Team Ai
Modelpublic

dataautogpt3/CALAMITY

sourceHugging Faceapache-2.0updated 2y agoView on Hugging Face
8likes49downloads
README.md107 linesDownload Raw Back to root
1---2pipeline_tag: text-to-image3widget:4- text: >-5    7-Vectors, glitched skyline, high rise ruins, retrowave, occult, glitching, retrofuturistic, glitch art, artifacts, circuitry demon, glitched screen, synthetic, english, scanlines, chromatic aberration, bad quality, red-pink, crt, vhs, noise, grain, highly detailed, 4k, synthetic anime6  output:7    url: assets/sythentic.png8- text: >-9     by Mike Mignola, (high contrast , Character Design Sheet but extremely beautiful:1.4), (intricate details, masterpiece, best quality:1.4), poster art style, Professional, modern, eye-catching, highly detailed, dark palette, looking at viewer, dark yellow hue, monster, by nixeu, 1boy, male focus, solo, disgust, black hair, messy hair, short hair, green eyes, zombie, green skin, black shirt, short sleeves, bandaged arm, goggles on head, portrait, red background, impasto, realistic10  output:11    url: assets/pose.png12- text: >-13     inkpunk three colors only, ink and paper illustration, inked, gunfighter in a crimson cape, desolate street of an old mining town, bleak, ink splashes, ink spots, ink smears, evoking loneliness and solitude, desolate, abandoned, finality, bold strokes, ink illustration, dust storm, very windy, night, very dark, dust billowing, powder in the air, vision obscured, rays of light, gray and black and deep crimson tones, light streaming thru window, linquivera14  output:15    url: assets/ink.png16- text: >-17     vibrant glowing purple drink, in the style of a product hero shot in motion, dynamic magazine ad image, photorealism, sleep and mystical elements around the background18  output:19    url: assets/drink.png20- text: >-21     big eyes, pale skin, glossy lipstick multiple  zombie, (guro:1.5), 1girl,  blood, violence, cannibalism, mutation, random background, maggot, saliva drip, graveyard, severed head22  output:23    url: assets/bloody.png24- text: >-25     hyper realism, IR photo, light leaks, (split screen overlay , x-ray view:1.3), double exposure, (glitched image:1.2), a glitched out portrait of a red skin (demon girl) walking through a graveyard, succubus girl, vaporwave aesthetic, pixelation, sinusoidal distortions, screen tear, post modern deconstruction, molecular view, vibrational spiral signals, (rose garden overlay, split screen image:1.1), abstractions, experimental, anime rage, oniric portrait of a tall cloaked figure (with a glitched mask) and green clothing, vaporwave aesthetic, glitched body effect, Azem the Traveler, yearning to explore the ends of the world to discover its wonders and help its denizens, by Andy Kehoe, a gradient masterpiece, blue cyan yellow, Rococopunk, luminism, seamless, China ink, Ink Bubbles, alcohol ink elements, chiaroscuro, Shadow play, Gold leaf small lines, bright splashes of alcohol ink puddles, volumetric light, auras, rays of sunlight, bright colors reflect, isometric, digital art, 3d render, chromatic aberration, octane render, volumetrics, by greg rutkowski26  output:27    url: assets/demon_girl.png28- text: >-29    1960s black&white portrait photo, feral spectral-alien with cybernetic-mutations, synthetic anime 30  output:31    url: assets/alien.png32license: apache-2.033---34<Gallery />35 36# CALAMITY: Horror Themed Text-to-Image Model v1.037 38CALAMITY is a specialized horror-themed text-to-image model built upon the Prometheus base model, which itself is a full rank finetune of the Playground v2.5 architecture. This checkpoint aims to generate haunting, surreal, and unsettling images with a distinctive aesthetic that sets it apart from mainstream models.39 40# Key Features41Unique focus on dark, creepy, and horror themed visuals42Trained on a curated dataset of horror imagery to capture genre-specific style43Seamless integration with the Prometheus base model's enhanced accessibility features44Advanced custom CLIP integration for improved text-to-image alignment45Optimized for generating high-resolution, evocative horror artwork46 47# Recommended Settings48 49Clip Skip: 250CFG Scale: 751Steps: 30-6052Sampler: DPM++ 2M SDE53Scheduler: Karras54Resolution: 1024x1024 55 56# Style Trigger word:"Sythentic Anime"57 58# Use Cases59 60Generating evocative, nightmarish artwork for games, films, and books61Concept art and visual brainstorming for horror-themed projects62Exploring surreal, uncanny aesthetics and compositions63Pushing creative boundaries in the horror genre through AI-augmented workflows64 65CALAMITY opens up new avenues for horror-themed image generation, providing creatives with a powerful tool for conjuring up unsettling and imaginative visuals. Experiment with different prompt combinations and settings to delve into a rich spectrum of macabre imagery.66 67# Use it with 🧨 diffusers68```python69import torch70from diffusers import (71    StableDiffusionXLPipeline, 72    KDPM2AncestralDiscreteScheduler,73    AutoencoderKL74)75 76# Load VAE component77vae = AutoencoderKL.from_pretrained(78    "madebyollin/sdxl-vae-fp16-fix", 79    torch_dtype=torch.float1680)81 82# Configure the pipeline83pipe = StableDiffusionXLPipeline.from_pretrained(84    "dataautogpt3/CALAMITY", 85    vae=vae,86    torch_dtype=torch.float1687)88pipe.scheduler = KDPM2AncestralDiscreteScheduler.from_config(pipe.scheduler.config)89pipe.to('cuda')90 91# Define prompts and generate image92prompt = "Sythentic Anime"93negative_prompt = ""94 95image = pipe(96    prompt, 97    negative_prompt=negative_prompt, 98    width=1024,99    height=1024,100    guidance_scale=7,101    num_inference_steps=50,102    clip_skip=2103).images[0]104 105 106image.save("generated_image.png")107```