Event sign up{:target="_blank" rel="noopener"}
🗓️ : March 06, 2025 Thursday
⏰ : 9 am PST / 11 am CST / 12 pm EST / 5pm GMT
Duration: 1 hour
- Quick intro about AI Alliance (5 mins)
- GneissWeb presentation (40 mins)
- Q&A (10 mins)
- Wrapup
Event recording will be available soon
Check resources - code, presentation slides ..etc
At IBM, responsible AI implies transparency in training data: Introducing GneissWeb (pronounced "niceWeb"), a state-of-the-art LLM pre-training dataset with ~10 Trillion tokens derived from FineWeb, with open recipes, results, and tools for reproduction!
In this session we will go over how we created GneissWeb and discuss tools and techniques used. We will provide code examples that you can try at your leisure.
👉 > 2% avg improvement in benchmark performance over FineWeb
👉 Huggingface page{:target="_blank" rel="noopener"}
👉 Data prep kit detailed recipe{:target="_blank" rel="noopener"}
👉 Data prep kit bloom filter for quick reproduction{:target="_blank" rel="noopener"}
👉 Recipe models for reproduction{:target="_blank" rel="noopener"}
👉 announcement{:target="_blank" rel="noopener"}
👉 Paper{:target="_blank" rel="noopener"}
Presentation
LLM app developers, data scientists, data engineers
Beginner - Intermediate
None
Presenation slides{:target="_blank" rel="noopener"}
Research Scientist @ IBM Almaden Research Center
Shahrokh Daijavad, a distinguished Research Scientist in the Watsonx Data Engineering group at IBM Almaden Research Center, has a rich background in Edge Computing and Data Engineering. He earned his B.Eng. and Ph.D. in electrical engineering from McMaster University and spent years at IBM T. J. Watson Research Center. His recent research focuses on AI@Edge and Data Engineering for IBM Watsonx AI offerings.
Linkedin{:target="_blank" rel="noopener"}
Please review the session recording