This is your work, valued

Singapore

Fangyuan Yu

Advanced
@fangyuan-ksgk

AI Researcher (LLM) @Thoughtworks || PhD KAUST

Tiny-GRPO. minimal GRPO implementation from scratch

104

Mini-LLaVA. A minimal implementation of LLaVA-style VLM with interleaved image & text & video processing ability.

99

Evolutionary-Model-Merge. Unofficial Implementation of Evolutionary Model Merging

42

CoT-Reasoning-without-Prompting. Unofficial Implementation of Chain-of-Thought Reasoning Without Prompting

35

repo-viewer. Visualize any repo or codebase into diagram or animation

24

selective-attention-transformer. Unofficial Implementation of Selective Attention Transformer

20

STaR. Implementation of STaR: Bootstrapping Reasoning With Reasoning

4

VocabularyCurriculum. Vocabulary curriculum improves scaling

4

paper-read-2024-Jan-Apr. Paper reading run from 17. Jan 2024 - 17. Apr 2024

3

CharacterLM. vocabulary curriculum + LLM

3

RuleAlign. Rule-Based Adherance evaluation for LLM Agent

3

llama3-sft. Supervised FineTuning with Llama-3

3

auto-test. Automatic Chat with LLM & Test

2

surfer. Voice Interface on Web Browsing

2

Multilevel-Ensemble-Kalman-Bucy-Filter. This repository contains an implementation of the Multilevel Ensemble Kalman Bucy Filter. Code for HPC implementation with MPI cores, designed to run on Supercomputer Shaheen is also included. The algorithm deals with linear Gaussian filtering problems in continuous time and has appealing performance in high dimensions.

2

RagAgent. Get some information locally with LLM

2

Iterative_graph_alignment. Official Implementation of Iterative Graph Alignment https://arxiv.org/abs/2408.16667

2

GeniusUpload. Why not ask genius mind instead of working on them ourselves?

2

waveletGPT. Experimental code for wavelet GPT and Fourier series expression

2

CS194-26-Computer-Vision-and-Computational-Photography. Jupyter Notebook

2

diamond. DIAMOND (DIffusion As a Model Of eNvironment Dreams) is a reinforcement learning agent trained in a diffusion world model. NeurIPS 2024 Spotlight.

1

GANs. Simple Implementation of implementations of Generative Adversarial Networks (GANs) using Tensorflow.

1

concept-encoding. experiment with concept encoding in LLM

1

abstraction-learning. Experiment for abstraction learning

1

Normalizing-Flows. This repository contains examples of simple implementation of NF. Normalizing Flows are generative models which produce tractable distributions where both sampling and density evaluation can be efficient and exact.

1

MiniAgent. Chinese ChatBot

1

c3po. Jupyter Notebook

1

MCVD. Video prediction with Diffusion Models

1

AnnoAttribute. Annotation Tool for Decomposable Attributes

1

FTchatter. Chat with FineTune model served with vLLM

1

DenseNet. Densely Connected Convolutional Neural Network implementation in Pytorch

1

vision-plus. Vision language model from scratch

1

human-eval. Code for the paper "Evaluating Large Language Models Trained on Code"

1

translate-thai. Translation for Thai Language

1

llama3-function-call. training function-calling llama-3 agent with mlx

1

CS231a-Computer-Vision-From-3D-Reconstruction-to-Recognition. Jupyter Notebook

1

detr-object-detection. Jupyter Notebook

1

PairRM-with-mps. PairRM but with MPS support

1

adaptor. Adapt LLM with human-level efficiency

1

CarND-Behavioral-Cloning-Project. Deep Learning for Self-Driving Cars

1

payflow. Handling your document for you.

1

Camera-Calibration. Jupyter Notebook

1

Chicken-Planner. Plan as a chicken

1

thought-compiler. Agent to help us in research

1

1bitNet. Experiments with 1Bit network

1

process-supervision. Experiments with Procedure Supervision

1

Paper-Read-April. Record of read paper in April

1

align-llama3-with-reft. Fast & Efficient Alignment with Representation FineTuning & Llama3

1

claude-agent. Claude Agent is built on Anthropic models to automate digital operation

1

Alignment-experiment. Alignment Experiments

1

fast_minbpe. Jupyter Notebook

1

Transformer-MachineTranslation. Jupyter Notebook

1

HPC_with_Python. Notebooks cataloging packages to enable using High Performance Computing on mulitcore and mulitnode architecture and clusters.

1

Bayesian-Neural-Networks. Pytorch implementations of Bayes By Backprop, MC Dropout, SGLD, the Local Reparametrization Trick, KF-Laplace, SG-HMC and more

1

Question-Answering-AI-with-BERT. Jupyter Notebook

1