@shaunliew20

Running Qwen3-VL on DGX Spark: Transformers vs vLLM vs SGLang

15h ago · 14 min read · I've been running Qwen3-VL locally for a while now, mostly with the standard from_pretrained() setup. It works, but it's slow. So, I kept wondering whether switching to vLLM or SGLang would actually m

Fixing CUDA PTX Error When Running Qwen3-VL with vLLM on H200

Dec 7, 2025 · 4 min read · Running vision-language models like Qwen3-VL with vLLM on high-end GPUs should be straightforward. Except when it's not. The Problem I was setting up Qwen3-VL-8B-Instruct on our H200 cluster (8x H200, 143GB VRAM each) when I hit this error: vllm ser...

Understanding neural style transfer

Jul 29, 2025 · 9 min read · Neural Style Transfer: Turning the Mona Lisa into a Picasso Have you ever wondered what the Mona Lisa would look like if Picasso had painted it? Neural style transfer makes this artistic dream a reality by combining the content of one image with the ...

Performing an adversarial attack on images

Jul 29, 2025 · 10 min read · Exploring adversarial attacks - where tiny, imperceptible modifications can completely deceive even the smartest neural networks From Autoencoders to Adversarial Attacks: A New Kind of Magic After my fascinating journey with Variational Autoencoders,...

From Noise to Images: Mastering Variational Autoencoders

Jun 29, 2025 · 20 min read · From Autoencoders to Variational Autoencoders: Solving the Generation Problem In my previous exploration of autoencoders, I discovered their remarkable ability to learn compressed representations of data and reconstruct images with impressive fidelit...