Sign In

LomoStyle

422
2.5k
74
Updated: Oct 16, 2023
base modelpersonfilmphotorealisticsexyfemale
Download (1.99 GB)
Verified:
SafeTensor
Type
Checkpoint Trained
Stats
2,491
Reviews
(423)
Uploaded
Jul 15, 2023
Base Model
SD 1.5
Training
Steps: 25,000
Epochs: 6
Usage Tips
Clip Skip: 1
Trigger Words
in lomostyle
Hash
AutoV2
C317220DD7
ndimensional's Avatar
ndimensional
0
0
0
0
License:
CreativeML Open RAIL-MAddendum

Lomostyle is the teaser for a future model. Trained on a subset of images and manual captions for the full dataset for testing, this model proved to be competent enough on it's own to justify a release.

So, what is it?
Lomostyle is nod to analog/lo-fi photography. Of which, the most consistent commonalities is the use of film over digital. This can be black and white film, reverse color, color negative, ect.. combined with cross processing techniques that would involve physical manipulation of the film. Overall, you can think of the aesthetic as being 'vintage'.

LomoXL Released here

This model, Lomostyle for SD1.5 will continue to updated.

Requirements :

VAE Required: Download here

Install :

stable-diffusion-webui -> models -> VAE

In Webui :

Under Settings -> Stable Diffusion -> SD VAE select the downloaded VAE and uncheck Ignore selected VAE for stable diffusion checkpoints that have their own .vae.pt next to them

Usage :

Prompt Format : a <angle> <shot size> photo of <<prefix> <nationality> <subject> with <hair length -> color -> style> hair with <skin color> doing <thing + setting>>, <face features: such as "perfect face", "detailed facial features", makeup, ect..>, looking <direction> viewer>, wearing <clothing: if any>, <body type>, <breast size>, <ass size>, <exposed <body part>>, <skin details>, <lighting>, film grain, photographed on a <camera>, <lens: mm + aperture>, with <specific film>, <indoors/outdoors>, in/at <location>, <specific city>, <country>, during the <time of day>, sfw/nsfw, in lomostyle

Note: If that seems confusing or convoluted; look at the prompts in the sample images for better guidance.

Note Deux : Not all tags need be added if not applicable to what you're trying to create.

Example of caption from database :

a closeup portrait photo of a Russian woman with pulled back black hair standing with one arm raised above her head, looking seductively at the viewer, makeup, wearing white striped v-neck shirt with many bead necklaces around her neck and bracelets on her wrist, studio lighting, backlit, green filter, film grain, photographed on a Holga 120 CFN, 120mm, with Fuji Provia film, 100 ISO, medium format, cross-processing, filter, indoors, in Russia, photo shoot, sfw, in lomostyle

Resolution :

The model was trained on 704x704 images. But feel free to use whichever resolution you'd like.

Tags :

Note, this is not an extensive list. The model recognize tags not listed here, and some tags listed might have more weight than others—some being barely noticeable.

NSFW/SFW :

  • nsfw

  • sfw

  • erotica (usually used when image was sfw but suggestive)

  • erotic (usually used in tandem with nsfw)

Cameras :

  • Asahi Pentax SV

  • Canon 3000N

  • Canon AE-1

  • Canon Canonet 28

  • Canon Elan 7

  • Canon EOS 3

  • Canon EOS 5

  • Canon EOS 5D

  • Canon EOS 33

  • Canon EOS 500

  • Cannon EOS Elan II

  • Canon EOS Series

  • Chamonix 045n-2

  • Contax G2

  • Fed 3

  • Fuji Instax Mini 90 Neo Classic

  • Fuji Instax Wide 210

  • Fuji Instax Wide

  • Hasselblad 500

  • Hasselblad 503 CWD

  • Hasselblad XPan

  • Holga 120 CFN

  • (to be continued)


Film :

  • Fuji Superia 400

  • Kodak Color Plus 200

  • Fuji Superia X-TRA 800

  • Fuji Superia 100

  • Fuji Superia 200

  • Ilford hp4 black and white film

  • Ilford Pan 100 black and white film

  • Kodak Elitechrome EBX

  • Kodak Gold 200

  • Lomography Color Negative 400

  • Lomography Color Negative 800

  • Photo service selection 200 film

  • Kodak Portra 800

  • Ilford Delta 3200 black and white film

  • Lomography Color Negative 100

  • Kodak VISION2 500T

  • Kodak VISION3 500T

  • Cinestill 800T

  • Fuji Provia 100F

  • Lomography XPro 200 ISO 35mm film

  • Kodak BW400CN black and white film

  • Fuji Velvia 100F

  • Kodak Ektar 100

  • Fuji Velvia Slide 100

  • Kodak T-Max 400 black and white film

  • Kodak Aerochrome III 1443 film

  • Lomography Metropolis 100-400

  • Ilford Delta 400 black and white film

  • Agfa RSX II

  • Fuji Superia X-TRA 400

  • Kodak UltraMax 400

  • Handmade film

  • Kodak Aerochrome EIR, infrared film

  • Lomography Redscale XR 50-200, redscale film

  • Lomography X-Pro Slide 200

  • (to be continued)

Angles :

  • Front view

  • Rear angle / From behind / Back turned towards viewer

  • Side view / Side angle

  • From above / Looking down at / Top down

  • 45 degree

  • Low angle

  • High angle

  • Eye Level

  • Hip Level

  • Cowboy shot

  • Knee Level

  • Ground Level

  • Over-the-shoulder shot

  • Combinations such as :

    • Rear angle Side view

    • 45 degree rear view

    • ect..

Shot Size :

  • Extreme closeup

  • Closeup

  • Portrait (Note : can be combined with other sizes)

  • Medium closeup

  • Medium shot

  • Medium-full shot

  • Full body

Lighting :

  • Natural Light - light from a natural source. Such as,

  • Sunlight - self explanatory I hope.

  • Ambient Light -Think in the vein of the soft glow from the LED/LCD lights of your computer monitor at night.

  • Cinematic Lighting - High contrast, dramatic, deep, combo of soft light and hard light.

  • Hard Light - Light that hits the subject directly from the source, creates sharp shadows.

  • Soft Light / Diffused Soft Light - Light that is diffused before hitting the subject, creates soft shadows

  • Reflective Light - Light that reflects off of the subject and is registered by the camera.

  • Tungsten Light - The color temperature associated with incandescent light, typically 3,200K.

  • Dappled Light - Light that typically comes through gaps in a tree canopy.

  • Eyelight - A small light dedicated to adding a glint to the subjects' eyes.

  • Florescent light - Type of ambient light - think, the lightbulbs in many homes.

  • Fill Flash Lighting - Using cameras flash in low-light environment

  • More lighting terms :

    • backlit

    • side lighting

    • silhouette

    • rim lighting

Misc. Tags :

  • Developed Film

  • Expired Film

  • Hand Developed Film

  • c41 process

  • Fashion

  • Selfie

  • Artistic Nude

  • Over Exposed

  • Flash

  • Film Still

  • Sprocket Holes / Sprockets (those holes in film)

  • Lens Flare

  • Light Leak

  • Bloom

  • Colorful

  • Seasons (Summer, winter, ect..)

  • Scan Lines

  • Cross Processing / Cross-processed

  • Overlay Text (format = Film Overlay text: ' ')

  • Shadows :

    • Sharp Shadows

    • Soft Shadows

  • Panorama / Panoramic

  • Waist up / Waist Down

  • Bodyscape

  • Photo Type (other than portrait/full body/ect.):

    • landscape, architecture, cityscape, seascape, seaside, artistic

  • Redscale

  • Infrared

  • Tungsten film

  • Time of day:

    • morning, dawn, day, afternoon, evening, dusk, night

Webui Settings :

Clip Skip : 1

  • Settings -> Sampler parameters -> Clip skip

ETA Noise Seed Delta : 31337

  • Settings -> Sampler parameters -> Eta noise seed delta

Face Restoration : Disabled (See ADetailer)

  • IF enabled : Settings -> Face restoration -> CodeFormer -> weight = 0.8

Parameters :

  • Sampler :

    • I used a combination of Euler a, DPM++ SDE Karras for many of the sample images, but this is just personal preference.

  • Steps : Depends on the sampler, typically 20 - 35 works well.

    • If DDIM use higher steps

Hires Fix (Optional) :

  • I used x1_ITF_SkinDiffDetail_Lite_v1 which you can find - here.

  • Note : You don't need to download the .yaml, just the .pth file.

  • Install to stable-diffusion-webui -> models -> ESRGAN

  • Feel free to use any other upscaler you prefer, this was just the one I used for the sample images.

Extensions :

ADetailer :

Extension for webui that uses ultralytics to 'fix' faces/hands/body after generation.

Download here - https://github.com/Bing-su/adetailer

Read the information in the Github repo linked above for more info and installation guide.

Questions/Feedback/Updates?

Visit my thread on the Unstable Diffusion Discord

Check out my other models :

SDXL

SD1.5

LoRA