The Future Of Coding-Code Interpreter With Chatgpt- Its Time To Make AI Your Friend

hello my name is krishnaik and welcome to my YouTube channel so guys in this video I'm going to talk about an amazing thing that has come up with from open AI that is code interpreter now now I'll tell you guys the things that I'm learning I'm preparing so many different content probably creating videos and again within that span of time more things are actually coming up right so the research that I've specifically happening in the AI field is quite amazing specifically in the generative AI field if I probably see the hype cycle generative AI is also in the peak you know a lot of startups are probably working into this now let me go ahead and talk about code interpreter what exactly is code interpreter okay and for that I'll share my screen okay the code interpreter the functionality from this open AI is that it is probably including a plugin for chargpt that will allow the AI that is present inside this chargpt to perform different different work right and that are all related to coding let's say I'll say I'll I'll upload a let's say I want to solve a problem statement I have this particular uh train.csv file what I will do I will try to upload it over here and I'll say hey perform the detailed exploratory data data analysis along with the visualization charts for this particular data set automatically what this entire charge GPT will now do is that will start writing the code before we do not had this functionality but now you have it and you will only be able to access it if you have that 20 subscription right so guys before I go ahead I really want to make a quick quick announcement regarding our full stack data science Pro batch with the job Assurance program which is starting from 12th August 2023 the class timings will be from 9 am to 12 pm every Saturday and Sunday and after that from 12 to 2 we'll be having the doubt clearing session so if you are really interested in becoming a data scientist or a data analyst here we have also included modules like llms open open a Ai and many more things along with that we are going to learn about Python programming language we are going to learn about machine learning deep learning competition NLP and all so if you are interested to become a full stack data scientist I will be providing all the information regarding this batch in the description of this particular video thank you right now in order to enable it what you need to do first of all we'll enable it okay then I will show you multiple examples and it is quite powerful guys you will love it right and out of all the points that I'm probably going to discuss it uh the third or fourth point is specifically for data scientists and what all things you can actually do okay so just let's have a look on to this okay first of all go to the settings here you have something called as beta features and here you just enable this code interpreter as soon as you enable it okay you will start using it okay and for this code interpreter has been included in gpt4 okay right now uh in GPT 3.5 you don't have this in GPT 4 just click on this it is uh like hovering to the the fault section over here instead of default you just go and click on code interpreter okay so this is right now in beta stage now see the power what all things I can probably do okay so here is an amazing research paper I know you don't like to read research papers right but uh this is what is the research paper that has probably led to the invention of Transformers and right now whatever GPT 3.5 for you have this all are based on the Transformer architecture now if you really want to understand this let's see how it is possible obviously by reading it uh it'll take more amount of time but if you really want to have a quick summary about it and all what I will do I'll go to the Chart GPT I will upload this PDF so let me just go ahead and upload this PDF so I've uploaded English PDF over here now uploading this PDF over here now what happens is that now since code interpreter already a plugin is there now I'll just ask the question let's say that uh can you explain explain me about the research paper about the research paper in simple words okay so this is what is the question and I'm putting up this question now what will do whatever the AI is probably there now I will first of all try to understand the task okay it what it will do from this particular PDF it may apply OCR it will read all the text and then probably it will understand all the text and then it will send all those texts to probably a specific model and through that you will probably get a summarized information about the explanation about this research paper right so I will just go ahead and execute it now just see the magic over here can you explain me about the research paper in simple words sure I'll read through the document first now here you can see the code will be written so here you can see the pi PDF PDF file reader is happening so it is probably creating a environment over there and then it is started writing a python script itself and suppose let's say if it executes some kind of code and it gets some kind of warnings and all see attention is all you need this all things are basically there the paper is titled as attention is all you need now you get the entire information it is written by a team of researchers mainly from Google brain Google research as well as the University of Toronto the peer pair represents there's all these things all the information are basically there right and as you know charge GPT 4.5 it will take usually some time so I think it's still generating the response right so here it has been basically Reddit it is providing you a simple information about it simple explanation about this specific research paper Okay isn't it amazing and the best thing is that it is understanding their task then it is probably writing this python script that is required for that task okay so any research paper or any mathematical equation let's say you have a specific mathematical equation in some PDF format upload it and tell that okay to solve the top five uh top five problem statement it will give you the answer with respect to that also okay now this is quite amazing okay now let me just go ahead and create a new chart and let me show you one more one more way okay so let's say over here I will probably say plot oh Parabola curve with the equation with the equation Y is equal to x to the power of 2 I think to the power of 2 I can basically write like this now let's see whether the code interpreter will understand it will start writing the python code over here and then you have this amazing thing right see this very nice here's the plot Parabola curve this is all information this is the code right now see let's say that this code does not work then what it will still go to backtrack away and it will probably try to find out another way and here you will be able to see even this chart GPT right AI has access to so many different kind of python libraries right uh can you show me the same example with c bond Library let's say whether it has access to c bond or not c bond Library amazing right so now you're telling the AI to specifically work for you okay and also validate the results and tell you okay so here you will be able to see even started and this is what is the c bond content right it is imported c bond libraries and if I probably say that if you have like how many packages you specifically have it has so many different different packages I was just seeing a Twitter account where it was showing that it has more than access to 3000 packages with respect to the python so definitely a lot of work can be actually done okay now this is just a start for the data scientist till there are a lot of many things with respect to a data scientist so obviously many people wants to perform machine learning deep learning or any kind of task but the most difficult thing is about the ADN feature engineering so I'm just giving one data set this data set is from this bike sharing uh data set bike sharing demand from kaggle now what I'll do is that please provide please perform a detailed exploratory data analysis data analysis along with and provide different visualization charts now so guys now what I'm I have given the message over here please perform a detailed exploratory data analysis and provide visualization charts okay so once I go ahead and probably click this now you will be able to see uh chart GPT will automatically take this training data and it will do all the necessary tasks okay it can give you the results it can give you the answers everything and the best thing about this is that it also gives you observations now this is super super important I have already taught you during Eda right the main aim specifically is about observations and you specifically give the good observations over here with respect to each and every task you know let's say that I'm just doing df.head I'm seeing all the details over here so these are my observations with respect to my visualization now this will continue to go ahead now the first step let's continue with more detailed analysis we'll check for missing values look at the summary statistics analysis of the data so powerful it is for data scientists just imagine we used to do all these things manually now here you are able to get all the informations and here see automatically they're saying there are no missing values in the data set which is great as it means we don't need to perform any imputation for the missing data so it will also provide you Guidance with respect to the imputation purpose now here you can see that how quickly you are probably able to get the results which is amazing which is just just just amazing right and this is super important for a data scientist now here also you'll be able to see I'm able to get all the information with code and it also does backtrack okay now let's proceed with the visual analysis of the data we'll generate the following plots all these things are automatically done and here also you can see the date time column has successfully got converted into date time so some amount of feature engineering is also done I think it also performs well with respect to the feature engineering and here is all your entire plots which looks absolutely good right now this is super super amazing guys just imagine uh I don't know just in some days you open AI may sell you Roberts robots right if the robots are connected to Wi-Fi you just need to ask them hey job is do this thing for me and automatically start doing it you know this is with respect to code interpreter I don't know what all products may come in the future but the future is really bright you know and here you are just telling the AI to do each and everything see how beautifully it has also constructed the box plot over here everything is there some kind of Errors if you are getting it is the kind of warnings and here you can see that it is also created a local environment is currently working with python 3.8 so here on it is executing it it is seeing what all visualization diagram it basically shows and here also you can see right the next uh the create a bar plots for the categorical column so each and every feature it is probably doing the detailed EDM right which is absolutely very very very good now this is definitely powerful guys with respect to the code interpreter now the first thing what people will start thinking well I lose my job because now the AI is doing everything by it is helping you to do most of the things now you can quickly complete your work uh always understand AI should not be a enemy of human AI should be a friend of human right and with the help of AI human is able to do its work very very much quickly now here you can see this correlation Matrix the correlation chart how beautiful it looks if you probably want to use it you can probably use this specific code so all the things is there this is what was when the code interpreter was not there right uh probably charge Deputy would give you some kind of summary everything let's say the code is not working but here the code is working and then the code interpreter is basically giving you all the things as suggested if somebody asks you and then interview also what is code interpreter it is a plugin for chart GPT that allows you to communicate with the AI to probably non-code and analyze code create charts files perform mathematical calculations many more things as such right so this looks absolutely fine here you can see time series bike rental I don't think so manually you'll be able to get better Eda than this where it has basically explored each and everything now let's try some more thing and uh not only this this this code interpreter is also very good at modifying images with respect to images with respect to color and all right so I will just try to use one of the image of mine so let's see so this is the image that I have over here okay and this image what I will do is that I will just say create a five seconds GIF okay with zoom in with zoom in with slow zoom in with slow zoom in okay now if I'm telling the Stars right it has the image now it will start writing the code here it is saying to create a gif with slow zoom in I will perform the following steps there's this the library that is to be used as image i o and here for the sake of example this please note that zoom in effect will will be created by cropping the images gradually now this starts writing the code right and it is also able to create the GIF and many things like you can probably read a PDF you can do multiple things you can practice multiple things this this entire uh code interpreter also has a feature of OCR so if you probably give a PDF it will be able to extract the details if you probably give a handwritten notes it will be able to extract the details and probably solve it to you so now here is your GIF here you can see let's see uh whether I'll be able to get the GIF or not and I will probably download the GIF so this is the gift that I have actually got and let's see this has got downloaded this is good and here you have see this five seconds that's very nice right quickly uh you're able to get this right I'll zoom in or zoom out feature also you can probably do it okay I can probably say zoom in or Zoom so code interpreter definitely check it out guys you should definitely explore this uh and you'll be able to explore this only if you have that 20 subscription with respect to the open AI uh but at the end of the day understand I'm really much excited what may happen in the future definitely open AI is doing an amazing amazing work um you know with how we will be doing coding in the future how helpful it will specifically specifically be for us you know and this is what uh I'm really much interested in and to see that what all things uh will happen or whatever things will basically come in the future it is providing you almost each and every functionality uh even the kind of python code that it requires that is also there right that is very very good that is very very much amazing so I hope you like this particular video this was it from my side I'll see you all in the next video have a great day thank you and all take care bye

Channel: Krish Naik

Views: 36,275

Keywords: yt:cc=on, open ai code interpreter, GPT -4, the future of coding using code interpreter, openai tutorials, ai tutorials, krish naik openai

Id: fbrPaKLGd1o

Length: 15min 19sec (919 seconds)

Published: Tue Jul 11 2023