Skip to content

api_documentation.md

Latest commit

 

History

History
140 lines (112 loc) · 3.28 KB

api_documentation.md

File metadata and controls

140 lines (112 loc) · 3.28 KB

API Documentation

Base URL

The base URL for all endpoints is: http://127.0.0.1:5000

Endpoints

1. Add Layer

Endpoint: /api/add_layer

Method: POST

Description: Adds a new layer to the Perceptron instance with the specified number of perceptrons and activation function. The weights are initialized randomly and can be trained using the train_perceptron endpoint. The Perceptron can be trained using the logistic_regression endpoint to perform logistic regression on the training data.

Request Payload:

{
    "size": 2,         // Number of perceptrons in the layer
    "function": "sigmoid",   // Activation function (sigmoid, relu, tanh)
}

Response:

{
  "message": "Layer added successfully",
  "weights": Object // json object representing the layers weights
}

Example Request:

curl -X POST http://127.0.0.1:5000/api/add_layer -H "Content-Type: application/json" -d '{"size": 2, "function": "sigmoid"}'

2. Set Input Size

Endpoint: /api/set_input_size

Method: POST

Description: Sets the input size for the Neural Network.

Request Payload:

{
    "size": 2  // input size (no. of columns/features)
}

Response:

{
  "message": "Input size set successfully"  
}

Example Request:

curl -X POST http://127.0.0.1:5000/api/set_input_size -H "Content-Type: application/json" -d '{"size": 2}'

3. Remove Layer

Endpoint: /api/remove_layer

Method: DELETE

Description: Removes the last layer from the Neural Network.

Response:

{
  "message": "Layer removed successfully",
  "weights": Object // json object representing the layers weights
}

Example Request:

curl -X DELETE http://127.0.0.1:5000/api/remove_layer

4. Set Train Data

Endpoint: /api/set_train_data

Method: POST

Description: Sets the training data and labels for the Neural Network.

Request Payload:

{
    X_train: [[0, 0], [0, 1], [1, 0], [1, 1]],  // Training data
    y_train: [[0], [1], [1], [0]]  // Labels
}

Response:

{
  "message": "{X_train}, {y_train}"
}

Example Request:

curl -X POST http://localhost:5000/api/set_train_data -H "Content-Type: application/json" -d '{"X_train": [[0, 0], [0, 1], [1, 0], [1, 1]], "y_train": [[0], [1], [1], [0]]}'

4. Train

Endpoint: train

Method: SOCKET

Description:

  • Trains the Neural Network using the specified training data and labels.
  • The training data and labels are sent to the server using the 'train' event.
  • The server will send real-time updates to the client using the 'weight_update' event.
  • The server will send 'loss_update' event to update the loss value.
  • Once the training is complete, the server will send the 'training_complete' event.

Request Payload:

{
    learning_rate: 0.01,  // Learning rate for the training (optional, default: 0.01)
    epochs: 1000          // Number of epochs to train the Neural Network (optional, default: 1000)
}

Real-Time Updates:

  1. Event: weight_update
{
  "data": Object  // Current weights of the Neural Network
}
  1. Event: loss_update
{
  "data": Object  // Current loss value of the Neural Network
}
  1. Event: training_complete