Configuring Models
Multiple models can be stacked together as a list in a single request. Each model has a
modelType and a modelConfig key. Key | Description |
modelType | Used to identify the type/name of model being enabled |
modelConfig | Used to send additional configuration specific to the model |
Some of the models are dependant on the outputs of other models. This dependency can be understood under the
Flow of Data and Stacking Models Section.Enable & Configure Models |
Emotion & Tone Detection |
Model Dependency and flow of data
As shown in the above Flow Diagram
question_response model is dependant on the outputs of emotion_analysis , sentiment_analysis, speech_type_analysis models and these models are dependants on the outputs of speech_to_text with speaker_seperation enabled.Therefore a sample request for
question_response modelwould look like this. (refer to the Example given below){
"txnId": "txn-dummy-1623245791219",
"enableModels":[
{
"modelType":"speech_to_text",
"modelConfig":{
"automatic_punctuation" : true,
"custom_vocabulary":[],
"speaker_seperation":{
"num_speakers":2
},
"enableKeywords":true,
"aggressiveness": 3,
"enableSuggestedIntents":false,
"topics": {
"threshold": 0.5
}
}
},
{
"modelType":"emotion_analysis"
},
{
"modelType":"sentiment_analysis"
},
{
"modelType":"speech_type_analysis"
},
{
"modelType":"action_items",
"modelConfig":{
"priority": 3
}
},
{
"modelType":"question_response",
"modelConfig":{
"quality" : 1
}
},
{
"modelType":"extractive_summary"
},
{
"modelType":"meeting_topics"
},
{
"modelType":"screengrabs",
"modelConfig":{
"ocr":{
"enable":true
}
}
},
{
"modelType":"screen_activity"
},
{
"modelType":"data_insights",
"modelConfig": {
"dead_air": {
"threshold": 3000
}
}
}
]
}
Last modified 1yr ago