FLUX.1-dev LoRA
According to the model card, this is a LoRA derived from black-forest-labs/FLUX.1-dev, matching the hub base-model tags.
Open Source Model Profile · LususNaturae
illuminations-lora
illuminations-lora is a text-to-image LoRA from LususNaturae for FLUX.1-dev. According to the model card, it is derived from black-forest-labs/FLUX.1-dev.
- Publisher
- LususNaturae
- Task
- text-to-image
- Model type
- Unknown
- License
- creativeml-openrail-m
- Library
- diffusers
- Publication status
- Accepted · not indexed
Model overview
illuminations-lora is published by LususNaturae as a diffusers text-to-image LoRA. Hub tags list black-forest-labs/FLUX.1-dev as base model and adapter. According to the model card, it is a LoRA derived from that base with 450 epochs and 300 documented training steps.
Recorded capabilities
Documented validation prompt
According to the model card, the main validation prompt was Photograph of an Irani man riding a horse on the moon at 1024 resolution and 20 steps.
Base text encoder reuse
According to the model card, the text encoder was not trained and the base model text encoder may be reused for inference.
Detailed training settings
According to the model card, training used 450 epochs, 300 steps, LoRA rank 16, dropout 0.1, and 16 images at 512 px.
Diffusers loading example
According to the model card, the publisher shows a DiffusionPipeline example loading the FLUX.1-dev base with LoRA weights.
Use cases in the source record
- Text-to-image generation on the tagged FLUX.1-dev base following the documented validation prompt pattern.
- Diffusers inference experiments reusing the base text encoder and loading the publisher LoRA weights.
Limitations and unknowns
- No evaluation results were extracted from this record.
- No parameter count was extracted for this LoRA record.
- Provider state is historical snapshot data, not independently refreshed current availability.
Source and provenance
Source: LususNaturae/illuminations-lora
Captured: Unknown. Processed: 2026-09-07T19:34:33.792823+00:00.
lora-training This is a LoRA derived from black-forest-labs/FLUX.1-dev . The main validation prompt used during training was: Photograph of an Irani man riding a horse on the moon Validation settings CFG: 3.5 CFG Rescale: 0.0 Steps: 20 Sampler: None Seed: 42 Resolution: 1024 Note: The validation settings are not necessarily the same as the training settings . The text encoder was not trained. You may reuse the base model text encoder for inference. Training settings Training epochs: 450 Training steps: 300 Learning rate: 0.0001 Effective batch size: 20 Micro-batch size: 10 Gradient accumulation steps: 2 Number of GPUs: 1 Prediction…
F001F002F003F005F006F007F008F009F010F011F012F013F014F015