Explore a curated, ranked list of 510 open-source atomistic machine learning projects scored by combined GitHub and package manager metrics — a model for scientific computing ecosystems.
Avatar Forcing implements diffusion forcing for causal, real-time multimodal input processing enabling expressive head avatars with ~500ms latency and 6.8X speedup over baselines.
A curated list of 24 rigorously selected books on LLM engineering, covering foundational theory to production deployment. Highlights a unique 6-step quality filtering process.
Google DeepMind’s representations4d bundles three self-supervised video learning approaches using transformers, including a novel object-centric tracking method with latent tokens moving off the pixel grid.
ForensiX combines ML-driven URL classification with browser artifact extraction for forensic analysis of Chrome and Brave data. Docker-based deployment included.
llm-madness offers a Python-built GPT-style transformer training pipeline with tokenizer training, memory-mapped datasets, and a unique web UI for per-layer attention inspection and loss visualization.
LTX Video Generator for Mac runs complex AI video generation entirely on Apple Silicon by bridging native SwiftUI with a Python subprocess. It manages large models, audio-video sync, and long tasks locally.
MegaTrain enables training 100B+ parameter LLMs on a single GPU by offloading all parameters to CPU RAM and streaming layers to GPU. Supports HuggingFace models and multi-GPU data parallelism without NCCL.
NAS3R enables self-supervised 3D geometry and camera parameter estimation without ground-truth data, using Gaussian splatting and a VGGT backbone. It supports multi-view setups and optional pretrained initialization.
OmniStream uses a multi-frame transformer to process continuous video streams with patch-level temporal indexing, supporting downstream vision-language-action tasks.
OpenMythos implements a recurrent-depth transformer that recycles layers via looped blocks, using input injection to prevent signal drift. It scales from 1B to 1T parameters with up to 1M token context.
SimRecon converts real-world videos into simulation-ready 3D scenes by combining geometry reconstruction, instance segmentation, viewpoint optimization, and semantic scene graph synthesis.
Voice Clone Studio unifies multiple voice AI engines in a modular Gradio web UI. Supports voice cloning, multi-speaker dialogs, speech-to-speech, and LoRA fine-tuning with GPU or Apple Silicon.
AgenticAiLabs AI Engineering Roadmap offers a modular, community-curated curriculum for self-taught AI engineers, covering from fundamentals to advanced AI topics like LLMs and RAG.
Explore an open-source course that teaches building a production-grade AI assistant using advanced retrieval-augmented generation, agent orchestration, fine-tuning, and LLMOps practices.
Explore a comprehensive LLM course with practical notebooks on fine-tuning (QLoRA, DPO), quantization (GPTQ), and tools like AutoEval and LazyMergekit. Ideal for aspiring LLM engineers.
This repo provides annotated PyTorch implementations of major deep learning papers with side-by-side explanations, aiding understanding and prototyping.
Microsoft’s “Generative AI for Beginners” offers 21 lessons with Python and TypeScript examples covering LLMs, prompt engineering, RAG, and AI app building.
LlamaFactory offers a modular Python framework for fine-tuning 100+ LLMs with diverse algorithms and optimizations, including LoRA, QLoRA, and reinforcement learning.
Microsoft’s ML-For-Beginners offers a 12-week, project-based classic machine learning course using Scikit-learn and Jupyter Notebooks, focusing on foundational concepts with interactive lessons and quizzes.
