SenseTime
商汤新模型一脑干四活：识图、读字、测距、抠图 — 方向对，但离落地还差口气
SenseTime's vision model on ByteDance's Bagel does detection, OCR, depth, segmentation. Tops benchmarks; 80GB GPU plus license keep it research-only.
Back to home
SenseTime
3 articles tagged with this topic
SenseTime
SenseTime Open-Sources U1.5-Lite: Train Specialized Experts, Ship One Model
SenseTime open-sources U1.5-Lite image model: per-task experts trained separately, distilled into one. Qwen-Image-Bench climbs 13 points to 60.18. Sin
LocalLLaMA
Open-Source LLM Releases Cluster Up — Chinese Players Take the Stage
LocalLLaMA's July open-source LLM tracker shows a release surge — Huawei Pangu, InternLM, and SenseTime all debut. China's first collective open-sourc