Create Job

Create job

curl --request POST \
  --url https://api.together.xyz/v1/fine-tunes \
  --header 'Authorization: Bearer <token>' \
  --header 'Content-Type: application/json' \
  --data '{
  "training_file": "<string>",
  "validation_file": "<string>",
  "model": "<string>",
  "n_epochs": 1,
  "n_checkpoints": 1,
  "n_evals": 0,
  "batch_size": 123,
  "learning_rate": 0.00001,
  "lr_scheduler": "none",
  "warmup_ratio": 0,
  "max_grad_norm": 1,
  "weight_decay": 0,
  "suffix": "<string>",
  "wandb_api_key": "<string>",
  "wandb_base_url": "<string>",
  "wandb_project_name": "<string>",
  "wandb_name": "<string>",
  "train_on_inputs": true,
  "training_method": {
    "method": "sft"
  },
  "training_type": {
    "type": "Full"
  },
  "from_checkpoint": "<string>"
}'

{
  "id": "3c90c3cc-0d44-4b50-8888-8dd25736052a",
  "training_file": "<string>",
  "validation_file": "<string>",
  "model": "<string>",
  "model_output_name": "<string>",
  "model_output_path": "<string>",
  "trainingfile_numlines": 123,
  "trainingfile_size": 123,
  "created_at": "<string>",
  "updated_at": "<string>",
  "n_epochs": 123,
  "n_checkpoints": 123,
  "n_evals": 123,
  "batch_size": 123,
  "learning_rate": 123,
  "lr_scheduler": {
    "lr_scheduler_type": "linear",
    "lr_scheduler_args": {
      "min_lr_ratio": 0
    }
  },
  "warmup_ratio": 123,
  "max_grad_norm": 123,
  "weight_decay": 123,
  "eval_steps": 123,
  "train_on_inputs": true,
  "training_method": {
    "method": "sft"
  },
  "training_type": {
    "type": "Full"
  },
  "status": "pending",
  "job_id": "<string>",
  "events": [
    {
      "object": "fine-tune-event",
      "created_at": "<string>",
      "level": "info",
      "message": "<string>",
      "type": "job_pending",
      "param_count": 123,
      "token_count": 123,
      "total_steps": 123,
      "wandb_url": "<string>",
      "step": 123,
      "checkpoint_path": "<string>",
      "model_path": "<string>",
      "training_offset": 123,
      "hash": "<string>"
    }
  ],
  "token_count": 123,
  "param_count": 123,
  "total_price": 123,
  "epochs_completed": 123,
  "queue_depth": 123,
  "wandb_project_name": "<string>",
  "wandb_url": "<string>",
  "from_checkpoint": "<string>"
}

POST

fine-tunes

Create job

curl --request POST \
  --url https://api.together.xyz/v1/fine-tunes \
  --header 'Authorization: Bearer <token>' \
  --header 'Content-Type: application/json' \
  --data '{
  "training_file": "<string>",
  "validation_file": "<string>",
  "model": "<string>",
  "n_epochs": 1,
  "n_checkpoints": 1,
  "n_evals": 0,
  "batch_size": 123,
  "learning_rate": 0.00001,
  "lr_scheduler": "none",
  "warmup_ratio": 0,
  "max_grad_norm": 1,
  "weight_decay": 0,
  "suffix": "<string>",
  "wandb_api_key": "<string>",
  "wandb_base_url": "<string>",
  "wandb_project_name": "<string>",
  "wandb_name": "<string>",
  "train_on_inputs": true,
  "training_method": {
    "method": "sft"
  },
  "training_type": {
    "type": "Full"
  },
  "from_checkpoint": "<string>"
}'

{
  "id": "3c90c3cc-0d44-4b50-8888-8dd25736052a",
  "training_file": "<string>",
  "validation_file": "<string>",
  "model": "<string>",
  "model_output_name": "<string>",
  "model_output_path": "<string>",
  "trainingfile_numlines": 123,
  "trainingfile_size": 123,
  "created_at": "<string>",
  "updated_at": "<string>",
  "n_epochs": 123,
  "n_checkpoints": 123,
  "n_evals": 123,
  "batch_size": 123,
  "learning_rate": 123,
  "lr_scheduler": {
    "lr_scheduler_type": "linear",
    "lr_scheduler_args": {
      "min_lr_ratio": 0
    }
  },
  "warmup_ratio": 123,
  "max_grad_norm": 123,
  "weight_decay": 123,
  "eval_steps": 123,
  "train_on_inputs": true,
  "training_method": {
    "method": "sft"
  },
  "training_type": {
    "type": "Full"
  },
  "status": "pending",
  "job_id": "<string>",
  "events": [
    {
      "object": "fine-tune-event",
      "created_at": "<string>",
      "level": "info",
      "message": "<string>",
      "type": "job_pending",
      "param_count": 123,
      "token_count": 123,
      "total_steps": 123,
      "wandb_url": "<string>",
      "step": 123,
      "checkpoint_path": "<string>",
      "model_path": "<string>",
      "training_offset": 123,
      "hash": "<string>"
    }
  ],
  "token_count": 123,
  "param_count": 123,
  "total_price": 123,
  "epochs_completed": 123,
  "queue_depth": 123,
  "wandb_project_name": "<string>",
  "wandb_url": "<string>",
  "from_checkpoint": "<string>"
}

Authorizations

Authorization

string

header

default:default

required

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

Body

application/json

training_file

string

required

File-ID of a training file uploaded to the Together API

model

string

required

Name of the base model to run fine-tune job on

validation_file

string

File-ID of a validation file uploaded to the Together API

n_epochs

integer

default:1

Number of complete passes through the training dataset (higher values may improve results but increase cost and risk of overfitting)

n_checkpoints

integer

default:1

Number of intermediate model versions saved during training for evaluation

n_evals

integer

default:0

Number of evaluations to be run on a given validation set during training

batch_size

default:max

Number of training examples processed together (larger batches use more memory but may train faster). Defaults to "max". We use training optimizations like packing, so the effective batch size may be different than the value you set.

learning_rate

number

default:0.00001

Controls how quickly the model adapts to new information (too high may cause instability, too low may slow convergence)

lr_scheduler

object

The learning rate scheduler to use. It specifies how the learning rate is adjusted during training.

Show child attributes

warmup_ratio

number

default:0

The percent of steps at the start of training to linearly increase the learning rate.

max_grad_norm

number

default:1

Max gradient norm to be used for gradient clipping. Set to 0 to disable.

weight_decay

number

default:0

Weight decay. Regularization parameter for the optimizer.

suffix

string

Suffix that will be added to your fine-tuned model name

wandb_api_key

string

Integration key for tracking experiments and model metrics on W&B platform

wandb_base_url

string

The base URL of a dedicated Weights & Biases instance.

wandb_project_name

string

The Weights & Biases project for your run. If not specified, will use together as the project name.

wandb_name

string

The Weights & Biases name for your run.

train_on_inputs

boolean

default:auto

Whether to mask the user messages in conversational data or prompts in instruction data.

training_method

object

The training method to use. 'sft' for Supervised Fine-Tuning or 'dpo' for Direct Preference Optimization.

Show child attributes

training_type

object

Show child attributes

from_checkpoint

string

The checkpoint identifier to continue training from a previous fine-tuning job. Format is {$JOB_ID} or {$OUTPUT_MODEL_NAME} or {$JOB_ID}:{$STEP} or {$OUTPUT_MODEL_NAME}:{$STEP}. The step value is optional; without it, the final checkpoint will be used.

Response

200 - application/json

Fine-tuning job initiated successfully

string<uuid>

required

status

enum<string>

required

Available options:

pending,

queued,

running,

compressing,

uploading,

cancel_requested,

cancelled,

error,

completed

training_file

string

validation_file

string

model

string

model_output_name

string

model_output_path

string

trainingfile_numlines

integer

trainingfile_size

integer

created_at

string

updated_at

string

n_epochs

integer

n_checkpoints

integer

n_evals

integer

batch_size

default:max

learning_rate

number

lr_scheduler

object

Show child attributes

warmup_ratio

number

max_grad_norm

number

weight_decay

number

eval_steps

integer

train_on_inputs

default:auto

training_method

object

Show child attributes

training_type

object

Show child attributes

job_id

string

events

object[]

Show child attributes

token_count

integer

param_count

integer

total_price

integer

epochs_completed

integer

queue_depth

integer

wandb_project_name

string

wandb_url

string

from_checkpoint

string

Get File Contents List All Jobs

Together APIs

Deprecated API

Command Line Interface

General

Authorizations

Body

Response