How to fix "ERROR: pip's dependency resolver does not currently take into account all the packages that are installed." while using Promt Flow.

Question

How to fix "ERROR: pip's dependency resolver does not currently take into account all the packages that are installed." while using Promt Flow.

Lovedeep Singh 230

I have made a custom image to use python's version 3.10.1 on Azure Ai Foundry 's Prompt Flow ( Earlier Ai Studio). The Flow is running successfully by ignoring the error.

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.mlflow 2.13.0 requires protobuf<5,>=3.12.0, but you have protobuf 5.29.1 which is incompatible.mlflow-skinny 2.13.0 requires protobuf<5,>=3.12.0, but you have protobuf 5.29.1 which is incompatible.

But, I according to me this causing trouble in the final production deployment. And, I have also checked the version of "protobuf" on my custom image but that was 4.25.5 and that should have worked in this scenario. Below is the screen capture of the error and docker container.
Possible_Reason_For_Deployment_Failure_ML_FLow_Docker_Check

The error which I am getting while doing the find deployment is as below:
Error_For_Random_Try_1

The Error is ResourceNotReady as per the documentation for troubleshooting this error 'https://learn.microsoft.com/en-us/azure/machine-learning/how-to-troubleshoot-online-endpoints?view=azureml-api-2&tabs=cli#error-resourcenotready'. there is a mention of score.py file. I want to know what is Score.py file, does this automatically get generated while doing deployment or there is a need to create this file separately while custom image. And ultimately how I can resolve this issue.

romungi-MSFT 49,096 Reputation points Microsoft Employee Moderator

2024-12-17T08:31:06.5866667+00:00

@Lovedeep Singh I think you need to add your customized environment inference_config to the flow.dag.yaml file of your flow. See this section here.

After making this change in the flow.dag.yaml file try to save and retry the deployment.

Lovedeep Singh 230

Can you please explain What I need to do or what point I can add or remove in my current flow.dag.yaml:

id: template_standard_flow
name: Template Standard Flow
inputs:
  Standard_1:
    type: string
    is_chat_input: false
  Kagan:
    type: string
    default: ""
    is_chat_input: false
  Grade:
    type: string
    default: ""
    is_chat_input: false
  Standard_2:
    type: string
    default: ""
    is_chat_input: false
outputs:
  joke:
    type: string
    reference: ${Semantic_Kernel_Code_Updated.output}
nodes:
- name: Semantic_Kernel_Code_Updated
  type: python
  source:
    type: code
    path: Semantic_Kernel_Code_Updated.py
  inputs:
    Grade: ${inputs.Grade}
    Kagan: ${inputs.Kagan}
    standard_1: ${inputs.Standard_1}
    standard_2: ${inputs.Standard_2}
  use_variants: false
node_variants: {}
environment:
  python_requirements_txt: requirements.txt
  image: <my_current_Azure_Container_registry>/fourth-custom-specified-image:v1.0
environment_variables:
  PF_DISABLE_TRACING: true

Lovedeep Singh 230 Reputation points

2024-12-18T06:02:59.07+00:00

Ok thanks a lot @romungi-MSFT

1 answer

Your answer

romungi-MSFT 49,096 Reputation points Microsoft Employee Moderator

2024-12-17T08:31:06.5866667+00:00

@Lovedeep Singh I think you need to add your customized environment inference_config to the flow.dag.yaml file of your flow. See this section here.

After making this change in the flow.dag.yaml file try to save and retry the deployment.
Lovedeep Singh 230 Reputation points

2024-12-17T10:43:20.0966667+00:00

Can you please explain What I need to do or what point I can add or remove in my current flow.dag.yaml:

id: template_standard_flow name: Template Standard Flow inputs: Standard_1: type: string is_chat_input: false Kagan: type: string default: "" is_chat_input: false Grade: type: string default: "" is_chat_input: false Standard_2: type: string default: "" is_chat_input: false outputs: joke: type: string reference: ${Semantic_Kernel_Code_Updated.output} nodes: - name: Semantic_Kernel_Code_Updated type: python source: type: code path: Semantic_Kernel_Code_Updated.py inputs: Grade: ${inputs.Grade} Kagan: ${inputs.Kagan} standard_1: ${inputs.Standard_1} standard_2: ${inputs.Standard_2} use_variants: false node_variants: {} environment: python_requirements_txt: requirements.txt image: <my_current_Azure_Container_registry>/fourth-custom-specified-image:v1.0 environment_variables: PF_DISABLE_TRACING: true
Lovedeep Singh 230 Reputation points

2024-12-18T06:02:59.07+00:00

Ok thanks a lot @romungi-MSFT

Answer 1

@Lovedeep Singh The first two conditions that should be satisfied are:

the docker image must be created based on prompt flow base image, mcr.microsoft.com/azureml/promptflow/promptflow-runtime-stable:<newest_version>. You can find the newest version here.
the environment definition must include the inference_config.

If your image is already created on top of base image, then the current flow.dag.yaml file should point to the correct location of this image.

The second requirement of inference_config this should be added after the image field in your current yaml.

inference_config:
  liveness_route:
    port: 8080
    path: /health
  readiness_route:
    port: 8080
    path: /health
  scoring_route:
    port: 8080
    path: /score

If this answers your query, do click Accept Answer and Yes for was this answer helpful. And, if you have any further query do let us know.

Share via

How to fix "ERROR: pip's dependency resolver does not currently take into account all the packages that are installed." while using Promt Flow.

1 answer

Your answer