input_tensors, out_ndims) -> out_shape_tensors ○ Data dependent (op_attrs, input_data, out_ndims) -> out_shape_tensors ○ Data independent (op_attrs, input_shapes, out_ndims) -> out_shape_tensors© 2019 input_tensors, out_ndims) -> out_shape_tensors ○ Data dependent (op_attrs, input_data, out_ndims) -> out_shape_tensors ○ Data independent (op_attrs, input_shapes, out_ndims) -> out_shape_tensors ● Why inputs[j][i] return out Shape function example Use hybrid script to write shape function Input shape tensors Type checking Data independent© 2019, Amazon Web Services, Inc. or its Affiliates. All rights reserved

classification algorithm 1. Load model and labels 2. Build interpreter 3. Allocate input and output tensors 4. Read image 5. Resize image 6. Copy resized image to input tensor 7. Run inference 8. Extract std::unique_ptr interpreter; 11 builder(&interpreter); 20 / 503. Allocate input and output tensors 1 // defined and properly initialized elsewhere: 2 // std::unique_ptr interpreter;

wavelength, position vector, ...) • Quantities may have different character – scalars – vectors – tensors CppCon 2024: Improving our safety with a quantities and units library Introducing quantity_spec wavelength, position vector, ...) • Quantities may have different character – scalars – vectors – tensors • Quantities may be defined as non-negative CppCon 2024: Improving our safety with a quantities

introduced. . . Stateless expression templates Old-school Expression Templates Data (vectors, matrices, tensors. . . ) serve as terminal symbols Expression templates are stateful Stateless Expression Templates

Miscellaneous P1416: Linear Algebra for Data Science and Machine Learning P1770: On vectors, tensors, matrices, and hypermatrices CppCon - Vincent Reverdy - September 17th, 2020 - Online CC0 1.0 Universal