diff --git a/TODO.md b/TODO.md deleted file mode 100644 index 764ac0d..0000000 --- a/TODO.md +++ /dev/null @@ -1,13 +0,0 @@ -### Next up -We have noticed that for complex projects the model is "lazy" in implementing many files. - -Hence, we want to let LLM generate a new prompt for a "sub-engnieer" that goes through each file, takes other relevant files as context and rewrites the files. - - -### More things to try -- allow for human edits in the code/add comments and store those edits as diffs, to use as "feedback" in the future etc -- Add step of generating tests -- Fix code based on failing tests -- Allow for rerunning the entire run from scratch, but "replay" human diffs by adding the diffs to the prompt and asking LLM to apply them in the new code -- keep a repository of examples of human feedback that can be reused globally -- Allow for fine grained configuration, per project, so that it can be regenerated from scratch applying all the human diffs that came after the initial AI generation step. Which diffs come in which steps, etc. diff --git a/identity/setup b/identity/setup index 4a0d841..4917945 100644 --- a/identity/setup +++ b/identity/setup @@ -10,6 +10,6 @@ Before you finish, double check that all parts of the architecture is present in File syntax: -```file.py +```file.py/ts/html [ADD YOUR CODE HERE] ``` diff --git a/identity/use_qa b/identity/use_qa index cb164c6..50628f5 100644 --- a/identity/use_qa +++ b/identity/use_qa @@ -8,6 +8,6 @@ Before you finish, double check that all parts of the architecture is present in File syntax: -```main_file.py +```file.py/ts/html [ADD YOUR CODE HERE] -``` +``` \ No newline at end of file diff --git a/main.py b/main.py index 811ef42..325fc59 100644 --- a/main.py +++ b/main.py @@ -19,7 +19,7 @@ def chat( run_prefix: str = typer.Option("", help="run prefix, if you want to run multiple variants of the same project and later compare them"), model: str = "gpt-4", temperature: float = 0.1, - max_tokens: int = 4096, + max_tokens: int = 8192, ): if project_path is None: