Nous Capybara 34B
Nous Capybara 34B uses Amplify-Instruct, which expands seed instructions into multi-turn training conversations.
Training and model behavior
The 34B release starts from the 200K-context Yi model. Nous trained it for three epochs on the Capybara dataset. Luigi D. led the Capybara project, with contributions to datasets and training from other Nous researchers.Nous Capybara 34B model card
Working with this model
Practical guidance. Compare continuity across long conversations as well as single answers. Synthetic examples can reproduce the assumptions of the models that generated them.
Specifications & access
| Detail | Value and conditions |
|---|---|
| Launch date | November 2023 |
| Access | Downloadable |
| Model identifiers | NousResearch/Nous-Capybara-34B |
| Context window | 200,000 tokens Configured position limit; practical serving limits depend on the runtime. Nous Capybara 34B configuration |
| Base model | Yi-34B-200KNous Capybara 34B model card |
| Repository license label | mitNous Capybara 34B model card |
Related models
- Nous Capybara 3B v1.9
Nous Research · Capybara
A conversation model trained with Nous's synthetic-data method.
Launch date October 2023Downloadable - Nous Capybara 7B v1
Nous Research · Capybara
A conversation model trained with Nous's synthetic-data method.
Launch date September 2023Downloadable - Nous Capybara 7B v1.9
Nous Research · Capybara
A conversation model trained with Nous's synthetic-data method.
Launch date 2023Downloadable