jimmycarter/flux-training-losercity-next-lycoris8

flux-training-losercity-next-lycoris8

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

The main validation prompt used during training was:

loona from helluva boss is eating a donut

Validation settings

• —CFG: 3.5
• —CFG Rescale: 0.0
• —Steps: 15
• —Sampler: None
• —Seed: 42
• —Resolution: 1024

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

You can find some example images in the following gallery:

<Gallery />

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

Training settings

• —Training epochs: 0
• —Training steps: 2700
• —Learning rate: 4e-05
• —Effective batch size: 16
• —Micro-batch size: 1
• —Gradient accumulation steps: 16
• —Number of GPUs: 1
• —Prediction type: flow-matching
• —Rescaled betas zero SNR: False
• —Optimizer: adamw_bf16
• —Precision: Pure BF16
• —Quantised: Yes: fp8-quanto
• —Xformers: Not used
• —LyCORIS Config:

{
    "algo": "lokr",
    "multiplier": 1.0,
    "linear_dim": 1000000,
    "linear_alpha": 1,
    "factor": 10,
    "full_matrix": true,
    "apply_preset": {
        "target_module": [
            "FluxTransformerBlock",
            "FluxSingleTransformerBlock"
        ],
        "name_algo_map": {
            "transformer_blocks.[0-7]*": {
                "algo": "lokr",
                "factor": 4,
                "linear_dim": 1000000,
                "linear_alpha": 1,
                "full_matrix": true
            },
            "transformer_blocks.[8-15]*": {
                "algo": "lokr",
                "factor": 6,
                "linear_dim": 1000000,
                "linear_alpha": 1,
                "full_matrix": true
            },
            "transformer_blocks.[16-18]*": {
                "algo": "lokr",
                "factor": 12,
                "linear_dim": 1000000,
                "linear_alpha": 1,
                "full_matrix": true
            },
            "single_transformer_blocks.[0-15]*": {
                "algo": "lokr",
                "factor": 8,
                "linear_dim": 1000000,
                "linear_alpha": 1,
                "full_matrix": true
            },
            "single_transformer_blocks.[16-23]*": {
                "algo": "lokr",
                "factor": 6,
                "linear_dim": 1000000,
                "linear_alpha": 1,
                "full_matrix": true
            },
            "single_transformer_blocks.[24-37]*": {
                "algo": "lokr",
                "factor": 4,
                "linear_dim": 1000000,
                "linear_alpha": 1,
                "full_matrix": true
            }
        },
        "use_fnmatch": true
    }
}

Datasets

defaultdatasetarb

• —Repeats: 9999
• —Total number of images: 41
• —Total number of aspect buckets: 1
• —Resolution: 1.33 megapixels
• —Cropped: False
• —Crop style: None
• —Crop aspect: None

defaultdatasetarb2

• —Repeats: 9999
• —Total number of images: 2565
• —Total number of aspect buckets: 1
• —Resolution: 1.33 megapixels
• —Cropped: False
• —Crop style: None
• —Crop aspect: None

defaultdatasetarb3

• —Repeats: 9999
• —Total number of images: 3220
• —Total number of aspect buckets: 14
• —Resolution: 1.33 megapixels
• —Cropped: False
• —Crop style: None
• —Crop aspect: None

default_dataset

• —Repeats: 9999
• —Total number of images: 42
• —Total number of aspect buckets: 1
• —Resolution: 1.048576 megapixels
• —Cropped: True
• —Crop style: center
• —Crop aspect: square

defaultdataset512

• —Repeats: 9999
• —Total number of images: 42
• —Total number of aspect buckets: 1
• —Resolution: 0.262144 megapixels
• —Cropped: True
• —Crop style: center
• —Crop aspect: square

defaultdataset640

• —Repeats: 9999
• —Total number of images: 42
• —Total number of aspect buckets: 1
• —Resolution: 0.4096 megapixels
• —Cropped: True
• —Crop style: center
• —Crop aspect: square

defaultdataset768

• —Repeats: 9999
• —Total number of images: 42
• —Total number of aspect buckets: 1
• —Resolution: 0.589824 megapixels
• —Cropped: True
• —Crop style: center
• —Crop aspect: square

defaultdataset896

• —Repeats: 9999
• —Total number of images: 42
• —Total number of aspect buckets: 1
• —Resolution: 0.802816 megapixels
• —Cropped: True
• —Crop style: center
• —Crop aspect: square

defaultdatasetuncaptioned

• —Repeats: 9999
• —Total number of images: 2565
• —Total number of aspect buckets: 1
• —Resolution: 1.048576 megapixels
• —Cropped: True
• —Crop style: center
• —Crop aspect: square

defaultdatasetuncaptioned_512

• —Repeats: 9999
• —Total number of images: 2565
• —Total number of aspect buckets: 1
• —Resolution: 0.262144 megapixels
• —Cropped: True
• —Crop style: center
• —Crop aspect: square

defaultdatasetart

• —Repeats: 9999
• —Total number of images: 2482
• —Total number of aspect buckets: 1
• —Resolution: 1.048576 megapixels
• —Cropped: True
• —Crop style: center
• —Crop aspect: square

defaultdatasetart_512

• —Repeats: 9999
• —Total number of images: 3193
• —Total number of aspect buckets: 1
• —Resolution: 0.262144 megapixels
• —Cropped: True
• —Crop style: center
• —Crop aspect: square

defaultdatasetart_640

• —Repeats: 9999
• —Total number of images: 3115
• —Total number of aspect buckets: 1
• —Resolution: 0.4096 megapixels
• —Cropped: True
• —Crop style: random
• —Crop aspect: square

defaultdatasetart_768

• —Repeats: 9999
• —Total number of images: 2989
• —Total number of aspect buckets: 1
• —Resolution: 0.589824 megapixels
• —Cropped: True
• —Crop style: random
• —Crop aspect: square

defaultdatasetart_896

• —Repeats: 9999
• —Total number of images: 2787
• —Total number of aspect buckets: 1
• —Resolution: 0.802816 megapixels
• —Cropped: True
• —Crop style: random
• —Crop aspect: square

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 = "loona from helluva boss is eating a donut"

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=15,
    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=3.5,
).images[0]
image.save("output.png", format="PNG")