encoders<ul><li><a target="_blank" rel="ugc" href="https://huggingface.co/Comfy-Org/ltx-2/resolve/main/split_files/text_encoders/gemma_3_12B_it_fp4_mixed.safetensors">gemma_3_12B_it_fp4_mixed.safetensors</a></li></ul><hr />tested on :ComfyUI version: 0.16.4, Python: 3.12.12, pytorch : 2.10.0+cu130Geforce RTX5060Ti16GB, 64GB RAMNote : For example, longer videos, such as 20 seconds long, resulted in an OOM error.<hr />In my method,For a consistent FLF image, I believe the key is to prepare a final image that maintains the consistency of the first image.First, prepare the first image.<img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/98f25e71-9b42-40a5-8b93-590c8eed65c5/original=true/98f25e71-9b42-40a5-8b93-590c8eed65c5.jpeg" />This qwen-image-edit2511 workflow creates a different situation.<a target="_blank" rel="ugc" href="https://civarchive.com/models/2252071">image_qwen_image_edit_2511_Multiangle Camera Workflow - v2.0 | Qwen Workflows | Civitai</a><img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/ed666c33-4a85-4d00-867b-b27fe0f84b2d/original=true/ed666c33-4a85-4d00-867b-b27fe0f84b2d.jpeg" />For now, I'll turn off camera control and keep Lora.<img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/3d864880-c03c-4225-bdc4-5672ca31eed1/original=true/3d864880-c03c-4225-bdc4-5672ca31eed1.jpeg" />Adjust Lora's strength as needed.prompt<pre><code>Black and white photograph, woman leaning forward over a wooden table at a street food stall, arched back, hands on the table, looking back over her shoulder with a happy open-mouth smile, In the background, a man standing very close behind her, his hand stroking and groping her buttocks, large close-up view of the man, he is smiling lewdly with his tongue sticking out, leaning his face cheek-to-cheek against the woman's cheek,
black and white</code></pre>We'll use FLF with these two images.<pre><code>- keeping the character in the frame centered.</code></pre><pre><code>- The Static camera, </code></pre><pre><code>- The woman is staring intently, smiling, she slowly grinds her hips seductively. The woman strikes a pose that emphasizes her buttocks.</code></pre><pre><code>- The fat man stroked his chin, let out an exaggerated sigh and made an OMG gesture.</code></pre><pre><code>- The man approaches, skipping along and says loudly, "Wow! Looks delicious sweet bun, hehehe"The man nodded while looking at the woman's buttocks.</code></pre><pre><code>- The man lewdly strokes and grabs the woman's buttocks and performs sexual acts on her, causing her to let out sexy moans.The woman lets out a sexy moan.The man was shaking his hips</code></pre><pre><code>- Then The woman whispers with sweet whisper,"It's soft and juicy inside..."</code></pre><pre><code>- They have an broken english with a chinese pronunciation.</code></pre><pre><code>- Music: traditional Chinese, Comedic, lighthearted music, trumpet</code></pre>

LTX2.3_FLF2V Simple Workflow v2 (distille model)

video_ltx2_3_flf2v_mod(Subgraph)_v2

Files with SHA256: 467d23d4153716a7e91b8da9c412fdbcd4061f1e6d3dfc2e8df59dab467d0d62