3 ms·
Is it worth it to host this on an EC2 which might take ~1.5$ per hour (on demand) than running GPT3.5 API for this purpose? What is the breakeven number of que
by noobcoder 4y ago
Is it worth it to host this on an EC2 which might take ~1.5$ per hour (on demand) than running GPT3.5 API for this purpose?
What is the breakeven number of queries (~2000 tokens/query) to justify the hosting of such model?