transformers

Biologically-grounded adversarial training platform: cyclic Wake/Dream/Nightmare/Compress phases that accumulate model robustness without catastrophic forgetting. Dockerized, EU AI Act compliance reporting.

This will contain the code for the 2nd edition of NLP with TensorFlow (Edition 2)

This includes the codebase for EntroPE (Entropy-Guided Dynamic Patch Encoder for Time Series Forecasting) paper.

Implementation of the paper "Last Query Transformer RNN for knowledge tracing" in PyTorch. (Kaggle 1st place solution)

A Named Entity Recognition + Entity Linker + Relation Extraction Pipeline built using spacy v3. Given a text, the pipeline will extract entities from the text as trained and will disambiguate the entities to its normalized form through an Entity Linker connected to a Knowledge Base and will assign a relation between the entities, if any.

Implementation of COCO-LM, Correcting and Contrasting Text Sequences for Language Model Pretraining, in Pytorch

Implementation of Kalmanformer, modeling the Kalman gain with a transformer

The official repository for our paper "Are Neural Nets Modular? Inspecting Functional Modularity Through Differentiable Weight Masks". We develop a method for analyzing emerging functional modularity in neural networks based on differentiable weight masks and use it to point out important issues in current-day neural networks.

Rhyme with AI

spaCy-wrap is a wrapper library for spaCy for including fine-tuned transformers from Huggingface in your spaCy pipeline allowing you to include existing fine-tuned models within your SpaCy workflow.

Tuning the Finetuning: An exploration of achieving success with QLoRA

Whispers in the Machine: Confidentiality in Agentic Systems

A practical guide to AI: from running your first local model to building your own agents. 52 files covering LLMs, Ollama, RAG, prompt engineering, machine learning, AI agents, and local AI tools.

Build LLM systems you actually control. A free, open engineering book + course — from tokenization to serving your own models. Mechanisms, trade-offs, and numbers, not prompt tips

Needle simplifies building RAG pipelines.

Contextualised Embeddings and Language Modelling using BERT and Friends using R

Code and Data for Text Classification of AI Related Patents Research Paper

A Bach music generator with Artificial Intelligence. This model is made by a VQ-VAE + Transformer (decoder-only). Sequences of midi 1 quarter length are compressed into 16 codebooks via VQ-VAE and a transformer learns how to generate the codebooks sequence to obtain a midi score.

[IEEE TMM 2022] Official implementation of YDTR: Infrared and Visible Image Fusion via Y-shape Dynamic Transformer

High-Performance Transformers for Table Structure Recognition Need Early Convolutions