spongebbob / tabular_automl_nni

PAKDD AutoML challenge 2nd Feature Engineering Part

ID： 包大人

关注我的知乎

tabular_automl_nni's Issues

Is it possible to add Hyperparameter tuning with autoEF?

Hi,

I am wondering if it is possible to also add hyperparameter tuning step with autoEF? Do you have suggestions how to do it?

Trial failed

I setup the environment with the file requirments.txt and run experiments with command provided in readme.md.
but the first trial failed and the dispatch log looks like
File "/Users/admin/Desktop/tabular_automl_NNI/./autofe_tuner.py", line 65, in generate_parameters sample_p = np.array(self.estimate_sample_prob) / np.sum(self.estimate_sample_prob) TypeError: unsupported operand type(s) for /: 'NoneType' and 'NoneType'

which data to use for criteo example

you give link to criteo example
https://labs.criteo.com/category/dataset/

but there are many data sets in this page
which one to use?

ERROR (nni.runtime.msg_dispatcher_base/Thread-1) unsupported operand type(s) for /: 'NoneType' and 'NoneType' Traceback (most recent call last):

Nonetype Error when running train.tiny.csv

Thanks for your sharing.

I have this problem when trying to run the demo code and train.tiny.csv. The program failed at the first time. When I switched to other sample data like heart or haberman, sample problems happened.
(I'm using windows、lightgbm 2.3.1、nni 0.9.1)

Could anyone tell how to solve this problem. Thanks!

Here is the dispatcher.log:
[05/02/2020, 12:13:50 PM] INFO (nni.msg_dispatcher_base/MainThread) Start dispatcher
[05/02/2020, 12:14:05 PM] ERROR (nni.msg_dispatcher_base/Thread-1) unsupported operand type(s) for /: 'NoneType' and 'NoneType'
Traceback (most recent call last):
File "D:\Anaconda3\lib\site-packages\nni\msg_dispatcher_base.py", line 102, in command_queue_worker
self.process_command(command, data)
File "D:\Anaconda3\lib\site-packages\nni\msg_dispatcher_base.py", line 160, in process_command
command_handlerscommand
File "D:\Anaconda3\lib\site-packages\nni\msg_dispatcher.py", line 106, in handle_request_trial_jobs
params_list = self.tuner.generate_multiple_parameters(ids)
File "D:\Anaconda3\lib\site-packages\nni\tuner.py", line 52, in generate_multiple_parameters
res = self.generate_parameters(parameter_id, **kwargs)
File "D:\git\tabular_automl_NNI.\autofe_tuner.py", line 65, in generate_parameters
sample_p = np.array(self.estimate_sample_prob) / np.sum(self.estimate_sample_prob)
TypeError: unsupported operand type(s) for /: 'NoneType' and 'NoneType'
[05/02/2020, 12:14:09 PM] INFO (nni.msg_dispatcher_base/MainThread) Dispatcher exiting...
[05/02/2020, 12:14:11 PM] INFO (nni.msg_dispatcher_base/MainThread) Terminated by NNI manager

Can't set trialConcurrency

If set trialConcurrency to any number > 1, the program would crush.

Error due to feature name

If some feature name in dataset include "_" ( like A_B), it will cause error.

solution: just rename your feature.