sagarsumant
commented
3 years ago Hello @jingwora - Sorry that you are facing this trouble. I tried reproing your problem like this
- creating vanilla conda environment and installing azureml-sdk[automl]==1.18.0
- install jupyter notebook(conda install nb_conda).
- Clone github notebooks and run them in the sequence you specified
Unfortunately I am not able to repro what you are seeing. So my guess at this point is when you updated your current environment it may have partial updates to 1.18.0 and hence there might be dependency mismatch leading to failures that you are seeing.
There are two options
- Do a pip freeze and ensure the dependency tree for azureml- packages has the right versions 1.18.
- Follow steps here to create new conda environment with the right dependencies.
jingwora
I have runned example as below. 00_Setup_AML_Workspace.ipynb 01_Data_Preparation.ipynb 02_AutoML_Training_Pipeline.ipynb
There is an error in 02_AutoML_Training_Pipeline.ipynb. ImportError: cannot import name 'list_remove_none_object'
Could you tell me how to fix it?
Environment: Azure compute instant
SDK version | 1.18.0
!pip install --upgrade azureml-sdk[automl] !pip install --upgrade azureml-pipeline-steps !pip install azureml.pipeline.steps
in 02_AutoML_Training_Pipeline.ipynb There are some warning as below.
from scripts.helper import get_automl_environment train_env = get_automl_environment(workspace=ws, automl_settings_dict=automl_settings)
WARNING:root:Received unrecognized parameter task WARNING:root:Received unrecognized parameter experiment_timeout_hours WARNING:root:Received unrecognized parameter time_column_name WARNING:root:Received unrecognized parameter max_horizon WARNING:root:Received unrecognized parameter group_column_names WARNING:root:Received unrecognized parameter grain_column_names
There is an error in below code:
from scripts.helper import build_parallel_run_config
PLEASE MODIFY the following three settings based on your compute and experiment timeout.
node_count=2 process_count_per_node=8 run_invocation_timeout=3700 # this timeout(in seconds) is inline with AutoML experiment timeout or (no of iterations * iteration timeout)
parallel_run_config = build_parallel_run_config(train_env, compute, node_count, process_count_per_node, run_invocation_timeout)
train_envtrain_env--------------------------------------------------------------------------- ImportError Traceback (most recent call last)