search

SoulScribe-AI / ml

0 stars 0 forks source link

Summarizing Entries #7

Open Jaishreebala opened 10 months ago

Jaishreebala commented 10 months ago

Summarize all the entries in a journal before inputting to the generative model to reduce the amount of text to be processed by the model. This can reduce tokens and potentially reduce cost of the generative model as well.

MeaganShim commented 10 months ago

look into this model and see how it performs on summarizing journal entries https://huggingface.co/Falconsai/text_summarization

Jaishreebala commented 10 months ago

Open AI Summarization Model: https://medium.com/muthoni-wanyoike/implementing-text-summarization-using-openais-gpt-3-api-dcd6be4f6933

MeaganShim commented 9 months ago

makes overall latency worse in general testing results were not detailed enough and a lot of context in the conversation was lost. The cost outwieghs the benefits at this moment in time Currently moving this to low priority.