-
Notifications
You must be signed in to change notification settings - Fork 161
Description
Can I use custom lora trained by myself other than the loras in XLabs-AI/flux-lora-collection?
But the result seems same with the output without my lora.
This is the script using my lora.
python3 main.py \ --prompt "fzh0326, a man in a cream silk embroidered linen vest and charcoal tactical pants with metallic buckles, shares an umbrella with a woman: long raven hair, off-shoulder chiffon blouse with lace trim, high-waisted denim A-line skirt adorned with floral embroidery. He looks down at her softly, texture0401, rain glistening" \ --image input.jpg --control_type openpose \ --local_path /home/data2/zyf/.cache/raulc0399/flux_dev_openpose_controlnet/model.safetensors \ --lora_local_path /home/data2/zyf/repo/ai-toolkit/loras/texture0401_flux_lora_6000.safetensors \ --use_lora \ --use_controlnet \ --model_type flux-dev \ --width 768 --height 1280 --timestep_to_start_cfg 2 \ --num_steps 20 \ --true_gs 4 \ --guidance 4 \ --save_path ./gen_imgs \
This is the script not using lora.
python3 main.py \ --prompt "fzh0326, a man in a cream silk embroidered linen vest and charcoal tactical pants with metallic buckles, shares an umbrella with a woman: long raven hair, off-shoulder chiffon blouse with lace trim, high-waisted denim A-line skirt adorned with floral embroidery. He looks down at her softly, texture0401, rain glistening" \ --image input.jpg --control_type openpose \ --local_path /home/data2/zyf/.cache/raulc0399/flux_dev_openpose_controlnet/model.safetensors \ --use_controlnet \ --model_type flux-dev \ --width 768 --height 1280 --timestep_to_start_cfg 2 \ --num_steps 20 \ --true_gs 4 \ --guidance 4 \ --save_path ./gen_imgs \
This is the result.

