Release 6.7 | Clarifai Docs

Skip to main content

GPU Shortage?Make the most out of your GPUs with Compute Orchestration. Learn More

On this page

Changelog 6.7 ​

New Feature Improvement Bug Fix Enterprise Only

API ​

Status Details
Don't allow updating task workers
Don't create duplicated task annotations
Fix detection evals showing and too many metrics calls
Fix pillow installs for webp
Add enum for embed model version id field type.
Don't show model types for backends that aren't responding.
Clean up output_info.data path
Add model_type_id to Model protos.
Add /models/types/{model\_type\_id} endpoint
Prevent models_versions.is_public from every being null.
model mode types that are internal only are being returned.
Create Labeling Order Object and send email to datalabeling@clarifai.com each time backend receives an Labeling Order Object & makes datalabeling a super user
App reindex
Patchable multi-embed workflows with re-index

Model ​

Status Details
Demographics model is now broken in model gallery.
Deprecate model.type from model mode
Update model gallery design
embed_model_version_id should be a dropdown

Portal ​

Status Details
Add versioning to repo, redux, on screen
Allow reviewer to modify annotations during review process.
Store entire canvas state in redux/context, and drive canvas updates by central store
batch v2 shape events by only 1 PATCH/DELETE request
Delete functionality v2
Completely detach labeler rendering from server syncing process to enable background syncing
Input navigating functionality in v2
Set new regionId as "selected" shape if user has selected a transient shape during async updates
Region edit + delete API sync
Fix annotation denormalizer to rehydrate actual concept value
Create new region in labeler v2
Logic to reduce batched drawing events to least number of API operations
Implement new selectors for regions in v2
Normalize Annotations & Regions data for redux storage
Nest labelerTasks reducer inside labeler reducer
Konva: Concept Region drawing implementation
Konva: Implement Rect Transformation
Implement a futureproof schema for labeler interaction events
Move to event-driven design & have the ability to batch updates using custom logic
Create a single Sidebar component for all Labeling types, make children configurable
Remove all props unnecessarily passed from LabelerPage to deep children and make components get props from Redux only
Remove all logic from components to sagas for higher level orchestration of features
Video selector improvements & test updation
Can only save 50 annotations on an image {Usability}
Konva: resizing BBox below minimum size and "crossing over" makes things awry
LabelerPage complete re-render of all components on mouseHover, mousMove (img attached)
Cypress script doesn't terminate webpack-dev-server child process
Cypress pre-run script doesn't check if dev server is already running
Create Unit+Integration testing framework
Integrate headless Cypress with build testing
Switch to react-konva for performant canvas rendering
Add task id to task list
Panning functionality improvements
Lock video playback and interpolation to fps
Annotations created with interpolation seem to have incorrect frame indices
Bounding Boxes and Concepts inconsistent during video playback {Usability}
While annotating video, interpolation freezes and all annotations disappear
Boxes/Interpolation objects are not saving after task submission
Display task instructions to workers in labeler mode
Add infinite scroll loading to labeler carousel
Not incrementing onNext and onPrev pages in Labeler Carousel
Display only minimal log in Portal react app
Update Model mode to use the GET /models/types endpoint
Add list/grid toggle in model mode on all view
Collectors UI should use the layout similar to ModellingMode/LabellerMode
Bulk add concepts to region annotations in app with multi-embed base workflow
Profile page crashes on load
Model mode array of concepts should be unique
Model creation/edit bugs
New Collector page not scrollable
App Workflows - Unable to update model version for custom models
Display Created At Date in App Grid View
Support .txt files from local file browser
Support uploading of multiple video assets as well as image and video assets within the same CSV file
Add better user feedback for uploading text assets
Upload Text by CSV for NLP
Final NLP MVP Feature Changes
Remove 0 area detection filtering from frontend code
Modify accepted CSV upload format so every column corresponds to a network request field
Resolve final bugs with bounding box indexes
Prevent uploading image and video asset types to Text apps
No Visual Feedback for Text input Upload
Support Uploading Files through the OS File Browser for NLP
Custom Model Prediction Bounding Boxes are misaligned from the Detections Bar
Workflow Tab should display and load on initial view for text apps
Unable to navigate between text assets within explorer's asset detail view
CSV uploads not parsing metadata and concepts
Explorer's Advanced Search does not support searching by concepts
Training a classification model no longer display anything within the Custom Model Predictions tab
create annotation CUD sagas for labeler v2
Don't create task annotations in frontend
Form: Input Source showing auto complete options from other apps
Rich text instructions icon bugs
Task create form doesnt force you to set a reviewer if you specify manual review
TypeError: val.add is not a function
Fix autocomplete when user selects "All inputs" for selecting inputs in task creation
Error pops up when collaborator tries to edit task
When I attempt to edit an existing labeling task t...
Page not responding [Usability]
[Explorer] concept thumbnails aren't displaying from model details view
Concept Detail View displays incorrect assets
Add all concepts button to model mode forms
Concept Autocomplete in Model Mode doesn't always display

Workflows ​

Status Details
Allow run workflow and search embedding from embed model in workflow

Applications ​

Status Details
Metadata Namespacing for Clarifai Apps
App details page should send a user to models page to create models rather than using modal