![Deepak John Reji](/img/default-banner.jpg)
- 114
- 150 145
Deepak John Reji
Ireland
Приєднався 5 сер 2012
I am an NLP practitioner & Researcher with experience in developing and designing solutions for data science products. I love working with environmental data, and I create videos on various prototypes that I experiment with NLP tutorials and podcasts with industry experts in the sustainability field.
I am an open-source contributor and I share my models in Huggingface and PyPI repositories. Recently I have been researching the topic "Bias & Fairness in AI Models" & "Environmental Remediation"
I am currently working on a Research thesis with the University of Limerick, on the topic "Immigration & Integration" and I teach Sociology bachelors students.
👋 Feel free to connect with me on Github and UA-cam 😀
I am an open-source contributor and I share my models in Huggingface and PyPI repositories. Recently I have been researching the topic "Bias & Fairness in AI Models" & "Environmental Remediation"
I am currently working on a Research thesis with the University of Limerick, on the topic "Immigration & Integration" and I teach Sociology bachelors students.
👋 Feel free to connect with me on Github and UA-cam 😀
Podcast #55 - Data as a Product and Data Products
This is a conversation with Supreet Kaur, where we dive into the fascinating world of data products and their growing significance in today's technological landscape. We explore what defines a data product and discuss how organisations can leverage these powerful tools to gain a competitive edge.
Join us as Supreet breaks down the key factors that make data products effective and examines their impact on traditional data management and governance practices. We'll highlight real-world examples of successful data products that have revolutionized industries and businesses, showcasing their transformative power.
Developing and monetizing data products comes with its own set of challenges and considerations. Supreet provides insights into overcoming these hurdles, ensuring that your organization can successfully harness the potential of data products.
We'll also discuss the integration of data products into the AI lifecycle, from data acquisition and preprocessing to model development and deployment. Learn how these products can streamline and enhance the entire AI process.
Ethical implications are a critical aspect of data products, Discover strategies for ensuring responsible and ethical use of data products while maximizing their value and impact.
Supreet Kaur
Supreet is an Assistant Vice President (AVP) at Morgan Stanley, where she leads the development of cutting-edge AI products that will define the next generation of financial technology. Her innovative approach to problem-solving and her expertise in leveraging the latest technologies have been instrumental in positioning Morgan Stanley as a leader in the fintech industry.
Prior to joining Morgan Stanley, Supreet honed her skills as a data science consultant at ZS Associates, where she automated complex workflows and created data-driven solutions for Fortune 500 clients. Her work on impactful use cases such as predicting SKUs for rare disease patients has been highly lauded by her peers and clients.
Supreet is a prolific speaker and has delivered nearly 50 talks on the intersection of data, AI, and business. She is passionate about democratizing AI and actively shares her insights and perspectives on industry trends through her engaging content on LinkedIn and her personal blog.
She is also a content creator on LinkedIn and shares insights and news about it on the platform
Links:
ksupreet1995.wixsite.com/brown-girl-quest
medium.com/@supreetkaur_66831
Join us as Supreet breaks down the key factors that make data products effective and examines their impact on traditional data management and governance practices. We'll highlight real-world examples of successful data products that have revolutionized industries and businesses, showcasing their transformative power.
Developing and monetizing data products comes with its own set of challenges and considerations. Supreet provides insights into overcoming these hurdles, ensuring that your organization can successfully harness the potential of data products.
We'll also discuss the integration of data products into the AI lifecycle, from data acquisition and preprocessing to model development and deployment. Learn how these products can streamline and enhance the entire AI process.
Ethical implications are a critical aspect of data products, Discover strategies for ensuring responsible and ethical use of data products while maximizing their value and impact.
Supreet Kaur
Supreet is an Assistant Vice President (AVP) at Morgan Stanley, where she leads the development of cutting-edge AI products that will define the next generation of financial technology. Her innovative approach to problem-solving and her expertise in leveraging the latest technologies have been instrumental in positioning Morgan Stanley as a leader in the fintech industry.
Prior to joining Morgan Stanley, Supreet honed her skills as a data science consultant at ZS Associates, where she automated complex workflows and created data-driven solutions for Fortune 500 clients. Her work on impactful use cases such as predicting SKUs for rare disease patients has been highly lauded by her peers and clients.
Supreet is a prolific speaker and has delivered nearly 50 talks on the intersection of data, AI, and business. She is passionate about democratizing AI and actively shares her insights and perspectives on industry trends through her engaging content on LinkedIn and her personal blog.
She is also a content creator on LinkedIn and shares insights and news about it on the platform
Links:
ksupreet1995.wixsite.com/brown-girl-quest
medium.com/@supreetkaur_66831
Переглядів: 81
Відео
Podcast #53 - Data-Driven Insights: Unleashing the Power of Measurement and Experimentation
Переглядів 27411 місяців тому
This is a conversation with Vanessa Pizante, where we delve into the realm of data-driven insights, uncovering the transformative power of measurement and experimentation. In this episode, we explore the profound impact of these practices on decision-making and overall business success. Discover how a scalable and adaptable measurement framework can be a game-changer for businesses, enabling th...
Podcast #52 - Fast Forward: How to Harness the Potential of AI for a Sustainable Future
Переглядів 117Рік тому
This is a conversation with Alice Schmidt, a prominent face in the realm of global sustainability, business consultancy, and authorship. With 25 years of experience at the crossroads of social, environmental, and economic affairs, Alice's insights are invaluable. Uncover the future possibilities of AI as we discuss harnessing its potential for a more sustainable future. Alice Schmidt Alice Schm...
Question Answering from PDF using OpenAI ChatGPT API and GPT4ALL Library
Переглядів 3,6 тис.Рік тому
In this comprehensive tutorial, we'll explore how to build an intelligent question-answering system that can extract answers from PDF documents. By combining the cutting-edge power of the OpenAI ChatGPT API and leveraging the open-source embedding capabilities of the GPT4ALL library, we'll show you how to create a robust and accurate solution to tackle this challenging task. 📋 Outline: 1. Setti...
Information Extraction with GPT4ALL Models and Langchain Components | Video Tutorial
Переглядів 2,5 тис.Рік тому
In this video tutorial, you will learn how to harness the power of the GPT4ALL models and Langchain components to extract relevant information from a dataset efficiently and with minimal lines of code. This tutorial will equip you with the knowledge to effectively retrieve valuable insights from your data. We will begin by introducing the GPT4ALL ecosystem (docs.gpt4all.io/index.html). You will...
Podcast #51 - The Future of Textile Manufacturing
Переглядів 3 тис.Рік тому
This is a conversation with Garett Gerson, the visionary founder and CEO of VARIANT3D. Join us as we delve into the future of textile manufacturing and explore how 3D knitting technology is revolutionizing the industry. Discover how VARIANT3D's disruptive platform paves the way for sustainable, customizable, and on-demand textiles with near-zero waste. From discussing the impact on the environm...
Podcast #50 - Empowering Data Science and Open-Source Python
Переглядів 89Рік тому
This is a captivating conversation with Sophia Yang, a Senior Data Scientist and Developer Advocate at Anaconda. Join us as we delve into the world of data science and open-source development with one of the industry's most inspiring voices. Sophia's extensive contributions to the Python open-source community, including her authorship of various libraries such as condastats, cranlogs, PyPowerUp...
Podcast #49 - Redefining Fashion: A Deep Dive into Circular Consumption and Sustainable Fashion
Переглядів 198Рік тому
This is a conversation with Sarah Garner, a visionary entrepreneur who is revolutionizing the fashion industry with her sustainable and circular fashion platform, Retykle. Join us as we sit down with Sarah to delve into her inspiring journey and the incredible impact she is making on the world. In this Podcast, we'll explore Sarah's mission to create a more sustainable future for fashion, tackl...
Train your custom Speech Recognition Model with Hugging Face models
Переглядів 14 тис.Рік тому
This tutorial will show you how to train a custom voice recognition model using Hugging face models. With the increasing popularity of voice-enabled devices and services, having accurate and reliable voice recognition is crucial. By training your own custom voice recognition model, you can improve the accuracy of your voice-enabled applications and services, and tailor them to your specific nee...
Podcast #48 - Beyond the Code: Navigating the Ethical Landscape of AI
Переглядів 392Рік тому
This is a conversation with Ravit Dotan, an expert in AI ethics and responsible AI governance. In this episode, we delve into the crucial topic of AI ethics implementation and ensuring it is grounded in genuine moral principles that respect all stakeholders. Ravit shares her insights on the need for a comprehensive and coherent framework of ethical guidelines, guidance for AI developers and use...
Podcast #47 - How Humans Perceive Their Relationship with AI
Переглядів 492Рік тому
This is a conversation with Marisa Tschopp, a researcher and expert on the intersection of technology and human psychology. In this episode, we explore the fascinating topic of how humans perceive their relationship with artificial intelligence. Marisa takes us on a journey through the world of AI, discussing the concept of anthropomorphism and how it shapes our perceptions of technology. We de...
Podcast #46 - Unpacking the Future: A Deep Dive into Gen Z in Tech
Переглядів 671Рік тому
This is a conversation with Brooke Joseph, a 17-year-old innovator and aspiring artificial intelligence expert at The Knowledge Society (TKS). In this episode, we explore the intersection of Gen Z and technology and how Brooke has found her passion for AI and Federated Learning. We dive into how the increasing use of technology in everyday life affects the way Gen Z interacts with and understan...
Podcast #45 - Digital Transformation: Challenges from the perspective of the public sector & society
Переглядів 666Рік тому
Podcast #45 - Digital Transformation: Challenges from the perspective of the public sector & society
Podcast #44 - Optimizing Job Shop Manufacturing with AI Scheduling
Переглядів 920Рік тому
Podcast #44 - Optimizing Job Shop Manufacturing with AI Scheduling
Podcast #43 - Revolutionizing Mental Health Care with Digital Technology
Переглядів 759Рік тому
Podcast #43 - Revolutionizing Mental Health Care with Digital Technology
Podcast #42 - Thought Leadership 101: How to Stand Out in Your Industry
Переглядів 719Рік тому
Podcast #42 - Thought Leadership 101: How to Stand Out in Your Industry
Podcast #41 - ChatGPT and the Future of the Legal Profession
Переглядів 560Рік тому
Podcast #41 - ChatGPT and the Future of the Legal Profession
Podcast #40 - Climate Optimism: Reimagining Our Future
Переглядів 556Рік тому
Podcast #40 - Climate Optimism: Reimagining Our Future
Podcast #39 - Grounded Language Understanding
Переглядів 857Рік тому
Podcast #39 - Grounded Language Understanding
Podcast #38 - Large language models, Applications & Implications
Переглядів 1,2 тис.Рік тому
Podcast #38 - Large language models, Applications & Implications
Podcast #36 - The role of open source in career development
Переглядів 447Рік тому
Podcast #36 - The role of open source in career development
Podcast #37 - Data-centric NLP in the era of LLMs
Переглядів 446Рік тому
Podcast #37 - Data-centric NLP in the era of LLMs
Create API for Question Answering pipeline using FastAPI
Переглядів 1,2 тис.Рік тому
Create API for Question Answering pipeline using FastAPI
Gibberish words detection with Python
Переглядів 1,2 тис.Рік тому
Gibberish words detection with Python
Build Question Answering pipeline with Transformers
Переглядів 813Рік тому
Build Question Answering pipeline with Transformers
Podcast #35 - Content Creation as a Viable Career: Tips, Challenges, and Success Stories
Переглядів 387Рік тому
Podcast #35 - Content Creation as a Viable Career: Tips, Challenges, and Success Stories
Podcast #34 - Azure Machine Learning - ML as a Service
Переглядів 608Рік тому
Podcast #34 - Azure Machine Learning - ML as a Service
Clustering with embed-clustering package
Переглядів 751Рік тому
Clustering with embed-clustering package
Biomedical Named Entity Recognition with Transformers
Переглядів 5 тис.Рік тому
Biomedical Named Entity Recognition with Transformers
if any model has for resume
great vid, I just have a quick question, How do I make a permanent link to share to others
is there any annotation tool for the preparing spacy ner data?
Thank you!
hello sir, does this model detect filler words ?
Its not working. ner_prediction(corpus=doc, compute='cpu') AttributeError: 'DataFrame' object has no attribute 'append'
instead of training an NER is there any way to pass a certain data into the spacy model i.e can we pass the custom data inside a spacy model?
Good video, so much tanks. You helped me a lot.
hey ,can you please provide the training notebook .thanks in advance
Hi, the notebook is not being shared, the research paper has its details for training the model
Hi, quick question: i had trained the model like you suggested. but when i loaded the best model and tested it on few docs, its returning the docs only instead of the entity. Can you suggest why this would be the case
Hi, have you used the model calling code correctly
Works great, but have a question. How can i calculate the metrics precision recall f1 accuracy scores
Good afternoon! Tell me, please, have you published a file with the training of the model? I really like your work and I want to develop in this field!
Hi, Thank you for watching, I haven't published the code files yet, the research paper has the details of the model: journals.plos.org/digitalhealth/article?id=10.1371%2Fjournal.pdig.0000152
Hello, I ran the codes and trained the model on the entire dataset, but when I run the inference code, the predictions are empty. Any idea why? Could it have anything to do with the fact that I don't have pytorch_model.bin in my model folder, but model.safetensors instead?
nope
Thx
Im just here to be nosy.. his man is intelligent and fine. Ok byee
Thanks Deepak for your guidance. I was struggling trying to get LocalDocs work with GPT4All and your method of using CSV files works better than LocalDocs. I have a question on further enhancing this. How do I get the model to do the following: 1) List more than one result? I tried k=4, but that did not give me any more than 1 result. 2) How do I get the model to summarize information? Example, I want it do simple things like 'How many novels were written by Tagore?' from a list of his works. Thanks in advance for your additional help.
the model you created is it based on the whole datasets?
Very useful work. But i am getting AttributeError: 'DataFrame' object has no attribute 'append'. can you pls recheck/update the code?
how can it be solved?
This is a good model that I've been using for my course project for some time. Your work is very much appreciated!
Thank you so much :)
Hi Brother How to annotate automatically for large text data by this i have to do it mannually
Hey, great work dude! I am wondering where can i access this Named Entity Spacy Tagger @ 1:46 Thank you
Thank you, That repo is down, unfortunately.
sir, I am note able to import huggingsound, Please help
its showing the following error :- ERROR: Ignored the following versions that require a different python version: 0.0.1 Requires-Python >=3.7,<3.10; 0.1.0 Requires-Python >=3.7,<3.10; 0.1.1 Requires-Python >=3.7,<3.10; 0.1.2 Requires-Python >=3.7,<3.10; 0.1.3 Requires-Python >=3.7,<3.10; 0.1.4 Requires-Python >=3.7,<3.10; 0.1.5 Requires-Python >=3.7,<3.10 ERROR: Could not find a version that satisfies the requirement torch!=1.12.0,<1.13.0,>=1.7 (from huggingsound) (from versions: 2.0.0, 2.0.1, 2.1.0) ERROR: No matching distribution found for torch!=1.12.0,<1.13.0,>=1.7
@@GouravKumar-qi5gt based on the error, it seems you are not using compatible python version, could you upgrade and check
guru@@deepakjohnreji , please play input and output sounds
The issue with this is, if we want the openai to set I don't know the answer if some irrelevant question is asked, it still answers. How to implement that
❤❤❤
Does this work on intranet environment
I really like the second cluster. Its funny to see twitter at its worst.
"promosm" ❤️
Is it required valid openain credential?
Yes, you need to have open ai credentials for this tutorial
@@deepakjohnreji so what is the position of GPT4All in this tutorial. I thought GPT4All as a openapi replacement
@@DefamsTVHere GPT4All embeddings are used, if you want to try out completely open source implementation please check out this tutorial ua-cam.com/video/1cx3wOhisTg/v-deo.htmlsi=opSsQCyhj0kg56iE
GPT4ALLEmbedding() throwong error "GGML assert "
github.com/imartinez/privateGPT/issues/428 could you check your system specification, and whether its supporting the mode loading
I am using huggingfacedmbedding for now
Thanks you for the amazing tuto. Can we annotat severals data in the same tIME ?
Thank you, Yes you could.
Excellent tutorial!!! It helped me to learn the custom NER, which otherwise looks difficult to follow in the spaCy documentation.
Thank you so much :)
Yes, Spacy documentation is poor
Such a lovely man
the real ones are here for a reason.
😂😂😂
The realest
The picture in the back :') cute
@@daria5846 OH MY GOD??
who's the picture of?@@daria5846
Good one,, i m trying to build the knowledge graph using this technique, but have got stuck into it. Would you please suggest me how to tackle it? 1- how to have the 2 edges from the same source node to destination node?I mean I have tried all possible ways best of my knowledge to build more than one transition edge from same source node to the same destination node in the same direction. 2- how to identify all the possible paths from the initial node to the final node, when there's a KG(knowledge graph) is available.
great video, I have a question tho let's say i trained a model with a TRAIN_DATA of 300 texts, now i have 200 more texts to train because the model was not accurate. is there a possibility to just train the same model with these 200 new texts or should i train a new model with all the 500 texts(it will take a long time)? if there is a way how pls ^^
Thanks, You could try training the model again on top of the 300 data sample model, I would say test that approach, if its not working out then better train with complete dataset again :)
Bring more of these amazing people loving these unscripted podcasts nowadays.
Great content sir . Thank you so much . Learnt a lot from this one . Keep growing 🎉
Thank you :)
how can i access the code?
I will soon publish it on my Github repo :)
@@deepakjohnreji thank you
Very good one.. i have a special use case.. let me know how to connect with you do discuss the usecase
Hi, you can reach out to me via my Linked In or email
Great tutorial!!! Very concise.
Thank you :)
Nice where we can see the code for above
I have uploaded the files on my GitHub page github.com/dreji18/GPT4ALL-Langchain
ValueError("[E024] Could not find an optimal move to supervise the parser. Usually, this means that the model can't be updated in a way that's valid and satisfies the correct annotations specified in the GoldParse. For example, are all labels added to the model? If you're training a named entity recognizer, also make sure that none of your annotated entity spans have leading or trailing whitespace or punctuation. You can also use the `debug data` command to validate your JSON-formatted training data. For details, run: python -m spacy debug data --help") I am getting this error.......
I guess it may be the spacy version and its dependencies, could you clean the current spacy and install it again.
@@deepakjohnreji Thank you Reji......but u taught well tho....:)
I have this thing as my training data drive.google.com/file/d/1ssBswos2TAh8OTpcdTz7iDNqU2jCti7V/view?usp=drivesdk How to train now?
I have requested for access to your training data.
I could access it now, please give more context about this data
@@deepakjohnreji you have to train your agent in such a way that when someone gives input from text1 and text 2 the agent should indicate the relevancy of the given sentences between 0&1 (0 if the sentences doesn't match and 1 if both the sentences are equal). I used spacy to do that but it was manual for example I used to manually write sentences and then used to check the accuracy of the two sentences. I never trained the algorithm to do that.
@@hmmmmn6770 This is a similarity check use case, for this you can use any of the embedding model and run similarity on it.
I am getting too many errors when I run the model_training code. I have tried running it on Google Colab, but I still cannot get any results. Can you please help me?
Is it working now?
Salam Mr. Deepak John Reji, i've tried to follow your video step by step, but when i reach the step 5 - run the training code i had a error massage "TF-TRT Warning : could not find TensorRT" , I have tried so many ways on the internet but until now I still haven't found the right one, can you help me, oh yes, I used google colab to do this coding.
Could you install spacy library again and try? In colab you shouldn't be getting these sorts of errors. Maybe opening a new kernel would help you fix the issue.
can this approach be useful to group google reviews into categories
Yes, it can be used.
Great work. Kindly provide the training notebook, codes on how to train the model. Thanks in advance
Thank you very much for this video. Step 3. Don't you think it should be i[1]['entities'] = [(0, 0, entity_name)] instead?
Thanks, I will look into it, could you share your file with me to check.
so great video but just 500 views is shocking me !!!
claps hands for thumbnail bait