Use embeddings create with API on my completions
Hello!
I'm working on creating an AI assistant similar to OpenAI's with Mistral. Initially, I'm focusing on replicating OpenAI's thread system, which involves using old message storage and embeddings. This approach is to avoid storing all messages in lengthy conversations.
I plan to use the completion API for this task. Hosting my own version of Mistral 7B / 8x7B is an option, but it seems unnecessary for my current needs. I can easily generate completions and embeddings using the platform API, but I'm unsure how to incorporate my embeddings into the completion requests.
Could anyone advise on how to do this? Am I understanding the use of embeddings correctly in the context of the platform API?
Thank you in advance for your help. I'm new to using AI models, and any guidance or resources for learning would be greatly appreciated. :)