shawntan

scattermoe

Triton-based implementation of Sparse Mixture of Experts.

theano_toolkit

Collection of useful, re-used routines.

icentia-ecg

Working on Icentia ECG data.

chicken-rice-nn

Miscellaneous code for doing NLP with Theano

SUT

Repository for Sparse Universal Transformers

theano-kaldi

Bunch of scripts for working with Kaldi.

stack-binary-recursive-nn

Parallelised implementation of Recursive Neural Networks for binary trees in PyTorch

lost-in-the-middle

Code and data for "Lost in the Middle: How Language Models Use Long Contexts"

rtdp

Code for the paper titled "Recursive Top-Down Production for Sentence Generation with Latent Trees"

stick-breaking-vae

Infinite MIxture Models with VAEs

awd-lstm-lm

LSTM and QRNN Language Model Toolkit for PyTorch

cv

CV, typesetted in Helvetica Neue, using XeTeX, TikZ and Biblatex

CategoricalNF

Official repository for "Categorical Normalizing Flows via Continuous Transformations"

compound-pcfg

IFT6135H19_assignment

Lasagne

Lightweight library to build and train neural networks in Theano

lexical

Lexicon Learning for Few-Shot Neural Sequence Modeling

libgpuarray

Library to manipulate tensors on the GPU.

life

Life - a timeline of important events in my life

lm-evaluation-harness

A framework for few-shot evaluation of autoregressive language models.

NAF

Experiments for the Neural Autoregressive Flows paper

nanotron

Minimalistic large language model 3D-parallelism training

neural_networks_chomsky_hierarchy

text

Data loaders and abstractions for text and NLP

Theano

Theano is a Python library that allows you to define, optimize, and evaluate mathematical expressions involving multi-dimensional arrays efficiently. It can use GPUs and perform efficient symbolic differentiation.

transformer-xl

transformer_latent_diffusion

Text to Image Latent Diffusion using a Transformer core

transformers

🤗 Transformers: State-of-the-art Natural Language Processing for TensorFlow 2.0 and PyTorch.

Triton-Puzzles

zoology

Understand and test language model architectures on synthetic tasks.