Eric-Wallace

open_llama

OpenLLaMA, a permissively licensed open source reproduction of Meta AI’s LLaMA 7B trained on the RedPajama dataset

distiller

Neural Network Distiller by Intel AI Lab: a Python package for neural network compression research. https://intellabs.github.io/distiller

trfl

TensorFlow Reinforcement Learning

loss-landscape

Code for visualizing the loss landscape of neural nets

EasyLM

Large language models (LLMs) made easy, EasyLM is a one stop solution for pre-training, finetuning, evaluating and serving LLMs in JAX/Flax.

BERT-related-papers

BERT-related papers

jiant

jiant is an nlp toolkit

scalable_agent

A TensorFlow implementation of Scalable Distributed Deep-RL with Importance Weighted Actor-Learner Architectures.

pytorch-maml-rl

Reinforcement Learning with Model-Agnostic Meta-Learning in Pytorch

FewRel

A Large-Scale Few-Shot Relation Extraction Dataset

finetune

Scikit-learn style model finetuning for NLP

few-shot-learning

Few-shot Learning of GPT-3

gensen

Learning General Purpose Distributed Sentence Representations via Large Scale Multi-task Learning

incoder

Generative model for code infilling and synthesis

universal-triggers

Universal Adversarial Triggers for Attacking and Analyzing NLP (EMNLP 2019)

interpretability-tutorial-emnlp2020

Materials for the EMNLP 2020 Tutorial on "Interpreting Predictions of NLP Models"

scpn

syntactically controlled paraphrase networks

WorldModels

Full World Models Implementation in Chainer

LM_Memorization

Training data extraction on GPT-2

sears

Code for "Semantically Equivalent Adversarial Rules for Debugging NLP Models"

long_tail_knowledge

Repo for the paper "Large Language Models Struggle to Learn Long-Tail Knowledge"

ContextualDecomposition

Demo for method introduced in "Beyond Word Importance: Contextual Decomposition to Extract Interactions from LSTMs"

Poisoning-Instruction-Tuned-Models

onestep-extraction

DifferentiableNeuralComputer

Optimized Differentiable Neural Computer In Chainer

aiorobinhood

Asynchronous Robinhood HTTP client built using asyncio and aiohttp

pathologies

EMNLP 2019 "Pathologies of Neural Models Make Interpretations Difficult"

kuro

Experiment manager for ML experiments

DPR

Dense Passage Retriever - is a set of tools and models for open domain Q&A task.

model-subspace

Pytorch hooks to save the subspace associated with a model so it can be used as input to SVCCA.