POST
/
evaluation
from together import Together
import os

client = Together(
    api_key=os.environ.get("TOGETHER_API_KEY"),
)

client.evaluation.create(
    type="classify",
    judge_model_name="meta-llama/Llama-3.2-3B-Instruct-Turbo",
    judge_system_template="You are a helpful assistant which can classify",
    input_data_file_path="file-1234-5678-abcd",
    labels=["Toxic", "Non-Toxic"],
    pass_labels=["Toxic"],
    model_to_evaluate={
        "name": "meta-llama/Meta-Llama-3.1-70B-Instruct-Turbo",
        "system_template": "You are a helpful assistant which can classify",
        "input_template": "please classify '{{prompt}}' into one of these two categories",
        "max_tokens": 512,
        "temperature": 0.7
    }
)
{
  "workflow_id": "eval-1234-1244513",
  "status": "pending"
}

Authorizations

​
Authorization
string
header
default:default
required

Bearer authentication header of the form Bearer <token>, where <token> is your auth token.

Body

application/json
​
type
enum<string>
required

The type of evaluation to perform

Available options:
classify,
score,
compare
Example:

"classify"

​
parameters
object
required

Type-specific parameters for the evaluation

Response

Successful response

​
workflow_id
string

The ID of the created evaluation job

Example:

"eval-1234-1244513"

​
status
enum<string>

Initial status of the job

Available options:
pending