I’ve just started using Chatgpt 4 to learn python and programming. Created a script allowing me to input a video, detect faces in the video, filter blurry + duplicate faces, output individual png frames for each individual. Thinking I could maybe sell this as a software to retail stores with high theft levels. Only took around 10 chats to get it to work as intended, starting with a simple (video2png script) it messed up a few times but after resubmitting the code and asking for it to fix it, it worked. It actually took longer just trying to figure out how to run a python script 😂 Overall I’m very impressed, went from knowing literally nothing about coding / programming, to successfully running an optimized custom script within an hour. it’s not perfect but it’s definitely beyond game changing.
Really interesting. I used ChatGPT to help me with a build project a few months ago. I was curious about your point at 4:45 about 'code style' and how you chain you code vs how ChatGPT formats it. Do you mind expanding on that?
Hi there! ChatGPT seems to take the "average" code style from example it sees across the web. I prefer to chain my Pandas expressions, where a dataframe gets passed from one function to the next rather than create intermediate variables that store the dataframe at various points. So instead of: df = pd.read_csv("retail.csv") df["date"] = pd.to_datetime([df["date"]) df.set_index("date", in_place=True) resampled_df = ... We could do: resampled_df = ( pd.read_csv("retail.csv", parse_dates=["date"]) .set_index("date) .groupby(["store_nbr", "family"]) .resample("M")["Sales"].sum() .reset_index ) . TL; DR: ChatGPT won't always align with your style/workflow preferences, so you may need to refactor code to get it into the shape you prefer. .
I’ve just started using Chatgpt 4 to learn python and programming. Created a script allowing me to input a video, detect faces in the video, filter blurry + duplicate faces, output individual png frames for each individual. Thinking I could maybe sell this as a software to retail stores with high theft levels. Only took around 10 chats to get it to work as intended, starting with a simple (video2png script) it messed up a few times but after resubmitting the code and asking for it to fix it, it worked. It actually took longer just trying to figure out how to run a python script 😂
Overall I’m very impressed, went from knowing literally nothing about coding / programming, to successfully running an optimized custom script within an hour. it’s not perfect but it’s definitely beyond game changing.
Really interesting. I used ChatGPT to help me with a build project a few months ago. I was curious about your point at 4:45 about 'code style' and how you chain you code vs how ChatGPT formats it. Do you mind expanding on that?
Hi there! ChatGPT seems to take the "average" code style from example it sees across the web. I prefer to chain my Pandas expressions, where a dataframe gets passed from one function to the next rather than create intermediate variables that store the dataframe at various points.
So instead of:
df = pd.read_csv("retail.csv")
df["date"] = pd.to_datetime([df["date"])
df.set_index("date", in_place=True)
resampled_df = ...
We could do:
resampled_df = (
pd.read_csv("retail.csv", parse_dates=["date"])
.set_index("date)
.groupby(["store_nbr", "family"])
.resample("M")["Sales"].sum()
.reset_index
)
.
TL; DR: ChatGPT won't always align with your style/workflow preferences, so you may need to refactor code to get it into the shape you prefer.
.
Can these codes be detected?