Families

Continue a family

Train a family’s next model from its latest, on the rows its dataset gained since the family last trained, up to dataset_version, its newest version unless you name an earlier one: add rows to the dataset first (POST /v1/datasets/{dataset_id}/versions). To start a separate line, create a family (POST /v1/families). One fine-tune of a family at a time. A deleted family gets 410 family_deleted. A family whose dataset was deleted gets 410 dataset_deleted: start a new family from its latest model. While a large dataset version’s training tokens are first counted, it answers 503 tokens_counting with Retry-After and starts nothing.

POST

families

{family}

continue

Continue a family

curl --request POST \
  --url https://api.sqwish.ai/v1/families/{family}/continue \
  --header 'Authorization: Bearer <token>' \
  --header 'Content-Type: application/json' \
  --data '
{
  "hyperparameters": {
    "steps": 2,
    "learning_rate": 0.00005,
    "lora_rank": 8,
    "seed": 42
  },
  "latest": "<string>",
  "dataset_version": 2,
  "max_price_usd": "<string>"
}
'

import requests

url = "https://api.sqwish.ai/v1/families/{family}/continue"

payload = {
    "hyperparameters": {
        "steps": 2,
        "learning_rate": 0.00005,
        "lora_rank": 8,
        "seed": 42
    },
    "latest": "<string>",
    "dataset_version": 2,
    "max_price_usd": "<string>"
}
headers = {
    "Authorization": "Bearer <token>",
    "Content-Type": "application/json"
}

response = requests.post(url, json=payload, headers=headers)

print(response.text)

const options = {
  method: 'POST',
  headers: {Authorization: 'Bearer <token>', 'Content-Type': 'application/json'},
  body: JSON.stringify({
    hyperparameters: {steps: 2, learning_rate: 0.00005, lora_rank: 8, seed: 42},
    latest: '<string>',
    dataset_version: 2,
    max_price_usd: '<string>'
  })
};

fetch('https://api.sqwish.ai/v1/families/{family}/continue', options)
  .then(res => res.json())
  .then(res => console.log(res))
  .catch(err => console.error(err));

{
  "id": "<string>",
  "dataset_id": "<string>",
  "model": "<string>",
  "method": "sft",
  "hyperparameters": {
    "steps": 123,
    "learning_rate": 123,
    "lora_rank": 123,
    "seed": 123
  },
  "family": "<string>",
  "family_deleted": true,
  "parent": "<string>",
  "dataset_version": 123,
  "status": "queued",
  "outcome": "joined",
  "created_at": 123,
  "updated_at": 123,
  "output_model": "<string>",
  "metrics": {},
  "error": "<string>",
  "progress": {},
  "limits": {},
  "split": {},
  "lineage": {},
  "price_usd": "<string>",
  "charge": "free",
  "dataset_evaluation": {
    "id": "<string>",
    "dataset_id": "<string>",
    "dataset_version": 123,
    "model_id": "<string>",
    "model_sha256": "<string>",
    "model_label": "<string>",
    "model_name": "<string>",
    "version": 123,
    "status": "queued",
    "created_at": 123,
    "updated_at": 123,
    "generation": 123,
    "training_job_id": "<string>",
    "dataset_sha256": "<string>",
    "evaluation_sha256": "<string>",
    "method": "sft",
    "evaluation_rows": 123,
    "decisions": 123,
    "completed_rows": 123,
    "metrics": {},
    "error": {},
    "named": "<string>"
  },
  "curve": [
    {
      "step": 123,
      "loss": 123,
      "learning_rate": 123
    }
  ],
  "rows_deleted": true
}

{
  "error": {
    "code": "<string>",
    "message": "<string>",
    "retryable": true,
    "request_id": "<string>",
    "details": [
      {
        "path": [
          "<string>"
        ],
        "message": "<string>",
        "type": "<string>"
      }
    ]
  }
}

{
  "error": {
    "code": "<string>",
    "message": "<string>",
    "retryable": true,
    "request_id": "<string>",
    "details": [
      {
        "path": [
          "<string>"
        ],
        "message": "<string>",
        "type": "<string>"
      }
    ]
  }
}

Authorizations

Authorization

string

header

required

Create an API key in the console.

Headers

idempotency-key

string | null

Reuse the same key and body after a lost response. A replay returns the job that key queued, with Idempotency-Replayed: true. A changed body conflicts.

Path Parameters

family

string

required

Body

application/json

A family's next fine-tune, from its latest model on its dataset's newer rows.

hyperparameters

Hyperparameters · object

Show child attributes

latest

string | null

The id of the family's latest model you saw, or null for none. If the family has moved on since, nothing is queued and the request is refused with family_changed.

Maximum string length: 140

dataset_version

integer | null

The version of the family's dataset to train on, its newest when left out. It trains on the rows added up to that version, which must be newer than the one the family's last finished job trained on, or the request is refused with no_new_data.

Required range: x >= 1

max_price_usd

string | null

The most you agree to pay for this job, in USD, such as the price you were shown. It is a decimal string, such as "2.500000000", never a JSON number, so no precision is lost: a number is refused with invalid_request, naming this field. If it costs more when it is admitted, it is refused with price_changed and nothing is held.

Pattern: ^\d{1,9}(\.\d{1,9})?$

Response

Successful Response

string

required

dataset_id

string

required

model

string

required

method

enum<string>

required

Available options:

sft,

reward

hyperparameters

ResolvedHyperparameters · object

required

What a job trains with, as its owner chose or the defaults gave it. The trainer's own settings stay on the server.

Show child attributes

family

string | null

required

The family the job adds a model to.

family_deleted

boolean

required

True once you deleted that family. The job and its charge stay.

parent

string | null

required

The id of the model it trains from, or null from a base.

dataset_version

integer

required

The version of its dataset it trains on.

status

enum<string>

required

Available options:

queued,

running,

cancelling,

succeeded,

failed,

cancelled

outcome

enum<string> | null

required

How it ended for its family: its model joined the family with the next number, or failed its gate and is not kept for you, or the run failed or was cancelled, or a rollback discarded the model that joined. Only a failed run or a cancellation is free and leaves the data for another try. Null while it runs.

Available options:

joined,

gate_failed,

failed,

cancelled,

rolled_back

created_at

number

required

updated_at

number

required

output_model

string | null

required

metrics

Metrics · object | null

required

Evaluation groups and the gate, when complete.

error

required

progress

Progress · object | null

required

limits

Limits · object

required

max_questions: the most decisions a row of it may ask.

Show child attributes

split

Split · object

required

train_rows and validation_rows, and trained_tokens: the training tokens the job trains, each decision of a row with the row's whole text, tokenized as input tokens are (an outcome at least 8, a row of n decisions at least 24 × n for each), each time a step takes the row. A priced fine-tune's fee is fine_tune_million_tokens_usd for them, at least fine_tune_minimum_usd.

Show child attributes

lineage

Lineage · object | null

required

price_usd

string

required

The fee held for this job, as exact USD.

charge

enum<string>

required

What the fee is doing: held while the job runs, charged if it succeeds, released if it fails or is cancelled.

Available options:

free,

reserved,

charged,

released

dataset_evaluation

Evaluation · object | null

Show child attributes

curve

CurvePoint · object[] | null

The training loss over steps, at most 300 points, each the means of a run of consecutive steps. Null until the job succeeds, or when its trainer kept none. Lists of jobs leave it out.

Show child attributes

rows_deleted

boolean

True once a dataset whose rows its specification held was deleted: the job, its model and its metrics stay, but its held-out records are gone.

Continue a family

curl --request POST \
  --url https://api.sqwish.ai/v1/families/{family}/continue \
  --header 'Authorization: Bearer <token>' \
  --header 'Content-Type: application/json' \
  --data '
{
  "hyperparameters": {
    "steps": 2,
    "learning_rate": 0.00005,
    "lora_rank": 8,
    "seed": 42
  },
  "latest": "<string>",
  "dataset_version": 2,
  "max_price_usd": "<string>"
}
'

import requests

url = "https://api.sqwish.ai/v1/families/{family}/continue"

payload = {
    "hyperparameters": {
        "steps": 2,
        "learning_rate": 0.00005,
        "lora_rank": 8,
        "seed": 42
    },
    "latest": "<string>",
    "dataset_version": 2,
    "max_price_usd": "<string>"
}
headers = {
    "Authorization": "Bearer <token>",
    "Content-Type": "application/json"
}

response = requests.post(url, json=payload, headers=headers)

print(response.text)

const options = {
  method: 'POST',
  headers: {Authorization: 'Bearer <token>', 'Content-Type': 'application/json'},
  body: JSON.stringify({
    hyperparameters: {steps: 2, learning_rate: 0.00005, lora_rank: 8, seed: 42},
    latest: '<string>',
    dataset_version: 2,
    max_price_usd: '<string>'
  })
};

fetch('https://api.sqwish.ai/v1/families/{family}/continue', options)
  .then(res => res.json())
  .then(res => console.log(res))
  .catch(err => console.error(err));

{
  "id": "<string>",
  "dataset_id": "<string>",
  "model": "<string>",
  "method": "sft",
  "hyperparameters": {
    "steps": 123,
    "learning_rate": 123,
    "lora_rank": 123,
    "seed": 123
  },
  "family": "<string>",
  "family_deleted": true,
  "parent": "<string>",
  "dataset_version": 123,
  "status": "queued",
  "outcome": "joined",
  "created_at": 123,
  "updated_at": 123,
  "output_model": "<string>",
  "metrics": {},
  "error": "<string>",
  "progress": {},
  "limits": {},
  "split": {},
  "lineage": {},
  "price_usd": "<string>",
  "charge": "free",
  "dataset_evaluation": {
    "id": "<string>",
    "dataset_id": "<string>",
    "dataset_version": 123,
    "model_id": "<string>",
    "model_sha256": "<string>",
    "model_label": "<string>",
    "model_name": "<string>",
    "version": 123,
    "status": "queued",
    "created_at": 123,
    "updated_at": 123,
    "generation": 123,
    "training_job_id": "<string>",
    "dataset_sha256": "<string>",
    "evaluation_sha256": "<string>",
    "method": "sft",
    "evaluation_rows": 123,
    "decisions": 123,
    "completed_rows": 123,
    "metrics": {},
    "error": {},
    "named": "<string>"
  },
  "curve": [
    {
      "step": 123,
      "loss": 123,
      "learning_rate": 123
    }
  ],
  "rows_deleted": true
}

{
  "error": {
    "code": "<string>",
    "message": "<string>",
    "retryable": true,
    "request_id": "<string>",
    "details": [
      {
        "path": [
          "<string>"
        ],
        "message": "<string>",
        "type": "<string>"
      }
    ]
  }
}

{
  "error": {
    "code": "<string>",
    "message": "<string>",
    "retryable": true,
    "request_id": "<string>",
    "details": [
      {
        "path": [
          "<string>"
        ],
        "message": "<string>",
        "type": "<string>"
      }
    ]
  }
}

Overview

Decisions

Models

Families

Datasets

Evaluations

Fine-tuning

Prompt tuning

Labeling

Accounts

Billing

Integrations

System

Continue a family

Authorizations

Headers

Path Parameters

Body

Response