Changelog

Changelog 6.3

New Feature

Improvement

Bug Fix

Enterprise Only

Applications

Status

Details

Send collaborator emails asynchronously

NLP bug fixes for non-text apps

Inputs

Status

Details

Consolidated input related status codes

Add frame.id to API as well as region.track_id

Granted select permission to clarifairead

Annotate

Status

Details

Added list annotations filter status

Added concept selection for tasks

Post/Patch annotations request now allow setting status

Changed task form options

Set annotation status to awaiting for review if the authorized user is not app owner

Return only input_level annotation in input.data

Drawing annotations: wrong embed model version id. Fixed

Eliminated error if no annotation to be deleted

Create one annotation for each bbox

Model

Status

Details

Added support for adding and training on text in the platform

Created a NLP mock prediction endpoint

Created test set to evaluate quick trained models or k-fold if no test search is specified

Added vocab_id for demographics model concepts

Fixed sorting of A.G.E. concepts in golang for demographics model so we don't chop off sets of them

Deprecated Face from javascript Client

Deprecated Face from Java Client

Confusion matrix predicted/true are swapped in evaluation results. Fixed

Explorer Image/Text Joint embedding

Fixed selectEmbedModelVersionId in detection apps

Fixed generalModel imports and optimized video click handlers with useCallback hooks

Status

Details

Persisted the saved search used in train a model version

Created log for annotation/search request/response

Region Searches within Search Bar still use crop coordinates instead of base64 bytes. Fixed

Portal

Status

Details

Created new Single Image View and Image Tools

Enabled Display Text Thumbnails in App Grid View and App Details View

Text Thumbnails display in Portal/Search Bar disabled

Enabled View Text Assets in Portal's Image View

Added Text Inputs To Explorer Apps

Imported new icons for Labeler Image Tools into the style guide

Added login tracking to analytics package in Portal

Allowed pasting into the add inputs text area and clear the text box after clicking submit

Search bar not visible. Fixed

Removed all instances of worker_id from Explorer

Fixed popover left/right overflow

Disabled all search by click handlers in Portal for Text Apps

Click Search button icons on Thumbs not working for localized search. Fixed

Fixed details page header missing description

Fixed demo font syntax

Workflow

Status

Details

Added a Range Slider to filter Workflow Predictions by Value

Updated Face workflow to include the detect faces as concepts for search

Changelog 6.2

New Feature

Improvement

Bug Fix

Enterprise Only

Accounts

Status

Details

Updated Privacy Policy URL

Fixed panic error in Signup

Applications

Status

Details

Ensured collectors are deleted when apps are deleted

View In Explorer button missing in app details. Fixed

Fixed failed to generate thumbnail

Fixed app duplication error when getting worker

Deleted collaborator should also mark application_worker to deleted. Fixed

Inputs

Status

Details

Inputs count stuck at > 0 after delete all, with all inputs seemingly deleted

Label

Status

Details

Task view UI for workers

Create task manager page and task creation page

New Icon for Task Manager/Task Viewer

Fixed POST annotations call on frontend to use correct embed model

Post annotations should include embed_model_version_id. Fixed

Model

Status

Details

Added Apparel Detection to Demo App

Created UI for creating knowledge graph concept relations relations

Create annotation writer model to write annotations to DB

Pass and use test and train data queries to trainer

Added migration to upgrade old model_type in DB

Depredated Face from Python client

Unified the TypeExt and Type fields in model object.

Deprecated facedetect* model types.

Unified FaceEmbedModel and DetectEmbedModel

Converted face.Identity responses to concepts like other detection models to be consistent

Fixed demo font syntax

Video Timeline does not display on the demo app

Fixed Range Slider Value/Text in Apparel Detection Demo

Fixed demographics model to return embeddings and work with auto-annotate

Adding collaborator model counter-intuitively requires ENTER in order to enable the submit button. Fixed

Predict

Status

Details

Validated that concept relation doesn't already exist on POST relations

Prediction requests are being fired too frequently instead of using cache. Fixed

postModelOutputs is not called for newly labeled assets without a manual refresh

Search

Status

Details

Return annotations posted by user in search results

Search by region not working for face detection. Fixed

Make “save” search button internal only

Saved Searches in Portal use the incorrect user ID

Fix crop search from single image view for faces/detections

Portal

Status

Details

UI for collector crud

Deprecate Face from Portal

Improve tabs UI

Video Predictions are failing in Portal

Fixed broken font syntax

Workflow

Status

Details

Video detection workflow prediction support

Public general v1.5 workflow

Allow Patching to existing public workflow

Can not train LOPQ if app base workflow is face. Fixed

Changelog 6.1

New Feature

Improvement

Bug Fix

Enterprise Only

Clients

Status

Details

Remove Feedback endpoints from Python client

Remove Feedback endpoints from Java client

Remove Feedback endpoints from Javascript client

Remove Feedback endpoints from Portal/demo

Remove image.crop field from Python API client

Remove image.crop field from Java API client

Remove image.crop field from Javascript API client

Model

Status

Details

Added detection evaluation in platform

Introduce concept mapping model that uses the knowledge graph relations, creating a path for users to eventually benefit from pool of networked data

Fix a bug that caused the new face predictions to have a huge performance drop

Train and eval worker didn't invalidate model related cache. Fixed

Fix bug in deleting a concept relation by ID

Portal

Status

Details

Bulk labelling can now be done from Explorer mode grid view.

Show Check/X on custom detection model predictions in Portal

Allow multi concepts per bbox

Negative tags not visible in Portal. Fixed

Predict

Status

Details

Remove extra round trip to storage in predict pathway

Remove the image.crop argument during predict and POST /inputs calls to simplify the API

Add region predictions from custom models to detections in videos

Search

Status

Details

Implement search by annotation.status in backend

Connect saved searches and annotation status

Changelog 6.0

New Feature

Improvement

Bug Fix

Enterprise Only

Accounts

Status

Details

Remove country field from signup form, simplifying new customer signups

Essential Plan User can't add collaborators. Fixed

API

Status

Details

Introduce new Python gRPC API client, enabling new features and performance enhancements across API

Introduce new Java gRPC API client, enabling new features and performance enhancements across API

Update API key type for "app_specific" for app-specific keys to be more clear to users

Applications

Status

Details

Allow Personal Access Tokens when calling /users/me (GetUsers)

[Frontend] Enable "Copy Application" from collaborated apps, making it easy to duplicate and build upon existing applications

Program to clean internal apps crashing. Fixed

Data Management

Status

Details

Added the ability to accept b64 Gifs

Functionality to upload pre-tagged images missing. Fixed

Images pre-tagged with concepts do not successfully upload into Clarifai UI On doing bulk uploads (>20-30 urls). Fixed

Bulk image upload issue. Fixed

"Download Failed" error when uploading images. Fixed

Issue with post inputs key being a PAT in a collector. Fixed

Annotate

Status

Details

Skip aligning landmarks if landmark points are out of range to avoid errors and unexpected behavior

Bounding Boxes and Cropped Regions aren't displaying on Videos with default runtime config. Fixed

Insert annotations and related data in batch to improve performance

Model

Status

Details

Evaluate new face embedding model workflow end to end for optimal performance

Validate that concept.app_id shouldn't be set when creating/patching models

Add new predicate to knowledge graph for "relates_to" to represent synonyms

Model training lag. Fixed

Model has missing inputs. Fixed

Submitted models becoming stuck in queue. Fixed

Custom training models when uploaded images are not fully pre-processed. Fixed

Custom facial recognition bboxes do not correspond with detection boxes/ Custom facial recognition prediction interval for video is still 1000ms for apps supporting 100ms runtime config. Fixed

frame_info time off by a factor of 10 for general detection model. Fixed

Detection Models throw error at end of video due to invalid index lookup. Fixed

Workflow

Status

Details

Deleting a workflow should clear or update localStorage. Fixed

Portal

Status

Details

Clean up app overflow UI, improving user experience

Improve Error boundary screen, improving user experience

Add sentry error Id to Error Screen

Images not loading. Fixed

Label and prediction on the right side under Custom Model Predictions section no longer shows up automatically. Fixed

Provide a way for user.metadata to be updated from portal when there are failing apps stuck in there. Fixed

Predictions for a detection model don't show properly in portal. Fixed

Custom facial recognition Predict Boxes not displaying. Fixed

Adding inputs in explorer redirects to explorer view with flashing images. Fixed

Predict

Status

Details

Run prediction by ID in small batch, improving performance

Custom model predictions not displaying. Fixed

Custom model detections not displaying. Fixed

Search

Status

Details

Add file upload input button to explorer search bar, simplifying the UX for file uploads

Filter custom facial recognition bboxes using a sliding bar, adding easy thresholding to custom facial recognition models

Search Bar allows file upload

Remove Explorer App Overflow Menu for improved UX

Changelog 5.11

New Feature

Improvement

Bug Fix

Enterprise Only

Accounts

Status

Details

Create a UI for personal access tokens making it easier for users to access their own apps and any apps where they have been added as collaborators

Updated /keys to work with PATs so that app-specific keys can be created programmatically.

Login (user/PW) has no rate limit/max attempts. Fixed

Remove all instances of worker_id from explorer

When email link to verify my email address clicked, still see "verify your email" banner. Fixed

API services do not function once Queue goes down and comes back up. Fixed. This makes on premise deployments more resilient to power failures.

Applications

Status

Details

Add apps and keys scopes so they can be created with personal access tokens

Copy app count and last_inputs added in app duplication

Fixed demo font syntax

Fixed details page header missing description

Added favicon for Portal

Unable to copy an app that has been shared via Collaborators. Fixed

Setting useCustomConfig isn't checked at login. Fixed

Collaboration apps have race condition where wrong user id is used

Stopped loading of collaborations for search demo/logged-out users

Return “All” scopes when listing available scopes so that you have that option when creating new keys.

Collaborators can not see workers. Fixed

Missing Apps_Get scope in session token auth caused creation of keys to fail temporarily. Fixed

List of missing scopes is not correct in error messages. Fixed

Data Management

Status

Details

Optimize video detection frame rate on Front end

Improve JSON serialization performance in our servers by using an optimized third party library

Able to overwrite default max conn for Citus

Rewrite input counting in the API to be more scalable and robust

Allow RegionInfo from SpireDetectEmbedResponse to contain Point when saving to DB

Unable to upload same file(s) through browse files. Fixed

ffmpeg can produce no frames for very short videos

Add Inputs/View Explorer does not display in new app anymore. Fixed

Clicking video thumbs in detail view does not reload a video. Fixed

Keyboard navigation in image details view highlights incorrect thumb

No Prompt when uploading an image to Explorer through URL. Fixed

Properly return error if AddAssets failed to insert into database

Annotate

Status

Details

Remove classification/detection toggle in image details view

Improved adding negatives to regions

Create one annotation for each bbox

Log capability added for annotation/search request/response

Eliminated error if no annotation to be deleted

Last concept used for bounding boxes is retained between apps. Fixed

The Add Positives / Add Negatives buttons on a Concept details view breaks portal

Custom facial recognition bboxes on grid view do not correlate. Fixed

Model

Status

Details

Ability to keep concepts sorted by alpha in Portal

Implement image crop model to make it possible to work in subregions of an image

Implement random sample model type, adding to fixed function feature set

Update training templates to have more straightforward names and more friendly defaults

Fix the WorkflowInput field name in proto to workflow_input

Allow models that need outputs from previous nodes in a workflow to have access to those outputs to support chaining complex graphs of models

Confusion matrix predicted/true are swapped in evaluation results. Fixed

Fixed generalModel imports and optimize video click handlers with useCallback hooks

Fix for selectEmbedModelVersionId in detection apps

Drawing annotations: wrong embed model version id

Made custom training evaluations for large models stable.

Training progress is saved too frequently, causing very slow training

Return friendlier errors for incorrect parameters passed to templates

Fixed a bug in tracing setup for custom trainer and evaluator

Some models were operating slowly because of lack of resources. Fixed

Training System failed to train some layers. Fixed

Prevent users from evaluating models that are not trainable

Fixed node ID validation logic in Bug in workflows

Predict

Status

Details

Add colors to differentiate region results

Cannot view workflow results in a face app. Fixed

Video spire tests are not running correctly. Fixed

Video processing fails with 'caseids' error. fixed

Search

Status

Details

Add click to search metadata attributes in image details sidebar

Implement visual search in another app as a model type you can add to a workflow

Search bar missing in some cases. Fixed

Region Searches within Search Bar still use crop coordinates instead of base64 bytes. Fixed

Click Search button icons on Thumbs not working for localized search. Fixed

Disable all search by click handlers in Portal for Text Apps

Disable "hide all positively labeled" inputs button for NLP until search works

Scroll active thumb into view in image details carousel

Render Video Assets in Search Bar

Editing geo/json search items no longer work after adding the search bar tooltip. Fixed

TypeError: Cannot read 'get' of undefined when clicking image thumbnails in Explorer search bar. Fixed

Explorer Visibility in small resolution screen improved

Changelog 5.10

New Feature

Improvement

Bug Fix

Enterprise Only

Accounts

Status

Details

Create delete email endpoints in v2 to finally get off old internal endpoints to streamline operations

Create Patch, Delete, Get CreditCards endpoint in v2 APIs to finally get off old internal endpoints to streamline operations

Improved billing for collaborators

PostVerifyEmail error causing some issues not being able to verify their email addresses upon sign-up. Fixed

Fixed flaky email verification integration test to provide more stability to sign-up process

Fixed a link to a non-public version of our API used for development purposes which led to a lot of login issues for users who landed there

Applications

Status

Details

Created display for scopes on collaborator invitations, allowing users to easily understand and control the scope of access allowed for app collaborators

Introduced Collaborators and Collaborations endpoints in API and UIs in Portal

Add ability to upload inputs from App Details screen in Portal

Created collaboration tab in Portal, making it easy to add collaborators to apps

Created display to show the user who invited you to collaborate on an app

Update email phrases for collaborator invitations. After successful sign-up, the user is now redirected to the app's dashboard in Portal

Fixed issue with concept counts in some apps

Clicking pencil icon to edit an API Key in Portal crashed apps. Fixed

Data Management

Status

Details

PATCH /inputs needs to check status of asset before patching

Removed sync DELETE /inputs after runtime config tested

Changed POST /inputs to be async always to simplify processing of workflows after API client tests updated

Added pagination to clusters making for easier data management

Sporadic inability to delete any inputs via Portal or in bulk via the API

Numerous third party security fixes under the hood during ongoing upgrades

Fix 40012 status caused by parallel deletes and adds having a race condition

Update status_changed_at when deleting inputs so we can better track changes

Cache the input counts so that apps can display them in Portal efficiently

Handle killing URL downloading if it is processing for more than 60s. This will make URL processing much more reliable

Return an error if a user sends YouTube video URL as that is not a valid URL to a video we can download

Prevent PostInputs from creating inputs with a user-provided Input.ID that contains a colon

Video calls failed if URLs contain parameters after the file type. Fixed

Failed to resolve DNS MX record in URL down-loader which effected some downloads. Fixed

Investigate why some re-hosted s3 links are no longer working

Getting input counts was broken in some apps, reporting zero, which caused Portal to add an input view to display always

Debug UnicodeErrors in URL downloading to fix URLs with Unicode characters

Fix the poor handling of video too large error message

Unable to batch delete inputs from time to time has been fixed

Media processor video handling was having errors with decoding some videos

Delete Image Button doesn't work in some scenarios

Fixed support for webp image format so it is available again

Annotate

Status

Details

Deploy General Detection Beta Model to recognize multiple objects with bounding boxes.

Deployed new face detector for improved face detection performance over images and video

Created custom training enhancements that handle negatives better for improved model performance

Created evaluation metrics for custom facial recognition in backend for improved facial recognition performance

Topological sort for workflows for scheduling a sequence based on dependencies

Cleaned up duplicate models in workflow model list

Deployed clarifai/main general v1.5 in concept model

Create Pixel Training Hyperparameter Help Guide

Improved accuracy of annotation counts, improving the user experience when annotating inputs

If an image is tagged with a concept that is not in the model, training fails due to KeyError, this is fixed

Fix detection labeling bug where previous images image ratio is used which would cause display issues

We have updated Portal to scale to a large number of concepts with much lower resource usage

Investigate face bounding box probabilities consistency to improve user experience

Bounding box creation canvas in Portal was breaking on resize of the window

Model

Cleaned up duplicate models in the workflow model list, so that you no longer see two General models

Unintended behavior for private model version IDs for certain customers has been fixed

Models referencing deleted backends should be marked as deleted

The latest version of our general model wasn't always default, now it is

Fixed a bug with face recognition evaluations.

Deleted Concepts Persisted in face recognition models. Not anymore!

Inability to see whether a large model is training and making progress, or hung has been addressed to better support our customers

Model won't train in some apps with no positive examples issue has been resolved

Fixed issues with color models failing for a short period of time

Fixed list of models available to workflows to only show a single General model

Predict

Status

Details

Return custom detection evaluations through the GO API

Improved cluster page performance

Investigate health checks killing a prediction backend service, which could affect some predictions in the API

Workflow predict sometimes was failing with 98012 status code. Many fixes here should reduce that

Workflow Predict called the wrong model sometimes. Not any more!

Video playback out of sync with detections in our demos

Fixed issues with regions predicted on inputs would be carried over between inputs in Portal

Fixed the flaky face recognition tests to ensure stability of our face recognition product

Face Detection backends were running out of memory for some predictions, this has been resolved

Return more descriptive error msg for post metric endpoint

Search

Status

Details

Added helper text/suggestions to improve Portal user experience

Header Search return app_owner's user info in collaboration endpoints

Explorer Search Bar - Clicking the green/red circle icons didn't reliably detect click, now it does!

Portal not showing the correct number of results in concept search. Fixed.

Left/right arrows in single image view don't switch between images with regions. Fixed

Fixed carousel thumbnail clicks wiping query params / trigger new search