AI Labs Debugging and Repairing TensorRT Inference Tracing and fixing a TensorRT FP16 inference bug after fine-tuning a classification model. Full walkthrough: diagnosing ONNX input issues, rebuilding pipelines, validating logits and softmax, and benchmarking model size and speed.
AI Labs Quantisation in Deep Learning: A Practical Lab Guide Explore INT8 and FP16 quantisation of a LoRA-fine-tuned BERT across PyTorch, ONNX Runtime and TensorRT. Compare model size, inference latency and nine-way F1 to discover which workflow best balances accuracy and performance on CPU and GPU deployments.
AI Labs Fine-Tuning a Model with LoRA Fine-tune BERT with LoRA adapters for custom category & headline classification. Walk through adapter integration, training loop, and evaluation... preparing your model for efficient quantisation.
AI Labs Machine Learning: Building Our First Classifier with Scikit-Learn Train your first ML model using Scikit-Learn and the classic Iris dataset. Learn how KNN works, evaluate model accuracy with a confusion matrix and classification report, and understand what features like petal length reveal about species prediction.
