This model can be created in a simpler way and likely in a SageMaker notebook first since that would be easiest. Explore a premade Hugging Face transformer and then explore a custom written transformer. Use transfer learning on the first Hugging Face one and then experiment possibly the idea of running a Hugging Face transformer on the text corpus and then a new one on top of the selection of that.
Research needs to be done into finding commonly used Verbs and Adverbs, first sentence transitions such as the first few words in different paragraphs. This is all about the science of what makes your writing yours, what tendencies are there, and how can we note cloud the mind of a traditional big corp transformer. Hypothetically this shouldn't be very hard. But further learning can be added underneath this issue, with equations in markdown and a roadmap through the SageMaker notebook.
BrevinS
commented
1 year ago
This model can be created in a simpler way and likely in a SageMaker notebook first since that would be easiest. Explore a premade Hugging Face transformer and then explore a custom written transformer. Use transfer learning on the first Hugging Face one and then experiment possibly the idea of running a Hugging Face transformer on the text corpus and then a new one on top of the selection of that.
Research needs to be done into finding commonly used Verbs and Adverbs, first sentence transitions such as the first few words in different paragraphs. This is all about the science of what makes your writing yours, what tendencies are there, and how can we note cloud the mind of a traditional big corp transformer. Hypothetically this shouldn't be very hard. But further learning can be added underneath this issue, with equations in markdown and a roadmap through the SageMaker notebook.