Nous ResearchHermes Vision
Nous Hermes 2 Vision Alpha
Nous Hermes 2 Vision Alpha combines OpenHermes 2.5 Mistral 7B with a SigLIP-400M image encoder.
Preview
Training and model behavior
The training mix includes image instructions, image conversations and private function-calling examples. The model uses a Vicuna-style conversation template. Function requests have a dedicated opening tag.Nous Hermes 2 Vision Alpha model card
Working with this model
Practical guidance. Verify both visual grounding and argument validity before accepting a function call. Keep the image encoder and preprocessing code paired with this release.
Specifications & access
| Detail | Value and conditions |
|---|---|
| Status | Preview |
| Launch date | 2023 |
| Access | Downloadable |
| Model identifiers | NousResearch/Nous-Hermes-2-Vision-Alpha |
| Context window | 32,768 tokens Configured position limit; practical serving limits depend on the runtime. Nous Hermes 2 Vision Alpha configuration |
| Base model | OpenHermes-2.5-Mistral-7BNous Hermes 2 Vision Alpha model card |
| Repository license label | apache-2.0Nous Hermes 2 Vision Alpha model card |