It looks like Wan2.2 For-Loop Scenario, but only 1 step. Then you can use output video as input and continuew it. At any step if you don't like something you can change and retake only 1 scene. Also you can use different step count for steps.
Description
Create a start video with T2V model
FAQ
Comments (21)
I have absolutely no idea what this is for.
Take let me start and create first video, then take let-me-continue and join second (3, 4, 5...) part and you have a long video
@thesharque588 Yes but what is fusionx and lighting? The look like nodes, but they are loras?
@thesharque588 Where is let-me-continue?
@Superelmostar675 they are various lora's. You dont need them, just select bypass for most of them except for the ligntning 4 step lora which is kind of essential.
If you find a lora you like somewhere, you can unbypass and load it using those nodes and wont have to configure much
@dominic1336756 There 2 of them but I will update soon
For anyone wondering, single step/lora workflow lets you drag videos in and continue from the last frame. Repeat as many times as you want for longer and longer vids.
after seeing the complex workflows for 'endless' gens, I appreciate the simplicity. Very easy chill workflow to continue video's that's easy to modify. Exactly what I was looking for TY
Also I would recomens looks at let-me-end to create a "perfect cycle" : )))
@thesharque588 I noticed after 30 seconds or so the entire video started to get burned/saturated and the lighting was effected. Any way to disable it from effecting already generated frames?
@OpoenBorders5 Im my case (in scene Egypt pricess and the snake) nothing helps me but "color match" at the end of all glued. So I did:
1. Get all scene 1 by 1
2. Glue it in 1 complete movie
3. Color match all the frame at once
4. upscale a bit + interpolate x4 and up to 50 FPS to have "soap move effect"
But this movie was banned : ))))) I can try to upload again but IDK : )
I have a 5-second clip. If I use it with this workflow, will it extend for another 5 seconds?"
@Saxlive it extend for any count of frame that you wish... And then glue your video with a new one
im pretty new to it so i think u got this asked alot, is there maybe a general Q&A? Because i have often the Problem that the face just suddenly changes and doesnt look like the person anymore. I take any Hints & Tipps, (it worked once with identical face but the movements were wayyyy to fast somehow)
Good Workflow and thanks for sharing!
Hi, I tried running your workflow but it seems to require some specific LoRA files that I don’t have. Because of that, the output doesn’t get saved.
Do I need to download any special models or nodes for this workflow?
These are the files it tries to load but are missing on my side:
Wan2.1_I2V_14B_FusionX_LoRA.safetensors
Wan2.2-Lightning_I2V-A14B-4steps-lora_LOW_fp16.safetensors
Wan2.2-Lightning_I2V-A14B-4steps-lora_HIGH_fp16.safetensors
Hi, just wanted to say thank you! 🙏 Your workflow is really well-designed and easy to use. I managed to get it working and successfully continued my scene thanks to your setup. Really appreciate your effort in making and sharing this!
Sorry for this dumb question, I did first image and load in let_me_start, so I load the second workflow and use the video, when I'm satisfied with the result Have I to toggle the bypass to "final video" and use the new fused composition as initial video to continue and reapeat that until I ve enough videos fused?Is it correct?btw very good workflows, tyvm!
I'm stuck on the single step workflow, it wont run. The FusionX and Lighting nodes in high and low are highlighted in red. I have the loras linked in other comments. Anyone know the solution?
Dunno if you still need it....but those nodes you are talking about are just fixed nodes...you can disable them and load the fusionX and lighting nodes/loras manually in conventional lora loaders...its a bit of an inelegant workaround...but it works.
hope that does explain it.
@martinr573 this is helpful, thank you.
For fast setup:
--------------------------
wan22 let-me-start i2v.json
--------------------------
- clip_vision/clip_vision_h.safetensors
- vae/wan_2.1_vae.safetensors
- text_encoders/umt5-xxl-encoder-gguf
- diffusion_models/wan2.2_i2v_high_noise_14B_Q3_K_S.gguf
- diffusion_models/wan2.2_i2v_low_noise_14B_Q3_K_S.gguf
- loras/wan2.2-i2v-high-pov-insertion-v1.0.safetensors
https://civitai.com/models/1855263/pov-insertion-wan-2x
- loras/wan2.2-i2v-low-pov-insertion-v1.0.safetensors
https://civitai.com/models/1855263/pov-insertion-wan-2x
- loras/Wan2.2-T2V-Femaled-Vaginus-High-v1.safetensors
https://civitai.com/models/1888857?modelVersionId=2137972
- loras/Wan2.2-T2V-Femaled-Vaginus-Low-V1.safetensors
https://civitai.com/models/1888857?modelVersionId=2137989
- loras/NSFW-22-H-e8.safetensors
https://huggingface.co/rahul7star/wan2.2Lora/blob/main/NSFW-22-H-e8.safetensors
- loras/NSFW-22-L-e8.safetensors
https://huggingface.co/rahul7star/wan2.2Lora/blob/main/NSFW-22-L-e8.safetensors
- loras/wan-i2v-squirt-k3nk-234epo.safetensors
https://civitai.com/models/1527558/female-ejaculation-squirt-hun-or-wan-video-lora-k3nk
- loras/Wan2.1_I2V_14B_FusionX_LoRA.safetensors
https://civitai.com/models/1678575?modelVersionId=1900322
- loras/Wan2.2-Lightning_I2V-A14B-4steps-lora_HIGH_fp16.safetensors
- loras/Wan2.2-Lightning_I2V-A14B-4steps-lora_LOW_fp16.safetensors
- loras/Wan21_I2V_14B_lightx2v_cfg_step_distill_lora_rank64.safetensors
--------------------------
wan22 let-me-continue.json
--------------------------
- loras/squirt-shura-final-000003.safetensors
https://civitai.com/models/1388704?modelVersionId=1572438
- loras/Wan2.2 - I2V - Orgasm - HIGH 14B.safetensors
- loras/Wan2.2 - I2V - Orgasm - LOW 14B.safetensors
- loras/Wan2.2 - creampie anime CRM-FULL-EPOCH-80-HIGH.safetensors
https://civitai.com/models/1941545?modelVersionId=2197409
- loras/Wan2.2 - creampie anime CRM-FULL-EPOCH-80-LOW.safetensors
https://civitai.com/models/1941545?modelVersionId=2215731
- loras/Wan2.1_genitals_helper_v1.0_e219.safetensors