fix: use Florence-2-base-ft with PR#6 revision for transformers 4.57 compat

feat: Florence-2 auto-captioning + ZIP download on Pose Variations tab too

feat: swap Image Edit tab to Qwen-Image-Edit 2511 (50 steps, proper instruction following)

feat: Florence-2 auto-captioning for dataset generator, ZIP with paired captions

feat: ZIP download with paired caption .txt files for LoRA training, remove consistency LoRA from poses

fix: remove Consistency LoRA from pose/dataset tabs - was destroying identity

rebuild: purpose-built tabs with optimal model/LoRA per task, Z-Image txt2img first tab

feat: multi-model (Klein/Z-Image/Qwen), merged 360 into dataset, strength slider, model selector on all tabs

feat: multi-reference support on Character Sheet and Dataset Generator tabs

fix: add LoRA controls to all tabs

feat: Quick Edit tab, LoRA Dataset Generator (50-150 captioned poses), fix LoRA reload for ZeroGPU

fix: remove unsupported show_download_button from Gallery

feat: advanced controls, LoRA strength, custom LoRA stacking, prompt templates, 360 character sheet

update app

update app

update app

add Flux2-Klein-9B-Consistency Adapter ✅ (#1)

update app

Update app.py

Update app.py

update app ✅

update app ✅

Update app.py

Update app.py

Update app.py

update app

Update app.py

Update app.py

update app

update app

update app

update app

Update app.py

update app

update app

Update app.py

update app

Update app.py

update app

update app

update app

update app

update app

update app

update app

Update app.py

update app

update app

update app

update app