
Learn how to use TensorFlow sparse tensors to efficiently represent tensors with many zeros by specifying indices, values, and dense shape, and convert with sparse to dense.
Learn to build a simple linear regression model in TensorFlow, tune weights and biases for data, and use normalization and dense layers in a sequential model.
The lecture explains how stochastic gradient descent updates model parameters to reduce loss, uses learning rate, and compares SGD with Adam, illustrating training across epochs and convergence.
Explore corrective measures for underfitting by diagnosing training and validation losses, and expand model depth with nonlinear activations like ReLU to improve performance.
Model sequential data with recurrent neural networks using embeddings and time-step hidden states, and explore bidirectional variants and their streaming limitations.
Master data preparation for sentiment analysis by tokenizing and numerically encoding text with TensorFlow's text vectorization layer, adapting a 10,000-word vocabulary and a 200-token sequence length.
Learn how word2vec converts words to dense embeddings, compare one-hot vs embedding layers, and train CBOW and skip-gram models with negative sampling to create transferable word representations for NLP tasks.
Integrate a pretrained word2vec embedding (10,000 by 300) from a 3 million word vocabulary via Gensim into a sentiment analysis model, trainable or frozen.
Visualize model embeddings with TensorBoard projector, projecting 300-dimensional vectors to three dimensions and exploring nearest neighbors using metadata and cosine similarity or euclidean distance.
Develop and implement a custom bleu score metric from scratch, converting targets to vectors, computing matches with a boolean mask, and integrating it into training in graph and eager modes.
Explore how Bahdanau attention overcomes fixed-length bottlenecks by linking each output to all input positions with learned attention weights. See how bidirectional encoders and additive scoring create context-aware translations.
Build and test Bahdanau attention by implementing the encoder with embedding and the attention layer, compute attention weights to a context vector, and decode with teacher forcing.
Explore visualizing transformer attention by computing and plotting attention scores from softmax across eight heads, revealing how each word attends to others in encoder–decoder contexts.
Deep Learning is a hot topic today! This is because of the impact it's having in several industries. One of the fields in which deep learning has the most influence today is Natural Language Processing.
To understand why Deep Learning based Natural Language Processing is so popular; it suffices to take a look at the different domains where giving a computer the power to understand and make sense out of text and generate text has changed our lives.
Some applications of Natural Language Processing are in:
Helping people around the world learn about any topic ChatGPT
Helping developers code more efficiently with Github Copilot.
Automatic topic recommendation in our Twitter feeds
Automatic Neural Machine Translation with Google Translate
E-commerce search engines like those of Amazon
Correction of Grammar with Grammarly
The demand for Natural Language Processing engineers is skyrocketing and experts in this field are highly paid, because of their value. However, getting started in this field isn’t easy. There’s so much information out there, much of which is outdated and many times don't take the beginners into consideration :(
In this course, we shall take you on an amazing journey in which you'll master different concepts with a step-by-step and project-based approach. You shall be using Tensorflow 2 (the world's most popular library for deep learning, built by Google) and Huggingface transformers (most popular NLP focused library ). We shall start by understanding how to build very simple models (like Linear regression model for car price prediction and RNN text classifiers for movie review analysis) using Tensorflow to much more advanced transformer models (like Bert, GPT, BlenderBot, T5, Sentence Transformers and Deberta).
After going through this course and carrying out the different projects, you will develop the skill sets needed to develop modern deep learning for NLP solutions that big tech companies encounter.
You will learn:
The Basics of Tensorflow (Tensors, Model building, training, and evaluation)
Text Preprocessing for Natural Language Processing.
Deep Learning algorithms like Recurrent Neural Networks, Attention Models, Transformers, and Convolutional neural networks.
Sentiment analysis with RNNs, Transformers, and Huggingface Transformers (Deberta)
Transfer learning with Word2vec and modern Transformers (GPT, Bert, ULmfit, Deberta, T5...)
Machine Learning Operations (MLOps) with Weights and Biases (Experiment Tracking, Hyperparameter Tuning, Dataset Versioning, Model Versioning)
Machine translation with RNNs, attention, transformers, and Huggingface Transformers (T5)
Model Deployment (Onnx format, Quantization, Fastapi, Heroku Cloud)
Intent Classification with Deberta in Huggingface transformers
Named Entity Relation with Roberta in Huggingface transformers
Neural Machine Translation with T5 in Huggingface transformers
Extractive Question Answering with Longformer in Huggingface transformers
E-commerce search engine with Sentence transformers
Lyrics Generator with GPT2 in Huggingface transformers
Grammatical Error Correction with T5 in Huggingface transformers
Elon Musk Bot with BlenderBot in Huggingface transformers
Speech recognition with RNNs
If you are willing to move a step further in your career, this course is destined for you and we are super excited to help achieve your goals!
This course is offered to you by Neuralearn. And just like every other course by Neuralearn, we lay much emphasis on feedback. Your reviews and questions in the forum will help us better this course. Feel free to ask as many questions as possible on the forum. We do our very best to reply in the shortest possible time.
Enjoy!!!