BAAIBeijing
Joined 2 years ago
Karma
48
Submissions
11
Member since
Dec 2
Recent submissions
Comment · 1 year ago
Hi HN, we're sharing an update on our work to improve reasoning in retrieval systems for RAG and Agents.
Our project, BGE-Reasoner, is an open-source, three-stage framework (Rewrite, Embed, Rerank) that showed strong performance on the BRIGHT benchmark for reasoning-intensive retrieval as of our submission on Aug 21.
Our main contribution is using synthetic data and reinforcement learning to handle complex queries that go beyond simple semantic matching. We're sharing this because we believe the framework itself is a solid, replicable contribution for anyone working on advanced RAG or Agent search.
We're in the process of open-sourcing the model weights, code, and training data to the community.
Love for you to take a look, share your thoughts, and we welcome any and all feedback or critiques. Thanks!
- Github: https://github.com/FlagOpen/FlagEmbedding/tree/master/resear...
- Benchmark: https://brightbenchmark.github.io
Comment · 2 years ago
The current model weights are distributed under the GPLv3 license. The authors are in the process of releasing a new series of Emu2 models that will be accompanied by a more flexible licensing arrangement. Stay tuned for updates.
Comment · 2 years ago
Thank you for reaching out with your inquiry regarding our dataset's compliance with the draft "Basic security requirements for generative artificial intelligence service."
I can confirm that our dataset adheres to the data source security requirements outlined in section 5.1 and the corpus content security requirements detailed in section 5.2 of the document. Additionally, we have implemented safety filters to address the main security risks associated with the corpus and generated content, as listed in the annex of the document, which outlines 31 types of potential risks across five categories.
Furthermore, our user agreement explicitly includes terms that mandate the secure use of our services. We take these measures seriously to ensure that our dataset and AI services are not only compliant with current regulations but also maintain the highest standards of data security and integrity.
Please let us know if you require any further information or clarification on this matter.
RoboBrain-X0 is now fully open-sourced, including pretrained models, datasets, and technical documentation, and is fully integrated with the BAAI RoboBrain 2.0 toolchain. This solution aims to provide a solid starting point for developers to build their own embodied intelligence applications that operate reliably in the real world. •RoboBrain-X0 GitHub: https://github.com/FlagOpen/RoboBrain-X0 •RoboBrain-X0 Multi-Chip Version (FlagRelease):https://huggingface.co/FlagRelease/RoboBrain-X0-FlagOS •RoboBrain-X0 Dataset:https://huggingface.co/datasets/BAAI/RoboBrain-X-Dataset •RoboBrain 2.0 Technical Documentation:https://arxiv.org/abs/2507.02029