qubee.configs.quantization_config.BitConfig Class Reference

qubee.configs.quantization_config.BitConfig Class Reference#

Mobilint SDK qb Compiler: qubee.configs.quantization_config.BitConfig Class Reference
Mobilint SDK qb Compiler v0.12.0.0
MCS002-EN

Configuration for bit precision. More...

Inheritance diagram for qubee.configs.quantization_config.BitConfig:

Public Member Functions

 __init__ (self, int query_act_bits, int key_act_bits, int value_act_bits, int output_act_bits, int ffn_act_bits, int head_act_bits, int query_weight_bits, int key_weight_bits, int value_weight_bits, int output_weight_bits, int ffn_weight_bits, int head_weight_bits, bool mixed_precision_apply, str save_path, str load_path, List[str] activation_16bits, List[str] weight_16bits, bool optional=optional)
 Initialize the BitConfig.
 default_config (cls)
 from_kwargs (cls, **kwargs)
 from_dict (cls, dict config_dict)
 Create BitConfig from dictionary (JSON structure)
 to_dict (self)

Public Attributes

 query_act_bits = query_act_bits
 key_act_bits = key_act_bits
 value_act_bits = value_act_bits
 output_act_bits = output_act_bits
 ffn_act_bits = ffn_act_bits
 head_act_bits = head_act_bits
 query_weight_bits = query_weight_bits
 key_weight_bits = key_weight_bits
 value_weight_bits = value_weight_bits
 output_weight_bits = output_weight_bits
 ffn_weight_bits = ffn_weight_bits
 head_weight_bits = head_weight_bits
 mixed_precision_apply = mixed_precision_apply
 save_path = save_path
 load_path = load_path
 activation_16bits = activation_16bits
 weight_16bits = weight_16bits

Static Public Attributes

bool optional = False
dict DEFAULTS
dict py_to_cpp_name
dict min_max_check

Detailed Description

Configuration for bit precision.

Defines bit-width parameterization for activations and weights used in mixed-precision quantization (e.g., attention and FFN components).

Definition at line 290 of file quantization_config.py.


The documentation for this class was generated from the following file: