Hugging Face 发布 30M 参数 GenZ 俚语翻译模型
原标题:Sankar-2910/genz-translator
AI 摘要
开发者 Sankar-2910 在 Hugging Face 发布 genz-translator v2,一个约 2965 万参数、从随机初始化开始训练的 Llama 风格解码器模型,专门把 Gen Z 网络俚语翻译成标准英语。模型使用自建 8000 词表 Byte-Level BPE 分词器,在约 100 万条清洗去重的指令-响应对上继续训练,验证损失从 2.137 降至 0.7386(困惑度 2.09)。该模型仅面向俚语翻译,不适用于通用对话、编程或数学,且尚未进行 chrF/BLEU 等任务级评测。
正文节选
--- license: apache-2.0 language: - en pipeline_tag: text-generation library_name: transformers tags: - gen-z - slang - translation - llama - from-scratch - decoder-only - english --- # 🚀 GenZ Translator > **A ~30M parameter language model trained from scratch to translate Gen Z slang into clear, standard English.** ## Why this model?? Most public language models start from billions of pretrained weights. **This one doesn't.** Every weight began as a random number and learned exclusively