Data science, career and other topics
20 April 2023
How Meta built all-purpose visual features by scaling self-supervised pretraining to a curated 142M-image dataset, producing models that outperform OpenCLIP without any fine-tuning.
17 April 2023
My review of the paper InceptionNeXt When Inception Meets ConvNeXt
02 April 2023
Bloomberg trained a 50B-parameter LLM on 363B tokens of proprietary financial data. It crushes existing models on finance tasks while holding its own on general benchmarks.
27 March 2023
My review of the paper ReBotNet Fast Real-time Video Enhancement
23 March 2023
A short guide on creating a personal website using Jekyll, Github Pages and custom domain
Type at least 2 characters to search...