Instructions to use google/gemma-4-31B-it with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use google/gemma-4-31B-it with Transformers:

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

pipe = pipeline("image-text-to-text", model="google/gemma-4-31B-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-4-31B-it")
model = AutoModelForImageTextToText.from_pretrained("google/gemma-4-31B-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
AMD Developer Cloud
Local Apps

vLLM

How to use google/gemma-4-31B-it with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "google/gemma-4-31B-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-4-31B-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-4-31B-it

SGLang

How to use google/gemma-4-31B-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-4-31B-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-4-31B-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-4-31B-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-4-31B-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-4-31B-it with Docker Model Runner:
```
docker model run hf.co/google/gemma-4-31B-it
```

Will there be a small model for speculative decoding?

#71

by Regrin - opened Apr 17

Discussion

Regrin

Apr 17

Hello!
I'm very concerned about the speed of models. The easiest way to improve it without sacrificing quality is speculative decoding.
I have a slow computer, and I'd like to try it, but the problem is that there's no small model for Gemma 4.
Could you perhaps train a microscopic model for speculative decoding?
That would be REALLY REALLY REALLY good.

MeganEFlynn

Apr 17

Hello! I'm not sure about Google's plans, but we have trained a model for use in vLLM, which is available here: https://huggingface.co/RedHatAI/gemma-4-31B-it-speculator.eagle3. We intend to keep iterating over the next week. Feel free to try it out!

qenme

Apr 20

Hello! I'm not sure about Google's plans, but we have trained a model for use in vLLM, which is available here: https://huggingface.co/RedHatAI/gemma-4-31B-it-speculator.eagle3. We intend to keep iterating over the next week. Feel free to try it out!

This is amazing if it works. Do you know if it's supported with pipeline parallelism on VLLM?

SporkySporkness

30 days ago

•

edited 30 days ago

My token generation speed shockingly doubled after I started using E2B for speculative decoding in llama.cpp. Tried a few quants and settled on IQ3_M.
Here's my launch command.

llama-server ^
--model "M:\VLM\Gemma-4-31B-it-GGUF\gemma-4-31B-it-Q5_K_M.gguf" ^
--model-draft "M:\VLM\Gemma-4-E2B-it-GGUF\google_gemma-4-E2B-it-IQ3_M.gguf" ^
--device-draft CUDA0 ^
--draft-max 8 ^
--draft-min 1 ^
--n-gpu-layers 999 ^
--ctx-size 98304 ^
--no-mmap ^
--flash-attn on ^
--context-shift ^
--jinja ^
--parallel 1 ^
--reasoning-budget 0 ^
--reasoning off

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