StableLM
The StableLM series of language models is Stability AI's entry into the LLM space. Trained on The Pile, the initial release included 3B and 7B parameter models with larger models on the way.
A side-by-side comparison of release history, capabilities, licensing and model sizes.
| Attribute |
 StableLM
|
 Vicuna
|
|---|---|---|
| Developer | Stability AI | UC Berkeley, CMU, Stanford, MBZUAI, UCSD |
| Current generation | – | – |
| Initial release | 2023-04-19 | 2023-03-30 |
| Instruct tuned | No | Yes |
| Coding | No | No |
| Finetuning | Yes | Yes |
| Open source | Yes | Yes |
| License | CC BY-SA 4.0 | Noncommercial |
| Model sizes | 3B, 7B | 13B |
| Links |
The StableLM series of language models is Stability AI's entry into the LLM space. Trained on The Pile, the initial release included 3B and 7B parameter models with larger models on the way.
Released alongside Koala, Vicuna is one of many descendants of the Meta LLaMA model trained on dialogue data collected from the ShareGPT website. According to the authors, Vicuna achieves more than 90% of ChatGPT's quality in user preference tests, while vastly outperforming Alpaca. As of May 2023, Vicuna seems to be the heir apparent of the instruct-finetuned LLaMA model family, though it is also restricted from commercial use.