Instructions to use google/gemma-3-27b-it with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use google/gemma-3-27b-it with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("image-text-to-text", model="google/gemma-3-27b-it")
messages = [
    {
        "role": "user",
        "content": [
            {"type": "image", "url": "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/p-blog/candy.JPG"},
            {"type": "text", "text": "What animal is on the candy?"}
        ]
    },
]
pipe(text=messages)

# Load model directly
from transformers import AutoProcessor, AutoModelForImageTextToText

processor = AutoProcessor.from_pretrained("google/gemma-3-27b-it")
model = AutoModelForImageTextToText.from_pretrained("google/gemma-3-27b-it")
messages = [
    {
        "role": "user",
        "content": [
            {"type": "image", "url": "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/p-blog/candy.JPG"},
            {"type": "text", "text": "What animal is on the candy?"}
        ]
    },
]
inputs = processor.apply_chat_template(
	messages,
	add_generation_prompt=True,
	tokenize=True,
	return_dict=True,
	return_tensors="pt",
).to(model.device)

outputs = model.generate(**inputs, max_new_tokens=40)
print(processor.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Inference
HuggingChat
Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use google/gemma-3-27b-it with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "google/gemma-3-27b-it"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "google/gemma-3-27b-it",
		"messages": [
			{
				"role": "user",
				"content": [
					{
						"type": "text",
						"text": "Describe this image in one sentence."
					},
					{
						"type": "image_url",
						"image_url": {
							"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
						}
					}
				]
			}
		]
	}'

Use Docker

docker model run hf.co/google/gemma-3-27b-it

SGLang

How to use google/gemma-3-27b-it with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "google/gemma-3-27b-it" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "google/gemma-3-27b-it",
		"messages": [
			{
				"role": "user",
				"content": [
					{
						"type": "text",
						"text": "Describe this image in one sentence."
					},
					{
						"type": "image_url",
						"image_url": {
							"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
						}
					}
				]
			}
		]
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "google/gemma-3-27b-it" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "google/gemma-3-27b-it",
		"messages": [
			{
				"role": "user",
				"content": [
					{
						"type": "text",
						"text": "Describe this image in one sentence."
					},
					{
						"type": "image_url",
						"image_url": {
							"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
						}
					}
				]
			}
		]
	}'

Docker Model Runner
How to use google/gemma-3-27b-it with Docker Model Runner:
```
docker model run hf.co/google/gemma-3-27b-it
```

Keeps repeating itself.

#53

by RealBiggly - opened Apr 13, 2025

Discussion

RealBiggly

Apr 13, 2025

Q4K_M. Simple role-play. LM Studio. RTX3090.
Every 4th line or so is basically a repeat of what it said in the previous response.
I thought at first it was because the context too long, when asking it to proof read a novel. But even a simple role-play, less than 10% of context used, with context set at 32K, it's repeating itself again.

When it works it's great, but at the same time it's totally trash and useless. Someone said there are known issues with the 3090? I have no idea how that would work? They also said there's an update, but from the looks of this it's not updated?

GopiUppari

Google org Apr 15, 2025

Hi @rdtfddgrffdgfdghfghdfujgdhgsf,

Even when only a small portion of the context window is used, this repetition issue still shows up and it’s not just limited to the RTX 3090. With different hardware setups have reported similar problems.

Could you please try to tweak the Sampling Settings:

Higher temperature values help make the responses less predictable and reduce repetition.
Adding a penalty (like 1.1) discourages the model from saying the same thing over and over.
Adjusting Top-k / Top-p can also help bring in more variety and control the randomness in the output.

Finally, make sure running the latest version of LM Studio because new releases can solve this type of issues. For more information, could you please refer to this reference.

Thank you.

DONGYUZHAO

Jun 10, 2025

•

edited Jun 10, 2025

I meet the same issue when trying to use gemma-3-27b to fine tune long texts. If the input text is longer than 3000 token, which is just half of typical length of a single novel unit in CJK, model will repeat if self after correct generate first 2500 token.

I tried to adjust temperature and penalty. Although the adjustment stopped the repetition, the quality dropped dramatically.

This make the model nearly unusable when handling CJK text work.

johnzhengaz

Aug 5, 2025

In my testing, I found it is resolved in latest image:

inferenceservice:
  predictor:
    containers:
      - name: kserve-container
        imageURL: vllm/vllm-openai:v0.10.0
        args:
          - --model=google/gemma-3-27b-it
          - --tokenizer=google/gemma-3-27b-it
          - --tensor-parallel-size=8
          - "--gpu-memory-utilization=0.9"
          - "--max-model-len=8192"
          - "--trust-remote-code"
          - "--enforce-eager"

huyhuynh1502

Aug 24, 2025

•

edited Aug 24, 2025

In my testing, I found it is resolved in latest image:

inferenceservice:
  predictor:
    containers:
      - name: kserve-container
        imageURL: vllm/vllm-openai:v0.10.0
        args:
          - --model=google/gemma-3-27b-it
          - --tokenizer=google/gemma-3-27b-it
          - --tensor-parallel-size=8
          - "--gpu-memory-utilization=0.9"
          - "--max-model-len=8192"
          - "--trust-remote-code"
          - "--enforce-eager"

Hi @johnzhengaz , what service did you use to run this container? Thank you for your help.

Upload images, audio, and videos by dragging in the text input, pasting, or clicking here.

Tap or paste here to upload images

· Sign up or log in to comment