I want to feed a file data into chat gpt and interact with chat gpt.
I tried to read a file and make it as array of string. Then tried to feed into chat completion API.
There was an rate limit error for this case as stringfied file contains too many tokens.
How do other people solve this and working on feeding data for chat completion?
Is there way to combine with file API?
Additional questions, My personal account has a paid plan but it seems it is not working with gpt-4 model for chat completion. What would I miss it here?
If you hit a rate limit on a single request, you are sending far far too much data. If you only exceed the model limit of 4k (or 16k for gpt-3.5-turbo-16k) but not going over the per-minute rate limit near 100000 tokens, you instead get a context length error so we know the data sent must have been huge.
The model can only accept and understand a small amount of data at once. About 2000 words, measured as tokens. You must find a way to use your data in small batches.
If you are using an account with monthly billing, but have never been billed because you barely use the service, that explains why you don’t have access to gpt-4 yet. You must be a “paying customer”.
Here’s a dangerous script to quickly spend $1 on API usage on your account before the end of the month. Then when you are billed to your payment method in a few weeks, you should shortly after be granted access.
It also depends on how large your file is. You can always make use of plugins where you can just drop a link of a file and then ChatGPT will be able to read it. Thereafter you can interact with the ChatGPT about the file data.