Thushan Ganegedara

Want to learn why #Transformer models are moving away from sinusoidal positional embeddings and embracing RoPE? Checkout the video. #machinelearning #ml #AI #LLMs #ChatGPT #deeplearning youtu.be/GQPOtyITy54

"You have to admit, seeing a comment like “This is super clean 😎” or “Didn’t know this could be done this way” on your code or pull request fills you with a wonderful feeling." Read more from @thush89's post. buff.ly/44RVokw

Python is a great option for ML practitioners due to its low barrier to entry combined with the massive ecosystem of scientific tooling, explains @thush89. buff.ly/44RVokw

Did you know about this new (unpopular) feature in the latest #PyTorch that allows you to scale Transformers to 10K / 100K long sequences? Watch the video to learn more. #machinelearning #deeplearning #LLMs #ml #ai youtu.be/eCWHbJpeil0

1) Natural Language Processing with #TensorFlow: amzn.to/43jEtpu 2) TensorFlow in Action: amzn.to/44hdcFP Books by @thush89 ———— #BigData #DataScience #NeuralNetworks #NLProc #MachineLearning #DeepLearning #AI #Transformers #MachineTranslation #ComputerVision

The animated Transformer: Learning about the Transformer and its internals in a fun and colorful way! #machinelearning #ML #AI #artificalintelligence #LLM #deeplearning youtube.com/watch?v=NzLwHc…

Hey GPU, what’s up with my matrix? by @thush89 buff.ly/42NlGmm

We now have open source LMMMs. In this video I try MiniGPT-4 in a Jupyter notebook to chat about images. It's quite impressive given that the model runs on a single modest 16GB GPU. Would this steal the spotlight from GPT-4? #LLMs #ChatGPT #GPT4 youtu.be/lbQC_pWPuVk

🏷️Deal of the Day🛒 Build, train, and deploy! Learn to create #deeplearning applications using Google's mighty TensorFlow 2 framework: mng.bz/XaVY @thush89 #NLP #machinelearning #python #Keras

If you think #LLMs are inaccessible due to their size and there's no hope of running them on your laptop, learn how you can do that with a few lines of code with @huggingface's awesome libraries. #deeplearning #MachineLearning #ChatGPT #GPT #llama youtu.be/CMVq48torQY

Interested in learning what's all the buzz around Llama and the profound implications it had over the last few weeks? Check out the video. #ChatGPT #chatgpt4 #LLMs #llama #artificalintelligence youtu.be/UFg0Xc08xLI

@huwprossercodes @TDataScience When the task is complex and presented from a mathematical sense, ChatGPT tends to get the final answers wrong. But you could formulate this problem in natural language stripping out mathematics, to which ChatGPT may respond better.

@TDataScience @thush89 I wonder, is it capable of recommending movies that are seemingly unrelated but you’d still enjoy.

Can ChatGPT recommend movies with machine learning by @thush89 buff.ly/41wmVWX

Can #chatgtp do movie recommendations from a dataset? Watch the video and find out what I unearthed during this fun exercise youtu.be/ks6qe3f3Fv8 #ml #ai #LLMs #nlp

Here's a playlist I created explaining how to get started on #gcp #vertexai #MachineLearning. If you haven't explored Vetex AI give it a go. You can do all exercises with a free tier account. Google Cloud Platform - Vertex AI: youtube.com/playlist?list=…

Now available for ACM Members: "TensorFlow in Action," by @thush89. This book teaches you to construct, train, and deploy deep learning models using TensorFlow. bit.ly/3JF9MVg

🏷️Deal of the Day🛒 Build, train, and deploy! Learn to create #deeplearning applications using Google's mighty TensorFlow 2 framework: mng.bz/XaVY @thush89 #NLP #machinelearning #python #Keras

📢Now in print!🍾 Taking on TensorFlow? Learn to design and deploy #deeplearning models with #TensorFlow through real world use cases: mng.bz/XaVY @thush89 #NLP #machinelearning #python #Keras

I put the book next to other books I already own. Attached you can see what it looks like. Welcome @thush89 to the book collection of @PacktAuthors! 5 of 6

2018 - ELMo - tldr; Contextualized word representations that are able to sense disambiguate (arxiv.org/abs/1802.05365)

2016 - FastText - tldr; Use sub-word information to generate embeddings robust at misspellings and can handle out-of-vocabulary words (arxiv.org/abs/1607.04606)

Evolution of word vectors (most notable milestones)🧵 #NLP #MachineLearning #DataScience