The Label Studio ML backend is an SDK that lets you wrap your machine learning code and turn it into a web server. The web server can be then connected to Label Studio to automate labeling tasks and dynamically retrieve pre-annotations from your model.
There are several use-cases for the ML backend:
- Pre-annotate data with a model
- Use active learning to select the most relevant data for labeling
- Interactive (AI-assisted) labeling
- Model fine-tuning based on recently annotated data
If you just need to load static pre-annotated data into Label Studio, running an ML backend might be overkill for you. Instead, you can import preannotated data.
Follow this example tutorial to create a ML backend service:
-
Install the latest Label Studio ML SDK:
git clone https://github.com/HumanSignal/label-studio-ml-backend.git cd label-studio-ml-backend/ pip install -e .
-
Create a new ML backend directory:
label-studio-ml create my_ml_backend
You can go to the
my_ml_backend
directory and modify the code to implement your own inference logic. The directory structure should look like this:my_ml_backend/ ├── Dockerfile ├── docker-compose.yml ├── model.py ├── _wsgi.py ├── README.md └── requirements.txt
Dockefile
anddocker-compose.yml
are used to run the ML backend with Docker.model.py
is the main file where you can implement your own training and inference logic._wsgi.py
is a helper file that is used to run the ML backend with Docker (you don't need to modify it)README.md
is a readme file with instructions on how to run the ML backend.requirements.txt
is a file with Python dependencies. -
Run the ML backend server
docker-compose up
The ML backend server will be available at
http://localhost:9090
. You can use this URL to connect it to Label Studio: Go to the project Settings > Machine Learning and Add a new ML backend.
This ML backend is an example provided by Label Studio. It actually doesn't do anything. If you want to implement the actual inference logic, go to the next section.
In your model directory, locate the model.py
file (for example, my_ml_backend/model.py
).
The model.py
file contains a class declaration inherited from LabelStudioMLBase
. This class provides wrappers for the API methods that are used by Label Studio to communicate with the ML backend. You can override the methods to implement your own logic:
def predict(self, tasks, context, **kwargs):
"""Make predictions for the tasks."""
return predictions
The predict
method is used to make predictions for the tasks. It uses the following:
tasks
: Label Studio tasks in JSON formatcontext
: Label Studio context in JSON format - for interactive labeling scenariopredictions
: Predictions array in JSON format
Once you implement the predict
method, you can see predictions from the connected ML backend in Label Studio.
You can also implement the fit
method to train your model. The fit
method is typically used to train the model on the labeled data, although it can be used for any arbitrary operations that require data persistence (for example, storing labeled data in database, saving model weights, keeping LLM prompts history, etc).
By default, the fit
method is called at any data action in Label Studio, like creating a new task or updating annotations. You can modify this behavior in Label Studio > Settings > Webhooks.
To implement the fit
method, you need to override the fit
method in your model.py
file:
def fit(self, event, data, **kwargs):
"""Train the model on the labeled data."""
old_model = self.get('old_model')
# write your logic to update the model
self.set('new_model', new_model)
with
event
: event type can be'ANNOTATION_CREATED'
, `'ANNOTATION_UPDATED', etc.data
the payload received from the event (check more on Webhook event reference)
Additionally, there are two helper methods that you can use to store and retrieve data from the ML backend:
self.set(key, value)
- store data in the ML backendself.get(key)
- retrieve data from the ML backend
Both methods can be used elsewhere in the ML backend code, for example, in the predict
method to get the new model weights.
Other methods and parameters are available within the LabelStudioMLBase
class:
self.label_config
- returns the Label Studio labeling config as XML string.self.parsed_label_config
- returns the Label Studio labeling config as JSON.self.model_version
- returns the current model version.
To run without docker (for example, for debugging purposes), you can use the following command:
pip install -r my_ml_backend
label-studio-ml start my_ml_backend
To modify the port, use the -p
parameter:
label-studio-ml start my_ml_backend -p 9091
Before you start:
- Install gcloud
- Init billing for account if it's not activated
- Init gcloud, type the following commands and login in browser:
gcloud auth login
- Activate your Cloud Build API
- Find your GCP project ID
- (Optional) Add GCP_REGION with your default region to your ENV variables
To start deployment:
- Create your own ML backend
- Start deployment to GCP:
label-studio-ml deploy gcp {ml-backend-local-dir} \
--from={model-python-script} \
--gcp-project-id {gcp-project-id} \
--label-studio-host {https://app.heartex.com} \
--label-studio-api-key {YOUR-LABEL-STUDIO-API-KEY}
- After label studio deploys the model - you will get model endpoint in console.