Together AI API
Together AI DocsGet an API key
HomeAPI ReferenceSDKsContext Plugin
R L

Weights sync

POST
/rl/training-sessions/{session_id}/operations/weights-sync

Submits a weights-sync operation that makes the session's current trained parameters available for sampling. Call this after optim-step when you want subsequent samples to use the updated policy.

Authorization

bearerAuth
AuthorizationBearer <token>

In: header

Path Parameters

session_id*string

Training session ID

Header Parameters

Idempotency-Key*string

Required key that makes retries return the original operation; use a new key for changed request bodies.

Request Body

application/json

Request body for publishing updated policy parameters for sampling.

weight_sync_type*RL.WeightSyncType

How updated parameters are made available for sampling. See WeightSyncType for accepted values.

Value in

  • "WEIGHT_SYNC_TYPE_SYNCHRONOUS"
  • "WEIGHT_SYNC_TYPE_BACKGROUND_PUBLISH"
  • "WEIGHT_SYNC_TYPE_PIPELINE"

Response Body

application/json

application/json

curl -X POST "https://example.com/rl/training-sessions/string/operations/weights-sync" \  -H "Idempotency-Key: string" \  -H "Content-Type: application/json" \  -d '{    "weight_sync_type": "WEIGHT_SYNC_TYPE_SYNCHRONOUS"  }'
{  "id": "string",  "status": "TRAINING_OPERATION_STATUS_UNSPECIFIED",  "output": {    "weights_version": "string"  },  "error": {    "code": "TRAINING_OPERATION_ERROR_CODE_UNSPECIFIED",    "message": "string"  }}

Optimizer step POST

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.

Sample POST

Submits a sample operation that will asynchronously generate text completions with logprobs.