Ahmad Farhan
Accelerating LLM Inference @EAGLE-Research
This profile was put together from Ahmad's public GitHub portfolio, and they haven't joined Vibetown yet, so reach them through their portfolio. Is this you? Contact us to update or remove it.
About
Ahmad Farhan works in AI engineering and ML & data science, based in Newcastle, United Kingdom. Accelerating LLM Inference @EAGLE-Research. Public projects on GitHub include Causal Synthesizer Multihead Attention (Synthesizer Self-Attention is a very recent alternative to causal self-attention that has potential benefits…), Machine-Translation-for-Endangered-Language-Revitalization (Cherokee is a highly endangered Native American language spoken by the Cherokee people) and minGPT-QA-teaching-transformer-models-facts (Teaching Transformer Models facts about the world throughpretraining, and accessing that knowledge through…), mostly written in Python, Jupyter and Go.
Projects 6
- github.com
Causal Synthesizer Multihead Attention
Synthesizer Self-Attention is a very recent alternative to causal self-attention that has potential benefits by removing this dot product.
- github.com
Machine-Translation-for-Endangered-Language-Revitalization
Cherokee is a highly endangered Native American language spoken by the Cherokee people. We provide a way to mitigate this issue by providing a Neural Machine Translation System.
- github.com
minGPT-QA-teaching-transformer-models-facts
Teaching Transformer Models facts about the world throughpretraining, and accessing that knowledge through finetuning.
- github.com
Exploring-generative-models
Image reconstruction and latent space interpolation
- github.com
Linear-classifiers-knn-NeuralNet-Numpy-Scratch
This is an implementation of two layered neural network, Linear classifiers(SVM + Preceptron) and K-Nearest-Neighbors from Scratch in numpy
- github.com
Skipgram Word2Vec
Iteration based method to learn Word Vectors. Word2vec is a method whose parameters are word vectors.This is an implementation of skipgram from scratch in numpy.