Amazon Transcribe: Custom Language Model or General model?

If you are using the Amazon Transcribe service for automated speech recognition (ASR) feature in your project (especially for the English language), you had to decide whether to build a custom language model or a general model provided by AWS transcribe service. It could also be the case that you tried both options in your application. As I had some experience in trying both options in my project, here I am going to share my two cents. You used the general model to transcribe your audio or video files. You noticed that Amazon Transcribe is not able to recognize certain not-so-frequent English words or phrases that have been pronounced by speakers in audio files.