Educational Implementation · Python

modded-GLM-5.3-flash-from-scratch

Pop Alexandru@pop123-ux · Atlas 70.7 · #4= builder

View on GitHub1 Open Pop Alexandru's Atlas

What it is

The repository is an educational implementation of a hybrid-attention Mixture-of-Experts language model, complete with training scripts, evaluation tools, and theory notebooks for learning and experimentation.

Scored from the code at its last scan, 20 Oct 2026. How merit is scored

Where it counts

Machine Learning85%
NLP & LLMs80%
Computer Vision15%
AI Agents15%
Data Science5%
MLOps5%

How much of the project is about each field, as scoring counts it. It ranks in the field it is most about.

Around it on the Machine Learning board

Whole board

RankProjectBuilderFieldMerit

1AECS-schedulerLibrary / Framework · PythonAECS is a PyTorch learning rate scheduler that adjusts the learning rate in real time based on training signals such as gradient norm and loss, switching between four modes (BASELINE, RECOVERY, EXPLORE, STABILIZE) using event-driven logic.67.6/100AdvancedRickMachine Learning
2ORBITLibrary / Framework · PythonThe ORBIT repository provides a Python implementation of the ORBIT optimizer for transformer models with rotary positional embeddings. It includes the optimizer, adapters for external models, baselines, training scripts, and detailed documentation for reproduction and use.67.4/100AdvancedPop AlexandruMachine Learning
3, shared rankORBIT-allTraining Experiment · PythonORBIT-all is a research repository implementing a function-aware optimizer for language model training, complete with a custom transformer-like model, experiment harness, and detailed protocol for ablation studies and validation.66.0/100AdvancedPop AlexandruMachine Learning
3, shared rankmodded-GLM-5.3-flash-from-scratchEducational Implementation · Python66.0/100AdvancedPop AlexandruMachine Learning
05trot-livive-netEducational Implementation · RustTROT is an educational implementation of a tiny neural network written from scratch in Rust, featuring hand-written forward and backward passes, deterministic data generation, and ONNX export with runtime verification, designed to teach the fundamentals of machine learning…65.5/100AdvancedJarad PeckenpaughMachine Learning

What it is

Where it counts

Around it on the Machine Learning board

More from Pop Alexandru