.. _auto_ml_image_classification_learn:

*CONNECTS TO A D3VIEW IMAGE CLASSIFICATION SERVER TO LEARN BASED ON ANNOTATIONS GENERATED FROM IMAGE ANNOTATOR*
===============================================================================================================

Connects to a d3VIEW Image Classification Server and triggers a training (learning) run using manually annotated images produced by the Image Annotator worker. Use this worker to train or fine-tune an image classification model within a d3VIEW workflow pipeline.

When to use
-----------

Tagged: ``annotation``, ``d3view_server``, ``image_classification``, ``ml_learn``, ``tfod``, ``training``, ``user``.

Inputs
------

.. list-table::
   :header-rows: 1
   :widths: 20 20 20 20 20 20

   * - Label
     - ID
     - Type
     - Default
     - Required
     - Description
   * - Base URL
     - base_url
     - text
     - http://10.1.10.63:8501/d3view_tfod_model_train/1001
     - ✓
     - Full HTTP URL of the running d3VIEW Image Classification training server (e.g., http://10.1.10.63:8501/d3view_tfod_model_train/1001); change only if the server is hosted on a different host or port than the default.
   * - LDWA
     - dataset
     - dataset
     - —
     - ✓
     - LDWA dataset containing manually annotated images produced by the Image Annotator utility worker; each record should carry image files and their corresponding annotation labels.
   * - Configuration Options
     - config
     - keyvalue
     - *(complex)*
     - ✓
     - Key-value pairs of training configuration options forwarded to the classification server (e.g., epochs, batch size, learning rate); leave default placeholder entries and update keys/values to match server-supported parameters.

Outputs
-------

.. list-table::
   :header-rows: 1
   :widths: 20 20 20 20

   * - Label
     - ID
     - Type
     - Description
   * - UID
     - uid
     - text
     - Unique identifier assigned by the classification server to this training job; use it to correlate status queries and downstream results.
   * - Model Path
     - model_path
     - textarea
     - Server-side file-system path (or URI) where the trained model artefacts are stored after the learning run completes.
   * - Base Url
     - base_url
     - text
     - Echo of the base URL used for this training job; useful for passing the server address to downstream inference or status workers.
   * - Status Url
     - status_url
     - text
     - Polling endpoint URL that can be queried to check the live training status and progress of the submitted job.
   * - Logs
     - logs
     - dataset
     - Dataset containing training log records (loss, accuracy, epoch metrics, error messages) emitted by the classification server during the learning run.

Disciplines
-----------

- ai_ml.supervised.classification
- ai_ml.vision
- platform.integration

.. raw:: html

   <hr style="margin-top:2em">
   <p style="font-size:11px;color:#888">
   Auto-generated from <code>platform</code> schema. Worker id: <code>ml_image_classification_learn</code>. Schema hash: <code>ec4b1629bc4f</code>. Hand-curated docs in <code>workerexamples/</code> override this page when present.
   </p>