File size: 760 Bytes
ad38611 5b54ca7 ad38611 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 |
---
language:
- fi
pipeline_tag: text-generation
---
# gpt3-finnish-8B-GGML
Based on initial release of FinGPT-3 monolingual GPT-3-like model for Finnish from [TurkuNLP Research Group](https://huggingface.co./TurkuNLP).
This model is quantized GGML format binary of [gpt3-finnish-8B](https://huggingface.co./TurkuNLP/gpt3-finnish-8B).
| File | Description |
| ---- | --- |
| gpt3-finnish-8B.bloomz.ggml.q4_0.bin | 4 bit quantized. For inference using [bloomz.cpp](https://github.com/NouamaneTazi/bloomz.cpp) |
```bash
bloomz.cpp/main -m gpt3-finnish-8B.bloomz.ggml.q4_0.bin -t 4 -n 200 -p "Lasten loru:"
```
*Lasten loru:*
*Miten monta kärpässilmää,*
*lakki punainen tai sininen.*
*Vielä reppukin soma pieni!*
*Pulkkakin on lautaan tehty.* |