model card: point widgets/thumbnails to validation images generated with final weights

remove validation error report (fixed)

validation image validation_04.png

validation image validation_03.png

validation image validation_02.png

validation image validation_01.png

validation kernel error report

model card: document v2 continuation (indoor/outdoor dataset, +2000 steps, stability guard)

king2-image continuous fine-tune (indoor/outdoor dataset)

continuation checkpoint 2000

continuation checkpoint 1500

continuation checkpoint 1000

continuation checkpoint 500

remove NaN checkpoints from failed continuation run

RESTORE original adapter_config.json - rollback NaN continuation run

RESTORE original (healthy) LoRA weights - rollback NaN continuation run

king2-image continuous fine-tune (indoor/outdoor dataset)

continuation checkpoint 2000

continuation checkpoint 1500

continuation checkpoint 1000

continuation checkpoint 500

add adapter_config.json for PEFT loading

chore: remove training-state files from checkpoints (optimizer/pickle) — keep LoRA weights only

docs: fix base_model metadata, VAE loading code, broken links; document stickman adapter

Add stickman LoRA adapter v1 (10k HumanML3D stick-figure dataset, 3500 steps)

Upload README.md with huggingface_hub

Upload examples/example_04.png with huggingface_hub

Upload examples/example_03.png with huggingface_hub

Upload examples/example_02.png with huggingface_hub

Upload examples/example_01.png with huggingface_hub

Upload README.md with huggingface_hub

KING2-IMAGE: SDXL LoRA trained on text-to-image-2M subset (Kaggle)

initial commit