README.md

---
license: other
base_model: "black-forest-labs/FLUX.1-dev"
tags:
  - flux
  - flux-diffusers
  - text-to-image
  - image-to-image
  - diffusers
  - simpletuner
  - safe-for-work
  - lora
  - template:sd-lora
  - standard
pipeline_tag: text-to-image
inference: true
widget:
- text: 'unconditional (blank prompt)'
  parameters:
    negative_prompt: ''''
  output:
    url: ./assets/image_0_0.png
- text: 'Photography of a beautiful 20-year-old woman in a yoga studio looking at the camera, wearing a full white body-hugging top with no bra and tight white yoga pants. She has long, dark brown hair with hints of red and large breasts that stretch her top to its limit.'
  parameters:
    negative_prompt: ''''
  output:
    url: ./assets/image_1_0.png
- text: 'A photorealistic full-body top-down angle Instagram influencer-style photo featuring a captivating 28-year-old woman with dark brown hair with hints of red and deep brown eyes. She wears an imaginatively stylized corset with a garter belt and delicate stockings, exuding a magical, alluring, and sultry charm. Her toned physique is emphasized with a narrow waist, long slim legs, and a thigh gap. She is lying in soft grass against a hyper-detailed and beautifully magical fantasy background. The overall scene is vibrant and dynamic, showcasing her perfect facial features and radiant eyes in an ultra-maximalist style.'
  parameters:
    negative_prompt: ''''
  output:
    url: ./assets/image_2_0.png
- text: 'A young woman stands on a rain-soaked street in a neon-lit cyberpunk city, her face fierce and determined. The woman is dressed in futuristic armor, glowing with electric lines as she stares down a dark alley. Her eyes reflect the neon glow, adding intensity to her gaze.'
  parameters:
    negative_prompt: ''''
  output:
    url: ./assets/image_3_0.png
- text: 'A heartbreakingly beautiful tall 44-year-old woman with long dark brown hair with hints of red styled over one eye. She has heavy makeup with dramatic eyeshadow, winged mascara, and glossy blood-red lips. Her piercing brown eyes smolder with confidence. She has a fit, toned, and very tanned body with 36D natural breasts. She is wearing a school uniform with a black pleated skirt and a white shirt tied in a knot to reveal her midriff. She has black stockings with lace tops and black high-heeled school shoes. She gives the viewer a flirty smile over her shoulder, lifting her skirt to show her booty from behind, with one hand on her peachy ass, revealing her tiny navy blue panties. She is in a highly detailed classroom with the backboard, school desks, and chairs visible. The view is from behind and below.'
  parameters:
    negative_prompt: ''''
  output:
    url: ./assets/image_4_0.png
- text: 'Realistic wide shot photo of woman posing in a luxurious satin lingerie set, featuring a plunging bra, delicate thong and a classic garter belt with black stockings. The satin lingerie shimmers softly in the light, and the cut emphasizes both sophistication and a hint of allure. The lingerie is detailed with fine lace edges, highlighting her alluring figure. She elegantly styled hair as if getting ready for a formal event. The photo has a cinematic quality with rays of light and dramatic play of shadow and light'
  parameters:
    negative_prompt: ''''
  output:
    url: ./assets/image_5_0.png
---

# reddy-v4

This is a standard PEFT LoRA derived from [black-forest-labs/FLUX.1-dev](https://huggingface.co/black-forest-labs/FLUX.1-dev).


The main validation prompt used during training was:
```
Realistic wide shot photo of woman posing in a luxurious satin lingerie set, featuring a plunging bra, delicate thong and a classic garter belt with black stockings. The satin lingerie shimmers softly in the light, and the cut emphasizes both sophistication and a hint of allure. The lingerie is detailed with fine lace edges, highlighting her alluring figure. She elegantly styled hair as if getting ready for a formal event. The photo has a cinematic quality with rays of light and dramatic play of shadow and light
```


## Validation settings
- CFG: `3.5`
- CFG Rescale: `0.0`
- Steps: `20`
- Sampler: `FlowMatchEulerDiscreteScheduler`
- Seed: `42`
- Resolution: `832x1216`
- Skip-layer guidance: 

Note: The validation settings are not necessarily the same as the [training settings](#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: 10
- Training steps: 2000
- Learning rate: 0.0001
  - Learning rate schedule: constant
  - Warmup steps: 500
- Max grad value: 2.0
- Effective batch size: 1
  - Micro-batch size: 1
  - Gradient accumulation steps: 1
  - Number of GPUs: 1
- Gradient checkpointing: True
- Prediction type: flow-matching (extra parameters=['shift=3', 'flux_guidance_mode=constant', 'flux_guidance_value=1.0', 'flow_matching_loss=compatible', 'flux_lora_target=all'])
- Optimizer: adamw_bf16
- Trainable parameter precision: Pure BF16
- Base model precision: `no_change`
- Caption dropout probability: 10.0%


- LoRA Rank: 16
- LoRA Alpha: None
- LoRA Dropout: 0.1
- LoRA initialisation style: default
    

## Datasets

### reddy-v2-512
- Repeats: 10
- Total number of images: 13
- Total number of aspect buckets: 1
- Resolution: 0.262144 megapixels
- Cropped: False
- Crop style: None
- Crop aspect: None
- Used for regularisation data: No
### reddy-v2-1024
- Repeats: 10
- Total number of images: 5
- Total number of aspect buckets: 1
- Resolution: 1.048576 megapixels
- Cropped: False
- Crop style: None
- Crop aspect: None
- Used for regularisation data: No


## Inference


```python
import torch
from diffusers import DiffusionPipeline

model_id = 'black-forest-labs/FLUX.1-dev'
adapter_id = 'Unmapped2895/reddy-v4'
pipeline = DiffusionPipeline.from_pretrained(model_id, torch_dtype=torch.bfloat16) # loading directly in bf16
pipeline.load_lora_weights(adapter_id)

prompt = "Realistic wide shot photo of woman posing in a luxurious satin lingerie set, featuring a plunging bra, delicate thong and a classic garter belt with black stockings. The satin lingerie shimmers softly in the light, and the cut emphasizes both sophistication and a hint of allure. The lingerie is detailed with fine lace edges, highlighting her alluring figure. She elegantly styled hair as if getting ready for a formal event. The photo has a cinematic quality with rays of light and dramatic play of shadow and light"


## Optional: quantise the model to save on vram.
## Note: The model was not quantised during training, so it is not necessary to quantise it during inference time.
#from optimum.quanto import quantize, freeze, qint8
#quantize(pipeline.transformer, weights=qint8)
#freeze(pipeline.transformer)
    
pipeline.to('cuda' if torch.cuda.is_available() else 'mps' if torch.backends.mps.is_available() else 'cpu') # the pipeline is already in its target precision level
model_output = 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(42),
    width=832,
    height=1216,
    guidance_scale=3.5,
).images[0]

model_output.save("output.png", format="PNG")

```