HuggingFaceTB/finemath-ablation-3plus-160B
TEXT GENERATIONConcurrency Cost:1Model Size:3.2BQuant:BF16Ctx Length:32kPublished:Dec 19, 2024License:apache-2.0Architecture:Transformer Open Weights Warm

HuggingFaceTB/finemath-ablation-3plus-160B is a 3.21 billion parameter Llama3-based causal language model, part of the FineMath ablation studies. It was pretrained on 160 billion tokens, with a significant focus on mathematical datasets (FineMath-3+ and InfiWebMath-3+), alongside FineWeb-Edu. This model is specifically designed for text completion in English with an emphasis on mathematical reasoning and performance comparison within the FineMath research framework.

Loading preview...