Model Description Master is a collection of LLMs trained using human collected seed questions and regenerate the answers with a mixture of high performance Open source LLMs. Master Yi 9B is trained using the ORPO technique. The model shows strong abilities in reasoning on coding and math questions. Quantized Version : Here Communitiy Quantization (Thanks to @LoneStriker) exl2: Master Yi 9B 8.0bpw h8 exl2, Master Yi 9B 6.0bpw h6 exl2, Master Yi 9B 5.0bpw h6 exl2, Master Yi 9B 4.0bpw h6 exl2 GGUFs: Master Yi 9B GGUF Master Yi 9B Vision : Coming Soon Prompt Template Examples Inference Code Benchmarks Nous Benchmark: Model AGIEval GPT4All TruthfulQA Bigbench Average : : : : : Master Yi 9B 43.55 71.48 48.54 41.43 51.25 AGIEval GPT4All TruthfulQA Bigbench Average score : 51.25% OpenLLM Benchmark: Model ARC HellaSwag MMLU TruthfulQA Winogrande GSM8K Average : : : : : : : Master Yi 9B 61.6 79.89 69.95 48.59 77.35 67.48 67.48 ARC HellaSwag MMLU TruthfulQA Winogrande GSM8K Average score : 67.48%
We use cookies for essential functionality and analytics. You can accept or reject analytics cookies.Cookie policy