Dr. Siladittya Manna

Whether you’re collaborating on a feature or just keeping your work in sync, pulling changes to a non-main branch is a daily task for most developers. Here is how to do it cleanly.

When you start serving large language models locally — like MedGemma-4B running behind a FastAPI server — you quickly encounter something that most ML engineers haven’t had to think about before...

Large language models are easy to run locally now. But running a model is not the same as serving it...

Unlike the Python package llama-cpp-python, the llama-server executable is not pre-installed anywhere. It is part of the C++ repository and must be compiled....

GGUF is a fully packaged, quantized model format designed specifically for inference....

Large multimodal models usually demand serious hardware. A 4B parameter model in full precision occupies roughly 8GB just for weights — and that’s before accounting for activations and KV cache during generation....

Segmentation masks are fundamental in computer vision applications, from medical imaging to autonomous vehicles. Visualising these masks...

with calflops and torchprofile - Learn how to measure computational complexity and efficiency of your PyTorch models...

In medical imaging, handling large datasets efficiently is crucial for storage and processing purposes. Neuroimaging Informatics...

On Multiple Nodes using SLURM