Instructions to use Alex995647/loras-wan-2.2 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Diffusers
How to use Alex995647/loras-wan-2.2 with Diffusers:
pip install -U diffusers transformers accelerate
import torch from diffusers import DiffusionPipeline # switch to "mps" for apple devices pipe = DiffusionPipeline.from_pretrained("Wan-AI/Wan2.2-T2V-A14B", dtype=torch.bfloat16, device_map="cuda") pipe.load_lora_weights("Alex995647/loras-wan-2.2") prompt = "Astronaut in a jungle, cold color palette, muted colors, detailed, 8k" image = pipe(prompt).images[0] - Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- Draw Things
Add 25 LoRA(s): lx-cut-anything-it-s-a-cake-wan-2-2-i2v-a14b, short-wetsuit-wetsuitbbs, vertigo-perfect-dollyzoom, nn-semi-realistic-wan2-2β¦
9da6bb7 verified Download scooby-doo-mystery-incorporated-style-lora-wan-2-2-14b-qwen/info.txt from Alex995647/loras-wan-2.2: direct link, hf CLI and curl.
- Browser
- Download file 30.4 kB
-
https://huggingface.co/Alex995647/loras-wan-2.2/resolve/main/scooby-doo-mystery-incorporated-style-lora-wan-2-2-14b-qwen/info.txt
- Command line
-
hf download hf://Alex995647/loras-wan-2.2/scooby-doo-mystery-incorporated-style-lora-wan-2-2-14b-qwen/info.txt
-
curl -L -o info.txt https://huggingface.co/Alex995647/loras-wan-2.2/resolve/main/scooby-doo-mystery-incorporated-style-lora-wan-2-2-14b-qwen/info.txt
30.4 kB
| βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ | |
| SCOOBY DOO MYSTERY INCORPORATED STYLE LORA WAN 2.2 14B + QWEN | |
| βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ | |
| βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ | |
| IDENTITY | |
| βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ | |
| Base model Wan 2.2 | |
| Generation mode T2V (Text-to-Video) | |
| Checkpoint T2V-A14B (MoE 14B, high- and low-noise experts) | |
| Category Visual style | |
| Author tazmannner379 | |
| Source https://civitai.com/models/2167027 | |
| Provenance CivitAI Β· model 2167027 Β· version 2483369 | |
| βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ | |
| WEIGHTS | |
| βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ | |
| File scoobydoo_wan22_high_e43.safetensors | |
| Version WAN 2.2 14B T2V-High v1.1 | |
| Size 292.6 MB | |
| Format SafeTensor / unknown precision | |
| SHA256 05b20ea6242084e070286794fec8d5879181f320e568c4d8bb29b0e2d865ca78 | |
| Published 2025-12-08 | |
| βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ | |
| TRIGGER WORDS | |
| βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ | |
| βΊ "Scoobydoostyle" | |
| βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ | |
| RECOMMENDED SETTINGS | |
| βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ | |
| Taken verbatim from the author's description: | |
| LoRA strength 1.5 | |
| Frame count 55 | |
| βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ | |
| ADOPTION & POPULARITY | |
| βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ | |
| Downloads 1,365 | |
| Rating 100% positive (69/69) | |
| Comments 5 | |
| Published 270 days ago | |
| Download rate 1.4/day since release | |
| Adoption rank #139 of 150 in this collection (Niche) | |
| Momentum rank #143 of 150 by download rate (Long tail) | |
| Adoption tier is the percentile of total downloads within this collection; | |
| momentum is the percentile of downloads-per-day since release. Momentum is | |
| ranked rather than measured against a fixed rate, because this base model | |
| is itself new and every LoRA looks fast on an absolute scale. Both numbers | |
| are CivitAI's, read on the scrape date at the foot of this file. | |
| βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ | |
| CHAINING / STACKING | |
| βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ | |
| Stacks with others not stated | |
| The author gave no chaining guidance. The rules below are the | |
| general Wan 2.2 ones, not anything specific to this LoRA. | |
| General Wan 2.2 rules that apply here (see REFERENCE.txt): | |
| β’ The high/low pair is not a stack. Every A14B LoRA that ships two files | |
| needs the high-noise file on the high-noise UNET chain and the low- | |
| noise file on the low-noise UNET chain β two LoraLoaderModelOnly nodes | |
| on two different models, never both on one model. | |
| β’ Never load a high-noise LoRA onto the low-noise expert or vice versa. | |
| It loads without a single warning because the two checkpoints are | |
| architecturally identical, and the damage shows up only as degraded | |
| output. Filename is the only guard. | |
| β’ A single-file A14B LoRA with no high/low marking (common for Wan2.1-era | |
| LoRAs and for LoRAs extracted from merged models) goes on BOTH experts | |
| at the same strength. Loading it only on the low-noise expert is the | |
| safer half-measure when it fights the composition. | |
| β’ Stack subject/style LoRAs after the accelerator LoRA in the chain: | |
| UNETLoader -> Lightning LoRA -> subject LoRA -> ModelSamplingSD3. Order | |
| within LoraLoaderModelOnly chains is additive so it rarely matters | |
| mathematically, but keeping the accelerator first matches every | |
| reference workflow. | |
| β’ Do not stack two step-distillation LoRAs on the same expert. One | |
| accelerator per expert. | |
| β’ When adding subject LoRAs on top of a 4-step distill LoRA, reduce the | |
| subject LoRA strength rather than the distill LoRA strength β the | |
| distill LoRA is what keeps 4 steps viable at cfg 1.0. | |
| β’ Keep the LoRA strength for the two experts independently tunable. | |
| Because the high-noise expert owns layout and the low-noise expert owns | |
| detail, the community routinely runs a style LoRA weaker on high-noise | |
| and full-strength on low-noise. | |
| β’ TI2V-5B LoRAs chain like any single-model LoRA and can never be mixed | |
| into an A14B chain, or the reverse β the two have entirely different | |
| tensor geometry. | |
| βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ | |
| PROS | |
| βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ | |
| + Well established β 1,365 downloads. [source: CivitAI stats] | |
| + Positively received β 69 thumbs-up with no down-votes. [source: CivitAI | |
| stats] | |
| + Explicit trigger word(s) β Scoobydoostyle β so the effect can be turned | |
| on and off from the prompt. [source: model metadata] | |
| + Compact at 293 MB β cheap to stack with other LoRAs. [source: file | |
| size] | |
| + Author published concrete recommended settings. [source: description] | |
| βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ | |
| CONS & CAVEATS | |
| βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ | |
| - Author references NSFW output β review before using in shared work. | |
| [source: description] | |
| βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ | |
| COMPATIBILITY WARNING | |
| βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ | |
| Three distinct answers. The structural half of each is verified from the | |
| shipped config.json and safetensors index files; the behavioural severity | |
| is community-reported and flagged in "unverified". (1) A14B <-> TI2V-5B: | |
| INCOMPATIBLE. The 5B model is dim 3072 / 30 layers / 24 heads / ffn 14336 | |
| against the A14B's dim 5120 / 40 layers / 40 heads / ffn 13824 (both | |
| configs fetched and compared). In ComfyUI native this is SILENT | |
| DEGRADATION, not a hard error: keys for blocks 30-39 hit | |
| logging.warning("lora key not loaded: {}") in comfy/lora.py, and keys for | |
| blocks 0-29 match by name but not by shape, so they hit the shape guard | |
| logging.warning("WARNING SHAPE MISMATCH {} WEIGHT NOT MERGED") in | |
| comfy/lora.py or the caught exception logging.error("ERROR ...") in | |
| comfy/weight_adapter/lora.py, which returns the weight unmodified. All | |
| three call sites were read in the current ComfyUI source. The queue | |
| completes, no exception reaches the UI, and the output is effectively that | |
| of the LoRA node bypassed. (2) T2V-A14B <-> I2V-A14B: STRUCTURALLY | |
| COMPATIBLE, behaviourally risky. Verified from the safetensors index files: | |
| both experts carry exactly 1095 tensors with identical key names, and the | |
| only tensor whose shape differs is patch_embedding.weight (in_dim 16 vs | |
| 36), which LoRA trainers normally exclude. So a T2V-A14B LoRA loads onto | |
| I2V-A14B and vice versa with no warnings. What you get is a real but off- | |
| target effect. Treat cross-loading as "loads and does something", never as | |
| "certified". (3) High-noise expert <-> low-noise expert (within one A14B | |
| family): the two checkpoints have identical architecture, so a LoRA trained | |
| on one loads onto the other with zero warnings. This is the most dangerous | |
| case in the family, because there is no signal at all that you mismatched. | |
| Filename is the ONLY discriminator. (4) Wan2.1 14B LoRAs. Verified: | |
| Wan2.1-T2V-14B's config.json is dim 5120 / ffn 13824 / 40 layers / 40 heads | |
| / in_dim 16 β the same block geometry as Wan2.2-A14B, so Wan2.1 T2V LoRAs | |
| load. Wan2.1-I2V-14B is only a partial match: its index has 1303 tensors | |
| including 8 img_emb.* keys and 200 cross_attn k_img/v_img keys for its CLIP | |
| path, which Wan2.2-I2V dropped entirely (0 such keys); those will be | |
| skipped with "lora key not loaded". | |
| βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ | |
| AUTHOR'S DESCRIPTION | |
| βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ | |
| About this lora This is a lora's purpose is to re-create the style of the | |
| 2010 show Scooby-Doo! Mystery Incorporated. It can also re-create any | |
| character that appears in the show. See the bottom of the description for | |
| how to prompt the characters. You do not need to prompt in anything | |
| regarding the style, the lora will do that on its own. | |
| About this version Trigger word: Scoobydoostyle | |
| Trained the high to 30K steps and the low to 17K steps. There is still some | |
| issues with the high not triggering the style without the characters in the | |
| prompt. In that case try 1.5 strength on the high model and adding | |
| "animated" to the prompt. But for the most part you do not need to mention | |
| the style to trigger it properly. It looks a lot closer to the real style | |
| without using lightx but the increased resolution from lightx makes it | |
| worth it to use it anyway. You can use my example workflow to test either | |
| way. | |
| The dataset is 107 videos, which I split from both seasons of the show | |
| using pyscene. I had them split out into 55 frames 16fps videos after | |
| selecting my clips. This allowed training efficiently. The images are from | |
| the qwen dataset trained earlier. I trained the images on 512x resolution | |
| and the videos on 256x resolution. | |
| I created an app using Google AI studio with vibe coding, where I could | |
| feed the dataset in and it followed the guide I created (linked in the | |
| suggested resources below) for captioning style. It allowed me to do all | |
| the captioning work in around 10 minutes for the videos. I did not modify | |
| any of the captions, You can use the tool yourself here: | |
| https://huggingface.co/spaces/comfyuiman/loracaptionertaz | |
| I followed the same principles as this guide to create this lora | |
| https://civitai.com/articles/20389 | |
| It can handle NSFW since that is in the image dataset. I haven't done much | |
| testing on lower body nudity on the video dataset yet though. Breasts are | |
| totally fine though. | |
| I'm going to do another round of training for a few thousand steps at 1024x | |
| resolution on both the high and low. I think some things aren't perfect but | |
| overall I'm pretty happy with the result. | |
| Below is how to prompt each character and their primary outfit with | |
| character tag. | |
| I am very happy with the output, you can simply put the character tag at | |
| the end of the prompt and paste as is the primary outfit (if you want to | |
| use that) and the Appearance text and it will fit it. Velma has the most | |
| issues, you may need to mention eyes visible with her glasses, and also | |
| "her mouth is a simple line, no lips" to get it not give her lipstick. | |
| Velma Dinkley - Character Tag: char_velmadinkley | |
| - Primary Outfit: A baggy orange turtleneck sweater, a red pleated skirt, | |
| orange knee-high socks, and red Mary Jane shoes. | |
| - Appearance: Fair skin, freckles, short reddish-brown bob hair with a | |
| small red bow, and thick black-framed square glasses with light blue lenses | |
| and visible eyes. Simple line for a mouth with no lips. | |
| Daphne Blake - Character Tag: char_daphneblake | |
| - Primary Outfit: A purple long-sleeved mini-dress with lighter purple | |
| stripes at the hem, a lime-green scarf tied around the neck, pink tights, | |
| and purple heels. | |
| - Appearance: Fair skin, shoulder-length wavy orange hair, a purple | |
| headband, purple eyes, and pink lipstick. | |
| Fred Jones - Character Tag: char_fredjones | |
| - Primary Outfit: A white long-sleeved collared shirt worn over a light | |
| blue collared shirt, a bright orange ascot tied around the neck, and blue | |
| pants. | |
| - Appearance: Blonde hair styled with a pompadour, fair skin, a square | |
| prominent jawline, and a muscular build. | |
| Shaggy Rogers - Character Tag: char_shaggyrogers | |
| - Primary Outfit: A loose-fitting lime green V-neck t-shirt and reddish- | |
| brown bell-bottom pants. | |
| - Appearance: Lanky and slender build, shaggy light brown hair, a scruffy | |
| goatee, and a long neck. | |
| Scooby-Doo - Character Tag: char_scoobydoo | |
| - Primary Outfit: A light blue collar with a yellow diamond-shaped tag | |
| engraved with "SD". | |
| - Appearance: A large brown Great Dane with black spots on his back, floppy | |
| ears, a long snout with a large black nose, and expressive eyebrows. | |
| Sheriff Bronson Stone - Character Tag: char_sheriffbronsonstone | |
| - Primary Outfit: A tan sheriff's uniform shirt with brown epaulets, a dark | |
| brown tie, a glowing gold star-shaped badge, and a brown cowboy hat. | |
| - Appearance: Broad-shouldered man, fair skin, grey hair at the temples, | |
| and a thick black mustache. | |
| Marcie Fleach (Hot Dog Water) - Character Tag: char_marciefleach | |
| - Primary Outfit: A light yellow top worn under a red vest, a tan skirt, | |
| and colorful barrettes in her hair. | |
| - Appearance: Messy dark brown hair, red-framed rectangular glasses with | |
| yellow lenses, and a thin face. | |
| Cassidy Williams - Character Tag: char_cassidywilliams | |
| - Primary Outfit: A white lab coat worn over a green collared shirt and | |
| pink pants. | |
| - Appearance: Dark skin, short black hair with a blue streak, green | |
| eyeshadow, and red lipstick. | |
| Professor Pericles - Character Tag: char_professorpericles | |
| - Primary Outfit: A dark purple scarf tied around the neck. | |
| - Appearance: An anthropomorphic parrot with purple and grey feathers, a | |
| scar near the eye, large white eyes, and a stern expression. | |
| Ricky Owens - Character Tag: char_rickyowens | |
| - Primary Outfit: A dark suit jacket over a white shirt with an orange | |
| ascot. | |
| - Appearance: Thin man with dark hair, sideburns, and large square-framed | |
| glasses. | |
| Brad Chiles - Character Tag: char_bradchiles | |
| - Primary Outfit: A white jacket with blue accents, a blue shirt, and a red | |
| ascot. | |
| - Appearance: Blonde hair, strong square jaw, distinct chin, and fair skin. | |
| Judy Reeves - Character Tag: char_judyreeves | |
| - Primary Outfit: A sleeveless peach-colored top with a high collar and a | |
| necklace of large white beads. | |
| - Appearance: Short stylish blonde bob, fair skin, and heavy blue | |
| eyeshadow. | |
| Mr. E - Character Tag: char_mre | |
| - Primary Outfit: A dark suit with a red tie. | |
| - Appearance: Gaunt man with a pale complexion, long dark messy hair, a | |
| goatee, and a weary expression. | |
| Mayor Fred Jones Sr. - Character Tag: char_fredjones_sr | |
| - Primary Outfit: A double-breasted maroon suit jacket, a pink shirt, and a | |
| yellow tie. | |
| - Appearance: Older man with white hair, a stocky build, and a square jaw. | |
| Angel Dynamite - Character Tag: char_angeldynamite | |
| - Primary Outfit: A lime-green 70s style outfit with a halter top and bell | |
| bottoms, large headphones, and a yellow headband. | |
| - Appearance: Dark skin, large black afro, and bright green eyeshadow. | |
| Alice May - Character Tag: char_alicemay | |
| - Primary Outfit: A casual 60s style outfit with a headband. | |
| - Appearance: Blonde hair, fair skin, blue eyes, and distinct facial | |
| features. | |
| Thorn (The Hex Girls) - Character Tag: char_thorn | |
| - Primary Outfit: A black dress with red accents and a high collar. | |
| - Appearance: Long black hair with red highlights, pale skin, and green | |
| eyes. | |
| Dusk (The Hex Girls) - Character Tag: char_dusk | |
| - Primary Outfit: A green and black goth-style dress with torn fishnets. | |
| - Appearance: Blonde hair in pigtails with green streaks, pale skin, and | |
| heavy makeup. | |
| Luna (The Hex Girls) - Character Tag: char_luna | |
| - Primary Outfit: A red and white goth-style outfit with a purple vest. | |
| - Appearance: Short white hair with red highlights, dark skin, and gold | |
| hoop earrings. | |
| Nan Blake - Character Tag: char_nanblake | |
| - Primary Outfit: A teal pantsuit with a matching bowtie and a white | |
| ruffled blouse. | |
| - Appearance: Older woman with styled red hair, an angular face, and a | |
| critical expression. | |
| Barty Blake - Character Tag: char_bartyblake | |
| - Primary Outfit: A pink suit jacket with a striped tie. | |
| - Appearance: Older man with short blonde hair, a mustache, and glasses. | |
| Angie Dinkley - Character Tag: char_angiedinkley | |
| - Primary Outfit: A yellow long-sleeved shirt. | |
| - Appearance: Older woman with short curly blonde hair and a round face. | |
| Dale Dinkley - Character Tag: char_daledinkley | |
| - Primary Outfit: A green jacket with breast pockets over a dark shirt. | |
| - Appearance: Balding man with brown hair on the sides, a beard, and | |
| glasses. | |
| Maile - Character Tag: char_maile | |
| - Primary Outfit: A light blue waitress uniform with a white collar and | |
| cuffs. | |
| - Appearance: Dark hair in braids, olive skin, and a light blue headband. | |
| Nova - Character Tag: char_nova | |
| - Primary Outfit: A pink collar with a large round tag. | |
| - Appearance: A Cocker Spaniel with tan fur, long dark ears, and red bows. | |
| Professor Emmanuel Raffalo - Character Tag: char_professoremmanuelraffalo | |
| - Primary Outfit: A white lab coat over a yellow shirt and tie. | |
| - Appearance: Older balding man with white hair on the sides, glasses, and | |
| a mustache. | |
| Vincent Van Ghoul - Character Tag: char_vincentvanghoul | |
| - Primary Outfit: A red jacket with a light blue ruffled cravat. | |
| - Appearance: Black hair with white streaks at the temples, a goatee, and | |
| dramatic eyebrows. | |
| Skipper Shelton - Character Tag: char_skippershelton | |
| - Primary Outfit: A blue coat over a striped shirt with a red neckerchief | |
| and a captain's hat. | |
| - Appearance: Older man with a white beard, weather-beaten face, and an | |
| eyepatch. | |
| Ed Machine - Character Tag: char_edmachine | |
| - Primary Outfit: A dark suit jacket with a white shirt and tie. | |
| - Appearance: Heavy-set bald man with a thick grey mustache. | |
| Blue Falcon - Character Tag: char_bluefalcon | |
| - Primary Outfit: A blue superhero suit with a cowl, a yellow belt, yellow | |
| gloves, and an "F" logo on the chest. | |
| - Appearance: Muscular older man with a square jaw. | |
| Dynomutt - Character Tag: char_dynomutt | |
| - Primary Outfit: A green superhero cape, mask, and boots. | |
| - Appearance: A grey robotic dog body with extended mechanical limbs. | |
| Randy Warsaw - Character Tag: char_randywarsaw | |
| - Primary Outfit: A flashy blue jacket with sparkles. | |
| - Appearance: Spiky orange hair, pink-tinted sunglasses, and a long face. | |
| Crybaby Clown - Character Tag: char_crybabyclown | |
| - Primary Outfit: A light blue onesie with a bib and diaper. | |
| - Appearance: Obese man with blue skin and clown makeup. | |
| Lord Valdronya - Character Tag: char_lordvaldronya | |
| - Primary Outfit: A black and white costume with a high collar and a cape. | |
| - Appearance: Vampire-like man with purple skin, glowing red eyes, and | |
| horns. | |
| Char Gar Gothakon - Character Tag: char_chargargothakon | |
| - Primary Outfit: N/A (Monster). | |
| - Appearance: A green Lovecraftian monster with a tentacled face and | |
| glowing red eyes. | |
| Aphrodite - Character Tag: char_aphrodite | |
| - Primary Outfit: A red and black outfit with a sweetheart neckline. | |
| - Appearance: Pale woman with horns, glowing eyes, and large wings. | |
| Hebediah Grimm - Character Tag: char_hebediahgrimm | |
| - Primary Outfit: A pilgrim-style outfit with a wide-brimmed black hat and | |
| a red cloak. | |
| - Appearance: A skeletal zombie-like figure with glowing yellow eyes. | |
| Dapper Jack - Character Tag: char_dapperjack | |
| - Primary Outfit: A cowboy duster coat, hat, and bandana. | |
| - Appearance: A skeletal cowboy with a glowing red cybernetic eye. | |
| Space Kook - Character Tag: char_spacekook | |
| - Primary Outfit: A bulky blue space suit with a clear helmet dome. | |
| - Appearance: A glowing skull inside the helmet with a red maniacal laugh. | |
| Odnarb (Green Orc Biker) - Character Tag: char_odnarb | |
| - Primary Outfit: A black leather biker vest with patches. | |
| - Appearance: Green skin, tusks, and pointed ears. | |
| Deputy Bucky - Character Tag: char_deputybucky | |
| - Primary Outfit: A tan deputy uniform with a tie and badge. | |
| - Appearance: Young man with brown hair and glasses. | |
| Janet Nettles - Character Tag: char_janetnettles | |
| - Primary Outfit: A professional red dress with light blue trim. | |
| - Appearance: Woman with long wavy brown hair. | |
| Obliteratrix - Character Tag: char_obliteratrix | |
| - Primary Outfit: A dark armored suit with a helmet featuring glowing red | |
| eyes. | |
| - Appearance: A robotic armored female figure. | |
| The Freak of Crystal Cove - Character Tag: char_freakofcrystalcove | |
| - Primary Outfit: Colorful mismatched patchwork clothing. | |
| - Appearance: Distorted face, wild hair, and a menacing grin. | |
| Dr. Phineus Phibes - Character Tag: char_drphineusphibes | |
| - Primary Outfit: A high-collared white coat and gloves. | |
| - Appearance: Black hair with a white streak and a mustache. | |
| George Avocados - Character Tag: char_georgeavocados | |
| - Primary Outfit: A green costume with a high collar. | |
| - Appearance: Man with short dark hair and a flat top. | |
| The Evil Entity - Character Tag: char_theevilentity | |
| - Primary Outfit: N/A (Monster). | |
| - Appearance: A massive glowing green entity with tentacles and evil eyes. | |
| βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ | |
| VERSION NOTES | |
| βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ | |
| Trained the high some more to 34,000 steps and on 1024x resolution on the | |
| images. I think it triggers more often now. Though there are still cases | |
| where you need to modify prompt or use "animated" trigger word. | |
| βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ | |
| EXAMPLE MEDIA | |
| βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ | |
| 6 preview(s) in ./example_images/ β community results for | |
| this LoRA. Videos keep a matching _poster.jpg still frame. | |
| No prompts are recorded below: CivitAI's public API flags these posts as | |
| having generation metadata but does not return it, so the prompts cannot be | |
| scraped. Open the model page to read them. | |
| 01_video.mp4 (video, 1248x720) | |
| still: 01_poster.jpg | |
| 02_video.mp4 (video, 1248x720) | |
| still: 02_poster.jpg | |
| 03_video.mp4 (video, 832x480) | |
| still: 03_poster.jpg | |
| 04_video.mp4 (video, 832x480) | |
| still: 04_poster.jpg | |
| 05_video.mp4 (video, 1248x720) | |
| still: 05_poster.jpg | |
| 06_video.mp4 (video, 1248x720) | |
| still: 06_poster.jpg | |
| βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ | |
| HOW TO USE (COMFYUI) | |
| βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ | |
| 1. Put every LoRA .safetensors in ComfyUI/models/loras/ (subfolders are | |
| fine and are used by reference workflows). Confirmed by the file-layout | |
| tree inside ComfyUI's own video_wan2_2_14B_t2v template note. | |
| 2. Install the base weights alongside them. A14B (from Comfy- | |
| Org/Wan_2.2_ComfyUI_Repackaged, split_files/): | |
| diffusion_models/wan2.2_{t2v,i2v}_high_noise_14B_fp8_scaled.safetensors AND | |
| the matching _low_noise_ file β you need both; | |
| text_encoders/umt5_xxl_fp8_e4m3fn_scaled.safetensors; | |
| vae/wan_2.1_vae.safetensors. TI2V-5B: | |
| diffusion_models/wan2.2_ti2v_5B_fp16.safetensors, the same umt5 text | |
| encoder, and vae/wan2.2_vae.safetensors. | |
| 3. Load the base model with UNETLoader ("Load Diffusion Model"), NOT | |
| CheckpointLoaderSimple β Wan2.2 ships as a bare diffusion model, with the | |
| text encoder loaded separately by CLIPLoader (type "wan") and the VAE by | |
| VAELoader. This is the node set used in every shipped Wan2.2 template. | |
| 4. For A14B, build two parallel chains, one per expert: UNETLoader | |
| (high_noise) -> LoraLoaderModelOnly -> ModelSamplingSD3(shift) -> | |
| KSamplerAdvanced, and separately UNETLoader (low_noise) -> | |
| LoraLoaderModelOnly -> ModelSamplingSD3(shift) -> KSamplerAdvanced. Use | |
| LoraLoaderModelOnly, not LoraLoader β Wan LoRAs patch the DiT only, there | |
| is no CLIP side to patch. | |
| 5. Put the high-noise LoRA file on the high-noise chain and the low-noise | |
| LoRA file on the low-noise chain. If a LoRA ships as a single file with no | |
| high/low marking, loading the same file on both chains is the standard | |
| community handling. | |
| 6. Wire the two KSamplerAdvanced nodes in series on one latent: the first | |
| (high-noise) runs add_noise=enable, start_at_step 0, end_at_step = | |
| boundary, return_with_leftover_noise=enable; the second (low-noise) runs | |
| add_noise=disable, start_at_step = boundary, end_at_step = total, | |
| return_with_leftover_noise=disable. Boundary is step 2 of 4 with a 4-step | |
| Lightning LoRA, step 10 of 20 undistilled. | |
| 7. For TI2V-5B there is only one chain and one KSampler: UNETLoader -> | |
| LoraLoaderModelOnly -> ModelSamplingSD3 -> KSampler. No expert split, no | |
| second sampler. | |
| 8. Fastest correct start: load ComfyUI's built-in templates | |
| video_wan2_2_14B_t2v / video_wan2_2_14B_i2v / video_wan2_2_14B_flf2v / | |
| video_wan2_2_5B_ti2v (Workflow > Browse Templates) and swap the LoRA | |
| filenames β the 14B templates already ship the two-expert plumbing and the | |
| Lightning LoRA nodes wired correctly. Note the 14B t2v/i2v templates are | |
| now packaged as a subgraph. | |
| βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ | |
| Scraped by lorakit on 2026-09-05 02:11 UTC. | |
| Stats and description are the author's; pros/cons are derived from the | |
| evidence tagged beside each line. | |
| βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ | |