datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
flux.2-dev-synthetic-2M
Flux2.dev Synthetic: 2.2M Text-to-Image Pairs at 512×512
Dataset Summary
This dataset contains ~2.2 million large-scale synthetic image–caption pairs generated using the FLUX.2-dev diffusion model:
Model: black-forest-labs/FLUX.2-dev
Caption source: Text2Image-2M
Total samples: 2,282,665 (571 shards × ~4000 samples)
Resolution: 512 × 512
Image format: PNG (lossless)
Total shards: 571
Samples per shard: 4000 (last shard: 2665)
Total size: ~865 GB
Each sample consists of:… See the full description on the dataset page: https://huggingface.co/datasets/KempnerInstituteAI/flux.2-dev-synthetic-2M.flux-dev-modelsFluxDevFP8fluxdev_controlnet_16kFluxflux-klein-datasetffhq-256___FLUX.1-dev_training_facesEurasian_Lynx-lynx_for_FLUX-DEV
Eurasian Lynx Dataset
This dataset contains images of the Eurasian Lynx with corresponding captions. As used to train Eurasian_Lynx-lynx_for_FLUX-DEV LoRA on civitai onsite trainer. Captioned with help of fancyfeast/joy-caption-pre-alpha with minor edits by hand
Dataset Structure
001.jpeg: Image of a Eurasian Lynx
002.jpeg: Image of a Eurasian Lynx
CSV File
The image_caption_dataset.csv file contains two columns:
image: Filename of the image
caption: Caption… See the full description on the dataset page: https://huggingface.co/datasets/idiootti-ilves/Eurasian_Lynx-lynx_for_FLUX-DEV.Flux-1-Dev-Images-1kThis dataset was generated by me using Flux 1 Dev at 50 steps
GenEval2_FLUX.2-dev_seed42_guidance4.0_steps50_resolution1024x1024_num_images1white_paper_holdout_3___FLUX.1-devffhq-256___FLUX.1-devGenEval2_FLUX.1-dev_seed42_guidance4.0_steps50_resolution1024x1024_num_images1white_paper_holdout_4___FLUX.1-devMS-COCO-unique___FLUX.1-devFLUX.1-dev-random
Flux 1 Dev Random Images
A dataset of over 1200 synthetic images (1024 x 1024) generated to explore the latent space of black-forest-labs/FLUX.1-dev in a completely randomized, unbiased way.
Generation Parameters
Model Checkpoint: flux1-dev-Q5_K_S.gguf
Sampler: euler_ancestral | Steps: 28 | CFG: 1.0 | Flux Guidance: 3.5
Seed: Random
Prompt Configuration
Prompt: Dynamically assembled string using up to the maximum limit of 512 random tokens.… See the full description on the dataset page: https://huggingface.co/datasets/agentlans/FLUX.1-dev-random.flux-dev-portrait
FLUX-dev portrait
This dataset is generated with FLUX.1-dev using default settings (guidance scale = 3.5, 50 steps) at 1024x1024 resolution. The prompts are taken from k-mktr/improved-flux-prompts-photoreal-portrait. The images are converted to uint8 and encoded with lossless WEBP, and packed in WebDataset format.
This can be used for distillation to another model or self-distillation when training new capabilities for FLUX.
Currently the generated faces are not diverse. Future… See the full description on the dataset page: https://huggingface.co/datasets/gaunernst/flux-dev-portrait.dwy-flux-dev-lokr-6-it
name
comment
img
Archway Elegance.jpg
In this photo, Deng Wenyi is standing elegantly under an antique-style archway. She is wearing a black and white color-block short-sleeved T-shirt paired with black athletic shorts, showcasing a sporty yet casual style. Deng Wenyi's long hair is slightly swaying, adding a touch of dynamism to her appearance. Her facial features are delicate, with clear and bright eyes, and a gentle smile playing on her lips, exuding confidence and composure.… See the full description on the dataset page: https://huggingface.co/datasets/likewendy/dwy-flux-dev-lokr-6-it.flux-1-dev-generated-10kbm-subnet-FLUX.1-devflux-dev-lorasFlux-dev-yellowffhq-256___0-to-9999___FLUX.1-dev_training_facesflux_dev_finetune_1000celeb-a-hq___FLUX.1-devceleb-a-hq___0-to-4999___FLUX.1-dev_training_facesFLUX.1-dev_woman_class_imagesaas_benchmark-flux_devbm-subnet-weekly-FLUX.1-dev-256flux-dev-oneimage