1 00:00:00,000 --> 00:00:01,000 Hello guys. 2 00:00:01,000 --> 00:00:06,000 So we are going to continue the discussion with respect to our end to end project, uh, gen AI project, 3 00:00:06,000 --> 00:00:10,000 which is nothing but Q&A conversation with PDF including chat history. 4 00:00:10,000 --> 00:00:15,000 Okay, now um, we will be implementing this completely step by step. 5 00:00:15,000 --> 00:00:19,000 First of all, let's go ahead and start importing all other libraries. 6 00:00:19,000 --> 00:00:26,000 Already we have discussed about this chat history, how to include that inside the chain in our previous 7 00:00:26,000 --> 00:00:27,000 module. 8 00:00:27,000 --> 00:00:28,000 And that module was nothing. 9 00:00:28,000 --> 00:00:32,000 But, uh, it was something related to conversational queue and a chat bot. 10 00:00:32,000 --> 00:00:36,000 So a detailed video in the previous module has been already uploaded. 11 00:00:36,000 --> 00:00:38,000 Now this will be completely an end to end project. 12 00:00:38,000 --> 00:00:41,000 So first of all we will go ahead and import Streamlit as St. 13 00:00:41,000 --> 00:00:44,000 Now this Streamlit that we are specifically going to use. 14 00:00:44,000 --> 00:00:48,000 Other than that I will be also importing some of the important libraries. 15 00:00:48,000 --> 00:00:52,000 So first of all, let me just go ahead and import two important libraries. 16 00:00:52,000 --> 00:00:56,000 One is called as create History aware Retriever and create retriever chain. 17 00:00:56,000 --> 00:00:57,000 Okay. 18 00:00:57,000 --> 00:01:00,000 So first of all this will basically be used to create the retriever chain. 19 00:01:00,000 --> 00:01:05,000 And uh if I want to add history along with the history prompt I'll be using this Create history aware 20 00:01:05,000 --> 00:01:06,000 retriever. 21 00:01:06,000 --> 00:01:07,000 Right. 22 00:01:07,000 --> 00:01:09,000 Uh, this is just like a retriever. 23 00:01:09,000 --> 00:01:14,000 This function will be just used to create a retriever with chat history functionalities. 24 00:01:14,000 --> 00:01:14,000 Along. 25 00:01:14,000 --> 00:01:18,000 Along with that I will be using this create stuffed document chain so that we can combine the entire 26 00:01:18,000 --> 00:01:20,000 document and send it to the context. 27 00:01:20,000 --> 00:01:23,000 Then we are importing our vector store db that is chroma. 28 00:01:23,000 --> 00:01:25,000 You can also go with files. 29 00:01:25,000 --> 00:01:26,000 It is up to you. 30 00:01:26,000 --> 00:01:32,000 Uh, then uh let's go ahead and import our chat history chat message history. 31 00:01:32,000 --> 00:01:36,000 So that will be available inside your Langton underscore community dot chat underscore message underscore 32 00:01:36,000 --> 00:01:38,000 history already. 33 00:01:38,000 --> 00:01:41,000 Uh, all these libraries has already been installed in this requirement dot requirements.txt. 34 00:01:42,000 --> 00:01:45,000 So please make sure that you have this requirement dot txt. 35 00:01:45,000 --> 00:01:47,000 And this is how the entire folder structure needs to be created. 36 00:01:47,000 --> 00:01:48,000 Like right? 37 00:01:48,000 --> 00:01:50,000 I have all the projects over here. 38 00:01:50,000 --> 00:01:53,000 This is my requirement dot txt with respect to all the files right? 39 00:01:53,000 --> 00:01:58,000 So please make sure that uh, so you don't have to even worry about any installation at arch. 40 00:01:58,000 --> 00:01:58,000 Okay. 41 00:01:58,000 --> 00:02:04,000 Then uh, you have this entire, uh, base chart message history over here. 42 00:02:04,000 --> 00:02:06,000 So I will also go ahead and import this. 43 00:02:06,000 --> 00:02:07,000 Okay. 44 00:02:07,000 --> 00:02:13,000 Then, uh, coming to the long chain underscore code or prompts, that is chat prompt template and message 45 00:02:13,000 --> 00:02:16,000 placeholder that also we will be using it okay. 46 00:02:16,000 --> 00:02:20,000 This message placeholder is to define the uh session key. 47 00:02:21,000 --> 00:02:25,000 Uh and what kind of session key I'm actually storing all the conversation hit okay. 48 00:02:25,000 --> 00:02:29,000 And this is my base chat message history, which we have already discussed. 49 00:02:29,000 --> 00:02:35,000 Then, uh, we will also be importing, uh, a couple of libraries, which is called as Chat Grok. 50 00:02:35,000 --> 00:02:40,000 So let's go ahead and uh, import this something called as chat grok over here. 51 00:02:40,000 --> 00:02:47,000 So chat grok will be specifically used uh, with respect to um, the OpenAI API key that I'm actually 52 00:02:47,000 --> 00:02:48,000 going to use. 53 00:02:48,000 --> 00:02:51,000 So sorry, not OpenAI grok API key that I'm actually going to use. 54 00:02:51,000 --> 00:02:54,000 Uh, and here I'm actually using grok. 55 00:02:54,000 --> 00:02:57,000 Uh, the model that is nothing but Google Gamma two. 56 00:02:57,000 --> 00:02:57,000 Okay. 57 00:02:57,000 --> 00:03:02,000 Along with this I will also go ahead and import Huggingface embeddings. 58 00:03:02,000 --> 00:03:07,000 So here is my Huggingface embeddings so that I will be able to import all my embeddings itself in order 59 00:03:07,000 --> 00:03:08,000 to import it. 60 00:03:08,000 --> 00:03:12,000 Uh, quickly I will go ahead and import two libraries that is import OS. 61 00:03:12,000 --> 00:03:12,000 Okay. 62 00:03:12,000 --> 00:03:19,000 And uh, I will say from dot env import load underscore dot env. 63 00:03:20,000 --> 00:03:20,000 Okay. 64 00:03:20,000 --> 00:03:23,000 And I will go ahead and initialize this load underscore dot env. 65 00:03:23,000 --> 00:03:28,000 And let me quickly go ahead and initialize this h.f underscore token which is present in my environment. 66 00:03:28,000 --> 00:03:28,000 Variable right. 67 00:03:28,000 --> 00:03:29,000 H.f underscore token. 68 00:03:29,000 --> 00:03:32,000 And this I have already shown you how to properly load it. 69 00:03:33,000 --> 00:03:37,000 Then, uh, I will go ahead and create my embeddings over here. 70 00:03:37,000 --> 00:03:40,000 So this is my embeddings that I'm actually going to use hugging face embedding. 71 00:03:40,000 --> 00:03:42,000 And the model name is this one right. 72 00:03:42,000 --> 00:03:45,000 All mini lm l6 v2 okay. 73 00:03:45,000 --> 00:03:49,000 So this is my model that we are specifically going to use. 74 00:03:49,000 --> 00:03:50,000 Perfect. 75 00:03:50,000 --> 00:03:52,000 Till here everything looks good enough. 76 00:03:52,000 --> 00:03:56,000 Uh, and now let us go ahead and see some more libraries that will be requiring. 77 00:03:56,000 --> 00:03:57,000 Okay. 78 00:03:57,000 --> 00:04:02,000 Uh, so I will be importing some three important libraries that I'm actually two important libraries 79 00:04:02,000 --> 00:04:04,000 that I'm actually going to use over here. 80 00:04:04,000 --> 00:04:08,000 One is recursive character text splitter and pi PDF loader. 81 00:04:08,000 --> 00:04:09,000 Right. 82 00:04:09,000 --> 00:04:11,000 This is basically to load the PDF itself. 83 00:04:11,000 --> 00:04:11,000 Okay. 84 00:04:11,000 --> 00:04:20,000 So uh, I think almost all the libraries has been, uh, taken place and has been, uh, imported. 85 00:04:20,000 --> 00:04:25,000 One final, uh, library that I want to import is with respect to the runnable history. 86 00:04:25,000 --> 00:04:25,000 Okay. 87 00:04:25,000 --> 00:04:29,000 So here we are going to use this runnable with his message history. 88 00:04:29,000 --> 00:04:34,000 The reason why we are using this is that we will be able to create our conversational queue and a chatbot 89 00:04:34,000 --> 00:04:37,000 with the help of, uh, the chat history that we have. 90 00:04:37,000 --> 00:04:38,000 Right. 91 00:04:38,000 --> 00:04:44,000 So here what we will be doing is that, uh, we will be using this to create our retriever, uh, when 92 00:04:44,000 --> 00:04:48,000 our, when we are creating, our retriever will be making sure that when we create that specific agent 93 00:04:48,000 --> 00:04:54,000 to run, the retriever will also be using this library to probably include the chat history. 94 00:04:54,000 --> 00:04:56,000 I'll be talking about more of this when we go ahead. 95 00:04:56,000 --> 00:04:57,000 Okay. 96 00:04:57,000 --> 00:05:00,000 Now, quickly, uh, let's go ahead and do this. 97 00:05:00,000 --> 00:05:02,000 And, uh, here I have already imported all the libraries. 98 00:05:02,000 --> 00:05:05,000 So first of all, we will go ahead and set up our Streamlit app. 99 00:05:05,000 --> 00:05:06,000 Okay. 100 00:05:06,000 --> 00:05:08,000 I will write s t dot title. 101 00:05:08,000 --> 00:05:19,000 And here I'm going to basically say conversational conversational rag with PDF uploads and chat history. 102 00:05:19,000 --> 00:05:20,000 Okay. 103 00:05:20,000 --> 00:05:24,000 And chat history. 104 00:05:24,000 --> 00:05:26,000 So everything is basically going to get included over here. 105 00:05:26,000 --> 00:05:29,000 Then I will go ahead and write SD Dot write. 106 00:05:29,000 --> 00:05:32,000 And here let's go ahead and include this. 107 00:05:32,000 --> 00:05:38,000 I will say upload PDFs okay. 108 00:05:39,000 --> 00:05:41,000 And chat with the content okay. 109 00:05:42,000 --> 00:05:44,000 Chat with the content. 110 00:05:44,000 --> 00:05:44,000 Perfect. 111 00:05:45,000 --> 00:05:54,000 Now, uh, let me go ahead and input the grok grok API key okay. 112 00:05:54,000 --> 00:05:59,000 And here I'm just going to go ahead and create my API key which I will be writing s t dot text underscore 113 00:05:59,000 --> 00:06:00,000 input. 114 00:06:02,000 --> 00:06:09,000 And enter your grok API key. 115 00:06:10,000 --> 00:06:13,000 And here specifically I am going to do one thing. 116 00:06:13,000 --> 00:06:17,000 I will just go ahead and write my type as password. 117 00:06:19,000 --> 00:06:19,000 Okay. 118 00:06:20,000 --> 00:06:21,000 Type as password. 119 00:06:21,000 --> 00:06:23,000 This will basically be my API key over here. 120 00:06:24,000 --> 00:06:24,000 Let's see. 121 00:06:24,000 --> 00:06:27,000 This looks fine or not okay I forgot to give comma. 122 00:06:27,000 --> 00:06:30,000 Okay so this is for my API key. 123 00:06:30,000 --> 00:06:38,000 And as usual I will go ahead and see check if grok API key is. 124 00:06:41,000 --> 00:06:43,000 Is provided okay so I'll write. 125 00:06:43,000 --> 00:06:49,000 Hey if API underscore key is equal to or double colon I will just go ahead. 126 00:06:49,000 --> 00:06:54,000 And uh if the API key is basically given I will just go ahead and initialize my LLM model. 127 00:06:54,000 --> 00:06:54,000 Right. 128 00:06:54,000 --> 00:06:56,000 So here I will go ahead and say chat. 129 00:06:56,000 --> 00:07:04,000 Grok the grok API key that I really want to give over here will be initialized to my API key. 130 00:07:06,000 --> 00:07:08,000 API key okay. 131 00:07:08,000 --> 00:07:11,000 The second parameter that I want to give is model underscore name. 132 00:07:12,000 --> 00:07:20,000 Now with respect to this particular model, uh, I will be using our Gamma2 model, uh, for the same 133 00:07:20,000 --> 00:07:20,000 purpose. 134 00:07:20,000 --> 00:07:25,000 So let's go ahead and go to my grok cloud. 135 00:07:27,000 --> 00:07:27,000 Okay. 136 00:07:28,000 --> 00:07:33,000 Grok cloud or I will just go ahead and click on this okay. 137 00:07:33,000 --> 00:07:35,000 I can use any of this model. 138 00:07:35,000 --> 00:07:35,000 It is up to you. 139 00:07:35,000 --> 00:07:38,000 So I will use gamma two nine be it okay. 140 00:07:38,000 --> 00:07:39,000 So gamma two. 141 00:07:41,000 --> 00:07:42,000 Gamma. 142 00:07:45,000 --> 00:07:46,000 Two. 143 00:07:48,000 --> 00:07:48,000 Nine. 144 00:07:48,000 --> 00:07:51,000 Be it. 145 00:07:54,000 --> 00:07:56,000 Okay, I'll use this specific model. 146 00:07:57,000 --> 00:07:57,000 Perfect. 147 00:07:58,000 --> 00:08:00,000 This will basically be my model name. 148 00:08:00,000 --> 00:08:05,000 So till here, everything looks absolutely fine and absolutely good itself. 149 00:08:05,000 --> 00:08:11,000 Uh, I think we should proceed, uh, in developing this particular entire application. 150 00:08:11,000 --> 00:08:11,000 So. 151 00:08:11,000 --> 00:08:17,000 So here, uh, as soon as I make sure that I initialize my LM model, I also make to probably create 152 00:08:17,000 --> 00:08:18,000 my session ID. 153 00:08:18,000 --> 00:08:23,000 So I will go ahead and write session ID is equal to SD dot text underscore input. 154 00:08:24,000 --> 00:08:30,000 And here I'm going to basically give my session ID. 155 00:08:31,000 --> 00:08:31,000 Okay. 156 00:08:31,000 --> 00:08:34,000 And here I'm going to basically go ahead and write. 157 00:08:34,000 --> 00:08:37,000 My value is equal to some session okay. 158 00:08:37,000 --> 00:08:42,000 You can put anything as you want but I will just keep it to default okay. 159 00:08:42,000 --> 00:08:46,000 Default default underscore session okay. 160 00:08:46,000 --> 00:08:49,000 So this will basically be my session ID okay. 161 00:08:49,000 --> 00:08:53,000 Now this is for my chat interface. 162 00:08:53,000 --> 00:08:53,000 Right. 163 00:08:53,000 --> 00:08:56,000 So this is my chat interface. 164 00:08:57,000 --> 00:08:59,000 What we are specifically using. 165 00:08:59,000 --> 00:08:59,000 Okay. 166 00:08:59,000 --> 00:09:06,000 In the next uh, what we are basically going to do is that I'm also going to state fully manage the 167 00:09:06,000 --> 00:09:07,000 chat history. 168 00:09:08,000 --> 00:09:11,000 I have to manage it with the help of session states. 169 00:09:11,000 --> 00:09:11,000 Okay. 170 00:09:14,000 --> 00:09:18,000 Now here I will say if I'll create a variable called a store. 171 00:09:18,000 --> 00:09:19,000 Okay. 172 00:09:19,000 --> 00:09:27,000 If store not in session, state not in session, you'll be understanding where we will specifically 173 00:09:27,000 --> 00:09:28,000 be using the store. 174 00:09:28,000 --> 00:09:33,000 If you remember in the Get session history in conversational Q&A when we discussed about. 175 00:09:33,000 --> 00:09:33,000 Right. 176 00:09:33,000 --> 00:09:40,000 So if I go ahead and write not in store dot session underscore state, I am just going to go ahead and 177 00:09:40,000 --> 00:09:42,000 write s t dot session. 178 00:09:43,000 --> 00:09:46,000 Session underscore state dot store. 179 00:09:46,000 --> 00:09:49,000 We'll go ahead and initialize to this empty dictionary. 180 00:09:49,000 --> 00:09:51,000 And inside this only we'll be having all the key value pairs. 181 00:09:51,000 --> 00:09:52,000 Right. 182 00:09:52,000 --> 00:09:57,000 Uh like of the messages of the I message content, human message content, all those things. 183 00:09:57,000 --> 00:10:00,000 Now I'll go ahead and create my uploaded files. 184 00:10:00,000 --> 00:10:03,000 So this will be upload button right. 185 00:10:03,000 --> 00:10:06,000 I will say hey SD dot file uploader okay. 186 00:10:06,000 --> 00:10:17,000 And I'll say hey choose choose a PDF file okay a PDF file file okay. 187 00:10:17,000 --> 00:10:21,000 And here I will say hey type is equal to PDF okay. 188 00:10:22,000 --> 00:10:24,000 And I will also say accept multiple files. 189 00:10:24,000 --> 00:10:28,000 I'll keep it to false because I just want to try with a single PDF. 190 00:10:28,000 --> 00:10:31,000 But if you want to try with multiple PDF, just go ahead and give this as true. 191 00:10:31,000 --> 00:10:33,000 Okay, perfect. 192 00:10:33,000 --> 00:10:36,000 Uh, so everything looks good till here. 193 00:10:36,000 --> 00:10:40,000 So I will go ahead and process my uploaded files. 194 00:10:40,000 --> 00:10:40,000 Okay. 195 00:10:40,000 --> 00:10:48,000 So here what I will do is that I will go ahead and write if uploaded underscore files. 196 00:10:48,000 --> 00:10:53,000 If the file is basically uploaded I will go ahead and create a document okay. 197 00:10:53,000 --> 00:10:57,000 This will basically be my document because in PDF I may have multiple files. 198 00:10:57,000 --> 00:10:58,000 Right? 199 00:10:58,000 --> 00:10:59,000 I may have multiple pages. 200 00:10:59,000 --> 00:11:04,000 So here I'm basically going to write documents and I'll say hey for uploaded files. 201 00:11:04,000 --> 00:11:12,000 So or uploaded underscore files in sorry file in uploaded files. 202 00:11:12,000 --> 00:11:18,000 So let's say even though I make this as true, let's make this as true I can have an option to probably 203 00:11:18,000 --> 00:11:20,000 select multiple files also. 204 00:11:20,000 --> 00:11:25,000 Okay, it is up to you whether you want to go with single or multiple, but I will write a generic code 205 00:11:25,000 --> 00:11:29,000 where I'll say that hey, for each and every file that is basically uploaded, we'll take that file. 206 00:11:29,000 --> 00:11:36,000 And first of all, we'll store that file file in our local folder because see when we upload it right. 207 00:11:36,000 --> 00:11:38,000 There is no specific, uh URL. 208 00:11:38,000 --> 00:11:39,000 Everything will be in the form of memory. 209 00:11:39,000 --> 00:11:45,000 So it is necessary that we create some temporary, uh, pdf file in our local. 210 00:11:45,000 --> 00:11:45,000 Right. 211 00:11:45,000 --> 00:11:51,000 So here I will go ahead and say temp uh, instead of writing temp temp temp pdf. 212 00:11:51,000 --> 00:11:53,000 Okay I'll go and write temp PDF. 213 00:11:53,000 --> 00:11:59,000 And this is a F string that I will be using and I'll upload I'll create this with the name of temp dot 214 00:11:59,000 --> 00:12:00,000 pdf. 215 00:12:00,000 --> 00:12:00,000 Okay. 216 00:12:00,000 --> 00:12:08,000 Now what we will do we will open this PDF I'll say with open temp PDF comma. 217 00:12:08,000 --> 00:12:11,000 I'll send right bytes mode okay. 218 00:12:11,000 --> 00:12:13,000 I'll open this PDF and I'll write as file. 219 00:12:13,000 --> 00:12:17,000 So once I open this I will go ahead and write file dot write. 220 00:12:17,000 --> 00:12:25,000 And here I'm just going to go ahead and create my uploaded file dot get value. 221 00:12:27,000 --> 00:12:28,000 Okay. 222 00:12:28,000 --> 00:12:35,000 So finally if you go ahead and see file underscore name is equal to uploaded file dot name okay. 223 00:12:35,000 --> 00:12:39,000 Whatever name will be there I will also take give that particular file name if I want okay. 224 00:12:39,000 --> 00:12:41,000 All the information from that. 225 00:12:41,000 --> 00:12:45,000 So in short what I'm doing is that I'm opening that file that I have uploaded, I'm reading all the 226 00:12:45,000 --> 00:12:47,000 values and I'm also reading the name. 227 00:12:47,000 --> 00:12:47,000 Okay. 228 00:12:48,000 --> 00:12:53,000 Now finally, in order to open this file and get that content, what I will do, I'll create a loader 229 00:12:53,000 --> 00:12:59,000 and I'll write, hey, let's go ahead and use this py pdf loader which will be useful in call loading 230 00:12:59,000 --> 00:13:00,000 your files. 231 00:13:00,000 --> 00:13:03,000 So here I'm going to write temp dir okay. 232 00:13:03,000 --> 00:13:05,000 Oh sorry temp PDF. 233 00:13:05,000 --> 00:13:06,000 It should not be Dir. 234 00:13:06,000 --> 00:13:07,000 It should be PDF. 235 00:13:07,000 --> 00:13:10,000 Whatever PDF file is there we are going to load. 236 00:13:10,000 --> 00:13:14,000 And then I have docs where I'll specifically write loader dot load. 237 00:13:14,000 --> 00:13:15,000 Okay. 238 00:13:16,000 --> 00:13:18,000 Uh so all these things is specifically there. 239 00:13:18,000 --> 00:13:20,000 Now, what I will do, I will go ahead and write documents. 240 00:13:20,000 --> 00:13:22,000 Dot extend. 241 00:13:22,000 --> 00:13:22,000 Okay. 242 00:13:22,000 --> 00:13:26,000 And here I'm going to go ahead and create my docs okay. 243 00:13:26,000 --> 00:13:31,000 So I'm saying hey let's go ahead and append all the docs over here right now. 244 00:13:31,000 --> 00:13:36,000 Uh, what we are basically going to do is that, uh, we are going to split and create embeddings for 245 00:13:36,000 --> 00:13:37,000 the documents. 246 00:13:37,000 --> 00:13:39,000 So here I will just go and copy and paste it. 247 00:13:39,000 --> 00:13:45,000 So instead of using OpenAI embeddings I'm going to use embeddings, whatever embeddings we have discussed 248 00:13:45,000 --> 00:13:46,000 over here. 249 00:13:46,000 --> 00:13:48,000 So let me just go ahead. 250 00:13:48,000 --> 00:13:50,000 But I think this is common code right. 251 00:13:50,000 --> 00:13:52,000 Split and create embeddings for documents. 252 00:13:52,000 --> 00:13:55,000 First of all we'll use recursive character text later. 253 00:13:55,000 --> 00:13:58,000 I'm using a chunk size of 5000 chunk overlap of. 254 00:13:58,000 --> 00:14:00,000 Let's make it to 500 at least. 255 00:14:01,000 --> 00:14:01,000 Okay. 256 00:14:01,000 --> 00:14:04,000 Then you have this text splitter dot split documents of documents. 257 00:14:04,000 --> 00:14:05,000 Then I have the splits. 258 00:14:05,000 --> 00:14:10,000 Then I'm using chroma to probably split all the documents and store it in my embedding vectors. 259 00:14:10,000 --> 00:14:10,000 Right. 260 00:14:10,000 --> 00:14:13,000 And finally we are converting this into a retriever. 261 00:14:13,000 --> 00:14:13,000 Okay. 262 00:14:13,000 --> 00:14:20,000 Now, uh, what I am actually going to do after this is that I am going to go ahead and I will just 263 00:14:20,000 --> 00:14:22,000 go ahead and create my new prompt. 264 00:14:22,000 --> 00:14:23,000 Okay. 265 00:14:23,000 --> 00:14:25,000 Now this new prompt will have context. 266 00:14:25,000 --> 00:14:31,000 So I'll say contextualize underscore cue system prompt. 267 00:14:31,000 --> 00:14:34,000 This is basically my system prompt that I really want to create. 268 00:14:34,000 --> 00:14:35,000 Okay. 269 00:14:35,000 --> 00:14:38,000 And here I will be writing this one. 270 00:14:38,000 --> 00:14:49,000 I'll say hey, given a chat history and the latest user question okay. 271 00:14:50,000 --> 00:14:51,000 Question okay. 272 00:14:51,000 --> 00:14:54,000 Along with this I will just put some information over here. 273 00:14:56,000 --> 00:14:58,000 I can also write all this like this. 274 00:14:58,000 --> 00:14:59,000 Right. 275 00:14:59,000 --> 00:15:06,000 So here you can see I'm saying hey, given a chat history and the latest user question which might reference 276 00:15:06,000 --> 00:15:11,000 context in the chat history formula to standalone application question which can be understood without 277 00:15:11,000 --> 00:15:17,000 the chat history, do not answer the question, just reformulate it if needed and otherwise return this. 278 00:15:17,000 --> 00:15:21,000 So this is a kind of a system problem that we are specifically giving to our LM model. 279 00:15:21,000 --> 00:15:25,000 And for this we will be creating a prompt template. 280 00:15:25,000 --> 00:15:28,000 So for this I will just go ahead and like like this. 281 00:15:28,000 --> 00:15:33,000 So I'll say contextualize Q prompt where I'm using chat prompt template dot from underscore message 282 00:15:33,000 --> 00:15:37,000 where I'll give my system which will be assigning to this particular system prompt. 283 00:15:38,000 --> 00:15:41,000 So this is the set of instruction that I'm giving to my LM model. 284 00:15:41,000 --> 00:15:45,000 Then I am going to create a placeholder which has this chart underscore history. 285 00:15:45,000 --> 00:15:48,000 So in my session I will go ahead and create this chart underscore history. 286 00:15:48,000 --> 00:15:50,000 Just in some time I will go ahead and create it. 287 00:15:50,000 --> 00:15:51,000 Don't worry. 288 00:15:51,000 --> 00:15:55,000 Then I have this human which respect to any input that I specifically give. 289 00:15:55,000 --> 00:15:55,000 Okay. 290 00:15:55,000 --> 00:16:00,000 Now after this what I will do, I will since I have created this placeholder for the chart history, 291 00:16:00,000 --> 00:16:04,000 I can go ahead and create my history underscore aware underscore retriever. 292 00:16:04,000 --> 00:16:05,000 Right. 293 00:16:05,000 --> 00:16:08,000 So that is where I am going to create my retriever. 294 00:16:08,000 --> 00:16:08,000 Right. 295 00:16:08,000 --> 00:16:10,000 So and this retriever will be nothing. 296 00:16:10,000 --> 00:16:14,000 But uh it will be a retriever with memory. 297 00:16:14,000 --> 00:16:14,000 Okay. 298 00:16:14,000 --> 00:16:16,000 So here I'm going to use this. 299 00:16:16,000 --> 00:16:21,000 I'll go ahead and call this create history retriever uh, aware retriever here I'm going to give three 300 00:16:21,000 --> 00:16:29,000 information LM retriever along with this whatever prompt that I have actually designed over here, this 301 00:16:29,000 --> 00:16:31,000 prompt and this chat history will be referring to that. 302 00:16:31,000 --> 00:16:38,000 So this basically means this retriever has a powerful feature of this particular chat history where 303 00:16:38,000 --> 00:16:40,000 it will be able to store all the information. 304 00:16:40,000 --> 00:16:41,000 Okay. 305 00:16:41,000 --> 00:16:45,000 So yes, uh, this was about this chat aware history retriever okay. 306 00:16:46,000 --> 00:16:52,000 Now finally what I will do, I will go ahead and create my answer question prompt okay. 307 00:16:52,000 --> 00:16:54,000 Answer question prompt. 308 00:16:54,000 --> 00:16:59,000 And let me just go ahead and write this particular prompt because this prompt will be based on the context 309 00:16:59,000 --> 00:17:00,000 okay. 310 00:17:00,000 --> 00:17:03,000 So here I have created a system prompt okay. 311 00:17:03,000 --> 00:17:03,000 Quickly. 312 00:17:04,000 --> 00:17:07,000 And this system prompt says that hey you are an assistant for the question answer task. 313 00:17:07,000 --> 00:17:11,000 Use the following pieces of retrieve context to answer the question. 314 00:17:11,000 --> 00:17:12,000 If you don't know the answer, say that you don't know. 315 00:17:12,000 --> 00:17:16,000 Use three sentence maximum and keep the answer concise. 316 00:17:16,000 --> 00:17:16,000 New line. 317 00:17:16,000 --> 00:17:19,000 And here we are passing this particular context. 318 00:17:19,000 --> 00:17:21,000 And this context will be replaced by the stuff document chain. 319 00:17:21,000 --> 00:17:21,000 Right. 320 00:17:22,000 --> 00:17:24,000 So this is my system prompt. 321 00:17:24,000 --> 00:17:28,000 Now I will go ahead and create this particular prompt for this kind of question. 322 00:17:28,000 --> 00:17:33,000 I will say queue a prompt and from message I'll be using this again. 323 00:17:33,000 --> 00:17:34,000 I'm using this placeholder. 324 00:17:34,000 --> 00:17:37,000 This and message that I'm actually giving is input okay. 325 00:17:37,000 --> 00:17:41,000 So uh, almost uh, this is specifically done. 326 00:17:41,000 --> 00:17:42,000 Okay. 327 00:17:42,000 --> 00:17:43,000 And it is up to you. 328 00:17:43,000 --> 00:17:45,000 Like how you really want to create this and all. 329 00:17:45,000 --> 00:17:46,000 Okay. 330 00:17:46,000 --> 00:17:51,000 But at the end of the day, what I will be doing is that, uh, I will make sure that I will integrate 331 00:17:51,000 --> 00:17:54,000 this particular prompt with this history aware retrieval. 332 00:17:54,000 --> 00:17:54,000 Okay. 333 00:17:55,000 --> 00:18:00,000 So now in the next step, what I will say, I will say, hey, let's go ahead and create our question 334 00:18:00,000 --> 00:18:02,000 answer chain okay. 335 00:18:02,000 --> 00:18:05,000 And I'll say create stop document chain. 336 00:18:05,000 --> 00:18:08,000 So here I'm going to use this here I'm going to give two parameters. 337 00:18:08,000 --> 00:18:12,000 One is uh lm and one is QA prompt. 338 00:18:12,000 --> 00:18:13,000 Right. 339 00:18:13,000 --> 00:18:17,000 Because this creates a document chain will be replacing, uh, the entire documents over here in the 340 00:18:17,000 --> 00:18:18,000 context. 341 00:18:18,000 --> 00:18:21,000 So the two information is basically given LM and QA prompt. 342 00:18:21,000 --> 00:18:21,000 Okay. 343 00:18:22,000 --> 00:18:27,000 Then finally I'll go ahead and create my wrap chain which will be nothing but create a retrieval chain 344 00:18:27,000 --> 00:18:28,000 okay. 345 00:18:28,000 --> 00:18:32,000 And inside this retrieval chain I will be using a history aware retrieval. 346 00:18:32,000 --> 00:18:35,000 And then you'll also be using this question answer chain. 347 00:18:35,000 --> 00:18:40,000 So in short I have combined this history aware retrieval with the recent QA prompt that I've actually 348 00:18:40,000 --> 00:18:41,000 used. 349 00:18:41,000 --> 00:18:41,000 Okay. 350 00:18:41,000 --> 00:18:45,000 So this basically gets my rag underscore chain. 351 00:18:47,000 --> 00:18:52,000 So guys now let's go ahead and uh you know we have already created this rack chain. 352 00:18:52,000 --> 00:18:59,000 But as you all far remember, you know in a conversational Q&A history or chat bot module, we had also 353 00:18:59,000 --> 00:19:02,000 discussed something about creating a get session history function. 354 00:19:02,000 --> 00:19:10,000 So here we are just going to go ahead and write get session history, get session history. 355 00:19:10,000 --> 00:19:12,000 And we will create the specific function. 356 00:19:12,000 --> 00:19:13,000 And the function. 357 00:19:13,000 --> 00:19:21,000 Main aim will be that whenever we give any session ID okay it should return back our base chat history. 358 00:19:21,000 --> 00:19:22,000 Message history. 359 00:19:22,000 --> 00:19:22,000 Right. 360 00:19:22,000 --> 00:19:24,000 So here what we'll do. 361 00:19:24,000 --> 00:19:29,000 We'll write if s t dot session or sorry. 362 00:19:29,000 --> 00:19:37,000 Or I'll just go ahead and write if session underscore id not in s t dot session underscore id state 363 00:19:37,000 --> 00:19:38,000 dot store. 364 00:19:38,000 --> 00:19:41,000 Because inside this only we'll go ahead and check it. 365 00:19:41,000 --> 00:19:46,000 And this is where we are going to probably put all our session ID automatically get stored inside this. 366 00:19:46,000 --> 00:19:47,000 Right? 367 00:19:47,000 --> 00:19:54,000 So I'll say PT dot session state dot store, dot store. 368 00:19:54,000 --> 00:20:01,000 And here I'm going to give my session ID which will be equal to our chat message history and this chat 369 00:20:01,000 --> 00:20:02,000 message history. 370 00:20:02,000 --> 00:20:02,000 What it will do. 371 00:20:02,000 --> 00:20:07,000 All the conversation that is basically happening with this particular session ID, it will get stored 372 00:20:07,000 --> 00:20:12,000 over here, and from here it will go to this particular, uh, you know, the session ID itself. 373 00:20:12,000 --> 00:20:13,000 It will get saved over there. 374 00:20:13,000 --> 00:20:18,000 Then finally we go ahead and return this s t dot. 375 00:20:20,000 --> 00:20:24,000 Session state dot store. 376 00:20:25,000 --> 00:20:29,000 And here I'm going to basically use my session underscore ID. 377 00:20:29,000 --> 00:20:30,000 Perfect. 378 00:20:30,000 --> 00:20:34,000 So this is the same store that we created on top right. 379 00:20:34,000 --> 00:20:37,000 So the same right which is a key value pairs. 380 00:20:37,000 --> 00:20:38,000 So yes. 381 00:20:38,000 --> 00:20:41,000 Uh, once this is done this is the function that we have actually created. 382 00:20:41,000 --> 00:20:46,000 Now I'm going to go ahead and create my conversational rag chain. 383 00:20:47,000 --> 00:20:53,000 So I'm going to write conversational rag underscore chain which will be equal to runnable message history. 384 00:20:53,000 --> 00:20:57,000 And here I will be giving my information like rack chain. 385 00:20:57,000 --> 00:21:01,000 Then I have my function which is called as get session history. 386 00:21:01,000 --> 00:21:05,000 The third thing that I'm actually going to give is my input key message or input message key. 387 00:21:05,000 --> 00:21:07,000 As you know what is the input that I'm giving. 388 00:21:07,000 --> 00:21:10,000 So I'm giving basically input right. 389 00:21:10,000 --> 00:21:15,000 And along with this you should also know which is my history message key. 390 00:21:15,000 --> 00:21:17,000 History message key is nothing but chat history. 391 00:21:17,000 --> 00:21:23,000 So automatically this basically comes to know that where it needs to save it, uh, all the chat session. 392 00:21:23,000 --> 00:21:27,000 And finally your output message key which will be assigned to your answer. 393 00:21:28,000 --> 00:21:28,000 Right. 394 00:21:28,000 --> 00:21:32,000 Your output message key means whichever response you will be getting that will be stored in the specific 395 00:21:32,000 --> 00:21:33,000 output. 396 00:21:33,000 --> 00:21:34,000 Okay. 397 00:21:34,000 --> 00:21:36,000 So till here, everything looks good. 398 00:21:36,000 --> 00:21:41,000 Uh, now let's start and write our code as soon as we get our input. 399 00:21:41,000 --> 00:21:44,000 Uh, that is the next thing that we really need to write. 400 00:21:44,000 --> 00:21:47,000 But I think till here, everything looks good. 401 00:21:47,000 --> 00:21:50,000 So this is my uploaded file. 402 00:21:51,000 --> 00:21:55,000 Uh, with respect to the uploaded file, we are able to see each and every thing. 403 00:21:55,000 --> 00:22:00,000 This is good enough, and we have our system prompt everything as such. 404 00:22:00,000 --> 00:22:01,000 Okay. 405 00:22:01,000 --> 00:22:05,000 I think one condition I may have missed somewhere. 406 00:22:05,000 --> 00:22:05,000 Okay. 407 00:22:06,000 --> 00:22:08,000 Uh, one, let me just go ahead and verify this here. 408 00:22:08,000 --> 00:22:11,000 I've given my API key here. 409 00:22:11,000 --> 00:22:12,000 I have checked the condition. 410 00:22:12,000 --> 00:22:16,000 Here is my uploaded file with respect to uploaded file. 411 00:22:16,000 --> 00:22:18,000 I have my text splitter then. 412 00:22:18,000 --> 00:22:19,000 Okay. 413 00:22:19,000 --> 00:22:19,000 Yeah. 414 00:22:19,000 --> 00:22:25,000 So this context splitter needs to be created all inside this if condition itself okay. 415 00:22:25,000 --> 00:22:30,000 So this is one of the changes that I will do I'll go ahead and press shift tab okay. 416 00:22:30,000 --> 00:22:33,000 So please make sure that you do these changes. 417 00:22:33,000 --> 00:22:35,000 It needs to be inside this right if condition. 418 00:22:35,000 --> 00:22:40,000 Because once I need to have the documents over here right then only I should be able to go ahead and 419 00:22:40,000 --> 00:22:41,000 create all this chain. 420 00:22:41,000 --> 00:22:43,000 If I do not have the documents, then it will be of no use. 421 00:22:43,000 --> 00:22:44,000 Okay? 422 00:22:44,000 --> 00:22:46,000 So please make sure that you make these changes. 423 00:22:46,000 --> 00:22:46,000 Okay? 424 00:22:46,000 --> 00:22:49,000 Just go over here and press select from here. 425 00:22:49,000 --> 00:22:51,000 This this entire the from the text editor code. 426 00:22:51,000 --> 00:22:54,000 It should be inside this if if block okay. 427 00:22:54,000 --> 00:22:55,000 Perfect. 428 00:22:55,000 --> 00:22:55,000 Until here. 429 00:22:55,000 --> 00:22:56,000 Everything looks good enough. 430 00:22:56,000 --> 00:23:02,000 And, uh, now, uh, the next thing that I am actually going to do is that after creating this conversational 431 00:23:02,000 --> 00:23:05,000 Q&A chatbot is that let's go ahead and create my user input. 432 00:23:05,000 --> 00:23:08,000 So I will say, hey, this is my user input. 433 00:23:08,000 --> 00:23:10,000 Whatever question I'm looking for. 434 00:23:10,000 --> 00:23:16,000 And I'm saying, hey, if user underscore input, I'll say create my session history. 435 00:23:16,000 --> 00:23:21,000 When my session history will be nothing but get session history. 436 00:23:21,000 --> 00:23:24,000 And here I'm going to give my session ID right. 437 00:23:24,000 --> 00:23:27,000 Whatever session ID I have the default session. 438 00:23:27,000 --> 00:23:29,000 That is what I'm actually going to give over here. 439 00:23:29,000 --> 00:23:36,000 And now I can probably go ahead and call my conversational rag Jane this one. 440 00:23:36,000 --> 00:23:36,000 Right. 441 00:23:36,000 --> 00:23:41,000 So I'll go inside this I will right dot invoke here input will be one key. 442 00:23:41,000 --> 00:23:47,000 And this configuration will be with for my session ID so that automatically once I use this config right 443 00:23:47,000 --> 00:23:55,000 inside my conversation Q&A, it knows right where I need to save all my session information, like all 444 00:23:55,000 --> 00:23:57,000 the chat right inside this chat history variable. 445 00:23:57,000 --> 00:23:58,000 This will be my input. 446 00:23:58,000 --> 00:24:03,000 This will be my, uh, chat history where all the information will be saved based on the session ID, 447 00:24:03,000 --> 00:24:03,000 right? 448 00:24:03,000 --> 00:24:05,000 Since that is the reason we are using this. 449 00:24:05,000 --> 00:24:10,000 Uh, say get session history now, Now, along with this, uh, I will quickly go ahead and display 450 00:24:10,000 --> 00:24:11,000 this. 451 00:24:11,000 --> 00:24:13,000 All the three information, like what is the store? 452 00:24:14,000 --> 00:24:19,000 Right then, uh, if you see over here session history, it should be session history. 453 00:24:19,000 --> 00:24:19,000 Okay. 454 00:24:19,000 --> 00:24:21,000 I will be displaying this store. 455 00:24:21,000 --> 00:24:24,000 I'll be displaying the answer and I'll be displaying my message. 456 00:24:24,000 --> 00:24:24,000 Okay. 457 00:24:24,000 --> 00:24:30,000 And here, uh, instead of getting the response, I will use store dot success over here so that it 458 00:24:30,000 --> 00:24:31,000 shows in green color. 459 00:24:31,000 --> 00:24:32,000 Okay. 460 00:24:32,000 --> 00:24:39,000 so yes, this was it, uh, all the information that I've actually coded in front of you and if the 461 00:24:39,000 --> 00:24:41,000 API key is not given. 462 00:24:41,000 --> 00:24:41,000 Right. 463 00:24:41,000 --> 00:24:44,000 So let me just go back over here, okay. 464 00:24:44,000 --> 00:24:50,000 And let me right else if when the API key is not given, I will give a warning saying that hey, please 465 00:24:50,000 --> 00:24:51,000 enter your. 466 00:24:52,000 --> 00:24:57,000 Please enter the grok API key. 467 00:24:57,000 --> 00:24:59,000 Okay I will go ahead and execute this. 468 00:24:59,000 --> 00:25:02,000 Now let me just go ahead and open this quickly. 469 00:25:03,000 --> 00:25:07,000 Uh, let me go back to this specific folder. 470 00:25:07,000 --> 00:25:13,000 So here I will write CD dot dot and the folder over here that I'm actually using if you see. 471 00:25:13,000 --> 00:25:14,000 Right. 472 00:25:14,000 --> 00:25:16,000 Uh which is the folder. 473 00:25:16,000 --> 00:25:20,000 So folder is nothing but 4.1 rag Q&A. 474 00:25:20,000 --> 00:25:20,000 right? 475 00:25:20,000 --> 00:25:22,000 So I will go ahead and write 4.1. 476 00:25:22,000 --> 00:25:33,000 CD 4.1 rag Q&A conversation I will clear the screen and next go ahead and run this Streamlit run app.py. 477 00:25:33,000 --> 00:25:39,000 Before that I will open my env file and copy this key so that it will be useful. 478 00:25:39,000 --> 00:25:40,000 Let's go ahead and press this. 479 00:25:40,000 --> 00:25:41,000 Enter. 480 00:25:42,000 --> 00:25:46,000 And here you have this entire website okay. 481 00:25:46,000 --> 00:25:47,000 So it is loading. 482 00:25:47,000 --> 00:25:54,000 Let's see I think 1 or 2 errors usually comes when I'm we are basically writing the code from here itself. 483 00:25:54,000 --> 00:25:54,000 Right. 484 00:25:54,000 --> 00:25:59,000 So let's go back over here and see whether we are getting any warning. 485 00:25:59,000 --> 00:26:01,000 Resume and download is procured. 486 00:26:01,000 --> 00:26:02,000 Okay. 487 00:26:02,000 --> 00:26:03,000 So it is first for the first time. 488 00:26:03,000 --> 00:26:06,000 It is just going to take some time okay okay. 489 00:26:06,000 --> 00:26:09,000 Password is not a valid input text okay. 490 00:26:09,000 --> 00:26:11,000 Let's see where is password. 491 00:26:12,000 --> 00:26:14,000 I'll go ahead and search for. 492 00:26:14,000 --> 00:26:16,000 So it should be double s okay. 493 00:26:16,000 --> 00:26:17,000 Perfect. 494 00:26:17,000 --> 00:26:19,000 Let's now see. 495 00:26:19,000 --> 00:26:20,000 Always rerun. 496 00:26:20,000 --> 00:26:22,000 Now I think it should see for the first time. 497 00:26:22,000 --> 00:26:27,000 It will take time because it needs to download the hugging face embedding technique. 498 00:26:27,000 --> 00:26:27,000 Okay. 499 00:26:27,000 --> 00:26:30,000 So please uh please make sure that you be patience for that. 500 00:26:30,000 --> 00:26:31,000 Okay. 501 00:26:31,000 --> 00:26:33,000 Now I have my environment key. 502 00:26:33,000 --> 00:26:36,000 Uh, then I will go ahead and paste it over here. 503 00:26:36,000 --> 00:26:37,000 Press enter. 504 00:26:37,000 --> 00:26:38,000 Okay. 505 00:26:38,000 --> 00:26:39,000 I think it should work. 506 00:26:40,000 --> 00:26:45,000 It has got entered now here also it is going to probably take some amount of time. 507 00:26:45,000 --> 00:26:46,000 No, I don't think so. 508 00:26:46,000 --> 00:26:49,000 It will take time because only for the first time when the hugging face embedding is basically required. 509 00:26:49,000 --> 00:26:51,000 At that time it will take. 510 00:26:51,000 --> 00:26:55,000 So now uh, here you can see it is running the embedding technique of hugging face is basically getting 511 00:26:55,000 --> 00:26:59,000 applied, you know, and, uh, here you have it. 512 00:26:59,000 --> 00:26:59,000 Right. 513 00:26:59,000 --> 00:27:05,000 So here, now let me just go ahead and ask about what is Transformers. 514 00:27:05,000 --> 00:27:06,000 Okay. 515 00:27:06,000 --> 00:27:09,000 If you use OpenAI embedding it will be much more faster. 516 00:27:09,000 --> 00:27:12,000 So here uh, I have searched for what is transformer. 517 00:27:14,000 --> 00:27:20,000 And uh, here you can see input to chat prompt template missing chat history. 518 00:27:20,000 --> 00:27:23,000 Uh expected chat underscore history. 519 00:27:23,000 --> 00:27:24,000 Received input. 520 00:27:24,000 --> 00:27:25,000 Chat history. 521 00:27:25,000 --> 00:27:31,000 Okay, I think this is some bit of mistakes that has been happened over here. 522 00:27:31,000 --> 00:27:33,000 Uh, we will just go ahead and check it out. 523 00:27:33,000 --> 00:27:35,000 Let's see which line it is. 524 00:27:35,000 --> 00:27:40,000 Basically saying, uh, input chat history chat underscore history is over here. 525 00:27:40,000 --> 00:27:43,000 Here chat underscore history is basically over here. 526 00:27:43,000 --> 00:27:50,000 and uh, down what I've used here also chart underscore history is there uh here okay. 527 00:27:50,000 --> 00:27:52,000 So here I have to use chart underscore history. 528 00:27:53,000 --> 00:27:54,000 So it's fine. 529 00:27:54,000 --> 00:27:56,000 Some smaller mistakes will happen. 530 00:27:56,000 --> 00:27:59,000 Now it is just going to give me the output okay. 531 00:27:59,000 --> 00:28:03,000 So I have to make sure that all these things are correct. 532 00:28:03,000 --> 00:28:03,000 Right. 533 00:28:04,000 --> 00:28:05,000 So one more error. 534 00:28:05,000 --> 00:28:09,000 Um response of answer. 535 00:28:09,000 --> 00:28:09,000 Okay. 536 00:28:09,000 --> 00:28:10,000 Start. 537 00:28:10,000 --> 00:28:11,000 Success. 538 00:28:11,000 --> 00:28:12,000 Assistant. 539 00:28:12,000 --> 00:28:13,000 Okay, I've given two parameters over here. 540 00:28:13,000 --> 00:28:14,000 Right. 541 00:28:14,000 --> 00:28:17,000 Take two positional arguments, but three is given. 542 00:28:17,000 --> 00:28:18,000 So let me do one thing. 543 00:28:18,000 --> 00:28:22,000 Instead of writing success, I'll just go ahead and write SD dot write for now. 544 00:28:22,000 --> 00:28:23,000 Okay. 545 00:28:23,000 --> 00:28:25,000 Now it should basically give me the output. 546 00:28:25,000 --> 00:28:29,000 And here you can see in the default session it is basically getting displayed. 547 00:28:29,000 --> 00:28:31,000 Uh transformers are type of neural network. 548 00:28:31,000 --> 00:28:33,000 Uh they use self-attention to weigh the importance. 549 00:28:33,000 --> 00:28:35,000 And all this information is basically there. 550 00:28:35,000 --> 00:28:36,000 Okay. 551 00:28:36,000 --> 00:28:38,000 Um, fine. 552 00:28:38,000 --> 00:28:41,000 Let's go ahead and ask about this. 553 00:28:41,000 --> 00:28:43,000 What is attention? 554 00:28:44,000 --> 00:28:55,000 Uh, or I'll just say provide a detailed summary of Transformers with attention is all you need. 555 00:28:57,000 --> 00:28:59,000 I'll go ahead and search for this. 556 00:28:59,000 --> 00:29:02,000 Let's see whether I'll be able to get the answer or not. 557 00:29:09,000 --> 00:29:10,000 So here is my entire output. 558 00:29:10,000 --> 00:29:15,000 This paper attention is all you need to introduce the transformer architecture with all the information 559 00:29:15,000 --> 00:29:17,000 you can see over here and in the chat history. 560 00:29:17,000 --> 00:29:18,000 Also you are able to see this. 561 00:29:18,000 --> 00:29:25,000 You can also get the detailed summary I can say provider, detail summary, provider, detail summary, 562 00:29:26,000 --> 00:29:32,000 provider detailed summary of the conversation we had. 563 00:29:32,000 --> 00:29:33,000 Okay. 564 00:29:33,000 --> 00:29:36,000 So here you can see the output. 565 00:29:37,000 --> 00:29:39,000 I hope you're enjoying this videos right. 566 00:29:39,000 --> 00:29:43,000 Because at the end of the day I know I'm creating this amazing modules. 567 00:29:43,000 --> 00:29:45,000 This can be integrated wherever you want, right? 568 00:29:45,000 --> 00:29:49,000 So here you can see once your conversation started, you asked me to ask about Transformers. 569 00:29:49,000 --> 00:29:51,000 I provided a concise explanation. 570 00:29:51,000 --> 00:29:56,000 I also outlined the advantages of transformer in emphasizing their palatability ability to handle. 571 00:29:56,000 --> 00:29:59,000 You can upload any PDF and you can start working, right. 572 00:29:59,000 --> 00:30:01,000 So I hope, uh, you understood this particular video. 573 00:30:01,000 --> 00:30:04,000 I hope you were able to understand the entire implementation. 574 00:30:04,000 --> 00:30:06,000 So yeah, this was it for my side. 575 00:30:06,000 --> 00:30:07,000 I hope you liked this particular video. 576 00:30:07,000 --> 00:30:08,000 I'll see you in the next video. 577 00:30:08,000 --> 00:30:09,000 Thank you. 578 00:30:09,000 --> 00:30:09,000 Take care. 579 00:30:09,000 --> 00:30:09,000 Bye bye.