FinGPT Medium
FinGPT Medium is a Finnish text-generation base model in TurkuNLP's Finnish GPT-3 series. It uses the BLOOM architecture and is not an OpenAI GPT-3 checkpoint.
Model design and training
24 layers, width 1,024 and 16 attention heads. The card reports 300 billion training tokens. The Finnish corpus combines web text, news, books, Wikipedia and discussion data. This is a pure language model, without dialogue or question-answering instruction tuning.gpt3-finnish-medium: official model card
Practical assessment
Practical guidance. Medium doubles the layer count of Small while remaining below one billion parameters. This makes it useful for studying depth and capacity within the same language family. Review long-distance agreement and named entities separately from general sentence fluency.
Specifications & access
| Detail | Value and conditions |
|---|---|
| Access | Downloadable |
| Model identifiers | gpt3-finnish-mediumFinnish GPT-3 Medium |
| Model size | 437 million parameters.gpt3-finnish-medium: official model card |
| License | Apache 2.0.gpt3-finnish-medium: official model card |
| Training state | Pretrained base; not instruction-tuned.gpt3-finnish-medium: official model card |
Related models
- FinGPT 13B
TurkuNLP · FinGPT
FinGPT 13B is a Finnish text-generation base model in TurkuNLP's Finnish GPT-3 series.
Launch date not confirmedDownloadable - FinGPT Small
TurkuNLP · FinGPT
FinGPT Small is a Finnish text-generation base model in TurkuNLP's Finnish GPT-3 series.
Launch date not confirmedDownloadable