Puro-2B: Poor Lab's Qwen2-1.5B Trained on RTX 5090 within $5090

Benchmark Model Rank Results
question-answering-on-boolqPuro-2B UD full#46Accuracy: 79.48
question-answering-on-piqaPuro-2B UD decay average#79Accuracy: 77.69
code-generation-on-humanevalPuro-2B UD Tulu3 SFT#107Pass@1: 42.27
common-sense-reasoning-on-winograndePuro-2B CMA215k average#109Accuracy: 61.48
sentence-completion-on-hellaswagPuro-2B UD1/2#137Accuracy: 63.62