Best fine-tune launch config for re-launch parity

datasets.get_best_launch_config() -> DatasetGetBestLaunchConfigResponse

GET/api/v1/datasets/{dataset_id}/finetune/best-launch-config

Returns the persisted launch snapshot for the winning AutoScientist job (terminal experiment with best_finetune_job_id) or the newest succeeded most recent succeeded standalone job when no experiment exists. Matches the launch payload plus original_model_name and persisted trained_model_name (output label / suffix). Use it to prefill a re-launch instead of reading config from GET /datasets/{dataset_id} or the list.

ParametersExpand Collapse

dataset_id: str

ReturnsExpand Collapse

class DatasetGetBestLaunchConfigResponse: …

best_job_config: Optional[BestJobConfig]

Launch-parity snapshot for the experiment best job (terminal experiment with best_finetune_job_id) or, when no experiment exists, the newest succeeded standalone job (training_experiment_id null). Null while an experiment is non-terminal, when no best job was chosen yet, or when no qualifying job exists.

finetune_job_id: str

Fine-tune job whose config is shown (experiment best or standalone).

formatuuid

training_experiment_id: Optional[object]

Training experiment when this snapshot is the AutoScientist best job; null for a standalone job.

formatuuid

original_model_name: Optional[str]

Base model id the job was launched with.

trained_model_name: Optional[str]

Output label / suffix for the trained model. Taken from the value recorded at launch when present; otherwise derived from the current dataset name, in which case it can differ from the label the job was submitted with if the dataset was renamed. Null only when no label was recorded and the dataset is unavailable.

training_method: Optional[Literal["sft", "dpo"]]

One of the following:

"sft"

"dpo"

training_type: Optional[Literal["lora", "full"]]

One of the following:

"lora"

"full"

data_format: Optional[Literal["chat", "instruction", "preference"]]

One of the following:

"chat"

"instruction"

"preference"

hyperparams: Optional[Dict[str, object]]

Hyperparameters the job was launched with.

Best fine-tune launch config for re-launch parity

import os
from adaption import Adaption

client = Adaption(
    api_key=os.environ.get("ADAPTION_API_KEY"),  # This is the default and can be omitted
)
response = client.datasets.get_best_launch_config(
    "dataset_id",
)
print(response.best_job_config)

{
  "best_job_config": {
    "finetune_job_id": "182bd5e5-6e1a-4fe4-a799-aa6d9a6ab26e",
    "training_experiment_id": {},
    "original_model_name": "original_model_name",
    "trained_model_name": "trained_model_name",
    "training_method": "sft",
    "training_type": "lora",
    "data_format": "chat",
    "hyperparams": {
      "foo": "bar"
    }
  }
}

Returns Examples

{
  "best_job_config": {
    "finetune_job_id": "182bd5e5-6e1a-4fe4-a799-aa6d9a6ab26e",
    "training_experiment_id": {},
    "original_model_name": "original_model_name",
    "trained_model_name": "trained_model_name",
    "training_method": "sft",
    "training_type": "lora",
    "data_format": "chat",
    "hyperparams": {
      "foo": "bar"
    }
  }
}