tokenizer - BPE 30_522 vocab size

model - Roberta

    trained using MLM 
    OSCAR dataset
    train data size 5000 lines olly

Convert 30+ 3D formats online: GLTF, GLB, GBX, OBJ, DAE, IFC, STEP, STL...

Unreal engine based photo realistic synthetic data generator for YOLO.

AI powered 3d texture generation and projection SDK for three.js.