Hermes Function Calling

Stale
GitHub Jupyter Notebook MIT

Description

Hermes Function Calling is Nous Research's open-source function-calling model and dataset, fine-tunable for local function-call LLMs.

Key Features

  • Open-source — model and dataset fully released
  • Function calling — strong function-call capability
  • Commercial-friendly — weights usable commercially
  • Multi-format — JSON, Python and JS function schemas
  • Fine-tunable — compatible with mainstream training
  • Agent integration — works with LangChain and similar

Use Cases

💡 Deploying a local function-call LLM
💡 Fine-tuning Hermes for enterprise APIs
💡 Commercial open-source alternative for function calling
💡 Dropping OpenAI for function calling to save cost

Quick Start

# 支持自定义函数 schema 微调
pip install transformers
from transformers import AutoModelForCausalLM
m = AutoModelForCausalLM.from_pretrained("NousResearch/hermes-function-calling-v1")
print(m.generate("What is the weather in SF?"))

Related Projects