Hello, what plan should I have to avoid getting an ‘exceeded quota’ in order to use the API with the course chatbot, incorporating an approximately 140-page PDF? Thank you.
I have the same issue. Any ideas on how to solve it?
I had the same problem as well. Not sure how to solve it yet or if there is a limit and/or what it is.
For Vectorstores and Embedding lesson, you can work around using a HF embedding to bypass
openAI one and run all the cells on the notebook.
For subsequent lessons after you may upload the HF embedding based chroma db created
for the Vectorstores/Embedding lesson to docs/chroma to override the pre-existing one.
This is needed because the pre-existing db is created with openAI embedding and with dimension
1536 and most likely will be different to one you may use from HF. The caveat is obviously the
source to create the db is the same among different lessons. I am still susing it out but at least
able to progress to Retrieval lesson now.
And I am still trying to work through all subsequent lessons issues where quota & rate limit giving
me issues and right now it’s one step at a time and one lesson at a time.
The other thing is obviously every lesson starts in a new env, so you will have to re-download
all python dependencies for use with non openAI equivalent.
Cheers,
MCW
I’ve solved downloading the notebook and run it locally with my openai key: the consumption is really small (<1$) and everything works perfectly.
1 Like