Model Pusher: Customer Segmentation

Key Takeaways

hello everyone welcome to the session so now let's talk about the very last part where we are trying to push our model pickle file inside the S3 bucket I just want to show you how is this happening again we need to specify the bucket name and all those stuffs let's get started and we'll try to see that how these things are happening and then you are good to go right with this particular project and once you will be able to understand this concept again I am saying you will be able to understand further projects very easily so let's get started I'm going inside the start model model Pusher here you can see we are initiating the model Pusher what we are passing we are passing the model trainer artifact the best model file and the model Pusher config file now here you can see that what we are trying to do we are trying to save the model and after that if you will observe here we are specifying the bucket name and the model path bucket name and the model path because we need to pass the that model file inside that particular bucket name which is present inside the S3 bucket so if if I just show you what is the bucket name or where it is specified and the model path you can see what is the model file name model. PK as well as uh if I'll go back and show you what is the bucket name if I'll just click on that and if I'll just show you uh let me show you the bucket name so let me go it manually if it is not clickable you will be easily able to get that uh yeah so here you can see right I have mentioned what is the name of a training bucket name and what is the name of a prediction bucket name although it is not specifically required as of now we are dealing with a training bucket name which is the customer segmentation bucket so we have already defined this bucket inside the S3 bucket so now that model. PK file will go inside this particular pickle file uh this particular bucket okay so let me go back again where was that start model Pusher let me again go inside this sometimes it's not clickable let me go it manually okay so now we will be able to understand right so we will be able to save the model and after that we will be able to push that model Pusher artifact inside the S3 bucket how we will Define the bucket name as well as we will Define the key path inside that that's simple thing that we are doing inside the model Pusher that's it very simple code right so we already have created the model we already have that model. pkl file real purposes of model pushing is that okay if my model is accepted if my model is giving a best predictions if my model is giving a best F1 score accuracy I need to save it somewhere so that I will use that Peele file with whatever bucket name I am specifying for the deployment predictions for the future predictions and that is what we really want to do here by specifying the bucket name and the S3 model key path very simple so this is what we are trying to do here and if you will see now if you will see now if I'll go back to the training pipeline if if finally if this uh this part will be done you can see that exited the Run pipeline method of train pipeline class and this is the real thing that we are doing real method that is running on whenever I'm hitting that so starting the model Pusher uh starting the model evaluation model trainer data transformation data validation data inje so you can see right for every particular uh you know starting from data injection to the model Pusher we are creating a CL class specifically and how we will be able to Define each and every class separately via different different components and then every component has its own purpose and now you clearly know that which particular component will have what kind of purpose so this is all about this project starting from first of all we talked about data inje how it is happened after that we have talked about data validation what's the real purpose of that after that we move towards the data uh I would say transformation where we have seen that clusters happen and we will be able to get a new Target column after that we have seen the model trainer where we will be able to get the best model now you can see how that that is happening after that model evaluation you can see that how the model evaluation happened and what's the real meaning of that and finally the model Pusher so that six in in a way we will be having a six classes six different different you know components where each component has its own specific task and that's the real you know no solution approach that we are trying to do to do the customer segmentation in this project now what I will do in the next upcoming part of the video I will try to demonstrate you that how you will run this file okay so obviously when you are running the uh when you are using mongodb I told you right you have to specify the mongodb URL when you are using S3 bucket right you have to use the Amazon AWS services so for that also you need to specify what is the access name what is the key password for that so I'll try to show you that how you will try to run this file and then what output are we getting that demonstration I will do in the next upcoming part of the video so until so far the videos are basically on the development part that how the major development happened internally now the next part of the video is the demonstration part where we will see that okay once we will be able to develop the files once we will be able to generate these things how our output will look like let's try to understand that part that that part as well and uh let's end this video with this particular note happy leing to all bye-bye everyone