Build an AI/NLP TV Series Analysis System with Hugging Face, Chatbots, Spacy, Gradio, and Python

แชร์
ฝัง
  • เผยแพร่เมื่อ 21 พ.ย. 2024

ความคิดเห็น • 63

  • @meu22422
    @meu22422 2 หลายเดือนก่อน +41

    JazakAllah brother. First the football one and now this, You have no idea how many jobless people you are helping to get a job.

    • @Tothefutureand
      @Tothefutureand 2 หลายเดือนก่อน

      R u using that in your resume !?

    • @meu22422
      @meu22422 2 หลายเดือนก่อน +2

      @@Tothefutureand why not?

    • @Filthbatman
      @Filthbatman 2 หลายเดือนก่อน

      @@meu22422 they will know 🤣 almost half a million people watched it so think abt it

    • @emileeid6347
      @emileeid6347 2 หลายเดือนก่อน +1

      Is it wrong to use those on the Resume

    • @TheCodingMathematician
      @TheCodingMathematician หลายเดือนก่อน

      @@emileeid6347 no as long as you understand it they will definetly qask about it and stuff. maybe add something once you finish which can make it unique

  • @emir5146
    @emir5146 2 หลายเดือนก่อน +2

    When you see the Activate Windows phrase, you know that it being such a wonderful video.

  • @DiyorbekTirkashov-e6v
    @DiyorbekTirkashov-e6v 2 หลายเดือนก่อน +2

    This is by far the most insightful content I have come across. It has enhanced my practical understanding of NLP. I can't wait to see what you share next! Many thanks man, wishing you the best of luck.

    • @cutieeeeeeeee-d5j
      @cutieeeeeeeee-d5j หลายเดือนก่อน

      Hey please reply! Does this chatbot work for any series or shows? Like, anything? Or if I just feed it one show, it only works for that?

  • @rafaychaudry320
    @rafaychaudry320 2 หลายเดือนก่อน +1

    One of the best single CS related videos on TH-cam. I have watched all of your computer vision videos so far. Amazing work keep at it. Maybe develop an AI Golf coach program where they give feedback on your posture when taking a swing. That would be a massive project😅

  • @PhamDuc8504
    @PhamDuc8504 2 หลายเดือนก่อน +4

    Thanks for your contribution. Hope you will come up with more projects about chatbot, or maybe image generation. I am looking forward to it!! 👏👏👏

  • @Тима-щ2ю
    @Тима-щ2ю หลายเดือนก่อน

    Bro, the amount of skills i gained from your 3 projects is impressive! You are helping people to convert their ml/dl knowledge into something "material". teniss/football analyzer, chatbot. Keep it that way.

    • @cutieeeeeeeee-d5j
      @cutieeeeeeeee-d5j หลายเดือนก่อน

      Hey please reply! Does this chatbot work for any series or shows? Like, anything? Or if I just feed it one show, it only works for that?

  • @salwamh2942
    @salwamh2942 2 หลายเดือนก่อน

    im so glad i found your channel! its a treasure honestly!

  • @sai_ai-z1c
    @sai_ai-z1c 2 หลายเดือนก่อน

    This video offers a fascinating dive into TV series analysis with a smorgasbord of AI techniques. From scraping data with Scrapy to crafting character networks and building chatbots, it’s a treasure trove for anyone eager to explore AI's creative potential.

  • @shotbotop3790
    @shotbotop3790 2 หลายเดือนก่อน +5

    Here before this blows up!! bro you are a legend🙌

  • @hemantrawat1576
    @hemantrawat1576 หลายเดือนก่อน

    Great Bro your channel is one of my favourite i'm always waiting for your videos!!!! God Bless you

  • @richardnwachukwu7664
    @richardnwachukwu7664 หลายเดือนก่อน

    Brooo thank you for blessing my weekend!!! I am gonna work on it this weekend!

  • @emir5146
    @emir5146 2 หลายเดือนก่อน +5

    Allah bless you my friend. Your channel is my favorite among my 3 favorite channels on TH-cam. Andrej Karpathy, Umar Jamil and You. How can donate you on the youtube?

  • @anandsharma1005
    @anandsharma1005 2 หลายเดือนก่อน +2

    just saw this in my feed, gonna complete this now

  • @ajaykumarreddy8841
    @ajaykumarreddy8841 2 หลายเดือนก่อน +2

    I just found out what I am doing this weekend :D. Thanks a lot brother!

  • @Ali-ll4vn
    @Ali-ll4vn 2 หลายเดือนก่อน

    Great video mate! Thanks for sharing your knowledge!

  • @vaishnavnegi9640
    @vaishnavnegi9640 2 หลายเดือนก่อน

    Thanks bro! Your tutorials are really gr8 and detailed. I always enjoy following them. More of the Gen AI and LLM projects. This is in vogue.

  • @shreyojitdas9333
    @shreyojitdas9333 2 หลายเดือนก่อน

    u a life saver in this tough job market

  • @Arthur12137
    @Arthur12137 2 หลายเดือนก่อน

    wanna give you a BIG THANKS FOR SHARING !

  • @truthseeker457
    @truthseeker457 2 หลายเดือนก่อน +2

    Assalamu alaykum brother. I was watching your tutorial on football detection. Thank you. If I don’t understand something, I will ask you.

  • @ingridakaz9097
    @ingridakaz9097 2 หลายเดือนก่อน +1

    Amazing as always!

  • @utkarshkapil
    @utkarshkapil 2 หลายเดือนก่อน

    ALways love unique projects. Thanks!

  • @akj3344
    @akj3344 2 หลายเดือนก่อน

    Incredible video. WiIl do this on a chill weekend. Would you consider doing a recommender system project?

  • @Lostveroo
    @Lostveroo 2 หลายเดือนก่อน +1

    Could you create one for boxing like you did for tennis? I would love a tutorial to follow along with!

  • @TRISITGANGULY
    @TRISITGANGULY 2 หลายเดือนก่อน +1

    you are a saviour bro

  • @pmnt98
    @pmnt98 2 หลายเดือนก่อน

    ur the best
    thx so much for this insights..

  • @senpai9790
    @senpai9790 3 วันที่ผ่านมา

    Timestamps (Powered by Merlin AI)
    00:05 - Build an AI/NLP TV Series Analysis System
    02:07 - Setting up data sets for TV series analysis
    07:11 - Ninjutsu, Genjutsu, and physical attacks are three types of combat techniques in the TV series analysis system.
    09:45 - Building an HTML page with tags
    15:29 - Creating and styling containers using classes and divs
    18:10 - Using Scrapy library for web scraping
    22:57 - Iterating over web pages to extract data using Python
    25:41 - Extracting website links and crawling relevant information using Python
    31:04 - Extract classification information from a specific div
    34:03 - HTML aside section and its components
    39:40 - Extracting Jutsu description from webpage
    42:13 - Running Scrapy for data extraction
    47:25 - Neural network predicting defined classes like cat or dog
    49:44 - Training involves forward and backward propagation and batch processing for efficient model improvement.
    54:31 - Analyzing text using mathematical functions in AI
    56:55 - Introduction to LSTM and challenges with RNN
    1:02:02 - Introduction to attention mechanism and transformation model
    1:04:29 - Training GPT model using semi-supervised learning
    1:09:11 - Implementing zero-shot classifiers for flexible theme classification
    1:11:21 - Setting up AI/NLP environment and Transformers installation.
    1:16:36 - Setting up Torch for GPU usage and loading AI model
    1:19:17 - Theme classification using zero-shot classifier
    1:25:03 - Cleaning text data and preparing it for classification
    1:27:50 - Cleaning and processing subtitles for TV series analysis
    1:33:35 - Demonstrating how to run NLP on Pandas data frame
    1:36:32 - Executing batched AI/NLP TV series analysis
    1:42:15 - Wrangling and calculating mean scores for themes using numpy
    1:45:01 - Using the apply function to run a function on all rows for a specific column.
    1:50:39 - Initializing theme list and loading model in AI/NLP TV Series Analysis System
    1:53:37 - Building a function called get themes for running inference on the whole Pandas data frame
    1:59:08 - Utilizing os.path.join for folder paths in a Python project.
    2:01:32 - Creating AI/NLP TV Series Analysis System
    2:07:20 - Creating a function to analyze TV series themes and generate plot output.
    2:10:02 - Setting up and visualizing a theme classifier in Python.
    2:15:41 - Demonstrating theme classification and visualization using Gradio
    2:18:17 - Setting up a vertical chart with specific dimensions
    2:23:07 - Setting up AI/NLP project in GitHub
    2:25:47 - Setting up AI/NLP analysis system on Google Colab
    2:31:14 - Uploading data set and running the application
    2:33:55 - Setting up and managing data and tools for analysis
    2:38:52 - Utilizing Spacy for word-level named entity classification
    2:41:41 - Loading and processing data sets in the analysis system
    2:47:34 - Extracting entity names and labeling in sentences
    2:50:14 - Creating a character network for TV series analysis.
    2:55:47 - Implementing entity relationship extraction with sorting for standardization
    2:58:24 - Creating a character relationship analysis system using Python libraries
    3:03:30 - Setting up a TV series analysis system in Python
    3:06:10 - Setting up NER model and defining functions in Python notebook
    3:12:10 - Creating a character network visualization in Python
    3:14:55 - Creating HTML and exposing modules outside the folder
    3:20:21 - Setting up character network analysis with Gradio and Python
    3:22:59 - Configuring paths and resolving errors in the script
    3:28:08 - Setting up and running NLP analysis using Hugging Face
    3:30:43 - Learned about NLP and building a text classification model
    3:35:48 - Simplifying Jutsu into different types for analysis
    3:38:16 - Implement a cleaner function to remove HTML tags and format text
    3:43:37 - Troubleshooting label encoding and mapping errors
    3:46:28 - Transforming text data into numerical format
    3:51:25 - Convert a Pandas dataframe to a Hugging Face dataset for analysis.
    3:54:05 - Creating a cleaner function and a Jutsu classifier file
    3:59:36 - Loading data and preprocessing with Hugging Face models
    4:02:26 - Import pre-processing and labels, train test split, and dataset
    4:07:27 - Obtaining class weights for data analysis
    4:10:13 - Building AI/NLP TV Series Analysis System
    4:15:01 - Building a custom trainer for AI model with Hugging Face.
    4:17:44 - Neural Network Training Process Overview
    4:23:03 - Training code is complete, now focus on loading and inference
    4:25:41 - Creating a text classification function with post-processing
    4:31:23 - Creating a hugging face token for model deployment
    4:33:41 - Create a token and add it to .env file for security
    4:38:45 - Fixing and updating the dataset path for training purposes
    4:41:35 - Setting up the AI/NLP TV Series Analysis System
    4:46:28 - Keep your model files up to date to avoid errors during training.
    4:48:50 - Building a character chatbot module using pre-existing models in Gradio
    4:53:03 - Building a character chatbot with Hugging Face and Python
    4:55:37 - Using Gradio for character selection and text processing
    5:00:43 - Creating a flag to determine relevant rows
    5:03:03 - Selecting indices for conversation analysis
    5:08:13 - Data wrangling and model training with Hugging Face
    5:10:31 - Creating a character chatbot in Python using Hugging Face
    5:15:43 - Creating data set and loading model in Python
    5:18:13 - Defining the train function with parameters and settings
    5:23:22 - Configuring and initiating the model for sequence classification
    5:25:52 - Setting up required configurations for Hugging Face and TRL
    5:31:04 - Setting up model training and saving process
    5:33:39 - Loading and configuring the pre-trained model and tokenizer.
    5:39:02 - Setting up the pipeline for text generation using Transformers
    5:41:37 - Adding system prompt and messages
    5:47:02 - Creating a character chatbot with Gradio
    5:49:54 - Setting up character chatbot and examining edits in the code
    5:55:22 - Ensure correct spelling in quantization error config
    5:58:42 - Creating a character chatbot using state-of-the-art models
    6:04:03 - Built AI/NLP TV Series Analysis System with various tools
    6:06:21 - Learned skills are important for NLP and AI.

  • @cutieeeeeeeee-d5j
    @cutieeeeeeeee-d5j 26 วันที่ผ่านมา

    00:05 Building AI/NLP TV Series Analysis System with multiple components
    02:07 Setting up data folders and acquiring data sets
    07:11 Ninjutsu, physical attack, and genjutsu classification
    09:45 Introduction to basic HTML tags
    15:29 Using classes to style divs
    18:10 Introduction to web scraping using Scrapy library
    22:57 Iterating over next pages using anchor tags
    25:41 Creating a web scraping function to extract data from a specific website
    31:04 Extract classification and relevant data from div using Beautiful Soup
    34:03 Exploring HTML structure of aside section
    39:40 Using BeautifulSoup to extract and manipulate text data
    42:13 Scrapy used to crawl websites and save data in structured format
    47:25 Neural networks predict defined classes like cat or dog based on input data
    49:44 Discussing the training process of AI models
    54:31 Neural network architecture for text analysis
    56:55 Introduction to the limitations of RNN in AI models
    1:02:02 Introduction to attention mechanism and transition to Transformer model
    1:04:29 AI models are trained using semi-supervised learning
    1:09:11 Zero shot classifiers allow input of any arbitrary number of classes for classification.
    1:11:21 Setting up the environment and installing necessary libraries.
    1:16:36 Loading AI model with Torch and using CPU
    1:19:17 Theme classification with zero-shot classifier
    1:25:03 Cleaning and combining sentences in NLP text processing
    1:27:50 Creating a function to load and clean subtitles data
    1:33:35 Demonstration on dividing script into sentences and batching for NLP tokenization
    1:36:32 Running and analyzing the batched sentences output
    1:42:15 Wrangling the output and calculating mean using numpy library
    1:45:01 Perform theme inference on TV show scripts using Python libraries
    1:50:39 Initializing AI/NLP TV Series Analysis System
    1:53:37 Setting up the classifier and necessary functions
    1:59:08 Setting up a GUI using Gradio for machine learning applications
    2:01:32 Creating a TV series analysis system using Hugging Face and Spacy
    2:07:20 Building a system to analyze TV series using AI/NLP tools
    2:10:02 Exposing classifier outside the folder
    2:15:41 Building and using a theme classifier with Gradio
    2:18:17 Customizing display settings and running local tests
    2:23:07 Configuring the system to avoid exposing sensitive information
    2:25:47 Setting up AI/NLP TV Series Analysis System
    2:31:14 Uploading and running the TV series analysis system
    2:33:55 Setting up output classifier and running it with Gradio
    2:38:52 Installing Spacy for named entity recognition
    2:41:41 Loading and processing data using Python modules
    2:47:34 Extracting and processing entity names using NLP
    2:50:14 Creating a character network using Python libraries
    2:55:47 Sorting entities for standardized representation
    2:58:24 Create a character relationship analysis system
    3:03:30 Visualizing network analysis in Python
    3:06:10 Implementing named entity recognition using Spacy in Python
    3:12:10 Creating a character network graph.
    3:14:55 Setting up HTML source code for character network
    3:20:21 Implementing AI/NLP TV Series Analysis System
    3:22:59 Setting up script and output paths for AI/NLP TV series analysis
    3:28:08 Building an AI/NLP TV Series Analysis System
    3:30:43 Setting up a text classification model for Jutsu techniques
    3:35:48 Simplify Jutsus into different categories for analysis
    3:38:16 Creating a cleaner function for scraping and cleaning data
    3:43:37 Creating and correcting labels for TV series analysis
    3:46:28 Data preparation for AI model training
    3:51:25 Convert Pandas DataFrame to Hugging Face Dataset for TV series analysis.
    3:54:05 Setting up the Jutsu classifier with parameters
    3:59:36 Loading dataset and preprocessing data using Python Pandas and Hugging Face
    4:02:26 Implementing pre-processing using labels and train test split
    4:07:27 Understanding Class Weights Calculation
    4:10:13 Building a function to train and load models for TV series analysis.
    4:15:01 Training strategies and custom trainer in AI/NLP model building
    4:17:44 Implementing forward and backward propagation in AI model training
    4:23:03 GPU memory management in training and model loading
    4:25:41 Creating a text classification function in Python
    4:31:23 Define and implement the classify text function
    4:33:41 Create and manage a secret token for Hugging Face integration
    4:38:45 Troubleshooting and correcting errors in the training data and code customization
    4:41:35 Setting up a text classification system using Hugging Face and Spacy
    4:46:28 Removing duplicate models and troubleshooting training issues
    4:48:50 Implementing Character Chatbot Module using Gradio
    4:53:03 Setting up a notebook to demonstrate data wrangling for an AI character chatbot
    4:55:37 Creating a chatbot without actions using Gradio
    5:00:43 Using Python to count and filter words in a dataframe
    5:03:03 Setting up prompts for AI chatbot to imitate Naruto character responses
    5:08:13 Data wrangling and model training process with Hugging Face
    5:10:31 Creating a character chatbot using Hugging Face and Python
    5:15:43 Creating a function to load data for AI/NLP analysis
    5:18:13 Define and explain the train function with key parameters
    5:23:22 Setting up AI/NLP TV Series Analysis System
    5:25:52 Installing necessary components for Laura config
    5:31:04 Setting up the AI/NLP TV series analysis system
    5:33:39 Setting up base model and tokenizer for AI/NLP analysis
    5:39:02 Setting up the pipeline for AI model with Hugging Face
    5:41:37 Adding system prompts and messages
    5:47:02 Setting up character chatbot in Gradio UI
    5:49:54 Setting up the AI system with Gradio and Python
    5:55:22 Troubleshooting and fixing issues during AI model training
    5:58:42 Creating a character chatbot with state-of-the-art models
    6:04:03 Built AI/NLP TV series analysis system with key tools
    6:06:21 Key NLP and AI skills are crucial for success.
    Crafted by Merlin AI.

  • @OmegaR4y
    @OmegaR4y 2 หลายเดือนก่อน

    Thank you! 🙏

  • @georgejere2306
    @georgejere2306 2 หลายเดือนก่อน

    You are a genius🤖

  • @spidey_24_web
    @spidey_24_web 2 หลายเดือนก่อน +1

    Do you guys actually put this in resume or is it just for learning purpose? I'm honestly confused.

    • @emir5146
      @emir5146 2 หลายเดือนก่อน +2

      I think both of them.

  • @judevector
    @judevector 2 หลายเดือนก่อน

    Wtf I am just seeing. This is amazing ❤

  • @cutieeeeeeeee-d5j
    @cutieeeeeeeee-d5j หลายเดือนก่อน

    Where is the website design in the start?? i can't see it in the end

  • @oforibright5628
    @oforibright5628 2 หลายเดือนก่อน +1

    another way of getting the texts in subtitles is spiltting by ",," and taking the last element

  • @rgolanng
    @rgolanng 2 หลายเดือนก่อน +1

    Please make a go/rust deployment project

  • @SM-vz1ek
    @SM-vz1ek 2 หลายเดือนก่อน

    Thank ypu so much ! I'm looking for an genAI project with Azure. Any advice?

    • @codeinajiffy
      @codeinajiffy  2 หลายเดือนก่อน

      I haven't worked with genAI with Azure. But there were some new releases in GCP and AWS that I played around with. They make utilizing and deploying LLMs very easy. I'm not sure whether Azure has similar solutions or note yet.

  • @oforibright5628
    @oforibright5628 2 หลายเดือนก่อน

    the selectors in the video dont work. how can i resolve. anytimes i call the selectors from smw-columnlist the output is empty

    • @codeinajiffy
      @codeinajiffy  2 หลายเดือนก่อน

      This is the the for loop line. It's a little different from what you wrote. The code is in the Github repo and I just reran and it's working fine. Maybe just copy paste it from there.
      for href in response.css('.smw-columnlist-container')[0].css("a::attr(href)").extract():

  • @jahidmdhasan1021
    @jahidmdhasan1021 2 หลายเดือนก่อน

    I have a Laptop configuration of 8 GB RAM and Nvidia RTX 1650, I am planning to upgrade the RAM to 16, will I be able to run these LLMs locally or should I try using a less powerful model. kindly, provide your suggestion, please

    • @codeinajiffy
      @codeinajiffy  2 หลายเดือนก่อน +1

      I think, RTX 1650 have 4GB of GPU memory which is enough for 3 of the 4 models I use. Actually those 3 models will work with the 8 GB of RAM. but Llama is big and requires at least 12 GB of GPU memory. I show you how to run this on Google Colab so you don't have to run it locally.

    • @cutieeeeeeeee-d5j
      @cutieeeeeeeee-d5j หลายเดือนก่อน

      @@codeinajiffy Hey please reply! Does this chatbot work for any series or shows? Like, anything? Or if I just feed it one show, it only works for that?

  • @PANDURANG99
    @PANDURANG99 2 หลายเดือนก่อน +1

    Need process from development to deployment

  • @salwamh2942
    @salwamh2942 หลายเดือนก่อน

    everytime i try to get a graph in the gradio app with colab i get this error : " error : Connection errored out." , idk if the problem is with my code or what

    • @NIKHILSAI-y6j
      @NIKHILSAI-y6j 5 วันที่ผ่านมา

      i also have the same problem

  • @NIKHILSAI-y6j
    @NIKHILSAI-y6j 5 วันที่ผ่านมา

    at 2:33:22 while running on the colab i am getting connection errored out error in my pc , is there any way to solve this issue,anyone please help me on this issue

  • @truthseeker457
    @truthseeker457 2 หลายเดือนก่อน

    Is it any possible to contact with you?

    • @codeinajiffy
      @codeinajiffy  2 หลายเดือนก่อน

      You have my email and LinkedIn profile linked in the TH-cam channel's information.

  • @Tothefutureand
    @Tothefutureand 2 หลายเดือนก่อน

    Like

  • @ranjan-p5s
    @ranjan-p5s 2 หลายเดือนก่อน

    i am new in ai its worth for me watching out not

  • @abdullaali8026
    @abdullaali8026 หลายเดือนก่อน

    مهندس محمد احتاج اتواصل معك عندي مشروع ونحتاجك معنا بمقابل مالي

    • @abdullaali8026
      @abdullaali8026 หลายเดือนก่อน

      ارسلت لك في الخاص في التيك توك

  • @jackymarcel4108
    @jackymarcel4108 2 หลายเดือนก่อน

    Anderson Gary Thomas Nancy Jackson Sarah

  • @northernswedenstories1028
    @northernswedenstories1028 2 หลายเดือนก่อน

    OSError: [WinError 127] The specified procedure could not be found. Error loading "C:\Users\....\AppData\Local\Programs\Python\Python312\Lib\site-packages\torch\lib\torch_cuda.dll" or one of its dependencies. unfortunately i had to give up with your tutorial as getting pytorch to work is a nightmare. it doesnt seem to work. i've tried different versions and different python kernals. i installed everything using pip. just doesnt work

    • @codeinajiffy
      @codeinajiffy  2 หลายเดือนก่อน

      I get the frustration of installing things. Especially on windows. I use WSL to simulate a Linux environment so that I don't have to deal with this.
      I recommend you to run it Google colab directly like I'm doing in the tutorial. You will just have to skip running locally.
      Another longer solution is to download WSL and set up your environment there. But that can be a long process.