lujie

emm, I curious about why Darknet model was generated by cfg-file, why not we write it in a normal as other basenets ???

Fuck PFE

2

there are too many bugs in PFE !!!

PFE-pytorch

1

this is my PFE-Pytorch-version[https://github.com/Ontheway361/pfe-pytorch], but it is something amazing that there is neg mls-loss during the training...

#!/usr/bin/env python3 #-*- coding:utf-8 -*- """ Created on 2020/04/23 author: lujie """ import torch import torch.nn as nn import torch.nn.functional as F from IPython import embed class MLSLoss(nn.Module): def __init__(self,...

#!/usr/bin/env python3 #-*- coding:utf-8 -*- """ Created on 2020/04/23 author: lujie """ import torch import torch.nn as nn import torch.nn.functional as F from torch.nn import Parameter from IPython import embed...

坦白来说，FSA-Net应该是我见过长得最丑的模型，没有之一。到目前为止，我手里的实验还没有完全验证完毕，但从已经有的结果来看： 1. 这里面哪些令人膈应的模块(two stream、Fine-grained structure mapping等)，其实是为了发论文而加上去的，去掉之后无伤大雅，预测效果几乎不受影响； 2. 科研工作者更应该提升自己的科研品味，更应该重视对“美”坚持不懈的追求。 谨以此总结，鞭策各位科研人员。

f**k-est implementation

3

DO U really Think this implementation is good ? this is the f**k implementation ever seen, the Logic and Organization of project is just like a shit, f**k the repo,...

pretrain model会release出来么 ？

1

Hi， yanxin， 想问下DCCRNpaper中用的模型可以作为pretrain model release出来么 ？

Hi, Guys, to be honest, i think this repo is just like a shit, if u also think so please feel free, the fuck dataset link is just here https://projects.csail.mit.edu/soundnet/

prior_box size is not ok

1

In my view, the vision of SSD512 is not the ok, becase thare are some errors in size between prior_box and loc, conf. Maybe you miss the layers in multibox...