Still preparing the dataset for financial data
Started preparing a dataset for experiments. Realised that using standard available data leads to conclusions already drawn by others
Finished 3Blue1Brown's series on GPT, played around with Ollama. Read some docs.
Understood LSTMs, Word2Vec and Seq2Seq
Implemented classification and generation of names in RNNs. Tried understanding a research paper, still a long way to go.
Understood many aspects of RNNs. Skipped some specialised RNNs for future. Preferred PyTorch over NumPy for this.
Learned about classification neural nets and different activation functions. Started with RNNs
Switched it up. Started with VikParuchuri/zero_to_gpt course to get my hands real dirty
Switched it up. Started with VikParuchuri/zero_to_gpt course to get my hands real dirty
Played with random forest technique from classical machine learning
Implemented a dead simple neural network from scratch in PyTorch and trained on the titanic dataset
Used huggingface's transformers library to build a basic NLP model, learned basics of numpy
Experimented with MNIST by myself using FastAI wrappers and understood what it takes to built basic functional things
Watched the first lecture on Getting started, setup the local dev environment and tried out basic deep learning models.
Why did I decide to start with AI? What am I planning to cover in the coming 60 days?