Pin Core ML fp16 to CPU_AND_NE (6.0x on s, 2.8x on x)

Pin Core ML fp16 to CPU_AND_NE (6.0x on s, 2.8x on x)

Core ML: drop fp32, fp16 is equivalent in use

Remove MLX artifacts: slower than Core ML on device

Rebuild mlx export with fork toolchain (ExecuTorch 1.4.1)

Rebuild mlx export with fork toolchain (ExecuTorch 1.4.1)

Add backend config (ExecuTorch 1.4.1)

Add backend config (ExecuTorch 1.4.1)

Add export (ExecuTorch 1.4.1)

Add export (ExecuTorch 1.4.1)

Update models

Update .gitattributes

fix(v0.9.0): switch to size-first layout (s/{xnnpack,coreml}/ + x/{xnnpack,coreml}/)

fix(v0.9.0): restore full 6-file matrix from pre-migration ref bc569f4c and apply canonical layout

Add stub root config.json for HF download counter

Add spec-compliant config.json files

Add spec-compliant config.json files

Restructure to MODEL_SPEC.md convention

Upload 2 files

Upload 2 files

Upload 2 files

Update README.md

initial commit