bitorch.quantizations

This submodule contains several quantization methods that can be used with our quantized layers to build quantized models.

If you want to implement a new function, use the Quantization base class as superclass.

Functions

quantization_from_name	returns the quantization to which the name belongs to (name has to be the value of the quantizations name-attribute)
quantization_names	getter for list of quantization names for argparse
register_custom_quantization	Register a custom (external) quantization in bitorch.

Modules

bitorch.quantizations.approx_sign	Sign Function Implementation
bitorch.quantizations.base	Quantization superclass implementation
bitorch.quantizations.config
bitorch.quantizations.dorefa	Dorefa Function Implementation
bitorch.quantizations.identity	Identity Implementation
bitorch.quantizations.progressive_sign	Progressive Sign Function
bitorch.quantizations.quantization_scheduler	Implementation of a quantization scheduler which replaces quantization functions inside a given model during training.
bitorch.quantizations.sign	Sign Function Implementation
bitorch.quantizations.ste_heaviside	Sign Function Implementation
bitorch.quantizations.swish_sign	Sign Function Implementation