Name: EditAnything
Rating: 5 (371 reviews)
Author: NRDX

For each model, see the version details for more information. All the models here are experimental; they often won't work perfectly, the prompt will often need improvement, the input will often need alteration, or the parameters will need adjusting. Therefore, don't treat these models as something that should work perfectly right away for everything. They will only cease to be experimental when we discover a better way to do this. <a target="_blank" rel="ugc" href="https://huggingface.co/Alissonerdx/LTX-LoRAs">Alissonerdx/LTX-LoRAs</a> <a target="_blank" rel="ugc" href="https://huggingface.co/Alissonerdx/EditAnything">Alissonerdx/EditAnything</a> <a target="_blank" rel="ugc" href="https://buymeacoffee.com/nrdx">If this model was helpful to you in any way, please consider helping me continue creating more model for the price of a coffee.</a>

These are experimental versions; this is not v2, a continuation of v1. These two new versions are experiments I've been doing over the last month. The models are far from perfect, so please keep that in mind. In terms of motion transfer, it doesn't work well with fast-paced videos or scene cuts, so it's considered an experiment.Read the README on HuggingFace <a target="_blank" rel="ugc" href="https://huggingface.co/Alissonerdx/EditAnything">Alissonerdx/EditAnything · Hugging Face</a>

v0.1 - Edit + Motion T.

These are experimental versions; this is not v2, a continuation of v1. These two new versions are experiments I've been doing over the last month. The models are far from perfect, so please keep that in mind. Obs.: It works better with objects than with people.Read the README on HuggingFace <a target="_blank" rel="ugc" href="https://huggingface.co/Alissonerdx/EditAnything">Alissonerdx/EditAnything · Hugging Face</a>

v0.1 - RV2V

This is not a direct continuation of v1.0. It's a fresh training run with a different recipe, done in two stages:- Stage 1 — pretraining on image edit pairs. Training a video model on stills is admittedly not ideal, but it was a way to expand the editing vocabulary well beyond what a small video-only dataset could teach (broader Add / Remove / Replace / Change coverage than v1.0 had).- Stage 2 — video fine-tune with first-frame conditioning enabled. This brought the temporal prior back.In theory v0.1 can do everything v1.0 did. In practice, temporal consistency may be weaker on some edits because most of stage 1 happened on still images. Whether it actually beats v1.0 on a given task is something we genuinely don't know yet.This is still experimental

v0.1 - 2 Stages

This model was trained on 8,000 video pairs, and training is still ongoing for a few thousand more steps. It is still experimental, not trained with a fully professional production target, and the model may be updated unexpectedly as new checkpoints.The current goal is not final polished production quality, but to explore:<ul><li>edit-anything behavior</li><li>prompt-following</li><li>inference tradeoffs</li><li>synthetic dataset building, especially for style data</li></ul>The model was trained around four main prompt patterns:Add <code>Add a/an [subject/object] with [clear visual attributes], [precise location in the scene].</code>Remove <code>Remove the [subject/object] [location or identifying description].</code>Replace <code>Replace the [original subject/object] [location] with a/an [new subject/object] with [clear visual attributes].</code>Convert / Style <code>Convert the video into a [style name] style.</code>Workflow URL: <code>https://huggingface.co/Alissonerdx/LTX-LoRAs/blob/main/workflows/ltx23_edit_anything_v1.1.json</code>One important thing during inference is CFG.A good starting point is testing a distilled setup with CFG = 1. If the edit feels too weak or the model is not following the prompt well enough, increasing CFG can be the key. In some cases, increasing the LoRA strength to around 1.2 can also help.The workflow is also not fully optimized yet. It still needs more testing to find the best combination of:<ul><li>CFG</li><li>LoRA strength</li><li>number of steps</li><li>model combinations</li></ul>It may also be interesting to combine this model with other models and see what kinds of results emerge.If you can test it, please share your findings. Feedback on prompt behavior, edit strength, consistency, style transfer, and failure cases would be very helpful while training is still in progress.Another very important thing is that the Removal task should have a very clear direction indicating where you want to remove what you want to remove.Examples:Remove the black robot sitting at the table.Remove the person riding the electric scooter on the left.Remove the person with glasses and the microphone in the foreground.Remove the image of the green trees on the top left.Remove the woman and the smoking bottle.For example, if the object are in front, use foreground ... background, left, right, top, bottom.Another way to remove things that don't want to be removed is to simply add a mask, for example magenta, over the object you want to remove, and use this video as a guide. When writing the prompt, you write something like: "Remove object masked with the pink color." Sometimes this is much more precise than waiting for it to recognize what actually needs to be removed, because in this case the biggest indicator is the magenta-colored object.

v1.0

EditAnything

1678540566871 (1).jpg

physical violence

weapon violence

wide hips

revealing clothes

thick thighs

downblouse

convenient censoring

huge breasts

pg-13

corpses

suggestive

oral invitation

pg13

sexy

sexual situations

male nudity

disturbing

male swimwear or underwear

female swimwear or underwear

partial nudity

undressed

female nudity

breasts out

exposed female nipple

breast out

lingerie

male underwear

hair over breasts

female swimwear

gigantic breasts

no panties

graphic violence or gore

covered nipples

huge butt

strapless leotard

sitting on face

emaciated bodies

one breast out

nsfw

female underwear

nude

graphic male nudity

adult toys

illustrated explicit nudity

nudity

graphic female nudity

hentai

futanari

porn

sexual intent

genitals

peeing

vore

oral

sexual activity

anal

blowjob

dildo riding

incest

hanging

hate symbols

nazi party

white supremacy

diapers

scat

self injury

hate speech

urine

extremist

child on child

latex clothing

swimwear

bukkake

fellatio

cumshot

implied fellatio

eat_cum

cumdrip

cum in pussy

cum on face

after fellatio

cum on hair

cum on body

cum on tongue

cum on hands

cum in mouth

triple fellatio

autofellatio

fucked silly

cum on pussy