feat(rvm): swap RKNN artifacts to 2026-06-03 batch (11 files, all no-resize)

docs(readme): strip private-repo references, sync RKNN list to 2026-06-03 batch

docs(readme): restructure into MODNet (§1) and RVM (§2) chapters; drop frontmatter

docs(rvm): add §7.5 graph-surgery op-count delta (353→317, −36 nodes)

docs: refresh last-updated date and add missing oplv3 RKNN entries

feat: add rvm_mobilenetv3_fp16_544x960 no-resize RKNN model

feat: optimization_level = 3

chore: README improve

feat: add RVM ONNX and RKNN model artifacts

docs: add RVM model registry (ONNX + RKNN) with quality summary

chore: rename rvm models

feat: new rvm models

chore: setup two directories MODNet and RobustVideoMatting

feat: add exported RKNN models for photographic MODNet (256x256, 384x384, 288x512, 576x1024)

docs: Reorganize and formalize MODNet README with comprehensive model registry

feat: fine-tune MODNet-BN on photographic dataset (epochs 1-15)

Put it in the InstanceNormalization OP folded photographic modified ONNX

Upload MODNet PretrainedModels From Author(ZHKKKe)`s Google Driver

initial commit