Create a fine-tune job

curl --request POST \
  --url https://dev.hub.oxen.ai/api/repos/{namespace}/{repo_name}/fine_tunes \
  --header 'Authorization: Bearer <token>' \
  --header 'Content-Type: application/json' \
  --data '
{
  "base_model": "<string>",
  "resource": "<string>",
  "script_type": "<string>",
  "is_public": false,
  "oxen_model_path": "<string>",
  "training_params": {
    "batch_size": 123,
    "caption_column": "<string>",
    "gradient_accumulation": 123,
    "image_column": "<string>",
    "learning_rate": 123,
    "lora_alpha": 123,
    "lora_rank": 123,
    "sample_every": 123,
    "samples": 123,
    "steps": 123,
    "timestep_type": "<string>",
    "use_lora": true
  }
}
'

{
  "fine_tune": {
    "base_model": "<string>",
    "created_at": "<string>",
    "id": "<string>",
    "name": "<string>",
    "status": "<string>",
    "updated_at": "<string>"
  },
  "status": "<string>",
  "status_message": "<string>"
}

POST

api

repos

{namespace}

{repo_name}

fine_tunes

Create a fine-tune job

curl --request POST \
  --url https://dev.hub.oxen.ai/api/repos/{namespace}/{repo_name}/fine_tunes \
  --header 'Authorization: Bearer <token>' \
  --header 'Content-Type: application/json' \
  --data '
{
  "base_model": "<string>",
  "resource": "<string>",
  "script_type": "<string>",
  "is_public": false,
  "oxen_model_path": "<string>",
  "training_params": {
    "batch_size": 123,
    "caption_column": "<string>",
    "gradient_accumulation": 123,
    "image_column": "<string>",
    "learning_rate": 123,
    "lora_alpha": 123,
    "lora_rank": 123,
    "sample_every": 123,
    "samples": 123,
    "steps": 123,
    "timestep_type": "<string>",
    "use_lora": true
  }
}
'

{
  "fine_tune": {
    "base_model": "<string>",
    "created_at": "<string>",
    "id": "<string>",
    "name": "<string>",
    "status": "<string>",
    "updated_at": "<string>"
  },
  "status": "<string>",
  "status_message": "<string>"
}

Authorizations

Authorization

string

header

required

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

Body

application/json

Create fine-tune request

Request payload to create a fine-tune job for a repository.

base_model

string

required

Canonical name of the base model to fine-tune

resource

string

required

Repository path to the training data file or directory

script_type

string

required

Name of the fine-tune script to run

is_public

boolean

default:false

Whether the resulting fine-tuned model should be public. Defaults to false.

oxen_model_path

string | null

Optional override for where the resulting model weights live in Oxen. Defaults to the fine-tune resource.

training_params

object

Training configuration parameters

Show child attributes

training_params.batch_size

integer

Per-device batch size used during training.

training_params.caption_column

string

Name of the column containing text captions or prompts.

training_params.gradient_accumulation

integer

Number of steps to accumulate gradients before an optimizer step.

training_params.image_column

string

Name of the column containing image file paths.

training_params.learning_rate

number<float>

Base learning rate for the optimizer.

training_params.lora_alpha

integer

LoRA alpha scaling factor for LoRA fine-tuning approaches.

training_params.lora_rank

integer

LoRA rank used for low-rank adaptation layers.

training_params.sample_every

integer

Interval (in steps) at which to generate samples.

training_params.samples

integer

Number of sample generations to produce at each sampling interval.

training_params.steps

integer

Total number of training steps to run.

training_params.timestep_type

string

Strategy for sampling timesteps during training (for example, 'weighted').

training_params.use_lora

boolean

default:true

Whether to use LoRA fine-tuning (true) or full fine-tuning (false). If omitted, defaults to true.

Response

Create fine-tune response

Standard wrapper for fine-tune responses.

fine_tune

FineTune · object

Fine-tune job resource

Show child attributes

fine_tune.base_model

string

Canonical name of the base model

fine_tune.created_at

string<date_time>

Creation time

fine_tune.id

string

Fine-tune ID

fine_tune.name

string

Fine-tune name

fine_tune.status

string

Current status of the fine-tune

fine_tune.updated_at

string<date_time>

Last update time

status

string

High-level status string (for example, 'success').

status_message

string

Human-readable status message (for example, 'resource_found').

List fine-tunes in a repository Get a fine-tune job

⌘I

Fine-Tuning API

fine_tunes

Authorizations

Body

Response