zhiningliu1998/mesa

NeurIPS’20 | Build powerful ensemble class-imbalanced learning models via meta-knowledge-powered resampler. | 设计元知识驱动的采样器解决类别不平衡问题

mesa's Issues

MESA训练集与其他模型的训练集不一样

为什么MESA模型的训练集是 train+valid,其他模型的训练集只有train,并没有用到valid? 这样还有可比性吗?

数据贴标签以及train_ir设置问题

[作者你好，很荣幸能够接触您整理和开源的这个项目。手上正好有一个数据不均衡的多分类项目，我尝试了imbalanced-ensemble中所有的方法，对于我的项目并没有太好的提升。所以想试试这个元学习，之前没有接触过元学习，因为项目急急可能没法去补充详细知识，问的问题可能有点。。。。

是这样的，我把我的数据转换成多个二分类来适应这个项目。但是在运行时，在train_ir有点问题。这个图片的代码是需要0是多数类，1是少数类嘛？不然得需要用train_ir来控制吧？同时这里的imbalance ratio是期望多数类比少数类吧？

同时想问下，对于离散特征数据，多分类数据重叠和不均衡有没有很什么好方法可以推荐下。

最后，再感谢大佬开源的项目。祝您身体健康，合家欢乐。

Could you please release the paper?

Hi, I'm very interested with MESA. But I'm still confused about some details. Could you please release the paper? Thanks!

Not able to run the model

Hi,

I've been trying to run your mesa_example notebook, however, I haven't managed to make it work. When the meta_fit() method is called, I'm getting a RuntimeError. I've tried debugging it, but without any luck. Here is a screenshot of the problem:

Have you encountered this? If yes, could you please tell me how to fix it?

Thanks in advance!

Add Contributor

在多分类任务下报错

我尝试直接修改源代码中的工具类utils.py，完全按照load_dataset的方式读取了自己的数据集，在Rater类下的score函数里加上了acc和micro f1_score尝试做多分类下的评估，用y_pred.argmax(axis=1)来确定预测类别，但是要么维度不匹配，要么就返回TypeError: Singleton array 4 cannot be considered a valid collection..，请问有没有多分类的usage案例可供学习使用？

Issue running model

Hi,

Thanks for the great work. I tried installing the dependencies as in explained in the last version of the ReadMe file and I got:

RuntimeError: one of the variables needed for gradient computation has been modified by an inplace operation: [torch.FloatTensor [50, 1]], which is output 0 of AsStridedBackward0, is at version 3; expected version 2 instead. Hint: enable anomaly detection to find the operation that failed to compute its gradient, with torch.autograd.set_detect_anomaly(True).

Alternatively, when installing the same version of pytorch 1.0.0 with GPU support I got this different issue:
https://discuss.pytorch.org/t/undefined-symbol-cblas-sgemm-alloc/32497

Is there any other way to build the dependencies?

请求修改代码以适配torch新的版本

查资料torch1.4.0以后的版本会报错
RuntimeError: one of the variables needed for gradient computation has been modified by an inplace operation: [torch.FloatTensor [50, 1]], which is output 0 of AsStridedBackward0, is at version 3; expected version 2 instead. Hint: enable anomaly detection to find the operation that failed to compute its gradient, with torch.autograd.set_detect_anomaly(True).

1.4.0是比较老的版本了 好像不支持python3.9 所以请求修改mesa的代码以适配新的torch
谢谢大神！！！

error : mesa.predict_proba

import time
from mesa import Mesa
from arguments import parser
from utils import Rater, load_dataset
from sklearn.tree import DecisionTreeClassifier

if name == 'main':

# load dataset & prepare environment
args = parser.parse_args()
rater = Rater(args.metric)
X_train, y_train, X_valid, y_valid, X_test, y_test = load_dataset(args.dataset)
base_estimator = DecisionTreeClassifier(max_depth=None)

# meta-training
print ('\nStart meta-training of MESA ... ...\n')
mesa = Mesa(
	args=args, 
	base_estimator=base_estimator, 
	n_estimators=args.max_estimators)
mesa.meta_fit(X_train, y_train, X_valid, y_valid, X_test, y_test)
mesa.predict_proba(X_test)

run....................

mesa.predict_proba(X_test)

File "D:\pyyj\mesa-master\environment.py", line 84, in predict_proba
if y_pred.shape[1] == 1:

IndexError: tuple index out of range