Gemma 2 with tool call support

tools 2b 9b 27b

ollama run cow/gemma2_tools:27b

curl http://localhost:11434/api/chat \
  -d '{
    "model": "cow/gemma2_tools:27b",
    "messages": [{"role": "user", "content": "Hello!"}]
  }'

from ollama import chat

response = chat(
    model='cow/gemma2_tools:27b',
    messages=[{'role': 'user', 'content': 'Hello!'}],
)
print(response.message.content)

import ollama from 'ollama'

const response = await ollama.chat({
  model: 'cow/gemma2_tools:27b',
  messages: [{role: 'user', content: 'Hello!'}],
})
console.log(response.message.content)

Details

Updated 1 year ago

1 year ago

8014fa906afd · 16GB ·

model

archgemma2

parameters27.2B

quantizationQ4_0

16GB

license

Gemma Terms of Use Last modified: February 21, 2024 By using, reproducing, modifying, distributing,

8.4kB

params

{ "stop": [ "<start_of_turn>", "<end_of_turn>" ] }

65B

template

{{- if or .System .Tools }}<start_of_turn>system {{- if .System }} {{ .System }} {{- end }} {{- if .

1.4kB

Readme

cow/gemma2_tools

This model is based on krtkygpta/gemma2_tools, with some optimizations. This model only modifies template files to support tools.

Known issues

Gemma 2 sometimes calls tools that don’t exist. If you encounter this problem, send a tool response that says “Tool not found.”