train

Endpoint Documentation: /v1/train

Use this API endpoint to train a model. This will train a model with the upload_file_path. The response will include a job id, dataset id, and the status of the job. You can monitor the job at https://app.lamini.ai/train.

Request

• HTTP Method: POST
• URL: https://api.lamini.ai/v1/train

• Headers:

  • Authorization: Bearer $LAMINI_API_KEY
  • Content-Type: application/json

• Example Body (JSON): To train on an already uploaded dataset

{
  "model_name": "meta-llama/Meta-Llama-3-8B-Instruct",
  "upload_file_path": "$YOUR_DATASET_PATH"
}

Parameters:

• model_name (string): The model you'd like to train on.
• upload_file_path (string): The path to the uploaded data.

Response:

The response will include a job id and the status of the job. You can monitor the job at https://app.lamini.ai/train. There are a number of statuses possible, each representing a different stage in the tuning process.

{
    "job_id": "<JOB_ID>",
    "status": "CREATED" | "LOADING DATA" | "TRAINING MODEL" | "EVALUATING MODEL" | "SAVING MODEL" | "COMPLETED" | "PARTIALLY COMPLETED" | "FAILED"
}

Example

Request

curl --location 'https://api.lamini.ai/v1/train' \
  --header 'Authorization: Bearer $LAMINI_API_KEY' \
  --header 'Content-Type: application/json' \
  --data '{
      "model_name": "meta-llama/Meta-Llama-3-8B-Instruct",
      "upload_file_path": "$YOUR_DATASET_PATH"
  }'

Response

{
  "job_id": "55555",
  "status": "CREATED"
}