David R. Winer – Medium

David R. Winer

Pinned

David R. Winer
in
Towards Data Science

Towards Generative AI for Model Architecture

How “MAD” AI will help us discover the next transformer

Nov 10, 2023

Towards Generative AI for Model Architecture

Nov 10, 2023

Pinned

David R. Winer
in
Towards AI

Walkthrough of Graph Attention Network (GAT) with Visual Implementation

Understanding Graph Neural Networks (GNNs) is increasingly relevant as transformers continue tackling graph problems like from the Open…

Sep 15, 2023

Walkthrough of Graph Attention Network (GAT) with Visual Implementation

Sep 15, 2023

David R. Winer

Graphbook takes the pain out of doing “surgery” to AI Models, like BERT, GPT, and T5

As language models get larger and more complex, they are also becoming more difficult to understand and edit. In order to convey the effort…

Oct 31, 2023

Graphbook takes the pain out of doing “surgery” to AI Models, like BERT, GPT, and T5

Oct 31, 2023

David R. Winer

AI Modeling as Diagramming: LucidChart to Cerbrec Graphbook

Recently, I was looking around some of my LucidChart diagrams and found some attempts to understand BERT from its original implementation…

Oct 30, 2023

AI Modeling as Diagramming: LucidChart to Cerbrec Graphbook

Oct 30, 2023

David R. Winer
in
Towards AI

Attention Sinks and Where to Cache Them: A Visual Walkthrough for the Streaming LLM Implementation

GPT-2 Rolling Cache Text Generator Block

Oct 11, 2023

Attention Sinks and Where to Cache Them: A Visual Walkthrough for the Streaming LLM Implementation

Oct 11, 2023

David R. Winer
in
Towards AI

Visual Walkthrough for Vectorized BERTScore to Evaluate Text Generation

AI-based text generation has clearly entered the mainstream. From automated writing assistants, to legal document generation, marketing…

Oct 8, 2023

Visual Walkthrough for Vectorized BERTScore to Evaluate Text Generation

Oct 8, 2023

David R. Winer
in
Towards AI

Walkthrough of LoRA Fine-tuning on GPT and BERT with Visualized Implementation

A visualized walkthrough for how (and where) in transformer architectures to swap dense self attention layers with LoRA layers.

Sep 26, 2023

Walkthrough of LoRA Fine-tuning on GPT and BERT with Visualized Implementation

Sep 26, 2023

David R. Winer

A Fully Visualized GPT2 Implementation and Comparison to BERT

Generative Pre-trained Transformer (GPT) is one of the key transformer architectures revolutionizing generative AI applications. There is a…

Sep 21, 2023

A Fully Visualized GPT2 Implementation and Comparison to BERT

Sep 21, 2023

David R. Winer

A Fully Visualized (And Honest) Implementation Of BERT

As a visual thinker, the first figure I look for in an AI paper (after reading the abstract) is the conceptual diagram. The typical diagram…

Sep 14, 2023

A Fully Visualized (And Honest) Implementation Of BERT

Sep 14, 2023

David R. Winer

David R. Winer

AI + ML Engineer with PhD, interested in transformer model architectures, graph neural networks, and AI model visualization. Co-creator of Cerbrec Graphbook.

Following

Help
Status
About
Careers
Press
Blog
Privacy
Terms
Text to speech
Teams