Joey Gao

XGBOD model file is too large

3

Hello! I spent an hour on my data set training an XGBOD model, and it worked well on the test set, but after saving the model file with PICKLE, I...

【feature】pyecharts作为pandas的画图后端

4

# 背景 之前写过一个`eplot`的包（https://github.com/pjgao/eplot ）模仿`pandas`的`df.plot`模块快速将`pyecharts`和`pandas`集成，直接调用`df.eplot()`/`df.eplot.bar()`...即可。但这种实现现在看来并不够优雅。 我们知道：`pandas`默认的画图接口是`matplotlib`，但在`pandas0.25`及之后的版本里也开放了第三方后端接口的支持。 比如我们把后端换成`plotly` ```python pd.set_option('plotting.backend', 'plotly') ts = pd.Series([1, 2, 3]) ts.plot() ``` 或者 ```python ts.plot(backend='plotly') ``` 这时调用`.plot()`时其实相当于 ```python import plotly plotly.plot(ts) ``` # 原理 从`pandas`的`_get_plot_backend`方法里可以看出：查找后端的逻辑就是先看是`setup`里的`entry_points`是否有等于`pandas_plotting_backends`的字段，没有的话再去直接尝试导入设置的后端名。 因此有两种方法：...

stacking中基模型fit时自定义fit参数

3

比如xgb/lgb early stoping参数

error params order for row and cols in ` plot_power_spectrum`

1

code this line [saliency.py#176](https://github.com/mbeyeler/opencv-python-blueprints/blob/15d7c3d7a06703caee74c5590f76c81b9345acbc/chapter5/saliency.py#L176) should be ```python frame = cv2.copyMakeBorder(frame, 0, nrows-rows, 0, ncols - cols, cv2.BORDER_CONSTANT, value=0) ``` not ```python frame = cv2.copyMakeBorder(frame, 0, ncols - cols, 0, nrows...

抄袭问题

3

https://pdf.hanspub.org/CSA20210700000_49286196.pdf  这个论文是抄袭了作者的项目吗？

print_config not work correctly when union dataclass-like class

3

## 🐛 Bug report ### To reproduce I have these classes,both `A` and `B` are inherited from `Base`，`C` is inherited from `pydantic.BaseModel` (dataclass-like class)： ```python from typing import Union from...

数字标注

1

请问训练数字是一个个截屏分割后保存下来然后手动标注的吗

Error assert when use lazy_instance

2

## 🐛 Bug report when use lazy_instance as default value, there is a assert cannot satisfy ```python assert name not in self.__dict__ ``` ### To reproduce 1. Using the CLI...