guo925658/linear-ttt-dclm-100m-600m
Qwen-LaCT：完整 Stage 1 100M + Stage 2 600M 训练数据 本目录包含实际冻结数据池的完整 token 数据，可复现训练输入。100M / 600M 是阶段预算的简称。 文件 完整输入 tokens 字节数 stage1-100m-99975168tokens.int32.bin 99,975,168 399,900,672 stage2-600m-599916544tokens.int32.bin 599,916,544 2,399,666,176 数据来源：mlfoundations/dclm-baseline-1.0，revision a3b142c183aebe5af344955ae20836eb34dcf69b。 格式：小端有符号 int32，tokenizer 为原生 Qwen2.5-3B-Instruct。文件是原始 master 的精确字节切片。 使用 将目录中的全部文件下载到同一目录。 执行 sha256sum -c… See the full description on the dataset page: https://huggingface.co/datasets/guo925658/linear-ttt-dclm-100m-600m.
This repository belongs to guo925658 on Hugging Face.
CoolFace never edits a repository it does not host. Visibility, licence, collaborators and gating are all managed at the source.