YZW-explorer

Does anyone know the source code of channel calculation in pytorch?

Dear author, I would like to confirm one thing with you. Does the quantization layer of Brevitas use the value after inverse quantization when calculating the convolution? (Floating point numbers),Shouldn't...

QuantRelu's forward reasoning process

6

Hello author, I am still trying to use pytorch to simulate the quantization and dequantization forward reasoning process of brevitas. I have the following two questions to ask you: 1....

Question about the matching of finn and onnx versions

2

Hello author, I am trying to implement this code https://github.com/Xilinx/finn/blob/master/notebooks/end2end_example/cybersecurity/2-export-to-finn-and-verify.ipynb I want to ask what the version of FINN is, my .onnx file is exported using onnx version 1.8.1

How can I quantify my model

Urgent unknown problem needs help！！！

1

I want to implement the deployment of 4bit quantization on the PYNQ board. I saw that the MobileNet-v1 example in the example is quantified to 4bit, but I saw it...

作者您好，很有幸使用了您的代码。我在尝试跑通您的代码，但是在运行的时候，终端疯狂有输出，请教一下您为什么，不胜感激！ 

关于损失值以及预训练次数的问题

1

各位大佬好！预训练次数是多少啊，以及，我如何通过生成器损失值,判别器损失值和内容损失值,这三个属性判断我的模型训练争取，我的数据如下所示，不胜感激！ Pre-training start! [1/50] - time: 340.47, Recon loss: 20.638 [2/50] - time: 337.64, Recon loss: 16.235 [3/50] - time: 337.67, Recon loss: 14.199 [4/50] - time: 337.59, Recon loss:...

FP8 PTQ With Physical Dependency 0515

2

### PR types New features ### PR changes APIs ### Description PaddleNLP设计逻辑： llm/fp8quant.py 定义FP8的量化逻辑，将FP8UniformObserver写入QuantConfig中 llm/fp8finetune_generation.py 调用llm/fp8quant.py中的量化逻辑，完成全部量化过程

FP8 PTQ With Physical Dependency 0515

2

描述：定义新的FP8 observer，目前只支持e4m3格式 功能：统计量化对象的abs_max，并根据abs_max和fp8格式的取值范围计算scale PaddleSlim设计逻辑： paddleslim/quant/observers/fp8uniform.py 定义FP8UniformObserverLayer类用以统计量化对象的abs_max以及根据用户选择的要量化到的FP8类型设置正确的量化区间，并最终计算出scale，方便FP8LinearQuanterDequanter进行量化和反量化操作 paddleslim/quant/observers/__init__.py 将FP8UniformObserver添加到__init__中方便外部调用