Differences Between h2o.gbm, h2o.xgboost and h2o4gpu.gradient_boosting_regressor
up vote
0
down vote
favorite
I would like to ask you a question about the different Gradient Boosting Machine functions of h2o package in R. In order to identify the speed difference between these functions; same parameters with same training data has been trained for h2o.gbm, h2o.xgboost and h2o4gpu.gradient_boosting_regressor. The models can be seen below;
model_cpu=h2o.gbm(x = x_col_names, y = y, training_frame = train, nfolds = 10, ntrees = 100, stopping_metric = "RMSE",max_depth = 20) #02:57.36
model_xgb=h2o.xgboost(x = x_col_names, y = y, training_frame = train, nfolds = 10, ntrees = 100, stopping_metric = "RMSE", max_depth = 20,learn_rate = 0.1) #06:31.41
model_gpu=h2o4gpu.gradient_boosting_regressor(n_estimators = 100, nfolds= 10, stopping_metric ="RMSE", max_depth = 20) %>% fit(x_gpu, y_gpu) #2:19.83
"#" sign after the commands indicates the run time of that command. As clearly can be seen; h2o4gpu is the fastest one when we compare. Then, I've decided to go on a more detailed model just increasing the ntree parameter with only h2o4gpu and h2o.gbm. Speed of h2o4gpu was amazing. When h2o.gbm finished in approx. 18 minutes, h2o4gpu finished in 3 and half minutes. Then; I just wanted to compare these models on test data. Result was shocking for me. There were an important difference between the results of these models.
cor_for_h2o.gbm=0.9294249, rmse_for_h2o.gbm=5.822826, mae_for_h2o.gbm=4.024654
cor_for_h2o4gpu=0.9182083, rmse_for_h2o4gpu=6.249201, mae_for_h2o4gpu=4.288272
As I understand, the algorithm behind these two models are different although the parameters are same. What might be the reason behind it? Should I continue to use h2o.gbm even though its slower? Moreover, why h2o.xgboost is much more slower than the others?
Btw, with its grid search option, I would prefer h2o.gbm to h2o4gpu even though it's slower. On the other hand, if you say h2o4gpu is better. Can you suggest any option for hyperparameter tuning in h2o4gpu?
r machine-learning h2o h2o4gpu
asked Nov 22 at 15:20
Cyric
105
add a comment |
up vote
0
down vote
favorite
I would like to ask you a question about the different Gradient Boosting Machine functions of h2o package in R. In order to identify the speed difference between these functions; same parameters with same training data has been trained for h2o.gbm, h2o.xgboost and h2o4gpu.gradient_boosting_regressor. The models can be seen below;
model_cpu=h2o.gbm(x = x_col_names, y = y, training_frame = train, nfolds = 10, ntrees = 100, stopping_metric = "RMSE",max_depth = 20) #02:57.36
model_xgb=h2o.xgboost(x = x_col_names, y = y, training_frame = train, nfolds = 10, ntrees = 100, stopping_metric = "RMSE", max_depth = 20,learn_rate = 0.1) #06:31.41
model_gpu=h2o4gpu.gradient_boosting_regressor(n_estimators = 100, nfolds= 10, stopping_metric ="RMSE", max_depth = 20) %>% fit(x_gpu, y_gpu) #2:19.83
"#" sign after the commands indicates the run time of that command. As clearly can be seen; h2o4gpu is the fastest one when we compare. Then, I've decided to go on a more detailed model just increasing the ntree parameter with only h2o4gpu and h2o.gbm. Speed of h2o4gpu was amazing. When h2o.gbm finished in approx. 18 minutes, h2o4gpu finished in 3 and half minutes. Then; I just wanted to compare these models on test data. Result was shocking for me. There were an important difference between the results of these models.
cor_for_h2o.gbm=0.9294249, rmse_for_h2o.gbm=5.822826, mae_for_h2o.gbm=4.024654
cor_for_h2o4gpu=0.9182083, rmse_for_h2o4gpu=6.249201, mae_for_h2o4gpu=4.288272
As I understand, the algorithm behind these two models are different although the parameters are same. What might be the reason behind it? Should I continue to use h2o.gbm even though its slower? Moreover, why h2o.xgboost is much more slower than the others?
Btw, with its grid search option, I would prefer h2o.gbm to h2o4gpu even though it's slower. On the other hand, if you say h2o4gpu is better. Can you suggest any option for hyperparameter tuning in h2o4gpu?
r machine-learning h2o h2o4gpu
asked Nov 22 at 15:20
Cyric
105
can you let us know big your training_frame was? And just as a quick note are you running h2o4gpu with gpus or cpus, because the other algorithms can only use cpus.
– Lauren
Nov 29 at 20:42
add a comment |
up vote
0
down vote
favorite
up vote
0
down vote
favorite
I would like to ask you a question about the different Gradient Boosting Machine functions of h2o package in R. In order to identify the speed difference between these functions; same parameters with same training data has been trained for h2o.gbm, h2o.xgboost and h2o4gpu.gradient_boosting_regressor. The models can be seen below;
model_cpu=h2o.gbm(x = x_col_names, y = y, training_frame = train, nfolds = 10, ntrees = 100, stopping_metric = "RMSE",max_depth = 20) #02:57.36
model_xgb=h2o.xgboost(x = x_col_names, y = y, training_frame = train, nfolds = 10, ntrees = 100, stopping_metric = "RMSE", max_depth = 20,learn_rate = 0.1) #06:31.41
model_gpu=h2o4gpu.gradient_boosting_regressor(n_estimators = 100, nfolds= 10, stopping_metric ="RMSE", max_depth = 20) %>% fit(x_gpu, y_gpu) #2:19.83
"#" sign after the commands indicates the run time of that command. As clearly can be seen; h2o4gpu is the fastest one when we compare. Then, I've decided to go on a more detailed model just increasing the ntree parameter with only h2o4gpu and h2o.gbm. Speed of h2o4gpu was amazing. When h2o.gbm finished in approx. 18 minutes, h2o4gpu finished in 3 and half minutes. Then; I just wanted to compare these models on test data. Result was shocking for me. There were an important difference between the results of these models.
cor_for_h2o.gbm=0.9294249, rmse_for_h2o.gbm=5.822826, mae_for_h2o.gbm=4.024654
cor_for_h2o4gpu=0.9182083, rmse_for_h2o4gpu=6.249201, mae_for_h2o4gpu=4.288272
As I understand, the algorithm behind these two models are different although the parameters are same. What might be the reason behind it? Should I continue to use h2o.gbm even though its slower? Moreover, why h2o.xgboost is much more slower than the others?
Btw, with its grid search option, I would prefer h2o.gbm to h2o4gpu even though it's slower. On the other hand, if you say h2o4gpu is better. Can you suggest any option for hyperparameter tuning in h2o4gpu?
r machine-learning h2o h2o4gpu
asked Nov 22 at 15:20
Cyric
105
I would like to ask you a question about the different Gradient Boosting Machine functions of h2o package in R. In order to identify the speed difference between these functions; same parameters with same training data has been trained for h2o.gbm, h2o.xgboost and h2o4gpu.gradient_boosting_regressor. The models can be seen below;
model_cpu=h2o.gbm(x = x_col_names, y = y, training_frame = train, nfolds = 10, ntrees = 100, stopping_metric = "RMSE",max_depth = 20) #02:57.36
model_xgb=h2o.xgboost(x = x_col_names, y = y, training_frame = train, nfolds = 10, ntrees = 100, stopping_metric = "RMSE", max_depth = 20,learn_rate = 0.1) #06:31.41
model_gpu=h2o4gpu.gradient_boosting_regressor(n_estimators = 100, nfolds= 10, stopping_metric ="RMSE", max_depth = 20) %>% fit(x_gpu, y_gpu) #2:19.83
"#" sign after the commands indicates the run time of that command. As clearly can be seen; h2o4gpu is the fastest one when we compare. Then, I've decided to go on a more detailed model just increasing the ntree parameter with only h2o4gpu and h2o.gbm. Speed of h2o4gpu was amazing. When h2o.gbm finished in approx. 18 minutes, h2o4gpu finished in 3 and half minutes. Then; I just wanted to compare these models on test data. Result was shocking for me. There were an important difference between the results of these models.
cor_for_h2o.gbm=0.9294249, rmse_for_h2o.gbm=5.822826, mae_for_h2o.gbm=4.024654
cor_for_h2o4gpu=0.9182083, rmse_for_h2o4gpu=6.249201, mae_for_h2o4gpu=4.288272
As I understand, the algorithm behind these two models are different although the parameters are same. What might be the reason behind it? Should I continue to use h2o.gbm even though its slower? Moreover, why h2o.xgboost is much more slower than the others?
Btw, with its grid search option, I would prefer h2o.gbm to h2o4gpu even though it's slower. On the other hand, if you say h2o4gpu is better. Can you suggest any option for hyperparameter tuning in h2o4gpu?
r machine-learning h2o h2o4gpu
r machine-learning h2o h2o4gpu
asked Nov 22 at 15:20
Cyric
105
asked Nov 22 at 15:20
Cyric
105
asked Nov 22 at 15:20
Cyric
105
asked Nov 22 at 15:20
Cyric
105
asked Nov 22 at 15:20
Cyric
105
105
can you let us know big your training_frame was? And just as a quick note are you running h2o4gpu with gpus or cpus, because the other algorithms can only use cpus.
– Lauren
Nov 29 at 20:42
add a comment |
can you let us know big your training_frame was? And just as a quick note are you running h2o4gpu with gpus or cpus, because the other algorithms can only use cpus.
– Lauren
Nov 29 at 20:42
can you let us know big your training_frame was? And just as a quick note are you running h2o4gpu with gpus or cpus, because the other algorithms can only use cpus.
– Lauren
Nov 29 at 20:42
can you let us know big your training_frame was? And just as a quick note are you running h2o4gpu with gpus or cpus, because the other algorithms can only use cpus.
– Lauren
Nov 29 at 20:42
add a comment |
active
oldest
votes
active
oldest
votes
active
oldest
votes
active
oldest
votes
active
oldest
votes
Thanks for contributing an answer to Stack Overflow!
- Please be sure to answer the question. Provide details and share your research!
But avoid …
- Asking for help, clarification, or responding to other answers.
- Making statements based on opinion; back them up with references or personal experience.
To learn more, see our tips on writing great answers.
Some of your past answers have not been well-received, and you're in danger of being blocked from answering.
Please pay close attention to the following guidance:
- Please be sure to answer the question. Provide details and share your research!
But avoid …
- Asking for help, clarification, or responding to other answers.
- Making statements based on opinion; back them up with references or personal experience.
To learn more, see our tips on writing great answers.
Sign up or log in
StackExchange.ready(function () {
StackExchange.helpers.onClickDraftSave('#login-link');
});
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
StackExchange.ready(
function () {
StackExchange.openid.initPostLogin('.new-post-login', 'https%3a%2f%2fstackoverflow.com%2fquestions%2f53434009%2fdifferences-between-h2o-gbm-h2o-xgboost-and-h2o4gpu-gradient-boosting-regressor%23new-answer', 'question_page');
}
);
Post as a guest
Required, but never shown
Sign up or log in
StackExchange.ready(function () {
StackExchange.helpers.onClickDraftSave('#login-link');
});
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
Sign up or log in
StackExchange.ready(function () {
StackExchange.helpers.onClickDraftSave('#login-link');
});
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
Sign up or log in
StackExchange.ready(function () {
StackExchange.helpers.onClickDraftSave('#login-link');
});
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
can you let us know big your training_frame was? And just as a quick note are you running h2o4gpu with gpus or cpus, because the other algorithms can only use cpus.
– Lauren
Nov 29 at 20:42