weights: add verified compact slim model (1.3 GB)

docs: add available model variants (Slim 1.3GB & Baseline 1.6GB)

docs: fix markdown formatting by replacing brittle KaTeX with clean ASCII diagrams

weights: add outlier-preserved AWQ INT4 binary weights

config: add quant_metadata.json

config: add servitor_config.json

docs: add Spark-Servitor model card and specifications

initial commit