finetrainers/CogView4-6B-rider-waite-tarot-v0

Prompt

TRTCRD a trtcrd of a knight mounting a running horse wearing an armor and holding a staff, \"knight of wands\"

Prompt

TRTCRD a trtcrd of a woman sitting on a throne, wearing a crown and holding a trophee, \"queen of cups\"

Prompt

TRTCRD a trtcrd of a person in a red robe holding a scale and giving coins to two kneeling figures, surrounded by six pentacles

This is a LoRA fine-tune of the THUDM/CogView4-6B model.

Code: https://github.com/a-r-r-o-w/finetrainers

Inference code:

import torch
from diffusers import CogView4Pipeline
from diffusers.utils import export_to_video

pipe = CogView4Pipeline.from_pretrained(
    "THUDM/CogView4-6B", torch_dtype=torch.bfloat16
).to("cuda")
pipe.load_lora_weights("finetrainers/CogView4-6B-rider-waite-tarot-v0", adapter_name="cogview4-lora")
pipe.set_adapters(["cogview4-lora"], [0.9])

image = pipe("<my-awesome-prompt>").images[0]
image.save("output.png")

Training logs are available on WandB here.

NOTE: this checkpoint uses sigmas logit_normal weighting. For shifted_sigmas logit_normal weighting, check https://huggingface.co/finetrainers/CogView4-6B-rider-waite-tarot-v0-shifted-sigmas