Pinned Loading

1. GPT-1B-From-Scratch GPT-1B-From-Scratch Public

   A complete 1B-parameter GPT implementation from scratch using PyTorch, DistributedDataParallel, Gradient Checkpointing, and 8× NVIDIA A100 GPUs.

   TypeScript

2. HPC_Infrastructure_Provisioning_xCat HPC_Infrastructure_Provisioning_xCat Public

   Complete xCAT deployment guide with a unified dashboard for centralized cluster provisioning, management, and HPC infrastructure administration.

3. HPC-Node-Cluster-Failure-Prediction-With-Prevention-CDAC-Project HPC-Node-Cluster-Failure-Prediction-With-Prevention-CDAC-Project Public

   Developed an AI-powered predictive maintenance system for HPC, leveraging Prometheus, Grafana, and ML models to analyze logs and real-time metrics (CPU, memory, latency, disk). Enabled anomaly dete…

   Shell

4. vllmops-exploration vllmops-exploration Public

   Exploring vLLM and vLLM-Ops for scalable LLM inference on NVIDIA DGX A100 (8 × 40 GB GPUs). Deployed eight independent model-serving endpoints, experimented with same-model and multi-model workload…

5. Vehicle-Detection-and-Counting-System-using-YOLOv5-Deep-SORT-and-OpenCV Vehicle-Detection-and-Counting-System-using-YOLOv5-Deep-SORT-and-OpenCV Public

   A real-time computer vision application that detects, classifies, tracks, and counts vehicles from video feeds. It uses YOLOv5 for object detection, Deep SORT for tracking, and Streamlit for an int…

   Python

6. distributed-llm-training-from-scratch distributed-llm-training-from-scratch Public

   A practical guide to distributed deep learning and LLM training with PyTorch, CUDA, DDP, FSDP, DeepSpeed, Megatron-LM, and multi-node GPU clusters.