CivArchive
    Zelda 64 Style - ZImageBase
    NSFW
    Preview 121229109
    Preview 121114869
    Preview 121114988
    Preview 121115085
    Preview 121115167
    Preview 121115280
    Preview 121115304
    Preview 121115364
    Preview 121115408
    Preview 121115492
    Preview 121115493
    Preview 121115705
    Preview 121115715
    Preview 121115738
    Preview 121115788
    Preview 121115903
    Preview 121116037
    Preview 121116256
    Preview 121117892
    Preview 121125409

    ZELDA 64 Style - Z-Image/Klein 9B

    This LoRA unleashes Nintendo 64 graphics, specifically textures, colors, and lighting from the Zelda 64 titles Ocarina of time and Majora's Mask. Works great for both Z-Image and Z-image-Turbo. I used a dataset of 78 screenshots I took while playing Ship of Harkinian (OOT and MM versions) and captioned them using natural language generated with Google Gemini. Trained with OneTrainer.

    The focus here is on Texture Variety. You'll have a hard time trying to create specific characters from the games. It can sometimes be achieved but will typically need detailed descriptions of the character you want to depict.

    Suggestions

    • Begin prompts with 'Zelda64 render' as your stylistic anchor for the most consistent look. For the standard Z-Image version, avoid using the low poly tag because the 64-bit geometry is already part of the base weights and adding it can negatively impact your composition. However, for Z-Image-Turbo, the high speed can sometimes cause the style to look too smooth or modern. In those cases, you should include the low poly tag and increase the LoRA strength to 1.1 or 1.2.

    • Generations may sometimes try to produce UI elements. Try using tags such as HUD, Health Bar, text, etc. in the Negative Prompt to help prevent this.

    • Using 'Anime Eyes' in prompts helps to trigger a flat, vibrant look to characters that stays true to the style if that is important for your goal.

    • LoRA Stacking is highly flexible. It plays well with other character, pose, or outfit LoRA's without shattering the N64 logic.

    • Responds well to both natural language and tags.

    Z-Image Settings

    • Strength: 0.8-1.0

    • Sampler: res_multistep or Euler

    • Scheduler: Simple

    • Steps: 25–35

    • CFG: 4.0 I found this to work best.

    • ModelSamplingAuraFlow: 3.0

    Z-Image-Turbo Settings

    • Strength: 1.0-1.2

    • Sampler: Euler

    • Scheduler: Simple

    • Steps: 8–12

    • CFG: 1.0

    • ModelSamplingAuraFlow: 6.0

    Klein 9B Settings

    • Strength: 1.0-1.2

    • Sampler: LCM/Euler

    • Scheduler: sgm_uniform/Simple

    • Steps: 5–9

    • CFG: 1.0

    Resolution: Tested mainly at 832x1216 and 1024x1536. It's very flexible across different aspect ratios.

    Feel free to experiment with these settings; they were what I settled on during my initial testing.

    Have fun and please share your creations!

    Description

    This LoRA work great with both Standard Z-Image and Z-Image-Turbo.

    78 Image dataset used from Ship of Harkinian (Ocarina of Time and Majora's Mask).

    FAQ

    Comments (3)

    sib0tmbFeb 16, 2026· 1 reaction
    CivitAI

    I'm seriously impressed by the results, congrats. Do you mind sharing your Onetrainer settings and how you captioned your images? The dataset would be neat to see but it may be too much to ask, I guess. Thank you for your work nonetheless.

    OddLogic
    Author
    Feb 16, 2026· 3 reactions

    Thanks, glad you like the results!

    For the OneTrainer settings using a 4070 Ti Super, I used Prodigy_Adv with 0.08 weight decay, Cautious WD on, and a constant learning rate. I also threw in a 150-step warmup and used multi-resolution (512, 768, 1024) with 32 rank and 32 alpha. In the concept settings, I turned Crop Jitter off and set Keep Tag Count to 0 since I didn't use a trigger word.

    I focused on capturing as many different texture types as possible while curating the dataset, splitting it about half-and-half between environment shots and character/enemy/boss portraits.

    Here are a few examples of exactly how I captioned them:

    "The character Sheik, blonde hair covering one eye, and deep red eyes, they are wearing a white tunic with a red magnifying glass symbol (the Eye of Truth) on the chest. Their arms are crossed over their chest, showing dark blue shoulder pads, a prominent purple glowing fairy is hovering over the right shoulder, and the background is a dark, rough-textured grey stone room."

    "A close-up view of the boss Volvagia, a long, thin, serpentine/dragon-like creature that appears to be flying against a dark red background. Its body is covered in a bumpy, textured, black and orange/yellow lava or flame pattern, its head has two dark antennae or ears and features large, bright green eyes. Orange flame-like effects are visible around the neck and head, and the body trails off into a long, striped curve."

    "A wide, low-light exterior scene dominated by a massive, dark mountain in the background. The sky is a stormy dark grey, and a swirling ring of intense orange and white fire or energy is hovering directly above the mountain's jagged peak. In the foreground are the brown and grey rooftops of a small village, including one building with a red roof and timber framing, and another stone building with a circular sign displaying a bearded figure."

    Hope that helps.

    sib0tmbFeb 16, 2026

    @OddLogic Thank you very much! Keep up the good work, I look forward for you next lora.

    LORA
    ZImageBase

    Details

    Downloads
    212
    Platform
    CivitAI
    Platform Status
    Deleted
    Created
    2/14/2026
    Updated
    6/10/2026
    Deleted
    6/4/2026

    Files

    Z-Image_Zelda64.safetensors

    Mirrors

    CivitAI (1 mirrors)
    TensorFiles (1 mirrors)