FINETUNE ('DESCRIBE') (SNOWFLAKE.CORTEX)¶

Describes the properties of a fine-tuning job. If the job completes successfully, additional details about the job are returned, including the final model name. Use this name when using the COMPLETE (SNOWFLAKE.CORTEX) function to make an inference on your fine-tuned model.

Syntax¶

SNOWFLAKE.CORTEX.FINETUNE(
  'DESCRIBE',
  '<finetune_job_id>'
)
Copy

Parameters¶

'DESCRIBE'

Specifies that you want to get the properties of the provided fine-tuning job.

finetune_job_id

The ID of the fine-tuning job that was generated when you created the job.

Output¶

Column

Type

Description

SNOWFLAKE.CORTEX.FINETUNE

OBJECT

An object containing the job status, progress and the tuning job ID. If the job status is SUCCESS, additional job information is returned.

id

Unique ID for the tuning job.

status

The status is one of the following:

  • PENDING

  • IN_PROGRESS

  • SUCCESS

  • ERROR

  • CANCELLED

progress

A number between zero and one that indicates the percentage of the job completed with 1.0 being 100%.

error

If the job has a status of ERROR, an object that contains the error message.

base_model

If the job has a status of SUCCESS, the name of the base model used for the fine-tuning job.

created_on

If the job has a status of SUCCESS, the timestamp of when the job was created.

finished_on

If the job has a status of SUCCESS, the timestamp of when the job finished.

model

If the job has a status of SUCCESS, the fine-tuned model name. Use this name when calling the COMPLETE function for inference.

training_data

If the job has a status of SUCCESS, the query used to retrieve the training data.

trained_tokens

If the job has a status of SUCCESS, the number of tokens used for training. This is calculated by the following formula:

trained tokens = number of input tokens  * number of epochs trained
Copy
training_result

If the job has a status of SUCCESS, the training result of the fine-tuning job.

validation_data

If the job has a status of SUCCESS, the query used to retrieve the validation data.

Access control requirements¶

For access requirements, see Access control requirements.

Examples¶

SELECT SNOWFLAKE.CORTEX.FINETUNE(
  'DESCRIBE',
  'CortexFineTuningWorkflow_f45b452d-1672-44fa-8516-d4cf772b16db'
);
Copy

Successful job output at various stages:

{"id":"CortexFineTuningWorkflow_6556e15c-8f12-4d94-8cb0-87e6f2fd2299","progress":0.0,"status":"IN_PROGRESS"}

{"id":"CortexFineTuningWorkflow_6556e15c-8f12-4d94-8cb0-87e6f2fd2299","progress":0.333,"status":"IN_PROGRESS"}

{
  "base_model":"mistral-7b",
  "created_on":1717004388348,
  "finished_on":1717004691577,
  "id":"CortexFineTuningWorkflow_6556e15c-8f12-4d94-8cb0-87e6f2fd2299",
  "model":"my_tuned_model",
  "progress":1.0,
  "status":"SUCCESS",
  "training_data":"SELECT prompt, completion FROM train",
  "trained_tokens":2670734,
  "training_result":{"validation_loss":1.0138969421386719,"training_loss":0.6477728401547047}
  "validation_data":""
  }
Language: English