Fine-Tune using OpenAI SDK

ScaleGenAI provides OpenAI SDK compatibility, which implies you can fine-tune models using ScaleGenAI in the same way you would using OpenAI GPT models, by simply swapping the base URL.

The OpenAI SDK has the following methods.

Function	Description
create	Create a new fine-tuning task.
get	Get fine-tuning job info.
finish	Finish a running fine-tuning task.
cancel	Cancel a running fine-tuning task.

---

Create Fine-Tuning Job

Endpoint: /openai/v1/fine_tuning/jobs

Description

This method is used to create a new finetuning task.

Request

• Method: POST
• Headers:
  • Content-Type: application/json
• Body:

  {
    "model": "string",
    "training_file": "string",
    "hyperparameters": {
      "batch_size": "auto",
      "learning_rate_multiplier": "auto",
      "n_epochs": "auto"
    },
    "suffix": "",
    "validation_file": "string"
  }

Parameter Description

• model :: string : The name of the model to be used for finetuning.
• training_file :: string : The ID of an uploaded file that contains training data.
• **hyperparameters ** [optional] :: dict : Hyperparameter configuration for the finetuning job.
• validation_file :: string : The ID of an uploaded file that contains validation data.

Response

Code: 200

{
  "id": "string",
  "created_at": 0,
  "error": {
    "code": "string",
    "message": "string",
    "param": "string"
  },
  "fine_tuned_model": "string",
  "finished_at": 0,
  "hyperparameters": {
    "batch_size": "auto",
    "learning_rate_multiplier": "auto",
    "n_epochs": "auto"
  },
  "model": "string",
  "object": "fine_tuning.job",
  "organization_id": "scaletorch",
  "result_files": [],
  "status": "validating_files",
  "trained_tokens": 0,
  "training_file": "string",
  "validation_file": "string"
}

Example

curl -X POST \
  -H "Content-Type: application/json" \
  -d '{
    "model": "meta-llama/Llama-3-70b",
    "training_file": "fileid01",
  }' \
  https://api.example.com/openai/v1/fine_tuning/jobs

---

Get Fine-Tuning Job

Endpoint: /openai/v1/fine_tuning/jobs/{fine_tuning_job_id}

Description

This method is used to get information regarding a fine-tuning task.

Request

• Method: GET
• Headers:
  • Content-Type: application/json

Parameter Description

• fine_tuning_job_id :: string : The job ID associated with the fine-tuning job.

Response

Code: 200

{
  "id": "string",
  "created_at": 0,
  "error": {
    "code": "string",
    "message": "string",
    "param": "string"
  },
  "fine_tuned_model": "string",
  "finished_at": 0,
  "hyperparameters": {
    "batch_size": "auto",
    "learning_rate_multiplier": "auto",
    "n_epochs": "auto"
  },
  "model": "string",
  "object": "fine_tuning.job",
  "organization_id": "scaletorch",
  "result_files": [],
  "status": "validating_files",
  "trained_tokens": 0,
  "training_file": "string",
  "validation_file": "string"
}

Example

curl -X GET \
  -H "Content-Type: application/json" \
  https://api.example.com//openai/v1/fine_tuning/jobs/test_job_id

---

Finish Fine-Tuning Job

Endpoint: /openai/v1/fine_tuning/jobs/{fine_tuning_job_id}

Description

This method is used to finish a running fine-tuning task.

Request

• Method: POST
• Headers:
  • Content-Type: application/json
• Body:

  {
    "result_model": "string",
    "result_files": [],
    "trained_tokens": 0
  }

Parameter Description

• fine_tuning_job_id:: string : The job ID associated with the fine-tuning job.
• result_model [optional] :: string : An openai file with resulting lora weights after finetuning.
• result_files:: list[string] : Other resulting files, like logs.
• fine_tuning_job_id [optional] :: int : Number of total consumed tokens.

Response

{
  "success": true,
  "message": "string"
}

Example

curl -X POST \
  -H "Content-Type: application/json" \
  -d '{
    "result_files": ["fileid01"]
  }' \
  https://api.example.com/openai/v1/fine_tuning/jobs/test_job_id

---

Cancel Fine-Tuning Job

Endpoint: /openai/v1/fine_tuning/jobs/{fine_tuning_job_id}/cancel

Description

This method cancel/stop a running fine-tuning task.

Request

• Method: POST
• Headers:
  • Content-Type: application/json

Parameter Description

• fine_tuning_job_id :: string : The job ID associated with the fine-tuning job.

Response

Code: 200

{
  "id": "string",
  "created_at": 0,
  "error": {
    "code": "string",
    "message": "string",
    "param": "string"
  },
  "fine_tuned_model": "string",
  "finished_at": 0,
  "hyperparameters": {
    "batch_size": "auto",
    "learning_rate_multiplier": "auto",
    "n_epochs": "auto"
  },
  "model": "string",
  "object": "fine_tuning.job",
  "organization_id": "scaletorch",
  "result_files": [],
  "status": "validating_files",
  "trained_tokens": 0,
  "training_file": "string",
  "validation_file": "string"
}

Example

curl -X POST \
  -H "Content-Type: application/json" \
  https://api.example.com//openai/v1/fine_tuning/jobs/test_job_id/cancel

---