search

kingfengji / gcForest

This is the official implementation for the paper 'Deep forest: Towards an alternative to deep neural networks'
http://lamda.nju.edu.cn/code_gcForest.ashx
1.31k stars 429 forks source link

gpu acceleration doesn't work for xgboost-only structure #67

Closed ZhouY-JustLL closed 1 year ago

ZhouY-JustLL commented 5 years ago 
I used only xgboost classifier for cascade forest, and used gpu for training by setting`"gpu_id": 0, tree_method: "gpu_hist"`. When the training was started, I checked the memory usage of gpu and it showed that the python script is using gpu. 
But when I compared the gpu method with the cpu method, the time consuming had no significant change, results are almost the same.
The xgboost version is 0.82. Nvidia 1080Ti gpu.
tutu96177 commented 5 years ago

Hi kingfeng, I hope we can comunicate by wechat. My wechat is chshichao。

At 2019-05-16 18:16:56, "ZhouY-JustLL" notifications@github.com wrote:

I used only xgboost classifier for cascade forest, and used gpu for training by setting"gpu_id": 0, tree_method: "gpu_hist". When the training was started, I checked the memory usage of gpu and it showed that the python script is using gpu. But when I compared the gpu method with the cpu method, the time consuming had no significant change, results are almost the same. The xgboost version is 0.82. Nvidia 1080Ti gpu.

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub, or mute the thread.