Sign in
RL

Optimizer step

Submits an optimizer step operation that will asynchronously apply accumulated gradients to update model parameters. Does not make the updated parameters available for sampling; call weights-sync afterwards when you want subsequent samples to use the updated policy.

post/rl/training-sessions/{session_id}/operations/optim-step

Path parameters

session_idstring required

Training session ID

Request body

Example request

{
  "adam_params": {
    "learning_rate": 0.0001,
    "beta1": 0.9,
    "beta2": 0.95,
    "eps": 1e-8,
    "weight_decay": 0.1,
    "grad_clip_norm": 10
  },
  "muon_params": {
    "learning_rate": 0.02,
    "momentum": 0.95,
    "newton_schulz_steps": 5,
    "grad_clip_norm": 10,
    "adam": {
      "learning_rate": 0.0001,
      "beta1": 0.9,
      "beta2": 0.95,
      "eps": 1e-8,
      "weight_decay": 0.1,
      "grad_clip_norm": 10
    }
  }
}

Response

Optimizer step operation details

idstring required

Operation ID

status'TRAINING_OPERATION_STATUS_UNSPECIFIED' | 'TRAINING_OPERATION_STATUS_PENDING' | 'TRAINING_OPERATION_STATUS_RUNNING' | 'TRAINING_OPERATION_STATUS_COMPLETED' | 'TRAINING_OPERATION_STATUS_FAILED' required

Example response

{
  "id": "550e8400-e29b-41d4-a716-446655440000",
  "error": {
    "message": "Operation timed out"
  }
}

Changes

Changed in 4 of the 15 revisions of this API.814

  • f26a23fab8e01See the full diff
    • ●

      removed the request property

      request-property-removed

  • ebd4756a41e41See the full diff
    • ●

      added the new TRAINING_OPERATION_ERROR_CODE_NON_FINITE_LOSS enum value to the / response property for the response status

      response-property-enum-value-added

  • 9ccc57e5e4b933See the full diff
    • ●

      removed the request property

      request-property-removed

    • ●

      removed the request property

      request-property-removed

    • ●

      removed the request property /

      request-property-removed

    • ○

      added the new optional request property

      new-optional-request-property

    • ○

      added the new optional request property /

      new-optional-request-property

    • ○

      added the new optional request property /

      new-optional-request-property

    This revision also has 11 changes that name no endpoint, such as unreferenced schemas being removed. See the revision's changelog

  • 0c4fdad77435311See the full diff
    • ●

      removed the request property /

      request-property-removed

    • ●

      removed the request property

      request-property-removed

    • ●

      removed the request property /

      request-property-removed

    • ○

      added the new optional request property /

      new-optional-request-property

    • ○

      added the new optional request property /

      new-optional-request-property

    • ○

      added the new optional request property /

      new-optional-request-property

    • ○

      the request property default value 1.0 was added

      request-property-default-value-added

    • ○

      the / request property format was generalized from float to no format

      request-property-type-generalized

    • ○

      the / request property format was generalized from float to no format

      request-property-type-generalized

    • ○

      the / request property format was generalized from float to no format

      request-property-type-generalized

    • ○

      the / request property format was generalized from float to no format

      request-property-type-generalized

    • ○

      the / request property format was generalized from float to no format

      request-property-type-generalized

    • ○

      the / request property format was generalized from int64 to no format

      request-property-type-generalized

    • ○

      the / request property format was generalized from float to no format

      request-property-type-generalized

    This revision also has 2 changes that name no endpoint, such as unreferenced schemas being removed. See the revision's changelog