Infinity08/KAWK500M-Korean-Pretraining-10B download history

Infinity08/KAWK500M-Korean-Pretraining-10B is a text generation dataset on the Hugging Face Hub. In the last 30 days it was downloaded 28 times (6 in the last 7 days), and 464 times in total. It ranks #289,747 among datasets by monthly downloads.

KAWK500M Korean Pretraining 10B 한국어 중심 500M급 decoder-only 모델을 처음부터 학습하기 위한 사전 토큰화 데이터입니다. 각 행의 input_ids는 uint16[2049]이며 앞의 2,048개 토큰을 causal-LM 입력·레이블로 사용합니다. 크기 train: 5,075,616 rows / 10,399,937,184 stored tokens / 10,394,861,568 effective 2048-token training positi

Models trained on KAWK500M-Korean-Pretraining-10B

3 models list it as training data.

• Infinity08/KAWK-500M-Korean-Instruct-v1 162 downloads in 30 days
• Infinity08/KAWK-1.5-500M-Korean-Base 151 downloads in 30 days
• mradermacher/KAWK-1.5-500M-Korean-Base-GGUF 83 downloads in 30 days

Open Infinity08/KAWK500M-Korean-Pretraining-10B on Hugging Face

Sister project: Paper Pulse, the upvote history of every Hugging Face Daily Paper.