mindspore_gs.ptq
Post training quantization algorithms.
import mindspore_gs.ptq as ptq
PTQ Config
Mode for ptq quantizer. |
|
Outliers suppression type for ptq quantizer. |
|
Precision recovery algorithms. |
|
Config for post trainning quantization. |
NetworkHelper
NetworkHelper for decoupling algorithm with network framework. |
|
|
Derived from 'NetworkHelper', a utility class for the MindFormers framework Llama2 network. |
|
Derived from 'NetworkHelper', a utility class for the MindFormers framework ParrallelLlamaForCasualLM network. |
Post Training Quantization Algorithm
Implementation of PTQ algorithm which supports the combination quantization of activation, weight, and kvcache. |
|
Native implementation for post training quantization based on min/max statistic values. |