Amber
Amber is an AI model developed by Mohamed bin Zayed University of Artificial Intelligence (MBZUAI), Petuum, University of Southern California, Carnegie Mellon University (CMU), University of Illinois Urbana-Champaign (UIUC), University of California San Diego and LLM360 (United Arab Emirates and United States), first published in December 2023. It works in the language domain, on tasks such as language modeling/generation and question answering.
Training it took an estimated 4.8×10²² FLOP of compute (estimation method: operation counting,hardware). The model has 6,700,000,000 parameters. Training ran on 224 NVIDIA A100 SXM4 80 GB for about 601 hours.
Access: Open weights (unrestricted). Its weights are openly available. Epoch AI rates the confidence of this record as confident.
- Organization
- Mohamed bin Zayed University of Artificial Intelligence (MBZUAI), Petuum, University of Southern California, Carnegie Mellon University (CMU), University of Illinois Urbana-Champaign (UIUC), University of California San Diego, LLM360
- Country of organization
- United Arab Emirates, United States
- Domain
- Language
- Task
- Language modeling/generation, Question answering
- Training compute
- 4.8×10²² FLOP
- Compute estimation method
- Operation counting, Hardware
- Parameters
- 6,700,000,000
- Training hardware
- NVIDIA A100 SXM4 80 GB
- Chips used
- 224
- Training time
- 601 h
- Training power draw
- 177.6 kW
- Numerical format
- BF16
- Model accessibility
- Open weights (unrestricted)
- Open weights
- Yes
- Epoch confidence
- Confident