Upload README.md with huggingface_hub

chore: remove GGUFs (moved to lthn/LEM-Gemma3-4B-GGUF)

docs: switch back to library_name mlx, link to GGUF repo

feat: add model-index with LiveBench eval results for widget display

fix: set library_name to gguf for quant picker widget

feat: add full GGUF range (1-bit to 16-bit, 17 quants) + updated model card

fix: relabel chart axes — positive framing, spectral stability

fix: use PNG for radar chart (SVG not rendered)

docs: move 19D radar chart to top of model card

feat: add GGUF formats, LiveBench scores, 19D radar chart

Initial upload: LEM-Gemma3-4B graduated model (CL-BPL P6 iter 4200)

initial commit