Values_1.ModelSpeculativeDecodingConfigSourceSettings for the model speculative decoding technique that's applied by a model optimization job.
type nonrec t = {technique : ModelSpeculativeDecodingTechnique.t;The speculative decoding technique to apply during model optimization.
*)trainingDataSource : ModelSpeculativeDecodingTrainingDataSource.t option;The location of the training data to use for speculative decoding. The data must be formatted as ShareGPT, OpenAI Completions or OpenAI Chat Completions. The input can also be unencrypted captured data from a SageMaker endpoint as long as the endpoint uses one of the above formats.
*)}val make :
?trainingDataSource:??? ->
technique:ModelSpeculativeDecodingTechnique.t ->
unit ->
tval to_value :
t ->
[> `Structure of
(string
* [> `Enum of string
| `Structure of
(string * [> `Enum of string | `String of Values_0.S3Uri.t ]) list ])
list ]