emfhal/gpt-oss

emfhal/ gpt-oss

434 Downloads Updated 2 months ago

CPU-optimized gpt-oss:20b for Kubernetes and VM environments. Runs without GPU, tuned for predictable CPU inference, balanced quality, and reduced memory usage.

tools thinking

ollama run emfhal/gpt-oss:20b-cpu-optimized

curl http://localhost:11434/api/chat \
  -d '{
    "model": "emfhal/gpt-oss:20b-cpu-optimized",
    "messages": [{"role": "user", "content": "Hello!"}]
  }'

from ollama import chat

response = chat(
    model='emfhal/gpt-oss:20b-cpu-optimized',
    messages=[{'role': 'user', 'content': 'Hello!'}],
)
print(response.message.content)

import ollama from 'ollama'

const response = await ollama.chat({
  model: 'emfhal/gpt-oss:20b-cpu-optimized',
  messages: [{role: 'user', content: 'Hello!'}],
})
console.log(response.message.content)