soyrsoyr/erebus-v2-training-data download history
soyrsoyr/erebus-v2-training-data is a dataset on the Hugging Face Hub. In the last 30 days it was downloaded 75 times (13 in the last 7 days), and 274 times in total. It ranks #188,946 among datasets by monthly downloads.
Erebus v2 Pre-tokenized Training Data Pre-tokenized training data used to train soyrsoyr/erebus-v2-1.5b-base, a 1.5B parameter Qwen3-architecture language model trained from scratch. Format Binary files containing packed token IDs as uint32 values (4 bytes per token), t
Open soyrsoyr/erebus-v2-training-data on Hugging Face
Sister project: Paper Pulse, the upvote history of every Hugging Face Daily Paper.