FinGPT-13B
FinGPT-13B is an AI model developed by University of California Los Angeles (UCLA), Columbia University and New York University (NYU) (United States), first published in October 2023. It works in the language domain, on tasks such as named entity recognition (ner), sentiment classification, language modeling/generation and financial management.
Training it took an estimated 1.6×10²³ FLOP of compute (estimation method: hardware). The model has 13,000,000,000 parameters. It was trained on roughly 76.8K datapoints. Training ran on 1 NVIDIA GeForce RTX 3090 for about 17.25 hours.
Access: Open weights (unrestricted). Its weights are openly available. It is built on top of Llama 2-13B. The reference paper has 113 citations. Epoch AI rates the confidence of this record as likely.
- Organization
- University of California Los Angeles (UCLA), Columbia University, New York University (NYU)
- Country of organization
- United States
- Domain
- Language
- Task
- Named entity recognition (NER), Sentiment classification, Language modeling/generation, Financial management
- Training compute
- 1.6×10²³ FLOP
- Compute estimation method
- Hardware
- Parameters
- 13,000,000,000
- Dataset size
- 76.8K
- Training hardware
- NVIDIA GeForce RTX 3090
- Chips used
- 1
- Training time
- 17 h
- Training power draw
- 382 W
- Numerical format
- FP16
- Model accessibility
- Open weights (unrestricted)
- Open weights
- Yes
- Base model
- Llama 2-13B
- Citations
- 113
- Epoch confidence
- Likely