For more information, see Using the default model.Īllowable values:
#Ibm speech to text prototype install
For Speech to Text for IBM Cloud Pak for Data, if you do not install the en-US_BroadbandModel, you must either specify a model with the request or specify a new default model for your installation of the service. The default model is en-US_BroadbandModel. See Using a model for speech recognition. The model to use for all speech recognition requests that are sent over the connection. For more information, see Authenticating to IBM Cloud Pak for Data. Pass an access token as you would with the Authorization header of an HTTP request. For more information, see Authenticating to IBM Cloud.
You pass an IAM access token instead of passing an API key with the call. Pass an Identity and Access Management (IAM) access token to authenticate with the service. After a connection is established, it can remain active even after the token or its credentials are deleted. You do not need to refresh the access token for an active connection that lasts beyond the token's expiration time. You remain authenticated for as long as you keep the connection open. After you establish a connection, you can keep it alive indefinitely. You pass an access token only to establish an authenticated connection. You must establish the connection before the access token expires. It is not available for previous-generation models.Pass a valid access token to establish an authenticated connection with the service. Then experiment with different values as necessary, adjusting the value by small increments.īeta: The parameter is beta functionality. To determine the most effective value for your scenario, start by setting the value of the parameter to a small increment, such as -0.1, -0.05, 0.05, or 0.1, and assess how the value impacts the transcription results. Positive values bias the service to favor hypotheses with longer strings of characters.Īs the value approaches -1.0 or 1.0, the impact of the parameter becomes more pronounced. Negative values bias the service to favor hypotheses with shorter strings of characters. The allowable range of values is -1.0 to 1.0. By default, the service is optimized to produce the best balance of strings of different lengths. Use caution when you set the weight: a higher value can improve the accuracy of phrases from the custom model's domain, but it can negatively affect performance on non-domain phrases.įor next-generation models, an indication of whether the service is biased to recognize shorter or longer strings of characters when developing transcription hypotheses. Assign a higher value if your audio makes frequent use of OOV words from the custom model. The default value yields the best performance in general. Unless a different customization weight was specified for the custom model when the model was trained, the default value is:Ġ.1 for next-generation English and Japanese modelsĪ customization weight that you specify overrides a weight that was specified when the custom model was trained.
You can use the customization weight to tell the service how much weight to give to words from the custom language model compared to those from the base model for the current request. If you specify a customization ID when you open the connection,
Pass a valid access token to establish an authenticated connection with the service.
0 kommentar(er)
