Text-bison model validation post fine tuning

[AviAIML]

Hi, Is there currently a good way to track and record training and validation loss during and/or after fine-tuning? It would be helpful for comparing different training data sets and/or parameter settings, and also deciding how long to continue training.

[Poala_Tenorio · Accepted solution]

Google Cloud Vertex AI does not provide built-in functionality specifically for tracking and recording training and validation loss during and after fine-tuning a Text-bison model. However, you can implement your own custom solution to track and record this information.

Within your fine-tuning code, implement custom logging to record training and validation loss, as well as any other relevant metrics. You can use standard Python logging libraries for this purpose.

You can also save the training and validation loss data to Google Cloud Storage (GCS) as CSV or JSON files. This allows you to store the data in a cloud-based storage solution for easy access and analysis.

Google Cloud Monitoring (formerly known as Stackdriver) provides monitoring and alerting capabilities through. You can set up custom alerts to be notified of any anomalies or issues during the fine-tuning process.

Google Cloud Logging can also be used to capture and store logs from your fine-tuning process. You can set up log sinks to export these logs to various destinations, including BigQuery for analysis.

[Poala_Tenorio · Accepted solution]

Google Cloud Vertex AI does not provide built-in functionality specifically for tracking and recording training and validation loss during and after fine-tuning a Text-bison model. However, you can implement your own custom solution to track and record this information.

Within your fine-tuning code, implement custom logging to record training and validation loss, as well as any other relevant metrics. You can use standard Python logging libraries for this purpose.

You can also save the training and validation loss data to Google Cloud Storage (GCS) as CSV or JSON files. This allows you to store the data in a cloud-based storage solution for easy access and analysis.

Google Cloud Monitoring (formerly known as Stackdriver) provides monitoring and alerting capabilities through. You can set up custom alerts to be notified of any anomalies or issues during the fine-tuning process.

Google Cloud Logging can also be used to capture and store logs from your fine-tuning process. You can set up log sinks to export these logs to various destinations, including BigQuery for analysis.

[AviAIML · Accepted solution]

Thanks for the response, it helps.

[Muskan778899]

i am unable to use my tuned version 

Is there a way to request access or permissions for the tuned model I'm trying to access in Google if I receive a 403 error?