Continue your Generative AI education
1. Continue your Generative AI education
Welcome back! Let's talk about what's next in your journey. Gen AI is a growing field that is changing how we do everything, and you have started on your learning path. Whether for professional, personal growth, or general interests, you have put in the time to complete the first step in the process of developing a skill that you can put to use in your own life or career. Let's remember what we covered in this course. We looked at what a LLM function is, and we followed that up with how to use the complete function, and after that, we learned how to use the task-specific functions, translate, sentiment, summarize, and classify text. We then looked at the helper functions, count tokens, and tryComplete. From there, we brought it all together and created a Streamlit app using LLM functions as building blocks for the app. We then learned about fine-tune, and how we can use this to specialize our LLMs to perform new tasks that out-of-the-box LLMs cannot. We learned how to prepare our data, splitting it into training and evaluation datasets, and how to start the fine-tuning job. From there, we started the job and learned how to manage this potentially long-running job using show and describe. From there, we looked at using the fine-tuned LLM for inference. We also learned how to build a Streamlit app that calls the LLMs to auto-generate custom responses for each customer ticket. Reflecting back on this whole course, we covered three main topics. Task-specific functions, the complete function, and the fine-tune function. I like to think of these as part of a sequence from simple to complex, depending on the needs of the task at hand. If you have a use case that can be solved with the task-specific functions, that's usually the best way to go. Use translate to translate text, summarize to summarize text, and so on. The functions are simple to use and out-of-the-box, but they don't offer flexibility to support other tasks. If you have a task that is not supported out-of-the-box by any of the task-specific functions, you can use complete with your choice of foundation model and some prompt engineering to solve it. Finally, if your use case is so specialized that it is not supported efficiently or adequately by the foundation models out-of-the-box, then you can use Cortex fine-tune function to fine-tune and create a custom model for it. As you can see, between the task-specific functions, complete, and fine-tune, Snowflake Cortex offers a powerful set of tools to solve many of your use cases. Congratulations again! You have covered the beginning of Gen AI in Snowflake's environment and started yourself on an exciting journey. Gen AI is changing everything, and those of us that are able to see and deploy the opportunities that Gen AI presents us with will leap ahead of others. Keep an eye out for further courses on Gen AI in Snowflake. It has been great spending this time working with you, and you have learned a lot and are probably excited to start putting these skills to work. Thank you for choosing to spend your time with us as you start your Gen AI journey. We look forward to seeing you in our other Snowflake courses. Goodbye for now!2. Let's practice!
Create Your Free Account
or
By continuing, you accept our Terms of Use, our Privacy Policy and that your data is stored in the USA.