This blog post focuses on new features and improvements. For a comprehensive list, including bug fixes, please see the release notes.
Auto Annotation
Introducing a new automated annotation experience for labeling image inputs. You can. Now leverage machine learning models and workflowsAutomatically annotate your entire image datasets with one click.
Clarifai’s advanced automated annotation feature comes with several ground-breaking capabilities, including:
- Annotate with one click.: Describe all inputs in your dataset with just one click, regardless of size.
- Fully automatic interpretation: Use a fully automated method without relying on a semi-automated, human-in-the-loop method. AI Assist.
- Set confidence limits.: Determine which concepts to annotate automatically and which to evaluate manually based on user-defined confidence limits.
- Targeted review: Only evaluate inputs that fall below your specified accuracy threshold.
- Custom model selection: Select the model or workflow that best suits your dataset during the labeling task creation process.
- Distributed review: Assign input to multiple reviewers, allowing for a faster review process and minimizing a single point of dependency.
- Effective review process: Easily review, accept or reject input annotated by a specific labeler.
- Improved user experience: Access key information such as “% reviewed” and “labeled” to streamline the user experience.
Check out the complete guide to auto-annotation with Clarifai here.
Integrated the EmbedChain framework with ClariFi.
- Embed chain is an open source framework that makes it easy to develop and deploy personalized AI-powered applications. We have integrated it with the ClariFi platform to streamline the creation of personalized LLM applications and offer a seamless process for managing a variety of unstructured data.
- You can check out the integration documentation. Here
Newly published models
Hosted by Clarify. Florence-2-Large, a lightweight, versatile vision language model from Microsoft, excels at multiple tasks.
Here are some key capabilities of the model:
- Handles a variety of vision and vision language tasks through a quick-based approach.
- Trained on the extensive FLD-5B dataset.
- Available in two variants: Florence-2-base (0.23B parameters) and Florence-2-large (0.77B parameters).
Wrapped up Claude 3.5 Sonnet, a fast, sophisticated AI model that excels at reasoning, knowledge, coding, and visual tasks. Ideal for complex applications.
- 2x faster than Claude 3 Opus.
- The Claude 3.5 Sonnet is now the strongest vision model in the Claude family, outperforming the Claude 3 Opus in all standard vision benchmarks.
- Sets new industry standards for Graduate Level Reasoning (GPQA), Undergraduate Level Knowledge (MMLU) and Coding Skills (HumanEval).
Extended organization functionality for non-enterprise plans
- Exclusively for First Enterprise subscribers, organization The functionality is now available for Community and PAYG billing plans. This provides more options for collaboration. However, advanced organization management features, such as roles and teams, will still only be available to Enterprise plan users.
Improved the form validation process on the “Create Your Profile” screen
- We have made sure that all fields are validated before moving to the next screen, by turning off the “Create Profile” button if the form is not valid, and by placing selected fields as input fields. Improved it by overriding the holder values.
Updated the UI design of several organization pages and tables.
- We improved the design of several pages, such as adding a new notification method for invitees to an organization, redesigning tables for members, and introducing new tables for teams and their members or apps. .
Fixed issues with predefined role scopes to enhance functionality and management of contributors and organization/team members
- These improvements ensure more accurate and flexible role assignments, resulting in better access control, collaboration and productivity within the organization.
Added ability to edit workflow name.
- Previously, Workflow Builder only allowed you to set workflow names when creating workflows. The styling was also outdated, and you couldn’t edit the workflow name while in edit mode. We improved this experience. Now, when in edit mode, the workflow name will be highlighted, allowing you to easily start typing a new name.
Made some improvements to the Python SDK.
- To prevent disk writes in environments where the file system may be protected, some features of the facility have been moved exclusively to the CLI. These features are only required for CLI usage and have been moved to this section of the package.
- Added ability to generate a random ID for uploaded text inputs if no input ID is provided in the data loader. Previously, uploading text data to a dataset that already had input would fail. Generating a random ID for new input solves this problem.
- introduced
BaseClient.from_env()And added some new endpoints. - Upgraded to the latest version of the Clarifai-gRPC package (version 10.5.0) to take advantage of the new gRPC functionality.
- Learn more about them. Here.
Released Node.js SDK v0.0.3.
- Added custom certificate support to enable the use of root SSL certificates for secure gRPC connections.
- Added new RAG capabilities.
- Introduced unified documentation for Python and Node.js SDKs.
- Learn more about it here.