dearwolf/aider
1
0
Fork 0
mirror of https://github.com/Aider-AI/aider.git synced 2025-05-28 16:25:00 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions 4

copy

Browse source
This commit is contained in:
Paul Gauthier 2023-07-08 09:23:05 -07:00
parent 5279f83a02
commit bcc8723fa7
1 changed files with 16 additions and 25 deletions
Download patch file Download diff file Expand all files Collapse all files

41
README.md
View file

@ -133,38 +133,29 @@ Aider has some ability to help GPT figure out which files to edit all by itself,
## GPT-4 vs GPT-3.5 ## GPT-4 vs GPT-3.5
Aider supports all of OpenAI's chat models, including Aider supports all of OpenAI's chat models,
the the brand new `gpt-3.5-turbo-16k` model. but you will probably get the best results with one of the GPT-4 models.
They have large context windows, better coding skills and
You will probably get the best results with one of the GPT-4 models, they generally obey the instructions in the system prompt.
because of their large context windows, GPT-4 is able to structure code edits as simple "diffs"
adherance to system prompt instructions and
greater competance at coding tasks.
The GPT-4 models are able to structure code edits as simple "diffs"
and use a and use a
[repository map](https://aider.chat/docs/ctags.html) [repository map](https://aider.chat/docs/ctags.html)
to improve their ability to make changes in larger codebases. to improve its ability to make changes in larger codebases.
The GPT-3.5 models are supported more experimentally
and are limited to editing somewhat smaller codebases.
They are less able to follow instructions and
aren't able to return code edits in a compact "diff" format.
So aider has
to ask GPT-3.5 to return a new copy of the "whole file" with edits included.
This rapidly uses up tokens and can hit the limits of the context window.
For more detailed information and a quantitative comparison, here are
[code editing benchmark results for GPT-3.5 and GPT-4](https://aider.chat/docs/benchmarks.html).
GPT-3.5 is supported more experimentally
and is limited to editing somewhat smaller codebases.
It is less able to follow instructions and
can't reliably return code edits as "diffs".
Aider disables the Aider disables the
[repository map feature](https://aider.chat/docs/ctags.html) repository map
when used with GPT-3.5 models. when using GPT-3.5.
The `gpt-3.5-turbo` context window is too small to include a repo map.
Evaluation is still needed to determine if `gpt-3.5-turbo-16k` can make use of a repo map. For a detailed and quantitative comparison, please see the
[code editing benchmark results for GPT-3.5 and GPT-4](https://aider.chat/docs/benchmarks.html).
In practice, this means you can use aider to edit a set of source files In practice, this means you can use aider to edit a set of source files
that total up to the sizes below. that total up to the sizes below.
You can (and should) add just the specific set of files to the chat Just add the specific set of files to the chat
that are relevant to the change you are requesting. that are relevant to the change you are requesting.
This minimizes your use of the context window, as well as costs. This minimizes your use of the context window, as well as costs.