Changelog
Acharya 0.3.4-alpha
Release date: 2023-07-29
Version: v0.3.4-alpha
Applicable Platforms
| Windows | Linux | OSX |
|---|---|---|
| ✔️ | ✔️ | ✔️ |
Bugs Fixed
AT-565 In training output, under per record evaluation scores, show the correct and wrong entry details in UI
AT-568 In workbench, highlight mode, disable mouse hover for grayed out tags
AT-578 Dictionary widget: Read more URL is incorrect
Tasks Completed
AT-213 Sort evaluated records based on accuracy score
AT-542 Create UI for data export
AT-564 In workbench have an option to show the Algo evaluated record side-by-side the annotated record
AT-573 Add links to the record numbers in accuracy per record in training reports to open the records in workbench
AT-576 Ability to pass parameters to Algo docker container
AT-579 Upload UI is confusing for first time users. Make it more user friendly
AT-580 In upload screen modify the form layout to make it easier to understand the screen
AT-581 Remove coreui dependency
AT-582 Remove ngx-monacoeditor dependency
Acharya 0.3.3-alpha
Release date: 2023-02-01
Version: v0.3.3-alpha
Applicable Platforms
| Windows | Linux | OSX |
|---|---|---|
| ✔️ | ✔️ | ✔️ |
Bugs Fixed
AT-540 Svg on login screens do not show up on cloud hosted service
AT-553 Edit entity color from dashboard does not refresh after update the second time
AT-560 Show the number of occurrences of a word in the datamap list. This helps the user understand how many times the word repeats in the record.
AT-569 In websocket connection ReadJSON can panic with error panic: repeated read on failed websocket connection
AT-570 InstanceManager should log to syslog
AT-571 Unicode conversion from IOBStyle and TSV style formats fails on certain text
Tasks Completed
AT-510 Regex search and add regex based tagging option
AT-572 Create a debug option for select background tasks
Acharya 0.3.2-alpha
Release date: 2022-11-11
Version: v0.3.2-alpha
Applicable Platforms
| Windows | Linux | OSX |
|---|---|---|
| ✔️ | ✔️ | ✔️ |
Bugs Fixed
AT-548 Binary build has not packed debug.sh used for training
AT-549 Caching error: Loading dataformat from cache causes data method to return empty data
AT-550 Dialing in teServer enters an infinite loop if no listening tcp socket daemon runs
AT-551 When running Eval from workbench, it should start the eval command if not running in teServer and also should copy NEREntities.JSON to the container
AT-552 Algorithm editor autocomplete is not working
AT-554 In Workbench when the data is edited/changed, eval of the data needs to be re-performed else the suggestions become skewed
AT-555 In Workbench have an option to turn-off eval of workbench data both when there is not algorithms trained and manually
AT-556 Train page auto-refresh - training details are not updated automatically after evaluation is done. One needs to refresh the page
AT-557 Make production build fails on windows
AT-558 Chrome Password forms autocomplete related fixes for admin settings
AT-559 In Light mode UI, MC classification entities are not displayed properly (are shown in white) on workbench
Acharya 0.3.0-alpha
Release date: 2022-08-26
Version: v0.3.0-alpha
Applicable Platforms
| Windows | Linux | OSX |
|---|---|---|
| ✔️ | ✔️ | ✔️ |
Bugs Fixed
AT-518 Upload details for failed upload should not appear on the Upload details table
AT-506 When a lot of entities are present, include a search box in the popup on workbench to search for a specific entity
AT-454 Work with Unicode chars
Tasks Completed
AT-539 Rename test_acharya_api to AcharyaApi
AT-538 Build process should name the binary and constants.
AT-526 In production when run behind a reverse proxy, UI should work with a custom pathname
AT-525 Entity distribution in dashboard should be sorted
AT-524 Cache the analytics output for a given classification and dataid so that it does not have to be recalculated
AT-521 Add ability to edit entities from dashboard Entity Distribution
AT-520 Check user config on login init and if not, then show create password page. Instead of creating password on first run
AT-479 Add Project Name in the Home UI
AT-458 Add ignore lists for anomalies and missing classifications
AT-432 Show details of the notification in the form of a modal when the details button is clicked
Acharya 0.2.0-alpha
Release date: 2022-03-01
Version: v0.2.0-alpha
Applicable Platforms
| Windows | Linux | OSX |
|---|---|---|
| ✔️ | ✔️ | ✔️ |
Bugs Fixed
AT-515 UI UT error fixes
AT-513 Loader total request and pending request does not reset while on the same page
AT-509 Unicode character offset between UI and API with TweeBankNLP dataset
AT-505 compare training command crashes
AT-504 Record next should not respond with status code 500 if records end has reached
AT-501 Import from singer does not update "Total records" in upload details
AT-428 Editing project details adds to Recent projects
Tasks Completed
AT-514 Move from packerd pkg to embed pkg
AT-499 Docker gpu support and deprecated runtime
AT-494 Option to edit tags and upload name on a previous upload
AT-403 Move loader texts to constants
Acharya 0.1.0-alpha
Release date: 2021-10-14
Version: v0.1.0-alpha
Applicable Platforms
| Windows | Linux | OSX |
|---|---|---|
| ✔️ | ✔️ | ✔️ |
Features
- Curate and annotate data
- Edit data and add custom data
- Present insights about the annotated data
- Export annotated data in multiple data formats
- Show annotated entities distribution on both training data and the data used for evaluating trained algorithms (evaluation data or test data)
- Display suggestions and previous classifications
- Auto-labeling with the best trained algorithm on that data
- Configure a custom api/web based dictionary
- Experiment with multiple algorithms with multiple libraries
- Connect the algorithm directly with the code's git repository, Acharya can train off of a custom branch making experimenting easy and integrable into the development cycle
- Train on a remote docker container on either cpu or gpu*
- Write algorithms independent of data format, Acharya will convert to a supported data format before the data is sent for training
- Track data changes and code changes between experiments
- Compare 2 different trainings of the same algorithm, Acharya tracks data changes which increases debug ability
- Help project managers track the progress of data annotation
- Track data overfitting or underfitting of specific entities based upon reports from periodic training and data classification
- Compare model deployed in production with a model freshly trained
- Tag and upload data hit in production and once annotated analyze the accuracy of the model deployed in production
Anything unclear or buggy in this doc, or you couldn't find what you were looking for? Please file an issue on GitHub!