Get training job

GET

training_jobs

{job_id}

Get training job

curl --request GET \
  --url https://api.veri.studio/v1/training_jobs/{job_id} \
  --header 'Authorization: Bearer <token>'

const options = {method: 'GET', headers: {Authorization: 'Bearer <token>'}};

fetch('https://api.veri.studio/v1/training_jobs/{job_id}', options)
  .then(res => res.json())
  .then(res => console.log(res))
  .catch(err => console.error(err));

{
  "object": "<string>",
  "id": "<string>",
  "status": "queued",
  "method": "<string>",
  "output_name": "<string>",
  "created_at": "2023-11-07T05:31:56Z",
  "updated_at": "2023-11-07T05:31:56Z",
  "base_model": "<string>",
  "dataset_id": "<string>",
  "dataset_snapshot_id": "<string>",
  "reward_count": 123,
  "hyperparameters": {
    "learning_rate": 123,
    "num_epochs": 123,
    "max_steps": 123,
    "rollouts_per_prompt": 123,
    "kl_coef": 123,
    "max_prompt_length": 123,
    "max_response_length": 123,
    "global_batch_size": 123,
    "seed": 123,
    "use_unsloth": true,
    "lora_rank": 123,
    "lora_alpha": 123,
    "load_in_4bit": true,
    "checkpoint_every": 123,
    "checkpoint_keep": 123,
    "harness_protocol": "<string>",
    "max_turns": 123,
    "harness_timeout_s": 123,
    "max_parallel_rollouts": 123,
    "vllm_gpu_memory_utilization": 123,
    "max_model_len": 123,
    "trajectory_retention_days": 123
  },
  "base_image": "<string>",
  "entrypoint": "<string>",
  "gpu": {
    "type": "<string>",
    "count": 123
  },
  "gpu_requested": {
    "type": "<string>",
    "count": 123
  },
  "provider": "<string>",
  "region": "<string>",
  "num_nodes": 123,
  "error": {
    "code": "<string>",
    "message": "<string>"
  },
  "dashboard_url": "<string>",
  "download_url": "<string>",
  "wandb_run_url": "<string>",
  "hf_push": {
    "repo": "<string>",
    "artifact": "<string>",
    "private": true
  },
  "hf_repo_url": "<string>",
  "current_step": 123,
  "total_steps": 123,
  "current_loss": 123,
  "cost_usd": 123,
  "duration_seconds": 123,
  "started_at": "2023-11-07T05:31:56Z",
  "completed_at": "2023-11-07T05:31:56Z"
}

Authorizations

Authorization

string

header

required

API key with the vk_ prefix. Create one from the dashboard.

Path Parameters

job_id

string

required

Training job ID

Response

The training job

object

string

required

string

required

status

enum<string>

required

Available options:

queued,

provisioning,

configuring,

running,

completed,

failed,

cancelled

method

string

required

output_name

string

required

created_at

string<date-time>

required

updated_at

string<date-time>

required

base_model

string | null

dataset_id

string | null

dataset_snapshot_id

string | null

The dataset snapshot this job was pinned to at submit ("<dataset_id>@snap-N"); set when dataset_id was given as …@snap-N or …@latest (or names a stream dataset). Null for legacy submits.

reward_count

integer<int32> | null

How many inline reward sources the job was submitted with (null for reward-free methods). Sources are viewable per job via the dashboard.

hyperparameters

object

Harness-in-the-loop GRPO (VS-372): the user's own agent harness (Claude Agent SDK / LangChain / OpenAI Agents SDK) drives multi-turn rollouts against the in-training policy. Shares GRPO's core knobs; the harness-specific ones tune rollout execution and the colocated vLLM policy server. The harness code itself arrives via the job's script payload (same shape as custom_script).

Show child attributes

base_image

string | null

entrypoint

string | null

gpu

null | object

Pydantic GPUInfo (with type field). Rust uses gpu_type internally but serializes as "type" for JSON parity.

Show child attributes

gpu_requested

null | object

Pydantic GPUInfo (with type field). Rust uses gpu_type internally but serializes as "type" for JSON parity.

Show child attributes

provider

string | null

region

string | null

num_nodes

integer<int32>

error

null | object

Show child attributes

dashboard_url

string | null

download_url

string | null

wandb_run_url

string | null

hf_push

null | object

The hf_push request echoed back (null when the job didn't opt in).

Show child attributes

hf_repo_url

string | null

Set after a successful end-of-training push to Hugging Face; the pushed repo's URL (mirrors wandb_run_url).

current_step

integer<int32> | null

total_steps

integer<int32> | null

current_loss

number<double> | null

cost_usd

number<double> | null

duration_seconds

number<double> | null

started_at

string<date-time> | null

completed_at

string<date-time> | null

Get training job

curl --request GET \
  --url https://api.veri.studio/v1/training_jobs/{job_id} \
  --header 'Authorization: Bearer <token>'

const options = {method: 'GET', headers: {Authorization: 'Bearer <token>'}};

fetch('https://api.veri.studio/v1/training_jobs/{job_id}', options)
  .then(res => res.json())
  .then(res => console.log(res))
  .catch(err => console.error(err));

{
  "object": "<string>",
  "id": "<string>",
  "status": "queued",
  "method": "<string>",
  "output_name": "<string>",
  "created_at": "2023-11-07T05:31:56Z",
  "updated_at": "2023-11-07T05:31:56Z",
  "base_model": "<string>",
  "dataset_id": "<string>",
  "dataset_snapshot_id": "<string>",
  "reward_count": 123,
  "hyperparameters": {
    "learning_rate": 123,
    "num_epochs": 123,
    "max_steps": 123,
    "rollouts_per_prompt": 123,
    "kl_coef": 123,
    "max_prompt_length": 123,
    "max_response_length": 123,
    "global_batch_size": 123,
    "seed": 123,
    "use_unsloth": true,
    "lora_rank": 123,
    "lora_alpha": 123,
    "load_in_4bit": true,
    "checkpoint_every": 123,
    "checkpoint_keep": 123,
    "harness_protocol": "<string>",
    "max_turns": 123,
    "harness_timeout_s": 123,
    "max_parallel_rollouts": 123,
    "vllm_gpu_memory_utilization": 123,
    "max_model_len": 123,
    "trajectory_retention_days": 123
  },
  "base_image": "<string>",
  "entrypoint": "<string>",
  "gpu": {
    "type": "<string>",
    "count": 123
  },
  "gpu_requested": {
    "type": "<string>",
    "count": 123
  },
  "provider": "<string>",
  "region": "<string>",
  "num_nodes": 123,
  "error": {
    "code": "<string>",
    "message": "<string>"
  },
  "dashboard_url": "<string>",
  "download_url": "<string>",
  "wandb_run_url": "<string>",
  "hf_push": {
    "repo": "<string>",
    "artifact": "<string>",
    "private": true
  },
  "hf_repo_url": "<string>",
  "current_step": 123,
  "total_steps": 123,
  "current_loss": 123,
  "cost_usd": 123,
  "duration_seconds": 123,
  "started_at": "2023-11-07T05:31:56Z",
  "completed_at": "2023-11-07T05:31:56Z"
}

Overview

Training jobs

Datasets

Deployments

Volumes

Models

Code artifacts

Regions

Billing

Workspace settings

Integrations

API keys

SSH keys

Webhooks

Authorizations

Path Parameters

Response