simplify: basic working version

fix: convert steps and guidance_scale to proper types

fix: proper string formatting for errors

fix: initialize generator before use

fix: convert seed and lora_scale to proper types

fix: validate LoRA repo ID format

feat: add LoRA support, higher resolutions, more settings

fix: use ZImagePipeline with guidance_scale=0.0

fix: convert image to RGB

fix: Z-Image-Turbo - 8 steps, no CFG, 1024x1024

fix: use Tongyi-MAI Z-Image-Turbo model

fix: use SDXL Turbo model (free, fast)

fix: gradio 6.5.1

fix: gradio 4.44 + spaces 0.40 for ZeroGPU compatibility

fix: import spaces before torch

fix: spaces version 0.3.0

fix: add @spaces.GPU decorator for ZeroGPU

fix: update sdk version to 3.36

fix: gradio 3.36 with older huggingface-hub

fix: gradio 3.50 without oauth

fix: use gradio 4.19 without huggingface_hub dep

fix: simplify for ZeroGPU compatibility

fix: use gradio 4.36 in SDK version

fix: use older gradio version without oauth dependency

fix: simplify requirements - remove huggingface_hub

fix: update huggingface_hub version for compatibility

Fix version compatibility for huggingface_hub

fix: add PIL import

feat: ZeroGPU version - local model inference

Switch to GuangyuanSD/Z-Image-Distilled (fast 5-15 steps, cfg 1.0-1.5)

Fix Gradio 6.x compatibility

Add README and gitignore

Add Gradio app with HF Inference API integration

Add .env.example template

Add requirements.txt