Multimodal and specialized AI models gain prominence
Microsoft releases Lens, a 3.8B text-to-image model
LLM
Microsoft published Lens, a 3.8B-parameter text-to-image model that matches or exceeds larger 6B+ parameter models while using only 19.3% of their training compute, leveraging dense captions and multi-resolution batching. Commerce teams can deploy faster, cheaper image generation for product catalogs and visual search without the infrastructure cost of larger models.
View full article →