Documentation Index
Fetch the complete documentation index at: /atlas/llms.txt
Use this file to discover all available pages before exploring further.
Call the chat models served by Gwenlake, with or without streaming.
for m in client.models.list():
print(m["id"])
gwenlake models.
messages = [
{"role": "system", "content": "You answer in one sentence."},
{"role": "user", "content": "Describe Argentina."},
]
response = client.chat.create(model="llama-3.1-8b", messages=messages)
print(response["choices"][0]["message"]["content"])
temperature defaults to 0. Raise it for more varied answers:
response = client.chat.create(model="llama-3.1-8b", messages=messages, temperature=0.9)
stream=True, the call returns an iterator of server-sent event lines as the model generates:
stream = client.chat.create(model="llama-3.1-8b", messages=messages, stream=True)
for chunk in stream:
print(chunk)
response_format to constrain the answer to JSON, or to a JSON schema:
response = client.chat.create(
model="llama-3.1-8b",
messages=[{"role": "user", "content": "Give the capital of France as JSON: {\"capital\": ...}"}],
response_format={"type": "json_object"},
)
from openai import OpenAI
oai = OpenAI(base_url="https://api.gwenlake.com/v1", api_key="sk-...")
oai.chat.completions.create(model="llama-3.1-8b", messages=messages)