# Standardized ODML Data Containers
Provides a set of standardized on-device ML (ODML) data containers, which could
be shared cross-platforms, among different ODML frameworks (TFLite, MediaPipe,
ML Kit), to increase the interoperability, share data conversions with hardwares
(CPU, OpenGL, OpenCL, NNAPI, AHardwareBuffer, Vulkan, etc.) and reduce data
copies.