Run Gemma 4 Locally

/set parameter num_ctx 32768     # longer memory for this chat
/set think                       # reason before answering
/set nothink                     # answer straight away
/clear                           # forget the conversation so far
/bye                             # leave

curl http://localhost:11434/api/chat -d '{
  "model": "gemma4:12b-it-qat",
  "messages": [
    {"role": "user", "content": "Why is the sky blue?"}
  ],
  "think": false,
  "stream": false
}'

# pip install openai
from openai import OpenAI

# Ollama: port 11434.  LM Studio: http://localhost:1234/v1
client = OpenAI(base_url="http://localhost:11434/v1", api_key="ollama")

reply = client.chat.completions.create(
    model="gemma4:12b-it-qat",
    messages=[{"role": "user", "content": "Why is the sky blue?"}],
)
print(reply.choices[0].message.content)

# pip install -U "sentence-transformers[image,audio,video]" transformers
from sentence_transformers import SentenceTransformer

model = SentenceTransformer("google/embeddinggemma-2")

query_emb = model.encode("ocean waves at sunset", prompt_name="SearchQuery")
doc_emb = model.encode("Photos from the beach trip", prompt_name="Document")
image_emb = model.encode({"image": "sunset_beach.jpg"})

# private question answering: both models on your machine
import ollama
from sentence_transformers import SentenceTransformer

embedder = SentenceTransformer("google/embeddinggemma-2")
passages = ["...your documents, split into passages..."]
passage_embs = embedder.encode(passages, prompt_name="Document")

question = "What did the contract say about renewal?"
q_emb = embedder.encode(question, prompt_name="QuestionAnswering")
best = passages[embedder.similarity(q_emb, passage_embs).argmax()]

reply = ollama.chat(model="gemma4:12b-it-qat", messages=[
    {"role": "user", "content": f"Answer from this passage only:\n{best}\n\nQuestion: {question}"}])
print(reply["message"]["content"])