The author is also considering the option of pre-training a custom model and instructing it. They are interested in hearing about others' experiences and strategies in dealing with similar data processing tasks. They are open to suggestions and keen to learn about the different methods used by others in the field.
Key takeaways:
- The author is seeking the most effective way to create a model that can produce great answers from a large corpus of unstructured data.
- They are considering whether to use RAG and whether fine-tuning the model on top of the data would be beneficial.
- They are also interested in finding out if there are any effective drag and drop solutions and how well the Open AI assistant API's built-in knowledge retrieval performs compared to other methods.
- Finally, they are contemplating whether it would be better to pre-train a custom model and instruct train it, and are seeking advice and experiences from others.