Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Package dependecies Issues for BYOM example #231

Open
chenyugaopwc opened this issue Jan 3, 2022 · 8 comments
Open

Package dependecies Issues for BYOM example #231

chenyugaopwc opened this issue Jan 3, 2022 · 8 comments
Assignees
@v-aangie

Comments

@chenyugaopwc
Copy link

According to the document: https://github.com/microsoft/PowerApps-Samples/tree/master/ai-builder/BringYourOwnModelTutorial
Also followed Matteo Iacomelli [email protected] instruction to update "pip==21.0.1" to "pip==20.2.4". Still got the following:

ERROR: pip's dependency resolver does not currently take into account all the packages that are installed. This behaviour is the source of the following dependency conflicts.
pyldavis 3.3.1 requires numpy>=1.20.0, but you have numpy 1.19.5 which is incompatible.
pyldavis 3.3.1 requires pandas>=1.2.0, but you have pandas 0.25.3 which is incompatible.
pycaret 2.3.5 requires lightgbm>=2.3.1, but you have lightgbm 2.3.0 which is incompatible.
pycaret 2.3.5 requires scikit-learn==0.23.2, but you have scikit-learn 0.22.2.post1 which is incompatible.
azureml-train-automl-runtime 1.36.0 requires azureml-automl-core~=1.36.0, but you have azureml-automl-core 1.19.0 which is incompatible.
azureml-train-automl-runtime 1.36.0 requires azureml-core~=1.36.0, but you have azureml-core 1.19.0 which is incompatible.
azureml-train-automl-runtime 1.36.0 requires azureml-dataset-runtime[fuse,pandas]=1.36.0, but you have azureml-dataset-runtime 1.19.0.post1 which is incompatible.
azureml-train-automl-runtime 1.36.0 requires azureml-telemetry=1.36.0, but you have azureml-telemetry 1.19.0 which is incompatible.
azureml-train-automl-runtime 1.36.0 requires azureml-train-automl-client~=1.36.0, but you have azureml-train-automl-client 1.19.0 which is incompatible.
azureml-train-automl-runtime 1.36.0 requires numpy<1.19.0,>=1.16.0, but you have numpy 1.19.5 which is incompatible.
azureml-responsibleai 1.36.0 requires azureml-core~=1.36.0, but you have azureml-core 1.19.0 which is incompatible.
azureml-responsibleai 1.36.0 requires azureml-dataset-runtime~=1.36.0, but you have azureml-dataset-runtime 1.19.0.post1 which is incompatible.
azureml-interpret 1.36.0 requires azureml-core~=1.36.0, but you have azureml-core 1.19.0 which is incompatible.
azureml-defaults 1.36.0 requires azureml-core~=1.36.0, but you have azureml-core 1.19.0 which is incompatible.
azureml-defaults 1.36.0 requires azureml-dataset-runtime[fuse]=1.36.0, but you have azureml-dataset-runtime 1.19.0.post1 which is incompatible.
azureml-contrib-notebook 1.36.0 requires azureml-core=1.36.0, but you have azureml-core 1.19.0 which is incompatible.
azureml-contrib-notebook 1.36.0 requires azureml-pipeline-core~=1.36.0, but you have azureml-pipeline-core 1.19.0 which is incompatible.
azureml-contrib-fairness 1.36.0 requires azureml-core~=1.36.0, but you have azureml-core 1.19.0 which is incompatible.
autokeras 1.0.16.post1 requires tensorflow<2.6,>=2.3.0, but you have tensorflow 2.1.0 which is incompatible.
@rhpoon
Copy link

rhpoon commented Jan 11, 2022

I think Are those from the install phase of your azureml-core in conda/pip/old environment? I came across similar problem (from an old aml sdk 1.19 to 1.36).
Install a new one 1.36. Basically i) complete whatever needed for conda install, incl extra packages pandas ii) then go pip for azureml-core and azure-sdk, iii) for further issues: if they are in conda, then get rid of pip installed packages in general..iv) occassionally i do manual deletion from anaconda......./site-packages/....... path as sometimes there are duplicated package issues.

quite a pain as anaconda don't have all those pipy packages there...

@omartin2010
Copy link
Contributor

omartin2010 commented Feb 11, 2022 •
edited
Loading

having similar issues right now, fixed it with changing the myenv.yml file to :

name: inference_environment
dependencies:
- python=3.6.8
- pip:
    - azureml-defaults==1.19
    - pillow==7.2.0
    - onnxruntime==1.4.0
    - inference-schema
    - azureml-contrib-services
    - numpy

(added the ==1.19 after azureml-defaults)

@rhpoon
Copy link

rhpoon commented Feb 12, 2022 via email

@rhpoon
Copy link

rhpoon commented Feb 12, 2022 via email

@omartin2010
Copy link
Contributor

omartin2010 commented Feb 12, 2022 via email

@rhpoon
Copy link

rhpoon commented Feb 13, 2022

Sorry my two posts above intended to reply to another thread with 'Pandas not found..." NOT this particular one.
I did comeacross above problems and I go to 'terminal' and use pip to update azureml-core to 1.38 which then solve the problem. (Python 3.8 azureml kernel)

@brendon-colburn
Copy link

I was able to fix the pandas not found issue by adding it to the env yaml, during registration I'm now seeing that aibuilder module is not found. I will try the same approach and see what mileage I can get... seems like the documented steps could use an update.

@iamramengirl
Copy link

In my case, I had to downgrade the azureml-core package to 1.38.0 and add pandas in the environment yml file and it resolved the problem.

@iamramengirl iamramengirl mentioned this issue Jul 5, 2022
Open
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees

@v-aangie v-aangie

Labels
None yet
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
6 participants
@omartin2010 @brendon-colburn @iamramengirl @rhpoon @chenyugaopwc @v-aangie