Skip to content

[Bug]: Embedding Evaluation experiment fails when using "Titan Embed Text V1" model #359

Description

@AkhilaTestAutomation

FloTorch Version

v3.0.1

Bug Description

When configuring an evaluation experiment and selecting the embedding model "Titan Embed Text V1", the experiment fails upon execution. This issue appears specific to this model and prevents evaluation from proceeding, even though the configuration is otherwise valid.

Steps to Reproduce

Steps to Reproduce
Login to FloTorch.ai
Click on Get Started button
click on Create Project
Enter the Project Name, select the region
Navigate to the Knowledge Bases selection section.
Select an AWS Bedrock Knowledge Base for embedding. Click on Next
Select the N shot prompt and attach the N shot prompt file
Select the KNN value
Select the any of the inferencing models
Select the inferencing model temperature
Click on Next button
Select the Ragas evaluation, embedding model and Inferencing model
Click on Submit

Expected Behavior:
The evaluation experiment should run successfully with the selected embedding model and display metrics/results.
Actual Behavior:
Experiment fails shortly after execution when "Titan Embed Text V1" is selected as the embedding model.

Browser and Version

135.0.7049.115

Runtime Version

python 3.12

Error Messages

Additional Context

No response

Metadata

Metadata

Labels

No labels
No labels

Type

No type

Projects

No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions