End-to-end Automatic Speech Recognition for Madarian and English in Tensorflow

BabyGPT: Build Your Own GPT Large Language Model from Scratch Pre-Training Generative Transformer Models: Building GPT from Scratch with a Step-by-Step Guide to Generative AI in PyTorch and Python

TensorFlow implementation of normalizations such as Layer Normalization, HyperNetworks.

Code for the paper "On the Expressivity Role of LayerNorm in Transformers' Attention" (Findings of ACL'2023)

Paper to Code automates the incorporation of research paper concepts into practical code using OpenAI's GPT models, bridging theory and implementation.

Layer normalization in PyTorch

End-to-end Automatic Speech Recognition for Madarian and English in Tensorflow

lstm with layer normalization

Time-Sensitive Deep Learning for ICU Outcome Prediction Without Variable Selection or Cleaning.

Fast-Slow Recurrent Neural Networks

The open source code for the paper "Block Attention and Switchable Normalization based Deep Learning Framework for Segmentation of Retinal Vessels"

The extension of torch.nn.LSTMCell

Layer normalization with einops semantics.

The implementation of Layer Normalization which follows paper "https://arxiv.org/abs/1607.06450".

RNNs with layer normalization; Prep package for tensorflow/addons, see v0.8.2, and commit https://github.com/tensorflow/addons/commit/cdb43ffd7e4b89d6ce8cdadcd62fec46c7f0f7fa

"A from-scratch implementation of a modern Large Language Model (LLaMA architecture). Features full backpropagation calculus in pure NumPy, followed by a GPU-accelerated PyTorch training loop."

A CUDA kernel playground for GEMM, Tensor Cores, Softmax, LayerNorm, validation, profiling, and performance benchmarking.

Pure C implementation of the Transformer Encoder Stack from Attention Is All You Need, composing multiple independent Encoder Layers into a complete Transformer encoder with efficient O(1) memory execution and zero external ML dependencies.

Pure C implementation of Layer Normalization with learnable γ and β parameters—built from scratch without ML frameworks or autograd.

Pure C implementation of the complete Transformer architecture from Attention Is All You Need, assembling embeddings, positional encoding, encoder stack, decoder layers, multi-head attention, and token prediction from modular components built entirely from scratch.