Dataloop announces its strategic integration with IBM’s watsonx.ai Foundation Models. This integration marks a significant step forward in simplifying and accelerating the deployment of RAG, RLHF and GenAI pipelines.
Many organizations face challenges in implementing advanced AI techniques efficiently. The Dataloop-IBM integration addresses these challenges head-on, offering a streamlined solution for AI developers and data scientists.
IBM’s watsonx.ai foundation models paired with the Dataloop platform allows users to access, deploy and run the IBM watsonx models, including Granite and Merlinite models, with unprecedented ease and speed and use it within their workflows and pipelines. This integration enhances the capabilities of Dataloop’s AI development environment and democratizes access to sophisticated AI techniques for a broader range of enterprises. The solution is designed to scale effortlessly, accommodating projects of various sizes and complexities and allows teams to focus more on innovation and less on technical implementation details.
See one of the many integrations in action: 5 Easy Steps to Deploy an RLHF Pipeline Using IBM watsonx.ai Foundation Models & Dataloop
To install IBM watsonx.ai foundation models on the Dataloop platform, follow these steps:
Step 1: Install 3 IBM watsonx.ai foundation models
-
Navigate to your project on the Dataloop platform.
-
Open the left-side Navigation Panel and select Marketplace.
-
In the Marketplace, click on the Models tab at the top of the page.
-
In the left search panel, under the Provider category, select IBM.
-
From the filtered models, click on Explore for the model requested to be installed.
-
In the model info panel on the right, click on Install.
Step 2: Insert project ID and region per model
After installing an IBM model from the Marketplace, its configurations must be updated. To do this, follow these steps:
- Open the left-side Navigation Panel and navigate to Models.
- Click on the name of the model you wish to use.
- On the model information page, under the Configuration section, update the project_id and region fields in the JSON with your IBM account’s project ID and region.
Step 3: Use RLHF template
To install and use the RLHF pipeline template, follow these steps within a project on the Dataloop platform:
Open the left-side Navigation Panel and navigate to Marketplace.
Select the Pipelines tab at the top of the page.
In the search textbox at the top right, type “RLHF”.
From the filtered pipelines, click on Explore for the RLHF pipeline.
On the right pipeline template info panel, click on Install, and then on Use Template.
Step 4: Replace template models with IBM watsonx.ai foundation models
After creating the RLHF pipeline from an installed template on the Marketplace, follow these steps:
Ensure the RLHF pipeline is opened in Edit mode.
On the pipeline canvas, select a Predict Model node.
In the right-side information panel, under the Model field, open the dropdown list and select the IBM watsonx.ai foundation model you want to use.
Repeat steps 2-3 for all other Predict Model nodes.
Step 5: Insert Key name and value
After updating the RLHF pipeline Predict Model nodes with the IBM watsonx.ai foundation models, their secrets must also be updated. To do this, follow these steps:
Ensure the RLHF pipeline is opened in Edit mode.
On the pipeline canvas, select a Predict Model node.
On the right pipeline node information panel, click on the Actions dropdown button and the top right side of the panel.
From the dropdown list, select Edit Service Settings.
In the Edit Service Settings panel, near the Secrets & Integrations section, select Edit Secret.
In the Secrets & Integrations panel, search for the IBM_API_KEY secret and click on the checkbox next to it.
Return to the Edit Service Settings panel by clicking on the Back To Service Setting button.
In the Edit Service Settings panel, near the Init Inputs Value section, select Edit Init Input.
In the Init Inputs panel, edit the value of the Init Input ibm_api_key_name field to be IBM_API_KEY.
Return to the Edit Service Settings panel by clicking on the Back To Service Setting button.
In the Edit Service Settings panel, click on the Save Changes button at the bottom right side.
Repeat steps 2-11 for all other Predict Model nodes.
The result: Human reviewers rank or rate the answers provided by the 3 IBM watsonx.ai foundation models based on quality criteria, resulting in a more precise and effective chatbot!
Watch the full video to see it in action.
Dataloop is excited for this strategic partnership with IBM watsonx.ai. This collaboration represents a significant milestone in our journey to speed up AI development and deployment. Integrating IBM’s watsonx.ai foundation models into our platform allows us to enhance our capabilities and open up new possibilities for our users. This partnership embodies the spirit of innovation and cooperation that drives the tech industry forward.
Dataloop is also part of IBM for Startups and IBM Partner Plus. We value the partnership with IBM and are committed to working alongside IBM to make advanced AI more accessible to businesses of all sizes.
