From 91fed81c31fb24989048eb3c03c76be54c6f5042 Mon Sep 17 00:00:00 2001 From: Paul Gauthier Date: Wed, 14 Jun 2023 06:21:00 -0700 Subject: [PATCH] copy --- README.md | 23 +++++++++++++---------- 1 file changed, 13 insertions(+), 10 deletions(-) diff --git a/README.md b/README.md index b15cb44bb..ed8724837 100644 --- a/README.md +++ b/README.md @@ -1,7 +1,7 @@ # aider is GPT powered coding in your terminal `aider` is a command-line chat tool that allows you to write and edit -code with gpt-4 or gpt-3.5-turbo. You can ask GPT to help you start +code with OpenAI's GPT models. You can ask GPT to help you start a new project, or modify code in your existing git repo. Aider makes it easy to git commit, diff & undo changes proposed by GPT without copy/pasting. It also has features that [help GPT-4 understand and modify larger codebases](https://aider.chat/docs/ctags.html). @@ -60,17 +60,20 @@ You can find more chat transcripts on the [examples page](https://aider.chat/exa Aider supports `gpt-4`, -`gpt-4-32k` -`gpt-3.5-turbo`, +`gpt-4-32k`, +`gpt-3.5-turbo` and the the brand new `gpt-3.5-turbo-16k` model. -You will probably get the best results with GPT-4, because of its large context window and -greater competance at code editing. +You will probably get the best results with GPT-4, +because of its large context window and +greater competance at coding +which will allow you to edit larger codebases. -The GPT-3.5 models are less able to follow instructions for -returning code edits in a diff-like format. -So aider needs to send GPT-3.5 the original code -and ask it to return a full copy of the modified code. -Both the original and modified copies of the code count towards the context window token limit. +The GPT-3.5 models are limited to editing somewhat smaller codebases. +They are less able to follow instructions for +returning code edits in a compact format, +so aider has +and ask GPT-3.5 to return a full copy of any code that needs to be edited. +This rapidly uses up the context window. In practice, this means you can only use `gpt-3.5-turbo` to edit files that are smaller than about 2k tokens (8k bytes). The new `gpt-3.5-turbo-16k` model should be able to edit code up to 8k tokens (32k bytes).