Qwen3.5-0.8B Instruction Following Prepared SFT Corpus
This is a public, tokenized/filtered derivative produced by the V5 Colab CPU-preparation notebook.
It contains int32 token IDs and training-boundary metadata, not the original raw download cache.
Base model: Qwen/Qwen3.5-0.8B
Base revision: f35bff78516d5b109092c941393bb6b124723ab1
Build ID: if-public-hf-v5_5-minimal-filters
Prepared examples: 1,317,837
Prepared bytes: 2,600,195,963
Upstream attribution /… See the full description on the dataset page: https://huggingface.co/datasets/MingSafeR/qwen35-08b-if-prepared-v5.
