firefunction-v2

162.2K Downloads Updated 1 year ago

An open weights function calling model based on Llama 3, competitive with GPT-4o function calling capabilities.

tools 70b

ollama run firefunction-v2

curl http://localhost:11434/api/chat \
  -d '{
    "model": "firefunction-v2",
    "messages": [{"role": "user", "content": "Hello!"}]
  }'

from ollama import chat

response = chat(
    model='firefunction-v2',
    messages=[{'role': 'user', 'content': 'Hello!'}],
)
print(response.message.content)

import ollama from 'ollama'

const response = await ollama.chat({
  model: 'firefunction-v2',
  messages: [{role: 'user', content: 'Hello!'}],
})
console.log(response.message.content)

Applications

Claude Code ollama launch claude --model firefunction-v2

Codex ollama launch codex --model firefunction-v2

OpenCode ollama launch opencode --model firefunction-v2

OpenClaw ollama launch openclaw --model firefunction-v2

Models

View all →

Name

17 models

Size

Context

Input

firefunction-v2:latest

40GB · 8K context window · Text · 1 year ago

firefunction-v2:latest

40GB

Text

firefunction-v2:70b

latest

40GB · 8K context window · Text · 1 year ago

firefunction-v2:70b latest

40GB

Text

Readme

Firefunction-v2 is competitive with GPT-4o function calling capabilities, scoring 0.81 on a medley public benchmarks vs 0.80 for GPT-4o.

Firefunction-v2 is optimized for real world scenarios including multi-turn conversation, instruction following and parallel function calling. It retains Llama 3’s multi-turn instruction capability (0.84 vs 0.89 on MT bench) while consistently outscoring Llama 3 on function calling tasks (0.51 vs 0.30 on Nexus parallel multi function eval)

References

Blog Post

Hugging Face