I am unable to deploy Multilingual text classification package. I have given the input dataset and followed the steps as per documentation. Any suggestions would be great

Ai Center deployment issues

Hi @Pooja_Shiv,

Welcome to community.

If you explain exactly what kind of error you encountered, I’m sure more support will be given. I am attaching the link so you can use it.

Regards,
MY

Hi @Pooja_Shiv

Can you please help to attach the error screen shot.

Thanks

Sure,

MTC

Please make sure you deploy ML Package and ML Package Together(ML Package first and then ML Skill one after other )

Once the status of ML package is deployed and the ML Skill is Available, then you can train your pipeline using the Datasets.

Thanks

@Pooja_Shiv The skill failed because the ML Package is not trained. For Multilingual classification model, the package must be trained first before deploying or else it will fail. Once you train the model, you will get a new minor version, Deploy the skill with the newer minor version (here you can see, you have tried to deploy version 4.0, 0 is the minor version, 0 means its not trained. Once you train with dataset, you will get the minor version 1, so you can deploy the skill for version 4.1).

Attaching the screenshot below.

Hi, Is it possible to share how to train the dataset? My dataset is an excel file.

Refer to this link which have the details on how to prepare data set and train model.

If you are still facing issues, let me know.

Hi, I am still not able to get how to train my dataset. It would be very helpful if you could help me with training the model.
Thanks in advance

@Pooja_Shiv Use a csv file with two cloumns (as mentioned in the above documentation) as the training data file.

  1. Now inside AI Center Project → Datasets, create a folder training and upload the csv file to this folder.
  2. Create a ML Package for Multilingual model.
  3. Inside Pipelines, create a train pipeline. Choose the ML package created in previous step as the ML Package, and choose the train folder created in step 1 as the Train Data, and start the pipeline.
  4. Once the pipeline execution is completed(When the status changes to Successful), you will get a new minor version (x.1)
  5. Create a new ML Skill by choosing the ML Package created and choosing the minor version as 1 (the trained version).