Nous ResearchEarly Nous instruction models
GPT4 x Vicuna 13B
This Nous release starts from Vicuna 13B v1.1 and uses roughly 180,000 GPT-4 instruction examples.
Training and model behavior
The dataset includes GPTeacher, role-play data, GPT4-LLM, WizardLM and Nous instructions. Training used five epochs on eight A100 80GB GPUs. The card specifies Alpaca prompt formatting.GPT4 x Vicuna 13B model card
Working with this model
Practical guidance. Treat this as an early research baseline. The name describes training data and ancestry; it does not mean that the model contains GPT-4 weights.
Specifications & access
| Detail | Value and conditions |
|---|---|
| Launch date | 2023 |
| Access | Downloadable |
| Model identifiers | NousResearch/GPT4-x-Vicuna-13b-fp16 |
| Context window | 2,048 tokens Configured position limit; practical serving limits depend on the runtime. GPT4 x Vicuna 13B configuration |
| Base model | eachadea/vicuna-13b-1.1GPT4 x Vicuna 13B model card |
| Repository license label | gplGPT4 x Vicuna 13B model card |