Poro 2 8B SFT
Poro 2 8B SFT is the intermediate instruction-tuned derivative of Poro 2 8B Base. It has not received the final preference-tuning stage.
Research release
Model design and training
The bilingual instruction collection includes Tulu-derived examples, generated multi-turn conversations, OpenAssistant material and translation pairs. Supervised training runs for two epochs. The card recommends the final Instruct model for most users and retains SFT for comparison of post-training methods.Llama-Poro-2-8B-SFT: official model card
Practical assessment
Practical guidance. This checkpoint isolates the effect of supervised instruction training before DPO. Keep prompts and decoding settings identical when comparing the two stages. It can also provide a starting point for preference training on a different application.
Specifications & access
| Detail | Value and conditions |
|---|---|
| Status | Research release |
| Access | Downloadable |
| Model identifiers | Llama-Poro-2-8B-SFT |
| Context window | 8,192 tokens Sequence length reported by this release's model card; this is not a guaranteed useful retrieval length. Llama-Poro-2-8B-SFT: official model card |
| Model size | 8.03 billion parameters; 32 decoder layers.Llama-Poro-2-8B-SFT: official model card |
| License | Llama 3.3 Community License, as specified for this derivative.Llama-Poro-2-8B-SFT: official model card |
Related models
- Poro 2 8B Base
LumiOpen · Poro 2
A Llama 3.1 8B derivative with continued pretraining for Finnish, English, code and mathematics.
Launch date not confirmedDownloadable - Poro 2 8B Instruct
LumiOpen · Poro 2
An 8B Finnish and English conversation model trained with supervised instructions and Direct Preference Optimization.
Launch date not confirmedDownloadable