Koolchh/AnimeBoysXL-v1.0

AnimeBoysXL v1.0

It takes substantial time and efforts to bake models. If you appreciate my models, I would be grateful if you could support me on Ko-fi ☕.

Prompt

2boys, multiple boys, yaoi, looking at another, hand on another's shoulder, smile, short hair, black hair, closed eyes, brown hair, blue eyes, shirt, lens flare, sky, cloud, blue sky, sweat, best quality, amazing quality, best aesthetic, absurdres, year 2023

Negative Prompt

lowres, (bad:1.05), text, error, missing, extra, fewer, cropped, jpeg artifacts, worst quality, bad quality, watermark, bad aesthetic, unfinished, chromatic aberration, scan, scan artifacts, 1girl, breasts, realistic, lips, nose

Prompt

1boy, solo, male focus, hoodie, grey hoodie, hood, hood up, close-up, sleepy, short hair, bara, undercut, stubble, black hair, green eyes, parted lips, white background, simple background, best quality, amazing quality, best aesthetic, absurdres, year 2023

Negative Prompt

lowres, (bad:1.05), text, error, missing, extra, fewer, cropped, jpeg artifacts, worst quality, bad quality, watermark, bad aesthetic, unfinished, chromatic aberration, scan, scan artifacts, 1girl, breasts

Features

• ✔️ Good for inference: AnimeBoysXL v1.0 is a flexible model which is good at generating images of anime boys and males-only content in a wide range of styles.
• ✔️ Good for training: AnimeBoysXL v1.0 is suitable for further training, thanks to its neutral style and ability to recognize a great deal of concepts. Feel free to train your own anime boy model/LoRA from AnimeBoysXL.
• ❌ AnimeBoysXL v1.0 is not optimized for creating anime girls. Please consider using other models for that purpose.

Inference Guide

• Prompt: Use tag-based prompts to describe your subject.
  • Append , best quality, amazing quality, best aesthetic, absurdres to the prompt to improve image quality.
  • (Optional) Append , year YYYY to the prompt to shift the output toward the prevalent style of that year. YYYY is a 4 digit year, e.g. , year 2023
• Negative prompt: Choose from one of the following two presets.
  1. Heavy (recommended): lowres, (bad:1.05), text, error, missing, extra, fewer, cropped, jpeg artifacts, worst quality, bad quality, watermark, bad aesthetic, unfinished, chromatic aberration, scan, scan artifacts, 1girl, breasts
  2. Light: lowres, jpeg artifacts, worst quality, watermark, blurry, bad aesthetic, 1girl, breasts
  • (Optional) Add , realistic, lips, nose to the negative prompt if you need a flat anime-like style face.
• VAE: Make sure you're using SDXL VAE.
• Sampling method, sampling steps and CFG scale: I find (Euler a, 28, 5) good. You are encouraged to experiment with other settings.
• Width and height: 832*1216 for portrait, 1024*1024 for square, and 1216*832 for landscape.

🧨Diffusers Example Usage

import torch
from diffusers import DiffusionPipeline

pipe = DiffusionPipeline.from_pretrained("Koolchh/AnimeBoysXL-v1.0", torch_dtype=torch.float16, use_safetensors=True, variant="fp16")
pipe.to("cuda")

prompt = "1boy, solo, male focus, hoodie, grey hoodie, hood, hood up, close-up, sleepy, short hair, bara, undercut, stubble, black hair, green eyes, parted lips, white background, simple background, best quality, amazing quality, best aesthetic, absurdres, year 2023"
negative_prompt = "lowres, (bad:1.05), text, error, missing, extra, fewer, cropped, jpeg artifacts, worst quality, bad quality, watermark, bad aesthetic, unfinished, chromatic aberration, scan, scan artifacts, 1girl, breasts"

image = pipe(
    prompt=prompt, 
    negative_prompt=negative_prompt, 
    width=1024,
    height=1024,
    guidance_scale=5,
    num_inference_steps=28
).images[0]

Training Details

AnimeBoysXL v1.0 is trained from Stable Diffusion XL Base 1.0, on ~516k images.

The following tags are attached to the training data to make it easier to steer toward either more aesthetic or more flexible results.

Quality tags

tag	score
best quality	>= 150
amazing quality	[100, 150)
great quality	[75, 100)
normal quality	[0, 75)
bad quality	(-5, 0)
worst quality	<= -5

Aesthetic tags

tag	score
best aesthetic	>= 6.675
great aesthetic	[6.0, 6.675)
normal aesthetic	[5.0, 6.0)
bad aesthetic	< 5.0

Rating tags

tag	rating
(None)	general
slightly nsfw	sensitive
fairly nsfw	questionable
very nsfw	explicit

Year tags

year YYYY where YYYY is in the range of [2005, 2023].

Training configurations

• Hardware: 4 * Nvidia A100 80GB GPUs
• Optimizer: AdaFactor
• Gradient accumulation steps: 8
• Batch size: 4 * 8 * 4 = 128
• Learning rates:
  • 8e-6 for U-Net
  • 5.2e-6 for text encoder 1 (CLIP ViT-L)
  • 4.8e-6 for text encoder 2 (OpenCLIP ViT-bigG)
• Learning rate schedule: constant with 250 warmup steps
• Mixed precision training type: BF16
• Epochs: 20