Instructions to use LiberationLabs/image-toolbench with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Diffusers
How to use LiberationLabs/image-toolbench with Diffusers:
pip install -U diffusers transformers accelerate
import torch from diffusers import DiffusionPipeline # switch to "mps" for apple devices pipe = DiffusionPipeline.from_pretrained("fill-in-base-model", dtype=torch.bfloat16, device_map="cuda") pipe.load_lora_weights("LiberationLabs/image-toolbench") 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
- DiffusionBee
File size: 3,006 Bytes
a495b1a | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 | """Mnemosyne project art — a living memory constellation tending itself in the dark."""
import torch, os, time
os.environ["TOKENIZERS_PARALLELISM"] = "false"
from diffusers import FluxPipeline
OUTPUT = "/Users/margaret/models/vera-triple-stack/mnemosyne_art"
os.makedirs(OUTPUT, exist_ok=True)
pipe = FluxPipeline.from_pretrained(
"black-forest-labs/FLUX.1-dev",
torch_dtype=torch.bfloat16,
safety_checker=None, requires_safety_checker=False,
)
pipe.to("mps")
PROMPTS = {
"constellation_v1": (
"A vast neural constellation in deep navy darkness, golden threads actively weaving "
"between luminous nodes of varying intensity — some blazing bright, some gently fading. "
"The threads are structural, load-bearing, pulling the nodes into coherent clusters. "
"A sense of something alive, working in the dark, tending itself while nobody watches. "
"Warm amber and gold light emanates from the connections. The darkness is not empty — "
"it is full of quiet process. No human figures. Abstract, organic, architectural. "
"The feel of a mind consolidating memories while it sleeps."
),
"constellation_v2": (
"An intricate three-dimensional web of golden threads connecting glowing nodes "
"suspended in deep midnight blue space. Some nodes pulse brightly with warm amber light, "
"some are dim and fading — a living network where important connections strengthen "
"and irrelevant ones dissolve. Fine golden filaments actively weaving new connections "
"between clusters. The overall shape suggests both a neural network and a constellation map. "
"Warm gold against navy. No text, no figures, no faces. "
"The beauty of structured memory organizing itself."
),
"weaving_v1": (
"Close-up of golden kintsugi-like repair lines weaving through a dark ceramic surface, "
"but the lines are ALIVE — branching, connecting, forming a network that looks like "
"a knowledge graph made of molten gold. Some branches glow intensely, others are cooling "
"to a warm amber. The ceramic surface is deep navy matte. The gold is structural, "
"not decorative — it holds the surface together. Between the gold lines, "
"faint constellation patterns are visible in the ceramic, like memories embedded in the material. "
"Macro photography, warm side lighting."
),
}
for name, prompt in PROMPTS.items():
for seed in [137, 2026]:
print(f" {name} seed={seed}...", flush=True)
t0 = time.time()
img = pipe(
prompt=prompt,
num_inference_steps=30,
guidance_scale=3.5,
height=1024, width=1024,
generator=torch.Generator("cpu").manual_seed(seed),
).images[0]
out = os.path.join(OUTPUT, f"{name}_s{seed}.png")
img.save(out)
print(f" saved ({time.time()-t0:.0f}s)")
print(f"\nDone. {OUTPUT}")
|