666DZY666 / micronet

micronet, a model compression and deploy lib. compression: 1、quantization: quantization-aware-training(QAT), High-Bit(>2b)(DoReFa/Quantization and Training of Neural Networks for Efficient Integer-Arithmetic-Only Inference)、Low-Bit(≤2b)/Ternary and Binary(TWN/BNN/XNOR-Net); post-training-quantization(PTQ), 8-bit(tensorrt); 2、 pruning: normal、regular and group convolutional channel pruning; 3、 group convolution structure; 4、batch-normalization fuse for quantization. deploy: tensorrt, fp32/fp16/int8(ptq-calibration)、op-adapt(upsample)、dynamic_shape
MIT License
2.21k stars 478 forks source link

能对shufflenet,mobilenet 这种轻量型网络8bit 量化吗? #1

Open Note-Liu opened 4 years ago

666DZY666 commented 4 years ago

可以,更换网络结构就行; nin_gc.py里模型就有用分组卷积结构,效果不错; 大一些的数据集还需要测试。

Note-Liu commented 4 years ago

可以,更换网络结构就行; nin_gc.py里模型就有用分组卷积结构,效果不错; 大一些的数据集还需要测试。

对于shufflenet包含的channel shuffle结构,也能处理吗?

666DZY666 commented 4 years ago

channel shuffle不影响量化,现在nin_gc.py里的模型就含有channel shuffle