romeerp/parameter-golf-caseops-v1-webish-ud24
收藏Hugging Face2026-04-30 更新2026-05-03 收录
下载链接：
https://hf-mirror.com/datasets/romeerp/parameter-golf-caseops-v1-webish-ud24
下载链接
链接失效反馈官方服务：
资源简介：
该数据集仓库包含用于参数高尔夫实验的重新标记化的CaseOps风格数据集和标记器。内容包括：标记器模型和词汇表文件，以及训练和验证用的二进制文件。标记器训练时添加了24个面向网络的用户定义符号，数据集被限制为80个训练分片以便记录跟踪式运行。
This dataset repo contains a retokenized CaseOps-style dataset and tokenizer for parameter-golf experiments. Contents include tokenizer model and vocabulary files, as well as binary files for training and validation. The tokenizer was trained with 24 added web-oriented user-defined symbols, and the dataset is capped to 80 train shards for record-track style runs.
提供机构：
romeerp
