Endlevel - GameMap

During the creation and training of the Steelworks LoRA, the idea arose to also develop the training data for a gaming map. The result is this LoRA.

Completely in gaming style and with many options. Water, glowing iron, molten iron, steel machines and stairs.

Trained with over 145 individual data, 983 training steps in four cycles and 11 epochs. This LoRA is very flexible and stable.

From a 50/50 setting, the game is loaded completely and faithfully to the prompt.

Settings:

• Base Model: Any Flux.1 D (trained with F Comp. fp16)

• Sampler: euler

• Scheduler: beta

• Steps: 17

• Seed: randomize

• Guidance: 0-7

• Latent: 768x1216

• Image Adjustment Nodes (contrast/sharpening/color correct)

• Triggers: game, gameworld, sifi, game envoirment

(Characters not in LoRA included - use your favorite Character Lora´s - will work)

Testing Prompt:

Create a contrast vivid photo-realistic image showing a grand hall from a game. half side view. ceilings. metal stairs. lights. sparks. melded metal. fire. maschines. highly detailed. Use high color contrast and soft lighting to accentuate the amount of details, creating a powerful, evocative atmosphere. dynamic angle view.

Have Fun while gaming and dont forget to eat, sleep an water your flowers! ;)