I Found The Best Alternative To ChatGPT- Get The Power Of ALL LLM Models

hello all my name is krishn and welcome to my YouTube channel so guys today in this particular video I'm going to talk about this amazing product which is called as chat llm team by aus.to chat GPT I'm not using chat GPT much but I'm using this because the cost is also less and it also has amazing features in this video I'm going to talk about all the features of this particular product one by one and I'll be showing you with practical demo so let's go ahead here you will be able to get all the different kind of models let it be a paid model open source model so like here you have gp4 oxs you have Lama 3xs you have Cloud Opus you have gp4 you have gini 1.55 Pro you also have Abacus dois smog now what exactly this specific model is all about this is a fine-tuned model of Lama 3 right so you'll be able to see that currently this is the best open source model the and in the world right now and it is based on Lama 3 finetune right so entirely Lama 3 model has been fine tuned and this is the model that you have actually got right and this will also be able to perform various task out there right and we will be discussing about what all task is there but if I just go ahead and compare with chat GPT in chat GPT you'll just be getting the variants of GPT models right but here you are getting variants for the open source models also you're also getting the uh all the variants of the model with respect to play paid s from chat GPT from Cloudy where whatever you specifically want right and uh all these open source models also that you'll be able to see over here this uh the backend inferencing is specifically Gro right so you'll be able to see the speed differences also whenever I try to use this kind of application so let's go ahead and let's see some of the application over here so first of all what I'm actually going to do I'm just going to go ahead and compare GPT 40 with Lama 3 okay so let me just go ahead and ask some question I'll just go ahead and ask hey uh what is generative AI okay now let's go ahead and press enter now here you'll be able to see that I'll be able to get the answer over here right and I'm able to see lot of uh okay the text is basically getting generated I'm able to see everything over here all the text is basically getting generated I'm also able to see the code so this is perfectly fine but now let me just go ahead and select some different open source model like Lama 3 okay now you know that Lama 3 is already like over here specifically in this uh app you know Lama 3 is hosted in that grock inferencing engine so uh this Lama 3 will be giving the response from Gro API itself right so now if I go ahead and ask okay what is generative AI okay now if I go ahead and ask this now here you'll be able to see the answer will be pretty much faster when compared to the GPT 4 oh why because Gro is completely based on lpu and this is the major difference if I compare chat GPT if I compare other llm applications that are available in the market right now this is the most versatile one because you have the combination of everything everything if you want open source if you want paid if you want different different paid Source models right like open AI cloudy everything is available over here not only that uh you'll be also able to see that all these models will also be able to execute code it will be able to provide you it comes with a code interpretor directly right so let's say if I go ahead and ask a question right uh provide me a code uh for playing playing snake game in Python right so if I go ahead and ask this question right I will be able to get the entire code itself right and uh I will be able to see this right not only that you'll also be able to see that I'm having this uh attachment button right so let's say that I'll will go ahead and add one file okay and uh let's say one of the file that I have uploaded over here is Krish resume I'll go ahead and select it okay so let me just go ahead and select this okay now after selecting it let me just go ahead and ask any question and for this let me just go ahead and select GPT 40 right now over here I have selected Chris resume let me just go ahead and ask a question hey uh from the resume provide me the summary of all the skill sets Okay and uh this is just for my resume I'm just trying to check it out and here you can see so nicely it is being able to see from the chat right the entire PDF and it is being able to give you this entire thing right and and the best thing is also like I'll also be discussing about the cost and this is far more cheaper than the chat GP things that we specifically use and that is the reason why I'm using this you know I've been exploring this because see in chat GPT anyhow they take $20 but this will be giving you just in $10 monthly subscription so guys I also want to show you one more feature that whether this chat GPT 4 or gini 1.5 Pro can analyze the image okay so for that what I will do is that I will just go ahead and upload from my computer okay so I I'll just try to upload some specific images and I've created a lot of images over here let's see I will go ahead and select this okay so let's say uh I've selected this specific image um this is from one of my thumbnail from my YouTube channel um can you let me know um the title from this thumbnail image I'm just going to search for it okay and if I go ahead and execute it you'll be able to see that yes I will be able to get the answer so the title from the image thumbnail is llm evaluation with mlow uh ml flow right so I had actually created a video regarding that recently and that is what is the thumbnail right uh can you provide some suggestion for the image for the image to look better to for can provide me some suggestion for the image let's see what kind of suggestion it'll be able to provide again for this you can also use uh um you know GPD 1.5 Pro it's up to you so now here here you can see hey these are some of the suggestion Ure the text is easily readable by increasing the contracts between the text and the background so all the information is specifically the faicial the person in the image has a friendly and engaging expression which is great ensure that the expression matches the tone of the video content so all those information you can probably see from here uh and uh I feel that yes this is actually doing all the task here you can actually use open source paid Source models whatever things you want along with this uh you also have features to inv invite team members so if I go ahead and probably click on this right and you want to probably add any of the team members so you can just go ahead and click on invite user provide the user email let's say that I want to go ahead and provide to my other email ID let's consider this I will just go ahead and send the invite automatically I'll be able to get so you can see okay it's already there in the organization so let me just give it on my other email ID so I can go ahead and give on this email ID and let's say if I'm sending the invite over here so I will specifically get the invite and once I probably confirm uh you know I will be able to see all the invites over there right and that is how we can work collaboratively in a team right so this is again one another amazing feature that we have right so uh this was uh about invite team uh other than that you'll also be able to connect with different different files because let's say you want to go ahead and create a chat bot where you can communicate with slack you can communicate with teams you can communicate with Google Drive so everything is probably available over here here and based on this you can probably do this let's say I want to go ahead and connect with my Google Drive right so here you'll be able to see this particular Services I can go ahead and click on add a new connector let's say all these connectors are specifically there right like bigquery uh Marketo mongodb sales force so that I get a kind of data source right from where my data is probably coming up so that at the end of the day if I just ask a question to my chatbot I should be able to do it let's say I want to go ahead with Google Drive so once you see this you just need to give your Google Drive link with respect to any PDFs or anything that you have actually stored in the Google drive folder once you give this particular link then just click on Save and automatically this becomes your data source and then you'll be able to communicate it so this is also one amazing uh feature that is actually provided over here and not only that you have other connectors as I've already shown you uh you can go with slack teams whatever things you specifically want so let me just go ahead and show you the most amazing feature in this chat llm uh that is the option of creating a new chatbot or your own custom llm now here what I will do is that I'll try to create a project with my own custom data or PDF file and then I'll try to create a document Q&A search from that just kind of chatbot which will be able to retrieve any results from those kind of PDFs so first of all I will just go ahead and create a uh project over here and let me just go ahead and click on custom chatbot you can also create an AI agent okay uh let's say that here I will just say PDF uh chatbot okay and I'll just go go ahead and click on finish now once I do this first of all I need to attach a doc data set so I will just go ahead and create or attach a do data set I have two options one is attaching an existing data set or creating a new data set I will just go ahead and create a new one let's say this will be my uh PDF data set that I really want to create okay or I'll just say my custom PDF and I'll click on continue the feature group type will be a list of documents so here I'll just click on continue now I will select on the file upload option now let me just go ahead and click on it here you have an option of uploading a zip file you have an option of uploading PDF file so I will just go ahead and click on one PDF file over here okay and that PDF file is also opened over here you can actually see that completely okay and then uh I will click on add data set see you also have an option of uh selecting read from external service you also have an option of selecting stream data set you can also check it out so I'll just go ahead and click on add data set now once I do that uh I will just go ahead and click on the next step now you'll be able to see that my entire project will get created in this specific way right so here you can see uh I'm setting up the data pipeline uh and it is also creating the feature group now let me just go ahead and click on the document retriever because uh once I probably get all my data in my feature store in the form of vectors I need to create a retriever so that it acts an interface with respect to any query that I asked from here it should be able to retrieve it then I'll go go ahead and select the feature group the same feature group that I had actually selected that is PDF uh data set here the chunk size I will keep it to 500 okay and the chunk overlap fraction I'll keep it to 10% okay and uh I think the options are over here I will just go ahead and click on create okay data set with table name uh PDF data set is still being processed so uh this will probably take some amount of time but let me just show you one of the project that I have already developed so I will click on the projects over here okay and let me click on my Pro chat chatbot project so here you can see I had uploaded one test data then this was my feature group I created my document rever then over here you can see manage deployment so you also have an option of managing the deployment over here so once you manage the deployment once you train the model see after step three you'll be able to train the model over here then you can also retrain it if you want and then you can actually deploy it so all those specific option is over there and obviously it is going to take some amount of time because here you'll be able to see that once I'm able to create this right it is telling me hey uh the PDF data set is still being processed so initially the upload and converting that into a vectors will take some amount of time but I've already done that specific project so I'm showing you over here itself so initially we go ahead and set up our data pipeline then feature group then document retrievers then we train the model with respect to that particular data set then we do the deployment and finally we have this kind of uh entire pipeline that is built okay now the other option is that I can just go ahead and click on ask AI to create create a uh custom chat bot so if I click this I just need to give my custom prompt what I really need to create it and just click on run then automatically this entire data pipeline will also get created now just to show you that how this things work right and how good this is right so what I will do I will go to deployments and I will this is the deployed project that I have I will click on dashboard okay and then I'll go to Abacus do a chat okay so here is my chatbot project deployment now let me just go ahead and ask any question that I want okay okay let's uh ask something like okay let's ask this okay uh how property is measured okay so let me just go ahead and ask this and this will show you how powerful this chatbot is and quickly in 10 to 15 minutes you'll specifically be able to create so how uh poverty how poverty is measured right so I'm just going to ask this particular question and let's go ahead and press enter so now what this is going to do is that it is going to probably communicate with the document retriever and from that feature store and it is going to get the result and it is going to display it so here you can see how quick it is how fast it is and you are able to get the answer prty status is determined by comparing an annual income and this entire information that you'll be seeing over here it will be coming from this particular PDF right see is determined by comparing annual incomes to a set of dollar values okay so this is how you'll be able to see that any question that you have specifically with respect to the PDF you have created your own custom Uh custom project and you are able to do it right and uh this is the most amazing thing over here right so uh you can create your own custom chatbot you can create your uh multiple you can see lot many features you can do the analysis of the code you can probably work with images analyze the image do multiple things now let me talk about the best thing about this Abacus uh you will be getting see the thing is that I have replaced completely from chat GPT I'm actually using this because I usually get the multiple llm capabilities and the price over here is just $10 see I'm not promoting anything uh the reason why I move from chat GPT to this particular platform because I have each and every platform I will be able to see this you know wherever I specifically require apis it also provides you apis which will you will be able to access it through it and you'll be able to execute your code um again the price is just $10 instead of giving $20 and the one one more amazing thing is that you'll also be able to get one month free trial okay so that is how you can also use it so I hope you like this particular video let me just see whether this is allowed me to do it or not right now so I'll just go ahead and click on create so it will still tell okay now it has started now see now here is your list of documents uh this is your feature group this is your document retriever now I'll just go ahead and click on train model now this is going to train the model completely on this particular document Retriever and this on this feature group right so if I just go ahead and click on train so this is my document retriever selected uh evaluation group I'll select on none what is the default I'll just go ahead and click on train model right now once you did see this right uh the next next thing is that you can manage deployment after at least one model has been completed training so I'll click on train model over here okay so this is basically getting trained okay sound Goods uh okay two times I have done retrain retrain no worries but anyhow it is just going to get retrained with respect to the data that I have okay let me go back to my this so once this is trained then this manage deployment will get enabled and you just need to do the deployment once you do the deployment you'll be able to see that uh you'll be finding this way right you'll be able to see your train models will get deployed right and after the deployment is done you just need to go to the deployment Tab and you'll be able to see the dashboard prediction API everything so see prediction API you have this you have the entire code also you'll be able to see this entire code right like if I ask what is the meaning of life what is poverty I can go ahead and ask over here and this will just give me the answer right I hope you are able to see this so if I go ahead and ask what is the poverty I will just go ahead and click on run and here you'll be able to see this amazing platform altogether and uh this is what I liked about it and you'll be able to see the response that we'll be able to get okay so definitely go ahead and try it out uh it is up to you to use it but again this is a better way to probably choose other than chat GPD because uh I find it out because I need to do it use it in my day-to-day task and obviously cost if one cost is one of the factor in my mind then definitely I'll go with this because here I get all the varieties of the model that I want like cloudy gp4 gin 1.5 everything at such so here this this was it for my side I will see you all in the next video thank you take care bye-bye

Channel: Krish Naik

Views: 35,285

Keywords: yt:cc=on, chatllm, chatgpt alternatives, generative ai tutorials, machine leanring tutroials

Id: PYvAfKNGVlg

Length: 16min 32sec (992 seconds)

Published: Sun Jun 09 2024