search

SeuTao / RSNA2019_Intracranial-Hemorrhage-Detection

Code for 1st Place Solution in Intracranial Hemorrhage Detection Challenge @ RSNA2019
411 stars 113 forks source link

can't reproduce the results #9

Closed ericzw closed 2 years ago

ericzw commented 4 years ago

hi,

Thanks for your work!

  1. I used the features provided by you to train the sequence model. I just obtain a score of 0.04578, rather than 0.04383. Can you provide the full used feature(0.04383)?

  2. I follow your instruction to train the 2DNet and extract the features. Then, I use the features to train the sequence model. I just obtain a score of 0.05026. It seems that the features provided by you can't be generated by the 2DNet. Are there some additional operations?

I would be very grateful if you can reply to me.

KarikalStrom commented 4 years ago

Hello We are a group of students creating an app based on this model unfortunately we couldn't produce this model can you please share your results with us it would be really appreciated here is my email : billel2020@live.fr

Manjurashu97 commented 3 years ago

Hi, I am also trying to reproduce your model, but unable to train the model.

Which dataset (stage1 train or stage2 train) is used to train the model? Have any papers been published based to this work? I got an error and I am confused with the training step. Could you please help me to train the model?

Thank you.

LLimerence commented 2 years ago

你好，

感谢您的工作！

  1. 我使用您提供的功能来训练序列模型。我只是获得了 0.04578 的分数，而不是 0.04383。你能提供完整使用的功能（0.04383）吗？
  2. 我按照您的指示训练 2DNet 并提取特征。然后，我使用这些特征来训练序列模型。我刚刚获得了 0.05026 的分数。看来你提供的特征是2DNet无法生成的。是否有一些额外的操作？

如果您能回复我，我将不胜感激。

Hello, is the NPY file you input in the training sequence model the original author's? Which of the three preprocessing methods does he use

ericzw commented 2 years ago

My best result is 0.04262