SpacecraftSchematics-Flux-ST-LoKr-bs6

This is a LyCORIS adapter derived from black-forest-labs/FLUX.1-dev.

No validation prompt was used during training.

None

Validation settings

• CFG: 3.0
• CFG Rescale: 0.0
• Steps: 20
• Sampler: None
• Seed: 42
• Resolution: 1024x1024

Note: The validation settings are not necessarily the same as the training settings.

You can find some example images in the following gallery:

Prompt

unconditional (blank prompt)

Negative Prompt

blurry, cropped, ugly

Prompt

In the style of SPCRFT technical blueprints: technical diagrams of a hamster, detailing its internal structure and biological systems

Negative Prompt

blurry, cropped, ugly

Prompt

In the style of SPCRFT technical blueprints, This illustration shows a detailed breakdown of a lunar landing module, highlighting various components for both ascent and descent stages. Labeled parts include environmental control sensors, communication systems, fuel tanks, and scientific instruments. Annotations identify specific functions, such as the egress platform for astronaut exit and the alignment optical telescope for navigation.

Negative Prompt

blurry, cropped, ugly

Prompt

In the style of SPCRFT technical blueprints: The depiction illustrates a detailed view of a spacecraft's components, including a launch escape assembly and service module. It highlights various sections such as the nose cone, main parachutes, crew compartment, and propulsion systems. Numerous labels point out specific features like helium tanks, electrical power systems, and environmental control systems, offering a comprehensive overview of the spacecraft's configuration.

Negative Prompt

blurry, cropped, ugly

Prompt

In the style of SPCRFT technical blueprints: Diagram of an octopus, highlighting its distributed neural network and adaptive problem-solving modules

Negative Prompt

blurry, cropped, ugly

Prompt

In the style of SPCRFT technical blueprints: Cutaway view of a beehive, detailing both biological systems and social hierarchy structures

Negative Prompt

blurry, cropped, ugly

Prompt

In the style of SPCRFT technical blueprints: Cutaway view of a robotic bird in flight, with detailed wing mechanisms

Negative Prompt

blurry, cropped, ugly

Prompt

In the style of SPCRFT technical blueprints: Technical diagram of a cat, showing internal cybernetic enhancements

Negative Prompt

blurry, cropped, ugly

Prompt

In the style of SPCRFT technical blueprints: Technical diagram of an owl, showcasing night-vision capabilities and sound-location systems

Negative Prompt

blurry, cropped, ugly

The text encoder was not trained. You may reuse the base model text encoder for inference.

Training settings

• Training epochs: 3
• Training steps: 2200
• Learning rate: 0.0008
• Max grad norm: 2.0
• Effective batch size: 6
  • Micro-batch size: 6
  • Gradient accumulation steps: 1
  • Number of GPUs: 1
• Prediction type: flow-matchingNone
• Rescaled betas zero SNR: False
• Optimizer: adamw_bf16
• Precision: Pure BF16
• Quantised: Yes: int8-quanto
• Xformers: Not used
• LyCORIS Config:

{
    "algo": "lokr",
    "multiplier": 1.0,
    "linear_dim": 10000,
    "linear_alpha": 1,
    "factor": 16,
    "apply_preset": {
        "target_module": [
            "Attention",
            "FeedForward"
        ],
        "module_algo_map": {
            "Attention": {
                "factor": 16
            },
            "FeedForward": {
                "factor": 8
            }
        }
    }
}

Datasets

SpacecraftSchematics-512

• Repeats: 17
• Total number of images: 38
• Total number of aspect buckets: 10
• Resolution: 0.262144 megapixels
• Cropped: False
• Crop style: None
• Crop aspect: None

SpacecraftSchematics-768

• Repeats: 17
• Total number of images: 38
• Total number of aspect buckets: 15
• Resolution: 0.589824 megapixels
• Cropped: False
• Crop style: None
• Crop aspect: None

SpacecraftSchematics-1024

• Repeats: 6
• Total number of images: 38
• Total number of aspect buckets: 16
• Resolution: 1.048576 megapixels
• Cropped: False
• Crop style: None
• Crop aspect: None

Inference

import torch
from diffusers import DiffusionPipeline
from lycoris import create_lycoris_from_weights

model_id = 'black-forest-labs/FLUX.1-dev'
adapter_id = 'pytorch_lora_weights.safetensors' # you will have to download this manually
lora_scale = 1.0
wrapper, _ = create_lycoris_from_weights(lora_scale, adapter_id, pipeline.transformer)
wrapper.merge_to()

prompt = "An astronaut is riding a horse through the jungles of Thailand."

pipeline.to('cuda' if torch.cuda.is_available() else 'mps' if torch.backends.mps.is_available() else 'cpu')
image = pipeline(
    prompt=prompt,
    num_inference_steps=20,
    generator=torch.Generator(device='cuda' if torch.cuda.is_available() else 'mps' if torch.backends.mps.is_available() else 'cpu').manual_seed(1641421826),
    width=1024,
    height=1024,
    guidance_scale=3.0,
).images[0]
image.save("output.png", format="PNG")