HAPPY NEW YEAR EVERYONE!!!!
I have tons to write about this release, so sit back and enjoy reading.
This version of Aniverse is my New Year gift to you all!
I have tons to write about this release, so sit back and enjoy reading.
This version was supposed to be born as "pure" at 1024px
Unfortunately it didn't go as I hoped... but it wasn't my fault.
The dataset was made up of over 5700 4K images and the training had to last for almost 1000 hours (about 40 consecutive days).
Unfortunately, having exceeded 40% (therefore with a training of approximately 2300 images), the electricity company had a blackout in my area.
Furthermore, I forgot to check the option to save various training points...
I won't tell you how many saints I mentioned...
Fortunately, a checkpoint beyond 680,000 steps had saved me (in total there must have been just over 1,700,000 steps).
So I made various attempts, so as not to throw away the work that had already been done, and this version of Aniverse came out.
I am really very satisfied with the details obtained, but is better if you think at this version like a "BETA"
I pushed the creations up to 1360X640px without using HiRes.Fix, obtaining good results, but with a slightly too high error rate (usually elongated bodies appear).
So I lowered the creations to 1280X720px, again without HiRes.Fix, and I think the error rate is absolutely acceptable (less than 5%).
The total safety zone is: 1024X576px
Unfortunately I also had to add some negative words compared to usual... and the hands (guys I give up, I've tried everything) are always a lottery, sometimes good, sometimes very bad.
Sampling method: DPM++ 2M SDE Karras
Sampling steps: 50+ (minimum, it is very important not to go below this value)
Width: 576 (o 768)
Height: 1024
CFG Scale: 6~7
MY FAVORITE PROMPT:
(masterpiece, best quality, highres:1.2), (photorealistic:1.2), (intricate and beautiful:1.2), (detailed light:1.2), (soft lighting, side lighting, reflected light), (colorful, dynamic angle), upper body shot, fashion photography, YOUR PROMPT, dynamic pose, light passing through hair, (abstract background:1.3), (official art), (perfect skin), (sharp)
NEGATIVE PROMPT:
(worst quality:1.8, low quality:1.8), moles, mole, tears, piercing, freckles, skindentation, cutoffs, shiny skin, lucid skin, pendant, scars on face, interlocked fingers,
YOU CAN ALSO USE THESE NEGATIVE EMBEDDINGS:
1) Easy Negative, negative_hand-neg, EasyNegative, negative_hand-neg, moles, mole, piercing, tears, face skin imperfection, freckles, skindentation, cutoffs, shiny skin, scars on face, bad-hands-5
2) Bad-Images-39000, negative_hand-neg, moles, mole, piercing, tears, face skin imperfection, freckles, skindentation, cutoffs, shiny skin, scars on face, bad-hands-5
3) ng_deepnegative_v1_75t, (worst quality:1.4), (low quality:1.4), (normal quality:1.4), lowres, bad anatomy, bad hands, normal quality, ((monochrome)), ((grayscale)), ((watermark)), negative_hand-neg, moles, mole, piercing, tears, face skin imperfection, freckles, skindentation, cutoffs, shiny skin, scars on face,bad-hands-5
4) FastNegativeV2, negative_hand-neg, moles, mole, piercing, tears, face skin imperfection, freckles, skindentation, cutoffs, shiny skin, scars on face bad-hands-5
5) For MEN images: girl, woman, female, tits, BadImage_v2-39000, negative_hand-neg, moles, mole, piercing, tears, face skin imperfection, freckles, skindentation, cutoffs, shiny skin, bad-hands-5, scars on face, bad-hands 5
As for the RESOLUTION, as I said, you can go up to 1280x720 (and vice versa), but in the general model tab, to be safe, I will lowered at 1024X576 resolution, because is totally safe!
As for STEPS, I highly recommend using 50 or more. This way you will get excellent details and fewer errors in the composition of the image.
As soon as I have time I will update with this new version of Anitoon, Animics, Animesh and Animerge and let's see what results I get.
As usual, the PRUNED version will be released next week.
Right now, I'm running a new Aniverse training.
Sorry for the length of this note.
I hope you have a wonderful holiday season with your loved ones.
Samuele
----------------------------------------------------------------
BUON CAPODANNO A TUTTIIIIII!!!!
Questa versione di Aniverse è il mio regalo di Buon Anno nuovo per tutti voi!
Ho tantissime cose da scrivere riguardo a questa versione, perciò mettetevi comodi e buona lettura.
Questa versione doveva nascere come "pura" a 1024px X 1024px.
Purtroppo non è andata come speravo... ma non per colpa mia.
Il dataset era composto da oltre 5700 immagini a 4K e il tranining, doveva durare per quasi 1000 ore (circa 40 giorni consecutivi)
Purtroppo, superato il 40% (perciò con un training di 2300 immagini circa), la compagnia elettrica ha avuto un blackout nella mia zona.
Oltretutto, mi ero dimenticato di spuntare l'opzione di salvare vari punti di allenamento...
Non vi dico quanti santi ho nominato...
Fortunatamente mi aveva salvato un checkpoint oltre i 680.000 steps (in totale dovevano essere poco più di 1.700.000 steps).
Così ho fatto vari tentativi, per non buttare va il lavoro che fin la era stato fatto, e ne è uscita questa versione di Aniverse.
Sono veramente molto soddisfatto dei dettagli ottenuti, ma è meglio se pensate a questa versione come una sorta di "BETA".
Ho spinto le creazioni fino a 1360X640px senza usare HiRes.Fix, ottenendo buoni risultati, ma con un tasso di errore un po' troppo alto (solitamente compaiono corpi allungati).
Perciò ho abbassato le creazioni a 1280X720px, sempre senza HiRes.Fix, e il tasso di errore penso sia assolutamente accettabile (meno del 10%).
La risoluzione di totale sicurezza è: 1024X576px
Purtroppo ho dovuto anche aggiungere degliu embedding negativi rispetto al solito... e le mani (ragazzi ci rinuncio, ho provato di tutto) sono sempre un terno al lotto, a volte bene, a volte malissimo.
Riguardo gli embedding negativi:
1) EasyNegative per i miei gusti rimane sempre il migliore, da un risultato più realistico, senza contrastare troppo l'immagine. Purtroppo però a volte crea degli artefatti sulla faccia.
2) Bad-Images-39000 mi piace tantissimo per l'uso più "fumettistico".
3) FastNegativeV2 funziona come sempre molto bene. Però per i miei gusti va a sparare un po' troppo i colori e ad appiattire l'immagine.
4) Sto anche testando questo prompt, che da risultati molto buoni e leggermente più realistici: (EasyNegativeV2:1.3), AloeCrapArt-1000-neg, moles, freckles, skindentation, cutoffs, shiny skin, bad-hands-5, scars on face, tattoos,
->
In ogni caso ecco i prompt positivi e negativi suggeriti:
Sampling method: DPM++ 2M SDE Karras
Sampling steps: 50+ (minimum, it is very important not to go below this value)
Width: 576 (o 768)
Height: 1024
CFG Scale: 6~7
MY FAVORITE PROMPT:
(masterpiece, best quality, highres:1.2), (photorealistic:1.2), (intricate and beautiful:1.2), (detailed light:1.2), (soft lighting, side lighting, reflected light), (colorful, dynamic angle), upper body shot, fashion photography, YOUR PROMPT, dynamic pose, light passing through hair, (abstract background:1.3), (official art), (perfect skin), (sharp)
NEGATIVE PROMPT:
(worst quality:1.8, low quality:1.8), moles, mole, tears, piercing, freckles, skindentation, cutoffs, shiny skin, lucid skin, pendant, scars on face, interlocked fingers,
YOU CAN ALSO USE THESE NEGATIVE EMBEDDINGS:
1) Easy Negative, negative_hand-neg, EasyNegative, negative_hand-neg, moles, mole, piercing, tears, face skin imperfection, freckles, skindentation, cutoffs, shiny skin, scars on face, bad-hands-5
2) Bad-Images-39000, negative_hand-neg, moles, mole, piercing, tears, face skin imperfection, freckles, skindentation, cutoffs, shiny skin, scars on face, bad-hands-5
3) ng_deepnegative_v1_75t, (worst quality:1.4), (low quality:1.4), (normal quality:1.4), lowres, bad anatomy, bad hands, normal quality, ((monochrome)), ((grayscale)), ((watermark)), negative_hand-neg, moles, mole, piercing, tears, face skin imperfection, freckles, skindentation, cutoffs, shiny skin, scars on face,bad-hands-5
4) FastNegativeV2, negative_hand-neg, moles, mole, piercing, tears, face skin imperfection, freckles, skindentation, cutoffs, shiny skin, scars on face bad-hands-5
5) For MEN images: girl, woman, female, tits, BadImage_v2-39000, negative_hand-neg, moles, mole, piercing, tears, face skin imperfection, freckles, skindentation, cutoffs, shiny skin, bad-hands-5, scars on face, bad-hands 5
Per quanto riguarda la RISOLUZIONE, potete spingervi fino a 1280X720px (e viceversa) con un tasso di errore abbastanza accettabile. Ma per stare sicuri vi consiglio di utilizzare la 1024x576 che è una risoluzione ultra sicura.
Per quanto riguarda gli STEPS, vi consiglio vivamente di usarne 50 o più. In questo modo otterrete ottimi dettagli e meno errori nella composizione dell'immagine
Appena ho tempo aggiornerò con questa nuova versione, sia Anitoon, Animics, Animesh che Animerge e vediamo che risultati otterrò.
Come al solito, la versione PRUNED, uscirà la settimana prossima.
Ora come ora, sto eseguendo un nuovo training di Aniverse.
Scusate per la lunghezza di questa nota.
Vi auguro di passare delle bellissime festività insieme ai vostri cari.
Samuele
## Overview - **Model ID**: `aniverse-v1-6-pruned` - **Category**: stable diffusion - **Provider**: modelslab - **Status**: model_ready - **Screenshot**: `https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/09078d28-f6df-43f4-8fce-5e1edcffe92b/width=576/4121900.jpeg` ## API Information This model can be used via our HTTP API. See the API documentation and usage examples below. ### Endpoint - **URL**: `https://modelslab.com/api/v6/images/text2img` - **Method**: POST ### Parameters - **`prompt`** (required): prompt help in image generation - Type: textarea - Example: Enter prompt - **`model_id`** (required): Enter model_id that can help in image generation - Type: text - Example: Enter model_id here - **`lora_model`** (required): - Type: multiple_models - **`width`** (required): width of the image - Type: number (range: 512-1024) - **`height`** (required): height of the image - Type: number (range: 512-1024) - **`negative_prompt`** (optional): Negative prompt help in avoid things that you do not want in image - Type: textarea - Example: Enter negative prompt that you do not want see in image - **`scheduler`** (optional): - Type: select (options: DPM++ 2M, DPM++ SDE, Euler, Euler a) - **`guidance_scale`** (optional): - Type: number (range: 1-10) ## Usage Examples ### cURL ```bash curl --request POST \ --url https://modelslab.com/api/v6/images/text2img \ --header "Content-Type: application/json" \ --data '{ "key": "YOUR_API_KEY", "model_id": "aniverse-v1-6-pruned", "prompt": "R3alisticF, hauntingly beautiful oriental necromancer, long flowing brown hair, bangs, darkly tanned skin, earrings, bone necklaces, dark eyeshadow, red lips, vibrant, front-laced transparent, filmy silk blouse, cleavage, holding skull, in a sandstone room lit by candles, High Detail, Perfect Composition, high contrast, silhouetted, chiascuro", "width": "1024", "height": "1024", "negative_prompt": "(worst quality:2), (low quality:2), (normal quality:2), (jpeg artifacts), (blurry), (duplicate), (morbid), (mutilated), (out of frame), (extra limbs), (bad anatomy), (disfigured), (deformed), (cross-eye), (glitch), (oversaturated), (overexposed), (underexposed), (bad proportions), (bad hands), (bad feet), (cloned face), (long neck), (missing arms), (missing legs), (extra fingers), (fused fingers), (poorly drawn hands), (poorly drawn face), (mutation), (deformed eyes), watermark, text, logo, signature, grainy, tiling, censored, nsfw, ugly, blurry eyes, noisy image, bad lighting, unnatural skin, asymmetry", "scheduler": "DPMSolverMultistepScheduler", "guidance_scale": "7.5" }' ``` ### Python ```python import requests response = requests.post( "https://modelslab.com/api/v6/images/text2img", headers={ "Content-Type": "application/json" }, json={ "key": "YOUR_API_KEY", "model_id": "aniverse-v1-6-pruned", "prompt": "R3alisticF, hauntingly beautiful oriental necromancer, long flowing brown hair, bangs, darkly tanned skin, earrings, bone necklaces, dark eyeshadow, red lips, vibrant, front-laced transparent, filmy silk blouse, cleavage, holding skull, in a sandstone room lit by candles, High Detail, Perfect Composition, high contrast, silhouetted, chiascuro", "width": "1024", "height": "1024", "negative_prompt": "(worst quality:2), (low quality:2), (normal quality:2), (jpeg artifacts), (blurry), (duplicate), (morbid), (mutilated), (out of frame), (extra limbs), (bad anatomy), (disfigured), (deformed), (cross-eye), (glitch), (oversaturated), (overexposed), (underexposed), (bad proportions), (bad hands), (bad feet), (cloned face), (long neck), (missing arms), (missing legs), (extra fingers), (fused fingers), (poorly drawn hands), (poorly drawn face), (mutation), (deformed eyes), watermark, text, logo, signature, grainy, tiling, censored, nsfw, ugly, blurry eyes, noisy image, bad lighting, unnatural skin, asymmetry", "scheduler": "DPMSolverMultistepScheduler", "guidance_scale": "7.5" } ) print(response.json()) ``` ### JavaScript ```javascript fetch("https://modelslab.com/api/v6/images/text2img", { method: "POST", headers: { "Content-Type": "application/json" }, body: JSON.stringify({ "key": "YOUR_API_KEY", "model_id": "aniverse-v1-6-pruned", "prompt": "R3alisticF, hauntingly beautiful oriental necromancer, long flowing brown hair, bangs, darkly tanned skin, earrings, bone necklaces, dark eyeshadow, red lips, vibrant, front-laced transparent, filmy silk blouse, cleavage, holding skull, in a sandstone room lit by candles, High Detail, Perfect Composition, high contrast, silhouetted, chiascuro", "width": "1024", "height": "1024", "negative_prompt": "(worst quality:2), (low quality:2), (normal quality:2), (jpeg artifacts), (blurry), (duplicate), (morbid), (mutilated), (out of frame), (extra limbs), (bad anatomy), (disfigured), (deformed), (cross-eye), (glitch), (oversaturated), (overexposed), (underexposed), (bad proportions), (bad hands), (bad feet), (cloned face), (long neck), (missing arms), (missing legs), (extra fingers), (fused fingers), (poorly drawn hands), (poorly drawn face), (mutation), (deformed eyes), watermark, text, logo, signature, grainy, tiling, censored, nsfw, ugly, blurry eyes, noisy image, bad lighting, unnatural skin, asymmetry", "scheduler": "DPMSolverMultistepScheduler", "guidance_scale": "7.5" }) }) .then(response => response.json()) .then(data => console.log(data)); ``` ## Links - [Model Playground](https://modelslab.com/models/community-model/aniverse-v1-6-pruned) - [API Documentation](https://docs.modelslab.com) - [ModelsLab Platform](https://modelslab.com)