Pinned Loading

1. V6-optimizer V6-optimizer Public

   A stability-first optimizer inspired by neural homeostasis and multi-scale memory, extending AdamW for heterogeneous architectures.

   Python 1

2. infersafe infersafe Public

   InferSafe is a production-grade, self-healing LLM inference infrastructure inspired by companies like Together.ai. Built with modularity, resilience, and scalability in mind

   Python 1

3. llmforge llmforge Public

   LLMForge is a modular, automated LLMOps pipeline for Reddit-based data curation and LoRA fine-tuning of TinyLLaMA-1.1B. It features real-world MLOps orchestration with Prefect and Modal, running ev…

   Python 6

4. finllama finllama Public

   FinLLaMA is a finetuned version of LLaMA 3 (8B) on financial instruction-style data. Trained using LoRA with the Unsloth framework, this model is optimized for reasoning, Q&A, and general understan…

   Jupyter Notebook

5. InfersafeV2 InfersafeV2 Public

   Production-style LLM serving pipeline with gateway validation, streaming, observability, and live benchmarking on a quantized vLLM backend.

   Python

6. The-drifter The-drifter Public

   Galaga-inspired arcade shooter with real-time hand tracking and computer vision controls.

   TypeScript