The Allen Institute for AI (AI2) has released an open-source text dataset named Dolma, designed to enhance the transparency and innovation of AI language models. As a centerpiece of AI2's Open Language Model (OLMo) initiative, Dolma will provide researchers and developers with free access to data resources, supporting a broader range of AI research. Not only is Dolma a vast open dataset with 3 billion tokens, but it also features straightforward usage and licensing terms. AI2 has adopted the "ImpACT License for Moderate-Risk Work" and encourages users to provide contact information and usage details. The openness of this dataset offers researchers and developers more resources, propelling the AI field towards a more transparent and collaborative future.