How do I in Python

Express things

• Implement an attention mechanism?
• Introspect or inspect or list model input variables?
• Interrogate the dimensions of internal layers of a network from within the Python API?
• Port projection of 1D input to 1D output from Python API to C++ API?
• Set the verbosity or traceLevel from Python?
• Restrict a prediction to a bounded interval?

Train models

• Relate alpha, beta1, beta2 and epsilon to learning rate and momentum in adam_sgd?
• Interpret the use of MinibatchSource.next_minibatch?
• Train two or more models jointly?
• Train with a weighted loss?
• Train a multilabel classifier in Python?

Evaluate models

• Evaluate a saved convolutional network?

Adapt models

• Read and modify the training weights from Python?
• Extract features from a specific layer using a trained model?

Read things

• Load model and access network weights / parameters?

Deal with errors

• Get things to work correctly when I take the last element of a sequence?