建立自訂語言模型

此範例使用create-language-model指令。如需詳細資訊，請參閱 CreateLanguageModel 及 LanguageModel。

aws transcribe create-language-model \ 
--base-model-name NarrowBand \ 
--model-name my-first-language-model \ 
--input-data-config S3Uri=s3://DOC-EXAMPLE-BUCKET/my-clm-training-data/,TuningDataS3Uri=s3://DOC-EXAMPLE-BUCKET/my-clm-tuning-data/,DataAccessRoleArn=arn:aws:iam::111122223333:role/ExampleRole \
--language-code en-US

下面是另一個使用create-language-model命令的示例，以及創建自定義語言模型的請求主體。

aws transcribe create-language-model \
--cli-input-json file://filepath/my-first-language-model.json

該文件 my-first-language-model.json 包含以下請求主體。

{
  "BaseModelName": "NarrowBand",
  "ModelName": "my-first-language-model",
  "InputDataConfig": {
         "S3Uri": "s3://DOC-EXAMPLE-BUCKET/my-clm-training-data/",
         "TuningDataS3Uri"="s3://DOC-EXAMPLE-BUCKET/my-clm-tuning-data/",
         "DataAccessRoleArn": "arn:aws:iam::111122223333:role/ExampleRole"
    },
  "LanguageCode": "en-US"  
}

這個範例會使用建AWS SDK for Python (Boto3)立 CLM，使用建立語言模型方法。如需詳細資訊，請參閱 CreateLanguageModel 及 LanguageModel。

如需使用 AWS SDK 的其他範例，包括特定功能、案例和跨服務範例，請參閱本章。使用 Amazon Transcribe 的程式碼範例 AWS SDKs

from __future__ import print_function
import time
import boto3
transcribe = boto3.client('transcribe', 'us-west-2')
model_name = 'my-first-language-model',
transcribe.create_language_model(
    LanguageCode = 'en-US', 
    BaseModelName = 'NarrowBand',
    ModelName = model_name,
    InputDataConfig = {
        'S3Uri':'s3://DOC-EXAMPLE-BUCKET/my-clm-training-data/',
        'TuningDataS3Uri':'s3://DOC-EXAMPLE-BUCKET/my-clm-tuning-data/',
        'DataAccessRoleArn':'arn:aws:iam::111122223333:role/ExampleRole'
    }
)

while True:
    status = transcribe.get_language_model(ModelName = model_name)
    if status['LanguageModel']['ModelStatus'] in ['COMPLETED', 'FAILED']:
        break
    print("Not ready yet...")
    time.sleep(5)
print(status)