Variety pack of 100 transparent images - captioned for LORA / machine learning training.

All images have been generated using ComfyUI and SDXL checkpoints

Followed by a manual process of removing most of the background for transparency.

Captions have been added by hand as well.

Prompt word: "l0r4" (change to your liking)

Themes include:

• Food

• Landscapes

  • Realistic & Cartoon

• Objects

  • Including weapons

• People

  • Male / Female / Ethnicities

  • Poses (Front / Rear / Side & Closeup)

• Transportation

  • Public & Military

Content:

• [ base_transparent_png_1024 ]

  • Contains all png images with transparency at 1024x1024 resolution.

• [ base_white_jpg_1024 ]

  • Contains all jpg images with a white background at 1024x1024 resolution.

• [ captioned_jpg_512]

  • Contains all jpg images with a white background, and captions, at 512x512 resolution.

• [ captioned_jpg_768]

  • Contains all jpg images with a white background, and captions, at 768x768 resolution.

• [ captioned_jpg_1024]

  • Contains all jpg images with a white background, and captions, at 1024x1024 resolution.

• [text_captions]

  • All captions as text files.