Assistance is very slow to get the answer from PDF file

API
,
1

Hi
The client.beta.threads.runs.retrieve() method takes minimum 8 seconds to answer the question.
Below is the code snippet.
Please note, as a knowledge base, I have pdf file of 100 mb of textual information.

while True:
run = client.beta.threads.runs.retrieve(thread_id=thread_id, run_id=run_id, timeout=5.0)
print('status: ', run.status)
if run.status == “completed”:
messages = client.beta.threads.messages.list(thread_id=thread_id)
latest_message = messages.data[0]
text = latest_message.content[0].text.value

        exit_time = time.time()
        loop_time = exit_time - enter_time
        print(f'Loop time: {loop_time} seconds')
        break
2

Hi and welcome to the Developer Forum!

If you PDF’s just contain text, it might be worth trying to extract just the text and upload that and see how it performs, but 8 seconds does not seem too far of what is expected, especially if you are on the first couple of account Tiers, i.e. low usage tiers.

See

https://platform.openai.com/docs/guides/rate-limits/usage-tiers?context=tier-free