Loading...
Searching...
No Matches
model_expression.h
Go to the documentation of this file.
51 void emit_c_calculate_outputs(ostringstream&, const string& expression, const vector<string>& lines, bool has_softmax) const;
63 void emit_python_calculate_outputs(ostringstream&, const vector<string>& lines, bool has_softmax) const;
69 void emit_js_runtime(ostringstream&, const string& expression, const vector<string>& lines, bool has_softmax, bool use_category_select) const;
75 static vector<string> fix_get_expression_outputs(const string&, const vector<string>&, const ProgrammingLanguage&);
76 static void apply_name_mapping(string&, const vector<string>& original, const vector<string>& mapped);
Abstract base class for all OpenNN layers; orchestrates operators and shape propagation.
Definition layer.h:106
ProgrammingLanguage
Output language for the generated expression source file.
Definition model_expression.h:25
@ JavaScript
Definition model_expression.h:25
ModelExpression(const NeuralNetwork *)
Builds an expression generator bound to the given trained neural network.
void save(const filesystem::path &, ProgrammingLanguage) const
Writes a self-contained source file implementing the network in the chosen language.
string build_expression() const
Returns a symbolic mathematical expression representing the network's forward pass.
Container of layers forming a feed-forward neural network, with parameter storage and I/O.
Definition neural_network.h:20
Definition adaptive_moment_estimation.h:14