Inspirational Post : Use ONNX Runtime in Flutter My .net Application Grab the simplest ONNX model on the shelf It's not a model package yet: it's python source, you need to compile it! To use python in Windows it's better for your health to use WSL (linux subsitem for windows) with any Ubuntu, follow readme for compile instructions. Inspect the resulting .onnx file with netron.app Tweak the model nodes and say "hey mama! I won the machile learning!" graph = helper.make_graph( [ # nodes helper.make_node("Add", ["A", "B"], ["C"], "Add1"), helper.make_node("Add", ["C", "E"], ["F"], "Add2"), ], "SingleAdd", # name [ # inputs helper.make_tensor_value_info('A', TensorProto.FLOAT, [1]), helper.make_tensor_value_info('B', TensorProto.FLOAT, [1]), helper.make_tensor_value_info('E',
TENSORE (ML) Let � be a field such as the real numbers � . A tensor � is an � 0 × � 1 × ⋯ � � array over � : � ∈ � � 0 × � 1 × … × � � . Here, � and � 1 , � 2 , … , � � are positive integers, and � is the number of dimensions. One basic approach (not the only way) to using tensors in machine learning is to embed various data types directly. For example, a grayscale image, commonly represented as a discrete 2D function � ( � , � ) with resolution � × � may be embedded in a mode-2 tensor as � ( � , � ) ↦ � ∈ � � × � . A color image with 3 channels for RGB might be embedded in a mode-3 tensor with three elements in an additional dimension: � � � � ( � , � ) ↦ � ∈ � � × � × 3 . In natural language processing, a word might be expressed as a vector � via the Word2vec algorithm. Thus � becomes a mode-1 tensor � ↦ � ∈ � � . The embedding of subject-object-verb semantics requires embedding relationships among three words. Because a word is itself a vector, subject-ob